“This transcripts has not been edited or corrected, but appears as received from the commercial transcribing service

P-R-O-C-E-E-D-I-N-G-S

8:10 a.m.

CHAIRMAN SALOMON: On the record. Good morning, everybody, on this post snow morning or pre snow morning depending on what happens with the weather in the few hours or day. For those of you who are here, I certainly want to say for the Committee and for the FDA that we very much appreciate your attendance, weather notwithstanding and welcome.

We started a little bit late just to give people a little extra time to get here given the roads. I assume the roads are on the icy side. Given that I haven't left the hotel, I can only surmise from past experience. I'm going to keep my introductory remarks short. We have a relatively long day today. The title of this talk is "Stem Cell Redux." That doesn't officially go anywhere. This is my challenge to Gail which is to at least have a real title for every meeting instead of Meeting No. 34 which just doesn't do anything for me. Welcome and with that I will turn to Gail to give an introductory statement.

EXEC. SECRETARY DAPOLITO: Good morning, Dr. Salomon and the Committee. This announcement is part of the public record for the Biological Response Modifiers Advisory Committee Meeting on February 27, 2003. Pursuant to the authority granted under the Committee Charter, the Director of FDA's Center for Biologics Evaluation and Research has appointed Drs. Mary Horowitz, Harvey Kline, Jeanne Linden and David Stroncek as temporary voting members for today's discussion.

Based on the agenda, it was determined that there are no products being approved at this meeting. The Committee participants were screened for their financial interests to determine if any conflicts of interests exist. The Agency reviewed the agenda and all relevant financial interests reported by the meeting participants.

The Food and Drug Administration prepared general matters waivers for special Government employees who required a waiver under 18 USC 208. Because general topics impact on so many entities, it is not prudent to recite all potential conflicts of interest as they apply to each member. FDA acknowledges that there may be potential conflicts of interest but because of the general nature of the discussion before the Committee, these potential conflicts are mitigated.

We also note for the record that Ms. Alison Lawton serves as the non-voting industry representative member acting on behalf of regulated industry. She is employed by Genzyme and thus has interest in her employer and other similar firms.

With regards to FDA's invited guests, the Agency has determined that the services of these guests are essential. The following interests are being made public to allow meeting participants to objectively evaluate any presentation and/or comments made by the guests. Dr. Nelson Chao is employed by Duke University. Dr. Pablo Rubinstein is employed by the National Cord Blood Program of the New York Blood Center. Dr. John Wagner is employed by the University of Minnesota.

Members and consultants are aware of the need to exclude themselves from discussions involving specific products or firms for which they have not been screened for conflicts of interest. Their exclusion will be noted for the public record. With respect to all other meeting participants, we ask in the interest of fairness that you state your name, affiliation and address any current or previous financial involvement with any firm whose product you wish to comment upon. Waivers are available by written request under the Freedom of Information Act. We also ask as a courtesy to the Committee discussion and your neighbors in the audience that you silence cell phones and pagers. Thank you.

CHAIRMAN SALOMON: So as usual given that the membership changes particularly with ad hoc members, I always like everyone to introduce themselves briefly and just explain in a sentence or two what your general area of expertise is. Richard, do you want to start?

DR. MULLIGAN: I'm Rich Mulligan from Harvard Medical School and Children's Hospital. My lab works on gene transfer and stem cells.

DR. ALLAN: I'm Jon Allan from Southwest Foundation for Biomedical Research. I work with SIV pathogenesis. I'm basically a virologist.

DR. KURTZBERG: I'm Joanne Kurtzberg. I direct the Pediatric Bone Marrow Transplant Program at Duke University and also the Carolinas Cord Blood Bank at Duke.

DR. TSIATIS: I'm Butch Tsiatis from North Carolina State University and I'm a biostatistician.

DR. STRONCEK: Dave Stroncek from the Department of Transfusion Medicine, NIH Clinical Center. I've been involved with unrelated stem cell donors on and off for a number of years.

DR. HOROWITZ: Mary Horowitz from the Medical College of Wisconsin, Bone Marrow Transplant Unit. I'm the Director of the International Bone Marrow Transplant Registry. I spend most of my life analyzing the outcomes of hematopoietic stem cell transplants.

MS. LAWTON: I'm Alison Lawton. I'm the industry rep. I'm also head of registry at Genzyme Corporation.

DR. RAO: My name is Mahendra Rao from the National Institute on Aging. I work on neural stem cells.

CHAIRMAN SALOMON: Dan Salomon, I'm from Scripps Research Institute. I'm working on organ and cell transplantation, gene therapy and xenotransplantation.

EXEC. SECRETARY DAPOLITO: Gail Dapolito, Executive Secretary of the Committee. I would also like to introduce Rosanna Harvey, the Committee Management Specialist.

MS. WOLFSON: Alice Wolfson. I'm the consumer representative on the committee.

DR. HARLAN: David Harlan. I work for National Institutes of Diabetes, Digestive and Kidney Diseases where I study diabetes and immunological therapies for transplant.

DR. HIGH: Kathy High. I am at the University of Pennsylvania and the Children's Hospital of Philadelphia. My interests are in the molecular basis of blood coagulation and gene transfer for hemophilia.

DR. LINDEN: Jeanne Linden from the New York State Department of Health. I'm a pathologist and transfusion medicine physician. In addition to many other facilities, we license stem cell banks to do business in New York.

DR. HARVATH: I'm Liana Harvath. I'm from the National Heart, Lung and Blood Institute, Division of Blood Diseases and Resources and have had a long standing interest in leucocyte biology and before going to NHLBI I worked at FDA for 19 years and worked in the stem cell policy area.

DR. LITWIN: I'm Stephen Litwin. I'm a medical officer at the Food and Drug Administration.

DR. LAZARUS: I'm Ellen Lazarus. I'm also a medical officer in the new Office of Cellular, Tissue and Gene Therapies. I'm a recent transplant from the Office of Blood and I'm at the FDA.

DR. SOLOMON: I'm Ruth Solomon. I'm from the Division of Human Tissues in the Office of Cellular, Tissue and Gene Therapies. I'm a pathologist.

DR. NOGUCHI: Phil Noguchi, Acting Director of the Office of Cellular, Tissue and Gene Therapies.

CHAIRMAN SALOMON: Thank you all. It just dawned on me as we went around here that this is more like the original BRMAC that I joined where it was a very strong group in stem cell transplantation and particularly at that time in bone marrow transplantation. That was a lot of what we were doing. It's interesting how the pendulum swings. It's fun to have everybody back again. Phil, I think you are up next.

DR. NOGUCHI: I'd like to ask Karen Weiss to join me at the podium. As with any committee, we like to think of the committee not just as a set of people to give us advice but as really a team of colleagues that takes time to develop. We need to have time to plan, a time to practice, a time to actually come together and then to make some decisions. Unfortunately as with all teams, we have rotations through this. Tomorrow is Dan Salomon's last meeting. Both Karen and I have enjoyed his leadership of this committee over a number of years and we would just like to say a few words but it won't be too embarrassing. Karen.

DR. WEISS: Good morning, everybody. My job this morning is fairly straight forward and quite pleasant for a change. I have the distinction of acknowledging and thanking you, Dr. Salomon, on behalf of the Office of Therapeutics Research and Review currently still in CBER for all that you have done for the agency, first as a member of the BRMAC and more recently over the many years as chair of this advisory committee.

During the many hours of deliberations at these meetings where the Agency sought and received the Committee's advice on important cutting edge areas of medicine and basic science, you have been a tireless champion of the public health. I specifically recall one meeting, it might have been the islet cell meeting or it might have been one of those many meetings on gene therapy and long term follow-up that never would end ? anyway, it was one of those meetings where another Committee member was rotating off the Committee. As typical, the Agency officials came up and acknowledged this member and presented him with a plaque.

Subsequently the meeting happened and as typical of all of the meetings I think that have characterized your term, the topic was quite complex and complicated and controversial. The FDA questions as typical were quite convoluted and long. As most people know, my former boss who has recently left the Agency, Dr. Jay Siegel, was a master at editing and making questions long and convoluted but also like you, a tireless champion of the public health. It was one of those types of meetings and I think somewhat frustrated, you kept pushing and challenging the Committee to be more responsive and more direct in their advice to the Agency. I think finally you said look guys at the end of the day when my term is up I want to get one of those plaques too so you have to do better.

Dan, you certainly have earned your plaque. But more than just the plaque, you have certainly earned mine and the Office and the Agencies' deep respect and appreciation for all that you have done over the years in your championship for what was right and for providing the best advice to the Agency and ultimately for enhancing the public health from in utero transplantation to islet cell transplantation to haplo-identical transplantation to stem cell factor for peripheral blood progenitor cell transplantation as well as for the more recent advice on gene therapy issues including and not last of course the long term follow-up where we kept pushing. We brought the issue back not once, not twice but three times. We finally got some really good solid advice. For all of that, I wish to personally thank you. You have left behind a legacy. It's been a great experience having you on the Committee and as the Chair. Thank you very much.

DR. NOGUCHI: I'm not going to spend too much more time on this. Karen has actually gone through many if not all the highlights over a relatively brief but an extraordinarily exciting time. I'll just give him my one little dig here. A number of years ago actually Dan had been on fellowship from his position at Florida to the NIH. Of course like any good FDA person, I tried to recruit him to FDA. He finally was gracious enough to tell me in person that well I have this fairly nice job at Scripps. It's in San Diego. My wife is from there. It's really hard to turn down the FDA but ?-

At that point, I said don't worry. You're always going to be working with us. Even though, Dan, this is your last meeting for the BRMAC as chair, don't worry. We'll keep asking you back.

We have here two letters for you. One is from the Agency representing our appreciation for just an outstanding job done. It's also a bill for what you are going to do for us in the future. First I would like to give you that. Then this is our plaque that you so richly deserve and wanted after all these years. Dan, my very best and thank you so much for helping us with all these difficult issues.

CHAIRMAN SALOMON: As my swan song here, all I can say is the privilege of serving as chair of this Committee over the last several years is something that I can't imagine having not had that experience. So for me, it's been a privilege. The quality of the people that I've met and friendships that I've made during that as well and just the intellectual stimulation and the feeling of doing something right, it's just a remarkable mix for me. I thank everybody here. The FDA has been so supportive and so flexible and so interesting.

All I can say is that if you think about it as a theme in America, the idea of voluntary service to the Government is something that began with the very start of this country. It is something that's extremely precious as part of being an American. So it's been a real privilege to participate in something as fundamental as that. Thank you.

What we can do now is get started with the FDA introduction to the topic. For that I introduce Ellen Lazarus.

DR. LAZARUS: Thank you, Dr. Salomon. Welcome everyone. As you know, the focus of the first meeting this morning is on unrelated allogeneic hematopoietic stem progenitor cells from placental umbilical cord blood for hematopoietic reconstitution hereinafter referred to as cord blood or in my slides UCB. The CBER presenters and the invited speakers will be presenting efficacy data for the use of minimally manipulated cord blood for hematopoietic reconstitution for recipients of particular age groups.

A brief overview of the field of cord blood research and transplantation. This history begins in the waning years of the 20^th century. In 1988, Dr. Gluckman and colleagues performed the first reported transplant of cord blood in a child with Fanconi's anemia using an HLA matched sibling donor. Then in the early to mid-90's, more reports were published demonstrating the feasibility and efficacy of related and unrelated HLA matched and mismatched cord blood transplantation.

By 1993, there were three public cord banks established located in New York, Dusseldorf and Milan. In 1997, Drs. Broxmeyer and Cooper published a report extending their earlier observations on the effects of frozen storage of cord blood. In this study, they assessed recovery of nucleated cells and proliferative capacity of the thawed hematopoietic cells in vitro demonstrating potential functionality of cord blood cryo-preserved for up to 10 years. They have actually recently published a report extending their observations to cord blood cryo-preserved for 15 years.

Then in 1998, Drs. Rubinstein and colleagues published their landmark study demonstrating that stored cord blood is a potentially useful source of hematopoietic stem cells for patients who lack an HLA matched related donor. Next slide.

Now more the 2000 unrelated donor cord blood transplants as estimated have been performed since the beginning of this history and evidence is accumulating in the published literature suggesting comparable efficacy between HLA matched bone marrow and cord blood for pediatric patients with acute leukemia. Some preliminary reports of successful unrelated donor cord blood transplant in pediatric patients with certain immune deficiency disorders and in-born errors of metabolism have been published. In contrast the experience with the use of cord blood for transplantation in adults with hematopoietic malignancies is more limited.

Now one of the potential advantages of cord blood as a source of hematopoietic stem cells for transplantation is the relatively rapid availability of a particular unit once it's been identified in a registry as being a good match for the patient. But it is difficult to exactly determine the number of facilities storing unrelated donor cord units and the exact number of units stored within. However we can use some Internet resources to come up with some reasonable estimates.

We have the very informative National Marrow Donor Program (NMDP) website that lists approximately 30 cord banks. Not all of these facilities participate in the program. Some of these listed banks store unrelated donor cord units exclusively while others store both unrelated donor units as well as units for family or autologous use.

Numerous cord banks have been established outside the U.S. as well as evidenced by the listing on the Bone Marrow Donors Worldwide website. Using these sources of information, it can be estimated that there are over 60,000 units of cord blood potentially available for unrelated transplantation in the U.S. There are over 130,000 cord blood units stored in facilities worldwide.

Now the quality of the stored cord blood products is essential to assure successful hematopoietic reconstitution which depends on the number of functional progenitor cells transplanted. The assurance of quality begins in the moments after delivery of the infant when first after assuring safety of the donor and mother. The collection staff must optimize collection of the cord blood.

Now many of the establishments that collect, process and store cord blood products follow voluntary standards published by AABB, FACT, NetCord, and NMDP. Nevertheless, the quality and characteristics of stored cord blood may vary due to both uncontrollable donor factors such as placental size and weight and controllable factors such as collection methods, product containers and different processing and storage systems. Next slide.

We are actually rather limited in the assays that we can use for determining the number and quality of the hematopoietic progenitor cells in each of these units. Nucleated cell counts are performed and these counts include all differentiated and undifferentiated nucleated cells. Viability assays are invariably performed either using dye exclusion methods or flow cytometry.

Colony forming unit (CFU) culture assays are performed on many of these products and in these assays colonies from individual cells grow. These cells are plated on the semisolid medium. The limitation of these assays we're all familiar with. The reading of them is somewhat subjective. They are difficult to standardize. They take 10 to 14 days approximately to grow. In contrast, CD34 cell enumeration by flow cytometry is fast.

Our speakers today will give more information about the role of these different assays in the manufacturing and selection of cord blood units. So that was just a very brief summary of medical and manufacturing issues that we need to keep in mind and will be used to inform our discussion today.

The purpose of our meeting is to review first CBER analyses of clinical outcome data that were submitted to the FDA regarding safety and efficacy of cord blood for hematopoietic reconstitution. We will have invited speakers present data from recent studies of cord blood transplantation, Drs. Wagner, Rubinstein and Chao. We will seek the Advisory Committee comments on factors the Agency should consider in determining safety and efficacy of cord blood and on the role of CD34 cell count in selection of cord blood and other measures of quality that should be considered. So I look forward to a very interesting discussion today. I'll turn the podium over to my colleague, Dr. Steve Litwin.

DR. LITWIN: Good morning. Transplants using allogeneic stem cells are a therapeutic adjunct in a large number of life-threatening diseases. First slide please. They include hematologic and non-hematologic malignancies and an increasing number of in-born errors of metabolism, phagocytic and histiocytic disorder storage diseases. The IBMTR, and this is an estimate from the year 2000, estimates that there are approximately 15,000 related and unrelated transplants during a year.

Despite the increasing activity which has leveled off somewhat, the restrictions on transplantation, particularly allotransplantation, remain on the supply side. What I mean by that is that the restrictions are on the availability of suitable well matched units for donor transplant which would in FDA terminology lower the risk. The emphasis then is on the sources of stem cells. Next slide please.

These are the classic sources of stem cells. With the development of growth factors, it was possible to mobilize stem cells from tissue sources into the blood and collect them by phoresis. The most recent addition to available sources of stem cells is umbilical cord blood. Next slide please.

This now serves as a useful and valuable alternative to bone marrow and peripheral blood. There is obviously much less patient experience to date. About two years ago, CBER set up a working group whose mandate was to look at the efficacy of stem cells from different sources with emphasis on umbilical cord blood and particularly with interest in the two problems which we are going to deal with today. That is the enumeration of umbilical cord blood in grafting cells and finally the age related risk of UCB transplants.

It is my task today to focus on the latter which is the age related risk. I would like to do that by presenting data that had been submitted to the docket from the New York Blood Center and which was analyzed by FDA statisticians. Before I do that, I would like to emphasize some salient features of umbilical cord blood and talk somewhat about the differences in engraftment kinetics. Next slide please.

There are three things I would like to emphasize. First of all, cord blood units have a smaller number of engrafting cells available than do the classic sources, bone marrow and peripheral blood. Secondly, cord blood may represent a different mix of stem cell populations. Finally, cord blood cells may have different biological properties mainly suggested by laboratory studies. Next slide please.

This is data taken from a study of umbilical cord blood from a single study by Wagner, et al. The total nucleated cells are listed on the top because this still is the method generally used by many centers to estimate the potential usefulness of a cord blood unit as a donor source. You look at the number of CD34 positive cells which are listed here, which may differ between studies, but are listed here as 280,000 per kilogram of body weight of the recipient. This can be compared to the number of CD34 cells which are usually included in the Infusaid in the peripheral blood transplant which is roughly a log larger. That is about 10 times larger between two and five million CD34 cells per kilogram.

Perhaps the good news is in the number of T-cells, CD3 positive cells, which are listed here for the cord blood unit in this study as eight million per kilogram of body weight of the recipient. That's five to ten times lower than the amount given in an T-cell non-manipulated peripheral blood transplant. Next slide please.

Because of this factor, that is the limitation on the number of available engrafting cells, the majority of the umbilical cord blood as a transplant source to date has been confined to younger recipients. These are data once again from the IBMTR. They are divided into the younger age group on the left, less than 20 years, and the older age group on the right.

If we just look at the period of time from 1998 to 2000, you can see in the younger group that there was a predominance of bone marrow transplants as opposed to peripheral blood, 69 percent vs. 17 percent. About 15 percent of the transplants and this is up to year 2000 were from cord blood sources. In the older age group there's a closer to equal number of sources from bone marrow and peripheral blood, 53 percent vs. 43 percent. A very small number were from sources from cord blood. Next slide please.

The second point is that umbilical cord blood may represent the different mixture of stem cells of different levels of maturity and different lineages. The incidents of graft-versus-host disease is suggestively lower after cord blood transplants. At this particular point in time because of the lack of comparative studies, there remain many questions about whether the anti-tumor that is allogeneic effect noted in allo-transplants is present to the same diminished or increased three in cord blood and particularly whether cord blood cells' ability to reconstitute the immune responses to infectious agents and to other wide variety of antigens needed in a reconstituted recipient. Next slide please.

The third point is based mainly on laboratory data which suggests that cord blood cells have different biological properties. They seem to have increased growth and engraftment potential. Colonies formed in culture are larger. Cord blood cells can repopulate immunodeficient mice without cytokine support in sharp contrast to peripheral blood as a source of stem cells in these laboratory studies. A subset of cells, CD34 positive, 38 negative, have a higher comparative cloning efficiency from cord blood in response to cytokines and there are several reports that suggest that cord blood cells may be more responsive to ex vivo expansion, a relatively critical area in experimental growth and experimental efforts to try to increase the supply. Next slide please.

The engraftment kinetics of cord blood cells also differ. In the first vertical column marked HLA mismatched greater than two, the percentages in each of three studies are shown. These are three illustrative studies of cord blood used as a source of stem cells. You can see that the percentages of mismatched greater than two mismatches at the HLA loci differ here. These studies are different.

Despite that, what is evident is the relatively high incidents of engraftment failure, the 12 percent, 29 percent and 19 percent, can be compared. These are only illustrative figures. I don't mean that they represent the whole field. They can be compared to the anticipated one to five percent one would expect in most homologous and well matched allotransplants. The neutrophil engraftment time, that's the fourth column, is about twice as long in cord blood as it would be anticipated from the historical literature from peripheral blood or bone marrow. The platelin engraftment time is two to three times as long as once again one would anticipate from historical literature.

Interestingly enough, the incident of acute graft-versus-host disease grades three and four are relatively similar than these three selected studies. At 20 percent, these figures are lower than one would ordinarily anticipate from studies of patients at this particular clinical status but there is no comparative study to make any definitive statements. Next slide please.

I would like to turn now to some of the statistical analyses that we performed. To begin with, these were done by Dr. Jean Wang of our Department of Bio Statistics with Dr. Peter Lachenbruch serving as a consultant. The source were primary data on 562 cord blood transplants which were collected at the New York Blood Center. The recipients for the most part were myeloablated and for the most part were under 20 years of age. The data was collected from 1992 to 1998. Next slide please.

There were four clinical outcome studied, neutrophil engraftment both the presence of neutrophil engraftment, yes, no and the time to neutrophil engraftment, platelin engraftment once again similarly, disease-free survival and severe acute graft-versus-host disease, grades three and four. The co-variate studies were multiple age, weight, the total number of nucleated cells per kilogram of body weight of the recipient and the HLA disparities. I bolded and enlarge the age because we spent most of our time focusing on this aspect of it. The remainder are competing risks and that is acknowledged. Next slide.

The goals were threefold: first of all, to compare the risk of bad clinical outcomes between the very youngest members of this 562 patient group and the oldest members in the study; and if there was an increased risk to see the trend or shape of that risk as to whether it was gradual cline or a very sharp point of demarcation by age that was a cut point; and finally to determine if different clinical outcomes, the four outcomes we looked at, showed the same trend. Next slide please.

The methodology for the most part that I will show you, there was a substantial number of analyses done and I have chosen three to try to illustrate the points I'm making, was simple logistic regression analysis and time-to-event analysis on the same clinical outcomes. Some of the other approaches were listed. Next slide please.

The strategy for the most part was to compare different age cohorts. I'll go into some of the limitations on that data in a moment. The data presented in three different types of analyses. The clinical outcomes were compared between the different age cohorts and trends searched for. The clinical outcomes in consecutive age cohorts were also looked at and sharp differences were searched for. Finally the clinical outcomes in dichotomous age cohorts, that is the population was divided into two parts and comparative differences were looked for. Next slide please.

This is the first set of data. These essentially are the clinical outcomes by the age cohort. I apologize for the slide. There was no other way I could get that much information on a slide. I know most of you in the audience cannot read that. That's why I bolded the top and the bottom, the youngest and the oldest, in red. You will have to take my word for it that there are differences but I will show you the graphs on the same data in just a moment.

If you look then at the youngest members of the group, the first column shows the age group in years. The first group is 0, 1 and 2 years of age. The second group is 3, 4 and 5 years of age, etc. Look at the clinical outcomes. Then compare them to the bottom which 18, 19 and 20 years of age and 21 on up. One can see that the poorer clinical outcomes are found in each of these four. That is 82 plus percent ANC engraftment as compared to the older group there were about 49 percent engraftment and so on.

The next thing we did besides the inspection and comparison of the upper and lower age groups is to look at the trend. Once again I know you can't see this but it appears to be in at least three of these four clinical outcomes a gradual change. I'll show the graphs in just a moment. Next slide please.

This is the same data. This is neutrophil engraftments starting in the youngest age group and going to the oldest age group bearing from 80 percent success rate to roughly a little over 40 percent success rate. Next slide please.

The same is true and perhaps even more dramatically in platelin engraftment where if you look hard enough you can imagine that there's a sharp cut point somewhere around the middle of that range. I don't want to make any claims for statistical significance to this except to say that it would seem apparent that the increase in risk from the youngest to the oldest is quite clear. Next slide please.

This is disease-free survival. Once again we see the increased risk from one to the other age cohort. Next slide please.

Acute graft-versus-host disease (aGVHD). Here the negative clinical outcome would be an increase in the number of patients with grade three and four. This is also evident but the pattern is clearly less marked here and very hard to interpret. Next slide please.

The second approach that we took was to look at consecutive age cohorts. If you look at the bottom of the slide, on the lower left side of each of the panels, we have the age cohorts which were compared. On the upper left panel to the left, it's age cohorts age zero through two compared to three through five. It's a three year age cohort. If we look along all of the panels we get comparisons of consecutive age groups looking for a change.

The most marked differences are in the first panel that is the ages zero through two, three to five. In the middle ages nine through 11 compared to 12 through 14. Then in the last one. The first two that I showed you are statistically significant. The last is not mainly because the numbers in this comparison are much smaller. Next slide please.

Finally, a third approach was to divide the population in half and to compare the two paths. In this case, we looked at the ages zero to 12 and ages 12 onward. The ages zero to 12 are here in red. The ages of 12 onward are in green. The difference between these dichotomous age cohorts is significant. The figures are once again hard to read but they are significantly different. The next slide please.

These are the same data from this approach. In addition to doing a Kaplan-Meier proportion and percent of these two dichotomous cohorts, and you see the figures here, we also did odds ratio. This is the technique frequently used in oncology. The odds ratio above two are generally considered significant.

As you can see the odds ratios are high with the exception of acute graft-versus-host disease which because it has a reversal that is a higher number here would be negative outcome is probably less than half. We did not invert these. Next slide please.

To finish up, we were concerned with whether comparing age cohorts was a fair tactic. That is whether the cohorts of age even from the same study might have different characteristics, whether they were balanced. We looked for a number of features. Here is a comparison between the dichotomous cohorts for the number of mismatches. There are differences just to illustrate one which was present in 44 percent of the younger age dichotomous cohort and present in 27 percent of the older age cohort. Next slide please.

The same comparison was done for diagnostic groups. The term here "acquired" is acquired bone marrow failures. Once again in the younger group, the major differences are the higher number of genetic disorders for which the transplant was performed 29 percent versus eight percent in the older age group and to the consequent lower number of leukemic states for which it was performed. Next slide please.

These data were extremely informative. However they were not definitive. There are a number of potential sources of bias. The retrospective analysis of data was done for data that were not essentially primarily structured for the questions that were posed when the study was designed. There are fewer patients in the older age groups. There are certainly some imbalances which we are still analyzing between the age cohorts. There is anticipated variability because the underlying transplantation protocols and the disease states are widely varied in the transplant field and this would be anticipated. Finally, we know that competing risks will exist for the weight of the subject or the number of total nucleated cells transplanted. If I showed you data that emphasized that, I'm sure it would look very similar.

In conclusion then, next slide please. The risks of negative clinical outcomes increase from young children to young adults with age. The comparative data between ages zero and 20 plus indicate a climb of increasing risk with age, most marked for engraftment, also present for event-free survival, less so for acute graft-versus-host disease. There is no sharp cut point demonstrable but there's a trend to change somewhere midway in this region of the zero to 20 age group and the presence of competing risks are observed. Thank you.

CHAIRMAN SALOMON: Thank you, Stephen and Ellen, for really nice introductory presentations. We are exactly on time. That's vaguely discomforting. I mean there is no triumph in finishing on time if you are on time the whole day. I initially said that we would have no questions or very limited questions after this.

I have to take the Chair's prerogative to ask one though. I think you were good, Stephen, pointing out all the different possible biases at the end of looking at outcome parameters like for example disease-free survival. Given that, one, it's very reasonable to presume that the disease states for very young children and 20 to 29 years old are dramatically different and therefore the risks are different. So I think there will be more discussion of that.

The other point that I would like to ask you about is the incidence of acute graft-versus-host disease. That's a little surprising in that it's essentially the same cord blood that you are getting. Of course, you are getting more total but the same amount per kilo. Really young children shouldn't have any reduced immune response since in organ transplantation we have the most trouble with five years and younger. Even if you throw out some renal vascular issues in the really young children, a lot of it is a very aggressive type of rejection. It's just a little bit surprising to see that kind of a difference. Do you or anyone else have any comments about that?

DR. LITWIN: We focused on the zero to 20 age group. I think there is a lot of literature suggesting that acute graft-versus-host disease is certainly at a much higher risk in an older age population.

But it would seem from this data, and this was only a touch of data essentially speaking, that that increased risk with age is not down in the zero to 20 range but maybe somewhat higher. The transplanters notoriously have been hesitant to transplant patients under 50 and 55 and now it's 60.

I would suggest that, and there are probably people who know much more about this than I do, that that risk continues to increase with older age. Whereas the risks in engraftment and in disease-free survival perhaps may be more focused at the younger age group. I think Dr. Rubinstein probably has some feelings about that. I know he is published on that issue.

DR. KURTZBERG: Two comments. First in response to that, I think it's well known even from marrow and stem cell transplantation that younger patients have a lower risk of acute and chronic GVHD and the younger the patient as well as the younger the donor the lower the risk. Second I wanted to ask Dr. Litwin. Did you do an analysis of chronic GVHD, because that's an important question to ask about cord blood?

DR. LITWIN: It's not finished. We did some analyses. They did not seem to be increased. We are still looking at that. There is data on that within this population group. It does not show much.

DR. KURTZBERG: I think it shows a lower incident compared to bone marrow transplants.

DR. LITWIN: We don't have many people over 20 so it's hard to do.

DR. KURTZBERG: No, I just mean in the entire cohort. Did you look at the incidence of chronic GVHD?

DR. LITWIN: Not yet, no.

DR. HARLAN: You said that if you made similar graphs showing weight versus success rate no matter what that you would imagine that they would look the same. I wonder within any single age group if you tried to break it down by weight as opposed to age to see if it's a weight thing like I suspect, in a dose of stem cell per weight as opposed to age.

DR. LITWIN: We tried to break that down but we found that the data were similar for weight. We did not focus on it obviously but we were not successful in dissociating the two.

CHAIRMAN SALOMON: I guess what I come away with is if it's true, particularly based on Dr. Kurtzberg's comments, that the incidence of acute GVHD follows the age of the recipient is not a direct factor of the use of the cord blood, then one of the questions that we're going to get to at the end of the day is what is the potential use of cord blood in older individuals.

One way of interpreting this data would be you go from about a 17 percent incidence of acute GVHD to about 40 percent in the 20 to 29 age cohort. So it would be bad to use cord blood in that group. But if it's really tracking with the recipient age, then this has nothing to do with cord blood. One could then take that data and make a different conclusion. Joanne.

DR. KURTZBERG: The comment that GVHD does track with age is true of any stem cell source, but you have to consider that these are mismatched transplants and to compare the incidence of GVHD as it would be to other adult stem cell sources. In that case, it is lower in both the younger children and the adults as compared to more traditional stem cell sources. So there are two points. GVHD does relate to age, but there's a difference here as you compare it to other stem cell sources.

CHAIRMAN SALOMON: And that's perfect. That's what I'm trying to set the stage for since that's where we want to go in the afternoon.

DR. HOROWITZ: I would say that the data in bone marrow transplants very much cuts at puberty. Because somewhere in adolescence it is not a continuous relationship between age and outcome. There is a threshold effect. There are fewer differences between 20 year olds and 40 year olds than there are between 10 year olds and 20 year olds. Many of these data could be reproduced in the population of patients who receive bone marrow transplants.

DR. LITWIN: I didn't mean to imply that there was a gradual change. I really meant to say that we failed to show a sharp cut point. I think that's somewhat different. What the truth may be may be lost in the variability which is inherent in this study and other limitations.

DR. TSIATIS: I have some questions about the analysis. Was the amount of follow-up the same for all the different age groups? That was one question and then I have a follow-up.

DR. LITWIN: Yes.

DR. TSIATIS: They all followed the same amount of time. The second question is it was hard sometimes looking at the graphs because they were small. When I looked at the nine to 11 group in those pair-wise comparisons, the nine to 11 group the survival curve in one comparison looked different than the nine to 11 group in the other comparison. When I actually just plotted the survival at 1,000 days, it looked to me like the biggest cut point was less than three and greater than three rather than 10 or 13 that you used.

DR. LITWIN: Dr. Wang was the statistician who did these studies.

CHAIRMAN SALOMON: Can you come to the microphone please.

DR. WANG: All the survival curves applied the same statistical methods, Kaplan-Meier survival curve. So the data just shows the group 10, 11, 12 is a significant difference from the consecutive age group 13, 14 and 15.

DR. TSIATIS: The question I had was that you had two comparisons with nine, ten and 11 comparing to a lower group and then to a higher group. The survival curve for the nine, ten and 11 in those two analyses looked different to me.

DR. WANG: No, I think that the analysis is applied the same.

DR. TSIATIS: I understand. I'm saying the pictures that were shown looked like one had a survival rate of about 40 percent and the other one looked like it had closer to 20 percent. I'm just wondering about the same cohort in two different curves looked different to me. The only thing is I'm a little concerned about the quality of the analysis.

DR. WANG: Can you show me the ones that you are talking about?

CHAIRMAN SALOMON: What I was going to suggest is ? that's a key point to bring up, Butch. Perhaps to give staff a chance between now and through the break to review the curves and maybe make a response to that when we come back. Would that be appropriate?

DR. LACHENBRUCH: This is Tony Lachenbruch. Butch, perhaps we could meet at the break and we'll settle it, I hope.

DR. WANG: Sorry about that.

CHAIRMAN SALOMON: No, it's okay. I think it's an important function of the group to do a reality check on the way the data looks. I think you need a little bit of time to go over it and give us feedback. Butch, if you would take responsibility for it and just give the group a response to that.

I had one last question. This is the dumb questions chairs get to ask. But if one looks at the clinical outcomes by age cohort in the zero through two group, you have a 58 percent platelet engraftment that then falls in the successive age cohorts. Does it really mean 40 percent of these children never get platelets and are life-long dependent on platelet transfusion? Of course not, so I just don't understand that number.

DR. LITWIN: A lot of them are censored because of events including death. That is a limitation of the data that there is no question that these are risky transplants.

CHAIRMAN SALOMON: I just wanted to correct an impression that someone might get just looking at these numbers that it's only effective in reconstituting the marrow in 60 percent of the kids. I don't think it's that. I think that if the kids survive, I would think that these are relatively effective in 100 percent of the cases. Is that fair? I just want to make sure that impression is correct.

DR. HOROWITZ: I think that's fair that most recipients who survive become platelet transfusion independent but it's a high competing risk environment. So in part with cord blood transplants, because the recovery is slower, patients succumb to the competing risks more frequently before they get platelet recovery, not that they succumb to the competing risk more frequently overall.

CHAIRMAN SALOMON: Steve, did you have a comment?

DR. LITWIN: No.

CHAIRMAN SALOMON: Joanne.

DR. KURTZBERG: Since I take care of several hundred surviving cord blood patients, mostly children but a few adults, I mean I can tell you that surviving patients fully engrafted are not transfusion independent at all and have normal counts. It may take up to even 18 months post-procedure to get to a normal platelet count of 150,000.

MS. LAWTON: I just have one question. Did you look at the two separate cohorts specifically by disease for the patients to see whether there were any differences in outcomes for the same disease in the two different age cohorts?

DR. LITWIN: No, we did not. We thought the numbers representing the different diseases were too small to get us the type of data that we were after. We did not.

DR. HOROWITZ: Related to that, did you make an attempt to identify patients by risk category? Even though there are multiple diseases, you can sort patients into low, intermediate, high risk of failure based on a wealth of data in other transplant setting.

DR. LITWIN: No, we did not.

CHAIRMAN SALOMON: I think that's a great start because what comes out of this first discussion is just how extraordinarily difficult this very important type of analysis is. You have to do this analysis, and yet we point out how difficult it is and then how careful one has to be about the interpretations one makes from it because of this complex and competing risk factor model that's operant here. So I think that's going to be a challenge for us to come to grips with later this afternoon. I think that was a very good beginning.

Now I'm comfortable. We're about almost 15 minutes running late. It's this idea of being on time for my last meeting. It was not just going to work. With that, I would like to introduce at this point John Wagner to make the first of three guest presentations on umbilical cord blood.

DR. WAGNER: I think it's important for me to go back and actually introduce to you some of the background information that really leads us to then have some sources of comparison by which to evaluate cord blood. Let me just start off because I have quite a few slides to show. Next slide.

Obviously when we are going back and thinking about how cord blood is successful or not successful, we have to do it in the context of what we know about bone marrow transplantation in general. Obviously, you have all seen this before that there are a number of obstacles that we are trying to overcome in terms of facilitating transplants for our patients.

One is that there is a restriction on the basis of HLA. There's a long search interval. Many patients will enter at either a higher risk category or will die while waiting for a transplant. For those that do have a transplant donor, there is still a high risk of non-relapse mortality as a result of graft-versus-host disease and opportunistic infection. There is poor survival in adults in most diseases with the exception of CML in chronic phase. When we are thinking about the success or failure of cord blood, we have to keep these parts in mind. Next slide.

The National Merit Donor Program is trying to overcome some of these obstacles by improving the level of HLA typing, by initiating a new study in terms of looking at the ultra-urgent searches and seeing if we can find donors more quickly, and that's been an overwhelming success. There's been a focus on research in non-relapse mortality. Obviously the reason why we are here today is that one of the things that has occurred as a result of these limitations with bone marrow transplants is we are exploring alternative stem cell sources. Next slide.

The first thing we need to do is we need to evaluate what are the risk factors that will predict outcome after bone marrow transplantation. Before I go on to the next slide, I would like to show one transparency because it relates to the conversations we just had. Although we certainly could provide you with a lot more information, let me first talk to you about survival.

We actually could do this for all factors whether it be graft-versus-host disease, engraftment or survival. Actually to make a point that was addressed a few minutes ago, is what do you expect in terms of engraftment? Dr. Litwin, the one thing that you said that was a bit misleading although not incorrect was that it depends upon how you evaluate engraftment. Now most of the reports today would record it as a cumulative incidence. There you would expect a graft failure rate with unmanipulated bone marrow of 11 percent, not one to five percent. So you have to have that in the context of what you expect with cord blood transplant. That's for all ages put together.

Now in terms of survival, we do a Kaplan-Meier probabilities, and I'm going to show the multi-variate in a second. However, this is a critical piece of information in which to evaluate the success of cord blood. As you can see here, there is a profound effect of age on the overall outcome. This is now mixing patients with all different types of malignant disease.

As reported by the National Merit Donor Program this was just done as of this past week, you can see here that there's a tremendous impact of age on the outcome in this analysis where you see those under the ages of 14 having the best outcome. I would suggest to you that there is a break point between 14 and 15 in that younger patients doing better than the older patients.

Then when you see between 15 and 45 there is not a tremendous difference at least in the way we have divided the patients up in survival. Then once again in those are over the age of 46, the outcome is inferior. This is mixing patients with CML as well as acute leukemia. CML patients will do better than those with acute leukemia.

This is also mixing patients with a variety of stages of disease at the time of transplant. If you would like, I have all the data here to be able to show you at the break perhaps which were the other variables that impacted on outcome.

We can turn this off and now go back to the results of the multi-variate analysis that was performed on the same dataset by the National Merit Donor Program. What you can do is see the relative risk in terms of age and its impact on survival. With the relative risk assigned to the younger age group of one, you can see with each increment we find that there is a significantly higher risk of death as a result of age alone.

You also see that when you look at the effective diagnosis, and although I didn't put it all out and I can show you the actual data, is if that you assign a relative risk of one to patients who have CML in chronic phase, every other disease and disease stage has a significantly worse outcome compared to CML in chronic phase. When you look at the impact of Karnofsky status and if you assign it a relative risk of one to those with the worst Karnofsky performance status you can see that those with the better Karnofsky performance will do better.

This is the result of a multi-variate analysis demonstrating that there are factors that we have to keep in mind: age, diagnosis, Karnofsky performance as well as HLA and CMV. We have to do this exact same analysis when we are looking at cord blood because what we are trying to do is not to compare really young patients to old patients, because old patients will always do worst regardless of the stem cell source. What we have to do is we have to compare the results of cord blood with bone marrow transplantations because bone marrow transplants are the standard of care, are the gold Standard by which we compare everything else. Next slide.

So graphically in this older analysis where we compared ?this was now the analysis of 5,000 and some bone marrow transplant patients performed several years ago, what you can see here that there is a tremendous impact of HLA mismatch and outcome. The more closely matched patients are on the top two lines. The more mismatched patients are on the lower two lines. This was done to compare T-cell depletion versus unmanipulated marrow. You can see the overall comparisons are really virtually identical between the two cohorts. Mismatch itself has a negative impact regardless of the GVHD prophylaxis. Next slide.

Then perhaps more relevant to the current discussion is that once again you see when you just use the break point of 18 above or below, you can see that there is an impact of age. Also patients don't do any differently comparing their GVHD prophylactic regimens if you segregate them on the basis of being older or younger. Next slide.

This is a result that I'm just going to show you, the multi-variate analysis of our randomized trial that was performed by the NHLBI and the analysis was just completed. I presented this at ASH.

Once again you can see that these are the risk factors that we have to look at in terms of predicting the ultimate outcome of transplant. We have to perform this exact same analysis when we are looking at cord blood again, not comparing younger versus older but comparing what results we have with the cord blood versus with marrow. So these are risk factors that we know exist. Next slide.

As a result of these outcomes, we think that there's been a variety of us, several of us in this room, that have looked at a variety of new directions in trying to improve the outcome of transplantation. As I mentioned, we are trying to shorten the search time. We're developing new preparative therapies. We're trying to figure out new ways of speeding immune recovery but we are looking at new stem cell sources.

The reason why we have chosen cord blood is because we have data now that suggest that there is less graft-versus-host disease. Yes, there is slower engraftment. There appears to be less HLA restriction and it's immediately available. So when we are talking about the outcomes, remember we are now comparing outcomes in our minds to what we expect with bone marrow transplants. Next slide.

Let me now review with you our past experience at the University of Minnesota. The reason for this analysis is specifically to address the issue of CD34. It will be self-explanatory as we go through. Next slide.

For purpose of this analysis, HLA typing is A and B serologic level and DRB-1 at high resolution. Next slide.

In this patient population there were 102 patients. This includes both adult and pediatric patients but the median age was 7.4 years. Obviously it's predominantly a pediatric population. Although there were approximately 25 patients that were in the adult age range within this group of patients. These are patients that only received a full prep. Next slide.

The median cell dose is 3.1 times 10⁷ nucleated cells per kilogram. Again to get to the issue, is CD34 analysis important, we now have looked at the infused CD34 cell dose at the time of thawing the cord blood unit. We then did a CD34 analysis so that we know exactly what went into the patient. This is not the data that was provided by the cord blood banks.

We similarly looked at the CD3 cell dose that was infused. You can see here that the incidence of engraftment is 88 percent. This is exactly the same incidence of engraftment which you would predict with unmanipulated bone marrow which is 89 percent. The median time to engraftment is 23 days as compared to 21 days with methotrexate cyclosporine we see in the unrelated marrow transplant setting. Next slide.

When we looked at the impact of age on ultimate outcome, you find that there may be a separation based on the older age patients compared to younger age patients, but there is nothing significantly different between these two populations. Next slide.

We also found no difference between those that received more matched grafts as compared to those receiving more mismatched grafts. What I should point out to you is that when we did this analysis, we looked at HLA looking at the graft rejection vector. Though I didn't show you the actual proportion of patients on this slide, which I probably should have done, approximately 45 percent of the patients received a two antigen mismatched graft. Approximately 40 percent received a one antigen mismatched graft. There is only a small proportion that received a matched graft in this analysis. Next slide.

However where we started seeing a difference is now looking at cell dose. Once again, it's not age specifically. It's related to the cell dose of the graft that goes in. Although this did not quite reach statistical significance, you do see a beginning of a separation in terms of the impact of nucleated cell dose on the incidence of engraftment and the speed of recovery.

This being the incidence and then ultimately the speed as you are comparing the graft what would happen at 50 percent incidence and what is the time to recovery. You can see there is a difference. Those receiving higher cell numbers have a more rapid recovery. Next slide.

Where we started seeing a dramatic difference in outcome and predicting the time to ultimate engraftment, you can see that CD34 has a very linear relationship with the time to recovery. I would suggest to you that this shows you the principle that CD34 is extraordinarily important in predicting outcome. However there is a problem with this analysis. That is that CD34 analysis in my institution may not be the same result as another institution. We'll get back to that in a second. Next slide.

Perhaps it's more easy to see the impact of CD34 on the outcome of engraftment when you look at this slide where the patients were divided into quartiles. Those receiving the lowest cell dose of 1.7 CD34 positive cells per kilogram, that is with a factor of 10⁵, you can see that this group of patient does exceedingly badly. These patients have a very delayed recovery with a median time of recovery of 32 days in comparison to 18 days in the highest cell dose group. The ultimate engraftment is only 70 percent. The incidence of graft failure then is 30 percent. I would suggest to you that this is an unacceptable outcome. Next slide.

Then we look at the impact of growth factor. The reasons why I'm highlighting are these are factors that actually fell out in the multi-variate analysis. You can see here that in this uni-variate presentation those receiving growth factor did better than those not receiving growth factor. Next slide.

As you can see in the multi-variate analysis, HLA matching, predominately we are comparing one antigen mismatch versus two antigen mismatch. It's misleading to believe that we are really comparing to the six-of-sixes, and perhaps Pablo Rubinstein will address this issue. So all I can tell you is that our data would suggest to us in our limited numbers of patients that there is no difference in neutrophil engraftment comparing one antigen versus two antigen mismatch.

There could be something very different in the six-of-six matches but I can't address that because we had too few at our institution. However when you do look at the impact of CD34 cell dose, you find that there is an extraordinary impact. The higher the cell dose could better the outcome. Next slide.

There appears to be a trend, although not quite significant in the multiple regression analysis, that growth factor seems to be helpful in predicting a better outcome in terms of engraftment. Next slide.

Now the second question that a number of you were asking beforehand is what is the impact of cord blood transplant on graft-versus-host disease. To put it in the context of what you expect, any pediatric population receiving an unrelated bone marrow, we would anticipate at our institution to have overall probability of grade three to four GVHD of approximately 20 percent. That's four patients receiving one antigen mismatch bone marrow transplant.

When you look at the impact of cord blood on GVHD, keep in mind that looking at the GVHD vector, approximately half the patients received a two or three antigen mismatch graft. Yet the ultimate incidence of grade three to four GVHD is only 10 percent. So it would suggest that the risk of GVHD is indeed lower. Next slide.

What is the impact of age? Now let me switch back. This is now grade two to four, not three to four as I was showing you previously. When looking at the risk of age, you find that patients that are older tend to have more graft-versus-host disease but no difference in that age zero to one. Now why is that the case? I can't really answer that. It may be a statistical issue or patient number issue but nonetheless there the p-value is still not significant. Next slide.

Then you look at the impact of HLA disparity. It's nothing that we can detect but once again it may be related to our patient numbers. Remember predominantly, we're comparing one versus two, but there's not even a suggestion of a difference within this cohort. The advantage of looking at a single institution is the fact that whether we grade GVHD the same as they grade it at Seattle or Duke or wherever, it doesn't matter because at least we are doing it consistently within the institution although it may be different between institutions. Next slide.

When we look at the multiple regression analysis, we find that no factor, and though I only highlight these because they are of interest, we found no factor that influenced or predicted the risk of acute graft-versus-host disease. Next slide.

As Joanne Kurtzberg asked, what is the impact upon chronic graft-versus-host disease? To put it in the context of bone marrow transplants and assuming I can show you data and I have that available of what you would expect, you would anticipate for a pediatric age group to be somewhere in the order of about 30 to 40 percent having chronic graft-versus-host disease even in children. Yet we find an incidence of 10 percent. I would suggest to you that at least within our dataset this is quite a significant difference. However soon we need to compare between bone marrow transplant in a more formal way. Next slide.

When you look at the impact by age, we find that adults have a higher risk of chronic graft-versus-host disease as compared to the younger age groups. But practically speaking, it's still quite low compared to what we would have expected with matched bone marrow and these are mostly mismatched unrelated cord bloods. Next slide.

We find no impact of HLA once again. Although again numbers may be prohibitive in doing this analysis. Next slide.

Importantly I think that the biggest obstacle to successful bone marrow transplants is non- relapsed mortality. The question is will cord blood reduce this risk. When we did the analysis of looking at treatment-related mortality, the first question was whether or not HLA mismatching had a negative impact. I would predict that yes there will be although not quite significant in this analysis.

With time, I'm sure we're going to show that a two antigen mismatch has a higher treatment-related mortality as compared to the one antigen mismatch group. Again remember there are too few in the HLA matched group to really have anything of statistical significance in this dataset. Next slide.

We find that age has a tremendous impact at least in this uni-variate analysis. Age has an independent risk in terms of its impact upon treatment related mortality. Next slide.

What I would suggest to you is that it certainly is not just being older. It's also having a lower cell dose. Although not all the patients receiving the lower cell dose were adults, and in fact half of them were children, you can find here that it's really cell dose that impacts upon the outcome and those receiving the higher cell dose have a better outcome with actually quite respectable, perhaps even lower than what we would expect in general, of an incidence of transplant-related mortality. It's on the order of 20 to 30 percent. However, when you get to this lower cell dose, there you have an unacceptable risk of treatment-related mortality. Next slide.

Survival by age is shown here. It's actually not a surprise compared to what the slide I just showed you of treatment related mortality. You find that those under the age of zero to one have very different survival, but at least within our dataset the reason why these results are so positive is because they are predominantly non-malignant diseases. So you can't look at these survival curves without having a comparable population. This population here is predominantly non-malignant diseases, whereas these two are predominantly malignancy. That by itself has a tremendous impact in addition to the other factors that we've presented. Next slide.

We find now a separation in terms of matching. Those with two antigen mismatch do have an inferior survival as compared to one antigen mismatches. Actually that has a p-value that is significant. Next slide.

Then you find that the impact of CD34 is obvious here. Those receiving the lowest cell doses have a markedly inferior survival. Next slide.

When we did the multiple regression analysis, several factors fell out as being important. This time HLA match is important. CD34 cell dose is important. Having had graft-versus-host disease is important in predicting the ultimate probability of survival. Next slide.

So now as a clinician, we have this problem of now trying to balance HLA matching and cell dose. How do you put the two together, because both are critical factors in choosing which is the appropriate cord blood unit for our patients needing transplantation? What I would suggest to you is that really what we observed -- there is a couple of points to make. One is that the curves are pretty similar in appearance overall. You find that those with the higher cell doses tend to do better than those receiving the lowest cell dose here.

But what is different is those receiving a two antigen mismatched graft the curves are just pushed down. So rather than having a one year survival of approximately 80 percent in the higher doses, it's 60 percent in the two antigen mismatch grafts. Those receiving a low cell dose in the one antigen mismatch group have an overall one year survival of 40 percent as compared to 11 percent. It tells us that both HLA as well as cell dose are two factors that we have to look at simultaneously in deciding which graft is appropriate.

The other thing that this tells me or at least suggests to me is that if you have a two antigen mismatch graft, part of that negative impact of being more mismatched can be in part overcome by increasing cell dose. Does that make sense? So even though HLA is important and cell dose is important, if you are forced to use an HLA mismatch graft, the higher the cell dose at least will help reduce the negative impact of HLA. Next slide.

Of those that died, what were the causes of death? It comes out just like in bone marrow transplant. Malignancy and infection are the two most common reasons for death in those that died. This is not the proportion of all patients but only for those that died and what were the causes. You get the overall impression that these are the two problems that are to be addressed. In part, the relapse issue is obviously related to patients, this is including both malignancy and non-malignancy patients, and so many of those analyses need to be done in a different way. Next slide.

What we need now is not to compare younger cord blood patients to older cord blood patients but to figure out ways of comparing bone marrow transplants and cord blood transplants, being able to match for those factors that we have already identified as being critical to the outcome after bone marrow transplantation. What we could do, and what's going to be presented this morning, is looking at these retrospective analyses, because obviously there is no perspective randomized trial to compare it to, all we can do is look at institutional matched pair data and we can look at registry data. Some of that will be presented by others, but let me just tell you what I can talk to you about which is our institutional matched pair data. Next slide.

What we did is we did two analyses. One is that we compared all of our cord blood patients to similar patients that had a matched bone marrow donor. We looked at engraftment, acute graft-versus-host disease and survival. Next slide.

What we did is we matched these patients on the basis of age, plus or minus three years. So we took all the cord blood patients and then we found a comparable bone marrow transplant recipient at our institution. We matched for age, their diagnosis and disease stage, and for patients with malignancy we matched them on the basis of their risk group. Again what we are doing is comparing predominantly mismatched core bloods only to patients receiving a matched bone marrow either with methotrexate cyclosporine as their GVHD prophylaxis or T-cell depletion by counterflow elutriation. Next slide.

To make a long story short, and I'll go through these slides pretty quickly, is that we find that there is a negative impact on engraftment by using cord blood, that patients receiving matched marrow will engraft more rapidly but ultimate engraftment is not that different. Next slide.

When we look at graft-versus-host disease, I was actually quite surprised to see that there was no difference in outcome. The advantage of this analysis is the fact that however we grade graft-versus-host disease, and it was actually Dr. Stella Davies that did this blindly, she took all of our patients and then she performed a retrospective grading of all the patients. What you find out is that there was no difference in the ultimate risk of grade three to four graft-versus-host disease. You can see for children predominantly we have a lower risk of graft-versus-host disease. Next slide.

So it was actually in contrast to what I expected before we did this analysis. I thought that cord blood would have a lower risk of graft-versus-host disease. When you look at overall survival, certainly cord blood was no different but it certainly was not inferior to what we expected with bone marrow transplants. Next slide.

Then we repeated the analysis but this time now looking at those receiving T-cell depleted grafts. Again what you find is that cord blood is slow in terms of recovery; however, ultimate engraftment is no different by day 45. Next slide.

You find that there is no difference in risk of graft-versus-host disease as you might have expected in this category. Next slide.

You find that overall survival is virtually superimposable between the two cohorts. Next slide.

I then repeated this analysis but this time another factor was matched between the bone marrow dataset and the cord blood dataset. Next slide.

What's different this time is I also matched on the degree of HLA disparity. So if the cord blood patient was mismatched, so was the bone marrow comparative group. Does that make sense? Otherwise they are all the same. Next slide.

Once again you find the same outcome in terms of incidence of neutrophil recovery. Next slide.

Here you now see a difference in terms of the probability of grade three to four GVHD with cord blood having a lower risk of GVHD as compared to mismatched marrow predominantly. Next slide.

And you find that there is a trend towards worst survival in the marrow recipients receiving predominantly mismatched grafts. So what you have done now is you have taken really the worst patients with bone marrow transplants and comparing those to cord blood. Next slide.

This is probably the more accurate comparative group. We did the same analysis now using T-cell depletion as our comparative group. Next slide.

This latter part data is not been published whereas the first part was published in Blood several years ago. You find no difference in GVHD. Next slide.

But you find a dramatic difference in terms of survival. Next slide.

At least my interpretation of this is that mismatched cord blood is associated with a rate of GVHD that is not too different compared to matched marrow but is superior to mismatched marrow. Umbilical cord blood is at least as good as HLA matched marrow in terms of survival but may be better than mismatched marrow. But remember that this is still predominantly a pediatric population. Next slide.

There is other data that we can look at. I'm going to briefly go over this. That is there is another retrospective analysis performed by Euro-Cord where they compared patients only with acute leukemia and they are all pediatric. They compared them to methotrexate cyclosporine with bone marrow transplants and T-cell depletion predominantly by Campak. Next slide.

Now, what we should point out to you, you find here now looking at the cord blood category there are some major differences. Patients receiving cord blood are predominantly more mismatches compared to bone marrow group. You can find where the mismatches are occurring are really not class one antigens but predominantly class two. Next slide.

When you look at overall recovery now looking at the Kaplan-Meier recovery here they compared 96 percent with methotrexate cyclosporine at 18 days median time to neutrophil of 500. When you compare that to cord blood once again it appears pretty consistent that cord blood is slower. In this analysis by Kaplan-Meier it was suggested indeed the risk of graft failure is higher. Again we need to be looking at cumulative incidence versus Kaplan-Meier results because their results are different. Next slide.

Can you go back one? One thing they bring up is the incidence of platelet recovery. You can see here to address your point. There is no difference in terms of probability of platelet recovery as compared to bone marrow transplants, but the time to get there is significantly delayed. Next slide.

In terms of graft-versus-host disease, you can see here that there is a difference between cord blood and methotrexate cyclosporine where the incidence is statistically lower using cord blood, about as good as T-cell depletion, but it's statistically lower in terms of grade three to four graft-versus-host disease as well to two to four as you can see here. Then the incidence of chronic GVHD is significant as with cord blood as compared to methotrexate cyclosporine. Next slide.

However the one thing that did fall out that was quite intriguing is that the early transplant-related mortality is significantly greater in patients receiving a cord blood transplant. This is driven by the higher number of HLA two antigen mismatch grafts being used with cord blood. So early death is higher, at least in this analysis, in recipients of cord blood. Next slide.

Even though you see the early deaths are different and then you look at those patients after day 100 and looking thereafter in terms of the impact of cord blood on survival, there is no difference. However what I should point out to you is that cord blood never quite makes up for that early death. So therefore that early death rate does have a prolonged impact on ultimate outcome for long term. Next slide.

One other issue is that because of the fact that there is a reduced risk of graft-versus-host disease there was great concern that there would be a higher risk of relapse. However in the IBMTR data analysis and in the Euro-Cord analysis, there is no difference in the risk of relapse between cord blood versus bone marrow at least in acute leukemia. There is not enough patients receiving transplants with cord blood in CML to make any firm conclusion at this point. However you can see that there is a higher risk of relapse in those receiving T-cell depletion. Next slide.

This is the dataset from the IBMTR comparing bone marrow versus cord blood, mixing the diseases together, but there is not even a suggestion that there is an increased risk of relapse, although that was of concern. Next slide.

This was just a summary which I have already gone through. Next slide.

I think that one thing we need to discuss perhaps is how do you interpret the data. But I think before we begin doing that and comparing what we observed at the University of Minnesota in a matched pair analysis, which was limited in patient numbers compared to the registry analysis, we should save that until after the subsequent presentations.

However we have to keep in mind the difficulties with a registry analysis that you have multiple institutions involved, multiple grading schemes for graft-versus-host disease; there is different cell dose requirements and limits between different institutions; and different degrees of HLA disparity between the different cohorts. It makes it difficult to interpret in some instances. Next slide.

In conclusion, unrelated donor bone marrow transplant is an effective treatment for young patients regardless of diagnosis. This is bone marrow transplant. Several factors clearly predict survival after unrelated donor bone marrow transplant, age, disease stage, HLA matched, performance status and CMV sero status. Therefore we have to use these same parameters in which to evaluate cord blood. Next slide.

Survival is significantly worse for each decade of age. However, there seems to be two critical break points in bone marrow transplants at the ages of 10 and 45. Next slide.

In terms of cord blood transplants, what I would say is my conclusion is that cell dose predicts engraftment, rate of hematopoietic recovery and survival regardless of whether you use nucleated cell dose or CD34. However nucleated cell dose does not have the same impact as CD34, but the problem with CD34 is that the way I do the analysis may be different than another institution.

What I have been able to show to you is the proof of principle. CD34 is a critical determinant, but I can't provide you with that absolute number that is applicable to every institution. It is only applicable to the University of Minnesota. So I would say to you that we need CD34 analyses being done but by the banks not by the institutions themselves.

HLA matched is the most significant predictor of survival and recipients with an adequate cell dose. This is not what I would have told you a year ago. I would have said to you it's cell dose. I think now what I would interpret this as is that if you have a cell dose that is adequate, for me greater than 1.7 times 10⁵ CD34 per kilo, then HLA becomes the most critical determinant. The risk of acute and chronic GVHD is low despite HLA disparity. Next slide.

The negative effect of HLA disparity on survival can be ameliorated by increasing the cell dose. This is a practical point in terms of choosing which graft to go after. Next slide.

Now how would I envision regulating unrelated donor cord blood? From my point of view, cell dose needs to be part of that regulation. But cell dose less than 1.5, I would think, is not acceptable for routine use regardless of the age. Mismatch of more than two HLA antigens is not acceptable for routine use regardless of age only because of the fact that we don't have enough data presented thus far, but perhaps Dr. Rubinstein will present more data that would tell us differently. But until I have seen that data with three and greater mismatches, I don't think that we can say that it should be used routinely. Until the method of CD34 quantitation is standardized, a specific dose cannot be stated as universal threshold. Next slide.

Age is a risk factor for survival regardless of stem cell dose. I think that the use of double cord bloods or co-infusion of haploidentical stem cells with cord blood or co-infusion of mesenchymal stem cells with cord blood or other cell populations or enhancing graft and GVLR immune recovery remain investigational or not for routine use, again regardless of age. I don't think that age is the important factor here. It's cell dose and HLA. Next slide.

In terms of activity in adults, don't be misled to think that cord blood transplants are not being done for adults. I polled a variety of cord blood banks and there are mothers that have not yet been presented here.

What you can see here is that there have been more than 1,103 transplants done in adults, and this is a dataset that we need to evaluate before we can make any great recommendations about the impact of cord blood transplantation in adults. Nonetheless, the type of analysis that needs to be done has yet to be done. Next slide. That's it. Thank you.

CHAIRMAN SALOMON: I think we'll do the questions after just because I think these are linked. I would like to invite Nelson Chao to do the next one. The only reason for the deviation from the printed schedule is that Pablo Rubinstein's slides have to be loaded I understand. We will do that during the break. Dr. Chao.

DR. CHAO: Thank you for the invitation to be here. As an adult transplanter looking at some of the kid's data, it makes us dream of getting data like that. What I wanted to present is the results of a prospective multi-institutional NHLBI sponsor trial that was presented at ASH. Next slide please.

The COBLT study is a prospective pilot analysis of unrelated cord blood transplantation in subjects requiring allogenated transplantation. There?s one stratum that was dedicated to adult patients and there were 34 patients enrolled. The primary endpoint of the study was to look at survival at 180 days post transplant. The secondary endpoints including engraftment, GVHD, relapse and long term survival.

The eligibility criteria for these patients were for malignant and non-malignant diseases suitable for allogenated transplantation. The subjects with relapse or refractured disease were eligible for enrollment. We needed a pre-cryo preserved cord blood unit containing 10⁷total nucleated cells. The recipient had to match at least four or greater HLA alleles. Subjects with active CNS disease, Kamofsky performance status of less than 70, primary myelofibrosis were suitable, but related donors were excluded. As we discuss this further later on, please keep this in mind that these were patients who did not have an HLA matched sibling and did not have an HLA matched unrelated donor either. Next slide please.

These are the baseline characteristics. The median age was 34.5 so clearly an adult population of the range of 18 to 55. Male/female was balanced. Ethnicity, three-quarters were Caucasian and 12 percent African-Americans.

Performance status again is important to note. Twenty-six percent had a 100 percent performance status. Twenty-one or 62 percent had performance status of 80 to 90 and 12 percent a performance status of 70. Next slide.

The primary diseases were primarily malignancies, hematological malignancies, 56 percent AML, 26 percent ALL, nine percent CML, one with non-Hodgkins lymphoma, one with myelodysplasia and one with PNH. Next slide.

Because these patients in a sense were in a pilot study, these were very poor risk patients, so of 94 percent of these patients, 32 of them had poor risk disease, 33 subjects with malignancies, 15 had refractory disease and 11 of them were second or greater remission. Again, keep this in mind as you look at this data that these are patients who really didn't have any other options for allogenated transplant and were in fact a very high risk group of patients. Next slide.

The conditioning regimen for these patients consisted of total body radiation, cyclophosphamide in the majority of the patients, 79 percent; and busulfan/melphalan in the other 21 percent. HLA matched were A and B with intermediate matching DRB1 at high resolution. There was only a single patient that was six-out-of-six HLA matched. Ten patients were five-out-of-six and 23 patients were four-out-of-six. Next slide.

These are the graft characteristics. At freezing, the median was 2.3 times 10⁷ per kilo. Infusion in terms of nucleated cells, 1.7 times 10⁷ per kilo of the recipient weight. Next slide.

This is a summary of the neutrophil engraftment to an ANC greater than 500 per cubic millimeter. Of the 34 patients enrolled, 29 were valuable. They had survived to day 14 after cord blood. So four died actually even before the transplant itself and one died on day six. This is a good example of the severity of this patient population in terms of disease status. Twenty-nine were valuable.

Deaths before engraftment were six. Deaths occurred between 15 and 40 days after cord blood engraftment and died before day 42 was a single patient. So 22 engrafted, so survived 42 days or greater. The number of primary graft failures were eight, so 28 percent. Secondary graft failure was a single patient. The time to get to 500 neutrophil was a median of 28.5 days with a range of 13 to 55. The cumulative incidence was 72 percent. Next slide.

This just shows graphically looking at the probability of the 29 patients are valuable to the cumulative incidence of neutrophil engraftment. Next slide.

In acute graft-versus-host disease, there were 13 patients with grade zero, five with grade one, five with grade two, six with grade three and none with grade four. Keep in mind that there was a single six-out-of-six HLA match patient population. Certainly for the adult transplanters, I think this is a remarkable and quite surprising low instance of acute graft-versus-host disease for HLA mismatched unrelated transplants. Next slide.

The primary cause of death is, as one would expect in these patients, infection in 39 percent, relapse in 35 percent. So again, the predominant cause of death was infection and the recurrence of the underlying disease. Seventy percent organ failure and three percent graft failure, three percent acute GVHD and hemorrhage in a patient. Next slide.

Survival, there are four subjects alive between 373 days and 743 days after transplant. The Kaplan-Meier survival estimates, excluding the four patients that died before they actually got the transplant, was 47 percent 100 days and 30 percent at 180 days. Next slide.

This is a Kaplan-Meier curve again excluding the four patients who died. Next slide.

Obviously there are only a very few patients in all this analysis, so its stratification looking at p-values, there is not a whole lot of power here. But if you look at age, degree of matching and cell dosage there really is no significant difference. Next slide.

Again there is no difference by survival by age, although there's a hint that perhaps some of the younger patients do better. There really was not enough patients. Next slide.

So cord blood was associated with acceptable engraftment and risk of graft-versus-host disease in adults. This was a multi-center study. There's a high treatment-related mortality, and relapse rates observed clearly reflect the poor risk subjects enrolled. These are all patients with fairly advanced leukemias. The role of cord blood will require additional studies and subjects with standard risk factors for transplantation. Next slide.

I just want to spend a couple of minutes asking whether we could do better. We clearly can do better I think if we start with a better patient population. A lot of the morbidity and the early deaths were related to the type of patients who enrolled.

There is a lot of research going on both in the pediatric group and the adult groups looking at ways to make cord blood work better. We could combine units for example. We could think of ex vivo expansion. A lot of the problems with the infectious complications we think are related to the low rate and the low pace of immune recovery. There are other things that one could think of such as IL-7 or LIF or adoptive T-cell transfer and so, as I mentioned earlier, better candidates. Next slide.

We actually have been looking at the idea whether one could use a non-ablative regimen in umbilical cord blood. A lot of the morbidity that we've seen in our adult patients had been related to the toxicity of the preparatory regimen and then the prolonged delay in terms of myeloid recovery. We have been exploring a regimen with fludarabine cyclophosphamide in ATG. Again these are non-myeloablative doses so the fludarabine dose and cyclophosphamide dose are taken from patients with CLL who can get this kind of a regimen monthly for over a year and the GVHD prophylaxis with cyclosporine and methoprednisone. Next slide.

The rationale was that we could actually get a lot of these cells into peripheral blood by mobilization; that anti-lymphocyte drugs can be used without too much toxicity; and large doses of hematopoietic precursors can lead to chimera. So that was the rationale in terms of using peripheral blood in these non-ablative transplants. Next slide.

You have a patient here with A type with leukemia cells which has this type of a pattern. You have donor cells and then you end up with an intermediate stage with a mixed chimera. Then in the peripheral blood in the bone marrow patients, we can actually given the donor leukocyte infusions and actually convert them over to complete chimera. Cord blood obviously you are really waiting for the cord blood itself to take over. Next slide.

In terms of engraftment, there's clearly a balance between a lot of these factors that we can control such as stem cell dose and T-cell dose, as well as in the host there is a degree of immuno-suppression you need to prepare them for and then the post transplant factors such as GVHD prophylaxis. Next slide.

If you look at the non-myeloablative in terms of stem cell peripheral blood, you can see that the conventional way of doing what the ablative regimen gives you about a two week period of aplasia. Whereas in the non-myeloablative, the host cells come down a little slower because the type of chemo is not as intense and the donor cells come back a little sooner so that the native period then is much shorter, and partly because the host cells are recovering as well because we have a mixed chimera state. As the donor cells come back later, you are actually being protected by the host cells that recover from the preparatory regimen. Next slide.

The potential benefits then are lower toxicity, lower anticipated late effect, potentially treat older patients, treat patients with co-morbid conditions. You can do this as an outpatient. The hope is that you have fewer complications and less infections. Next slide.

I'm going to show you some of the preliminary data in cord blood patients. You can see in fact that this does happen. If you look at these four cord blood patients, this patient really had really a transient engraftment and then lost the graft.

You can see at two weeks, rather than waiting for the 26 or 28 days of engraftment, you have a white count now that's already 3.5. You have a platelet count of 29. This is still transfused. You have 6.5 percent donor by chimerism studies in this patient. By three months you can see that the patient is now 100 percent donor and the patient really didn't suffer much from neutropenia.

There is another patient at two weeks who is still neutropenic but actually by six weeks is 100 percent donor ready. The fourth patient here again is 37 percent donor, and this patient by six months becomes 100 percent donor. It's a way of trying to potentially look at using the recipient's own cells to recover partially as the endogenous cord blood grows. Next slide.

Part of the reason for this really is to try to get these patients immune recovery to come back sooner. We had done a previous study comparing adult cord blood, which is shown in these diamonds, with the pediatric recipients shown in these open squares. As you can see the pediatric patients always recover faster.

If you look at absolute lymphocyte count, the adult patients are not recovering to about a year out. Whereas in these orange dots here these non-ablative recipients clearly by a year have recovered their absolute lymphocyte count in the normal range. Next slide.

If you look at naive phenotype for T-cells or memory T-cells, again a very rapid engraftment of these. These are donor cells because we know at these time points that they are 100 percent donor. Next slide.

If you look at spectrotyping looking at all their T-cell receptors, what you see normally is a nice Gaussian distribution with about eight peaks in each of these T-cell receptors. Next slide.

You can see in the immunoscope recovery comparing a patient with myeloablated cord blood versus a non-myeloablated cord blood. that there's a very nice and rapid response of the T-cells. Again as I mentioned, these are all donor T-cells because by chimerism studies at this point they are all 100 percent donor. If you go back to the cord blood and you look what you put in, you get the same type of repertoire. Next slide.

This is just an example of two patients. This is pre-transplant at three months, six months and a year. This is very different from what we see from unrelated cord blood with ablative regimen and much more similar to what we see with like an HLA match sibling. Next slide.

The other thing we can do actually is look at these T-cell receptor incision circles which are a marker for new thymic immigrants. Again these are the pediatric population looking months after transplant. You can see that the kids recover very nicely these TREC-positive T-cells. Whereas many months after transplant, the adult patient population, it's really not until two years that we start to see a few patients with positive TREC signal. It's not until three years when a lot of these patients recover on the adult side. But within these non-ablative transplants, actually we're starting to see some signal actually at the one year mark. Next slide.

We think that the T-cell recovery following cord blood transplantation occurs to the peripheral expansion of these adoptively transferred donor cells. The recipients of these non-myeloablative regimen have a much more rapid recovery in all the cell lines including T-cells and that these non-myeloablative regimens may be one of the ways at least that we can use cord blood in these patients who don't have donors. Thank you.

CHAIRMAN SALOMON: Thank you very much. Can I take some questions for both Drs. Wagner and Chao at this point? Dr. Wagner, if you can step up to the podium that might make things easier. I had one question just to start things for Dr. Wagner, though Dr. Chao may comment as well, just to finish this data gathering piece. You stopped short of telling us exactly why the cord blood transplant patients died. What was it that killed them in the group that didn't survive?

DR. WAGNER: For those patients that died it was predominantly relapse and infection.

CHAIRMAN SALOMON: So just to make a point that will come back in the discussion later, the delay in immune reconstitution, T-cell and neutrophil, that occurs with the cord blood transplants as compared to the bone marrow which was consistent, is it fair then to say that's one of the major Achilles? heels of cord blood transplantation as that specifically increases the risk?

DR. WAGNER: No, so far we have not presented any data this morning that actually compares immune reconstitution between bone marrow versus cord blood. You've not seen any of that. However, what we do is if you look at cord blood patients, at least within our own dataset at University of Minnesota, and you look at those with an adequate cell dose, that is excluding that cohort that has less than 1.7, the risk of transplant-related mortality is very low which is the infection is very low. But there has been no comparison between the two arms of marrow versus cord blood. That's yet not been done.

CHAIRMAN SALOMON: But I thought you showed data that compared the time to ANC reconstitution comparing bone marrow and cord blood to make sure I'm clear on this.

DR. WAGNER: That's true but that's not immune recovery. That's neutrophil recovery. So neutrophil recovery and immune recovery -- at least in my mind I think of it two different ways. One is looking at lymphocytes and one is looking at neutrophils. I didn't present anything on immune recovery. However it is indeed true that neutrophil recovery is indeed delayed.

When you look at those patients, the fact is that some of those patients will die of infection. You are right. But most patients don't die that quickly. They die later on, as do bone marrow transplant patients, because of their immune recovery. Their lymphocyte recovery is markedly delayed. I can't tell you and I don't believe that it's any different between marrow versus cord blood.

So yes, delayed neutrophil recovery occurs up front but most of the patients don't die of infection during that period of time. They die later after they have neutrophil recovery because that's not all that you need. You need the lymphocytes as well.

CHAIRMAN SALOMON: Again just in terms of adjectives, immune recovery, you made the definition that it's lymphocyte. I happen to think of the innate and adaptive immune system and I'm comfortable thinking about neutrophils as being apart of the immune system.

DR. WAGNER: You are not incorrect except that -- so the definition is clear. When I talk about immune recovery, I'm talking about lymphocyte and lymphocyte subset recovery because that's where patients die. It's because of that delayed lymphocyte recovery that causes death. It's not really deaths up front, because with the antibiotics we have now even though a neutrophil recovery is slow, that's not where they die.

CHAIRMAN SALOMON: Dr. Chao, do you want a chance to respond?

DR. CHAO: So in the adult population, I think all of us would be much happier if we got a neutrophil recovery in ten days. Although I would like to say if you look at a different group of patients who we are studying where we're looking at a non-myeloablative regimen using Campak in a matched sib setting looking at a very similar patient population with very advanced disease, that the neutrophil recovery really doesn?t -- we have many patients who recover their neutrophils and go on to die from complications of either fungal infection or other complications which are much more related I think to the disease that they are entering with rather than necessarily just the graft effects.

DR. WAGNER: I just want to make one more comment to really finalize this point. Though I have not shown you that and though I could and it was introduced by Nelson, it is that looking at the non-myeloablative approach at our institution we have done 47 patients using non-myeloablative approach. The median date of recovery after cord blood transplantation is day 11.

However, you have a mixture of both host and cord blood neutrophils. So they have neutrophils very early and actually never become fully neutropenic. Yet the single most important cause of death is infection because of the lymphocyte recovery.

CHAIRMAN SALOMON: Butch, Joanne and then Mahendra.

DR. TSIATIS: This is a question for Dr. Wagner. First of all, I just wanted to say that I really enjoyed his analysis. I think with the kind of observational data that you have, this mass analysis is exactly the right thing to do to make these kind of comparisons.

This question I have is a minor issue. I was intrigued by your multi-variate analysis where you looked at some of the different factors that affect survival. But I was wondering. Did you include a multi-variate analysis where you had age, cell dose and HLA matching together?

DR. WAGNER: They all were. Age and HLA were always forced into the analysis.

DR. TSIATIS: When you forced age with cell dose and HLA matching, did the effect of age get diminished?

DR. WAGNER: As you notice, in terms of cord blood results, age fell out.

DR. TSIATIS: So it did fall out.

DR. WAGNER: Yes.

DR. TSIATIS: So it was no longer significant.

DR. WAGNER: I only showed you those that were significant.

DR. TSIATIS: And it became insignificant when you included cell dose and HLA matter.

DR. WAGNER: It became insignificant in the survival.

DR. TSIATIS: In survival.

DR. WAGNER: However in the transplant-related mortality, age fell out as an independent risk factor but it didn't fall out in survival. I would bet that as we enlarge the number of patients in the dataset, age will fall out as being an independent risk factor as it would with marrow.

DR. TSIATIS: Thank you.

DR. KURTZBERG: I wanted to go back and comment on immune reconstitution because it really is different than neutrophil recovery and it relates to endogenous host thymic function. It's influenced by age strongly. As well, it's influenced by preparative regimen. So patients getting TBI are going to have more injury to the thymus and more delayed immune reconstitution. Patients who are younger have a more active thymus and can contribute more quickly to immune reconstitution.

Much of the data, you?re seeing particularly in the very young kids where they get high cell doses, they also don't get TBI because the disease there transplants -- Remember the younger group had more genetic and non-malignant diseases and most groups do not radiate those patients. There are just many factors contributing to this that are host- or treatment-related rather than graft source related.

CHAIRMAN SALOMON: Following up on that, Dr. Chao's data on TREC analysis would fit that really nicely where there was reconstitution forming quite late.

DR. RAO: I was curious about the HLA mismatch data which fell out as an absolute criteria in the survival data. What was the speculation of what the mismatch didn't seem to affect the graft?

DR. WAGNER: I think that the answer is going back to the issue of infection. I believe that the HLA mismatch has probably an impact upon the speed of immune recovery, at least the way I define immune recovery. That is I think what's happening is that -- this is speculation because we have not proven this -- but I would speculate that with higher levels of HLA mismatch because we did not see an impact on the other outcomes like graft-versus-host disease and engraftment. Again that's really comparing one versus two in our dataset. I bet it's related to delayed immune recovery and that's what is impacting the overall survival and non-relapse mortality.

DR. RAO: Do you think that's just a general consensus in the field?

DR. WAGNER: We can ask the people in here, but I believe that's true because that's clearly true for bone marrow transplants. So why would we believe that it would be different for cord blood?

DR. HOROWITZ: I think definitely it's true in general for transplantations. Recipients of HLA mismatch transplants do worse than recipients of HLA match transplants in general, and it's hard to ascribe it to any one single transplant-related outcome like the acute GVHD, chronic GVHD, engraftment, immune reconstitution. They do worse in everything and it ends up to a significant effect on the bottom line, which is survival. Certainly slower immune reconstitution contributes importantly to it.

CHAIRMAN SALOMON: Dr. Harvath and then Dr. Harlan. And then I think we have someone from the --

DR. HARVATH: John, I wanted to get clarification from you about the data that you presented with adult cord blood transplant experience at Minnesota. Is there any overlap in the patients in your analysis with the patients that were analyzed by FDA? I mean, were any of the same patients analyzed in both?

DR. WAGNER: Yes, I'm sure they were because the dataset that the FDA has is from the New York Blood Center, I believe. That would have incorporated at least a proportion of our patients, although the majority of our patients were done after 1998. So remember the dataset that you are reporting is from 1994 to 1998.

DR. HARVATH: Right.

DR. WAGNER: Which is very different.

DR. HARVATH: Did your conditioning regimen and use of growth factors change, let?s say, from the experience presented and analyzed by FDA with the experience after that time?

DR. WAGNER: That's indeed true and I hadn't thought about that until now. In March 1998, we did not use growth factor. Which we then subsequently after comparing our results with Duke, we realized that growth factor may have an important impact upon outcome, and then we modified our use after March 1998.

DR. HARVATH: I had one other question. This is really whether you are planning an analysis or perhaps have already done it on the importance of the antigen matching and, for example, class one versus class two. Would you believe that's an important consideration that should be done with cord blood transplants?

DR. WAGNER: Absolutely it's an important consideration to be done. I have not yet done it. But really when you consider the numbers of patients required for such an analysis, only the registries could do that at this point. That's a critical piece of information, but there's no point in me doing it.

DR. HARVATH: Mary, are you going to do that?

DR. HOROWITZ: In the study that we did in collaboration with the New York Blood Center comparing bone marrow and cord blood transplants in children, we did not see a difference in disparity of class one versus class two. That study had about 300 cord blood transplants. That's totally inadequate to answer that question.

Trying to answer that question in the National Marrow Donor Program with bone marrow transplants is difficult even with thousands of patients, trying to sort out the effect of one locus versus another. I would say that we haven't found it, but that's not a definitive statement that it doesn't exist.

CHAIRMAN SALOMON: Dr. Harlan and then I think Dr. Rubinstein.

DR. HARLAN: I'm going to ask that hard question, maybe a stupid one because I want you to speculate biologically what you think is happening based on three different observations that you have shown us. One is the CD34 stem cell dose. While it quite clearly from your empirical data is important, it would only be two cell cycles to make up the difference between the low dose and the high dose. 1.7 to 5.4 would be made up with two cell divisions.

The second observation is the relevance of the mismatch. The third observation is the importance of this non-myeloablative regimen. It seems to allow for a quicker recovery. So I'm wondering what you think biologically if there's a grand unifying hypothesis for that. Basically what you want is that stem cell that's marked within the CD34 population to start dividing and to repopulate the bone marrow. I'm asking for you to connect those dots for us if you can. I'm sure you have thought about it a lot.

DR. WAGNER: I think that in terms of the first question about why is it that cord blood all we are asking for it to have one or two cell divisions to go into cell cycle and then give you more rapid recovery. At least I can tell you in culture, because obviously like others we are interested in ex vivo expansion. Even if we can get a two-fold expansion, then we have addressed the issue of cell dose. But number one, remember that CD34 only provides us limited information. It's not truly the stem cell.

Secondly there is something biologically different between cord blood and marrow. It takes significantly longer to get the cord blood cells into cell cycle. So yes, you're right. It only needs several cell divisions but it takes a while to get there. That's an artificial system but still it's different between marrow versus cord blood. It's probably about three times longer for cord blood as compared to marrow. That may in part relate to it.

I think that what we are seeing is that also when you look at the cells on day 21 bone marrows, early days we have actually had people report back that they had leukemia when in fact they were transplanted for non-malignant diseases. If you look at the bone marrow specimens, there is a high proportion of very primitive cells in cord blood recipients. I can actually show you some slides of that a little bit later.

There is something very different. We reported on that to show that the cord blood even when it recovers even morphologically it doesn't differentiate as quickly as do bone marrow cells. It may be related to what you are putting in rather than anything else. What was your second question?

CHAIRMAN SALOMON: Actually he asked about myeloablative related to the non-myeloablative. I wonder if Dr. Chao would want to take that one.

DR. CHAO: There is two bits of data I think that are helpful. We have mouse model for cord blood as well as a model where we used KTLS cells. These are KIT lineage negative Scouwin positive cells but that's assorted cells where we can put ten to 50 cells in a mouse and reconstitute them.

If you look at those animals, there really is a tremendous delay in the time to get to lymphocyte counts which are normal. If you use a congenic strain of 45.1 versus 45.2, you can very nicely see that the donor cells ultimately take over and the majority of the cells you get early on are the residual resecant cells which are radio resistant. That's one piece of information. If you compare that instead of using a cord blood source, what we use is fetal blood day 20 from the animals, if you use a marrow source, those cells seem to come up faster.

I think the other piece of clinical data that's useful to think about is the SCID transplants that Dr. Buckley does without any ablation. If you look at the time it takes to first start to see lymphocytes, it's about 100 days in these SCIDs. There is going to be a delay in that time. I think what we would like to see with these non-ablation transplants is this bridging with the residual host cells which may allow you to overcome some of the lack of responder cells in the host.

DR. RUBINSTEIN: This is in respect to the use of CD34 as a marker. I have a brief comment and a question for Dr. Wagner on this. We have seen and looked at the correlation between CD34 and total nucleated cells and also the correlation between colony-forming cells and total nucleated cells. These correlations are very strong. It seems that total nucleated cells are also perhaps not as close to the true stem cell as CD34 but it is also an portent of engraftment ability.

The question is really whether you have looked at the correlation between the total nucleated cells you recover after thawing the cord blood and the more specific CD34 cell. This is a critical issue for us from the point of view of banking because it relates to our ability to give an estimate that is meaningful for the transplant physician. Clearly the performance of CD34 counting varies quite a bit by technical reasons and other conditions that Dr. Wagner alluded to. This would be extremely useful from a practical point of view.

DR. WAGNER: To answer the question what is the correlation between total nucleated cells and CD34 cells, there is a good correlation between the two. But as you can guess, with any other biological correlations, there is a great spread. For every CD34, it's not a one-to-one correlation.

However when you look at the uni-variate analyses that we have done, total nucleated cell dose does not predict neutrophil recovery. It does not come out as quite significant. It does predict survival but does not predict nucleated cell recovery. It's at a p-value of 0.06. I showed you that slide.

However, if you look at CFUGM and you look at CD34, though I've not shown you the GMs, they are both very profound predictors of the time to neutrophil recovery. Both of them do correlate with total nucleated cell dose, but the other two are much more predictive of what you can expect in terms of days to recovery.

Now CFUGM is also a good predictor. I didn't show that to you because you either do a CD34 or you do a CFUGM. The two are virtually superimposable if I showed you the curves, except that CD34 is rapidly available. CFUGM takes 14 days.

DR. RUBINSTEIN: Yes. Our data concurs exactly on that but there is a slight difference, and that may be related to the numbers. We do show significant effect of total nucleated cell dose on the rapidity of engraftment. Again, we also can show with the multi-variate analyses, colony-forming ability displaces total nucleated cell dose from significance.

DR. WAGNER: Correct. Yes, and I agree with that. It's just that I think that if you have to choose one, there's more variability with total nucleated cell doses compared to CFUGM in terms of predicting outcome. So I think it's a better predictor.

CHAIRMAN SALOMON: We'll get back to this point this afternoon where one of the specific questions is using this.

DR. CREER: We actually have some data that will show exactly that correlation. The difference is really post-processing versus post-thaw. The correlation is significantly different between CD34 and TNC and data obtained post-thaw. That's what most of the studies are coming out from the transplant center perspective that have been correlated with clinical outcomes and not post-processing data coming out of the bank. There is a big issue there. But my question really is -?

CHAIRMAN SALOMON: Before you ask your question, for anyone stepping up to the mike just identify yourself so that the transcriptionist can tell who you are.

DR. CREER: I'm Michael Creer. I'm the Director of the St. Louis Cord Blood Bank. The question I really have is when we looked at outcomes, we focused a lot on survival as the main thing. We haven't really asked the question so much about the quality of that survival.

I'm wondering if anybody has ever looked at the quality of life following survival from a cord blood transplant as opposed to either bone marrow or peripheral blood. It seems like the lower incidence of graft-versus-host disease, the fact that once stable in graft in supported care seems to be diminished, there is no transfusion dependency, that these people are probably much better off in the long term. Is there any evidence to suggest that that?s really indeed the case?

DR. WAGNER: No one has actually formally done that analysis, at least as it has not been reported yet. It's certainly a key question that we need to address. However one point I can make to you is that we clearly demonstrated in the randomized trial with T-cell depletion that we were able to markedly reduce the risk of acute graft-versus-host disease. We found at least nothing that we can tell so far at one year and perhaps at three years that there is any difference in quality of life.

What you did was you changed what the problems were. Rather than just chronic GVHD, you still have opportunistic infection as a barrier. The fact that you get rid of graft-versus-host disease doesn't mean that you'll have improved quality of life, but still the question needs to be asked and it's not yet been done formally as far as I'm aware of.

MR. COELHO: My name is Phil Coelho. I'm the Chairman of ThermoGenesis Corporation. I'm aware of a study also on adults, the use of cord blood adults, done at the University of Tokyo's Medical Center and Transplant Center. It's 46 adults, a little higher than the COBLT study, 29 of which were high risk, 19 low risk. I know the study has been provided to the FDA and some of the other panel members here.

In that case, the survivals were much better. The high [sic] risk patients, all 17 of them, survive at one year. Fifty-seven percent of the high risk patients remain alive at the end of one year. I'm not sure what were the differences in the regimen utilized under the COBLT study and this study, but the results were so different. Perhaps the lower size of the patients helped the cell dose, but I'd be curious for those who are aware of this study if they would have some comments in that regard.

CHAIRMAN SALOMON: Before they answer when you say "much better," you're not comparing to a control group in that study. You're just comparing one study to the COBLT study.

MR. COELHO: They are similar in that respect. There is no control group. They are just simply the use of cord blood on adults, but the percentage of high risk is not too dissimilar. The cell dose was a little higher in Tokyo because the patients were a little smaller, but they were all adults in both studies and neither of which are published.

CHAIRMAN SALOMON: Dr. Wagner, I want Dr. Chao to comment. You had a slide that had Japan on it with the adult cord blood. Was that this study or is it a different study?

DR. WAGNER: No, all I did a couple of weeks ago was just contacted them asking what was the total number of adults that have been transplanted. So all I have is just the total number. I have a much higher number than what's been reported.

CHAIRMAN SALOMON: Dr. Chao, do you want to comment?

DR. CHAO: I'm aware that the study exists although it's not published. It's hard for me to answer because I can't compare what the differences are. One can speculate that the HLA pool is different. Out of these 36 patients, 35 were high risk, so they were all high risk. It's hard to know HLA, GVH prophylaxis, cell dose, disease status.

MR. COELHO: These actually were not good matches and most of them were four-out-of-six. Again there, like here in the United States, most of these patients are waiting for a bone marrow match. So with a homogenous population, you would of course expect the bone marrow.

So those who are unable to get a bone marrow match tended to go to the cord blood. These were predominantly four-of-sixes. I think a few five-of-sixes. No perfect matches. Again others here are aware of this study and could comment more knowledgeably than myself.

CHAIRMAN SALOMON: Dr. Kurtzberg.

DR. KURTZBERG: I actually had the ability to review both studies as a reviewer for one of the major journals. I know one of them is in press. But there are two Japanese studies that look at adults, one that Phil is referring to and a second with 13 adults with myelodysplasia. In both studies, the median cell dose was between two and three times 10⁷per kilo versus in the COBLT study of a number below two.

Otherwise the matching was predominantly four-of-six. There weren't any differences in the preparative regimens or anything that was obvious from what was presented in the data. But the event-free survivals are markedly different. In the MDS paper, it was 85 percent at one year for 13 patients. In the this paper, it was 57 percent with the high risk patients and 100 percent in the low risk patients. This is a small number but the trends are certainly superior to what we saw in COBLT. It may be that cell dose threshold is really breaking somewhere between 1.5 and 2.5 and that's the crux of the matter.

CHAIRMAN SALOMON: Just for the record, you are talking about nucleated cell dose not CD34.

DR. KURTZBERG: Yes, I'm talking about nucleated cell dose of the graft from the cryo-preserved dose that the bank provides.

CHAIRMAN SALOMON: Do you have a specific comment?

DR. LACHENBRUCH: I had one comment and a couple of questions. Comment one refers to all of the papers where many tests were being done. We all are aware that multiple testing is a potential hazard in the sense that we get too many false leads. It might be appropriate to consider some adjustments for the multiple tests.

For Dr. Wagner, in your matching you had UCB versus bone marrow. There might be some reasons for getting one or the other that I'm sure you people here are far more aware of than I that might have an effect on survival. Essentially, you are creating a comparison group in this way.

For Dr. Chao, it appeared that there's a very rapid failure rate in the early period of time followed by something else. This might suggest a mixture of survival distributions which becomes considerably more complicated but certainly not outside the realm of possibility. I would just offer that as a possible suggestion.

DR. WAGNER: You're right in that when you do a retrospective analysis comparing patient populations, even though it's a matched paired analysis there could be different reasons for going to cord blood. If anything because cord blood is the newcomer, typically they would be higher risk patients. If anything, it actually makes their argument in a stronger way, but that's pure speculation. The data is the data and the data shows that there is no significant disadvantage and, if anything, an advantage to bone marrow at least within our dataset. In fact, it is so compelling to our group that we have all but stopped doing bone marrow transplants in the adults because the results are so positive. That's for acute leukemia.

It would be a different story if we were still doing patients with CML because there, bone marrow transplants still is the standard by which not everything else needs to be compared. There needs to be more patients to be done with cord blood to be able to say that it's as effective.

I have to make one comment about the COBLT study which I was a participant of and still am a participant of. There is high risk and then there is really high risk. These patients were the worst risk patients you could ever have chosen for a transplant procedure simply because when we designed the study back in 1998, it was appropriate then to put the high risk adults in the cord blood trial. We have learned a lot since that study has come on based in part we know cell dose effect and we know HLA effect. But those patients, as Nelson pointed out, four died before the transplant even took place. That's extraordinary. That doesn't happen that commonly if you take the typical high risk patient.

Nonetheless you can just say this is your speculation and that I'm biased and that's probably indeed true, but I can tell you -- and in fact at the break I'll make another transparency and show you those same results with bone marrow transplants. You pick these really bad patients. You get a really bad outcome. The outcome with bone marrow transplants is 14 percent.

DR. CHAO: I just have one comment that I want to make. As a transplanter taking care of these adult patients, in many ways you can think of these patients who went on to the cord blood as patients who really didn't have a chance to fail. In other words, if you look for a donor and you have somebody who has his third relapse of leukemia and you know that remission is going to last two or three weeks and you look for an unrelated donor at that point, you know you're not going to find one because it takes a while to find a donor.

Certainly three or four years ago when we started to do this, there was a long delay in trying to get to an unrelated donor transplant. Whereas in the cord blood there really was no delay. You end up actually pulling a larger group of patients who in fact may be relapsing as you are trying to get them on to the study, which I think reflects the poor status of these patients. That early drop-off is in fact the reflection of the type of patients that were picked.

DR. McNIECE: Ian McNiece from Johns Hopkins. Considering that we are talking about how to evaluate products and which product to choose for a transplant, John, on your comparison of the six-of-six marrow to a six-of-six cord blood, if you only had a five-to-six bone marrow available, wouldn't a better comparison be six-to-six cord blood to the outcome of five-to-six marrow? I think the analysis probably would come out in better favor of the cord obviously.

We have to evaluate from a patient perspective if we are going to be choosing products what's actually available for that patient. I would like your comment now.

But in the same light, if we are going to evaluate CD34 levels, what percentage of products actually have that available? If you are going to consider putting that as a criteria, if those data are unavailable, then we need to be very careful in terms of listing that as a minimum requirement.

DR. WAGNER: Let's go to the first one first because that's easier. The first question, that analysis that you are asking for is really a detailed analysis comparing bone marrow versus cord blood. Matched and mismatched bone marrow versus matched and mismatched cord blood has been done. At least it's been done and reported by Mary Horowitz at ASH a year ago.

Pablo, will you presenting that data today? Yes, so I purposely didn't address that detailed analysis because it will be presented shortly. It's a very good question. It's a pivotal question, but just wait a little bit. In terms of your second question, what was it again?

DR. McNIECE: If we are going to list CD34 levels as a criteria for choosing a product, what percentage of products would actually have that available, and is that a reasonable criteria to put down if it's not available to the majority of products, which I presume it's probably not?

DR. WAGNER: There are two parts to that question. One is how many would be excluded from cord blood transplants based on that cell dose. The way I did that analysis, basically it would come out that 75 percent of our patients would have met that higher dose limit. Also, as you know at our institution, that's the pivotal reason why we are now doing double cord blood transplants and why we are doing non-myeloablative therapy.

It's because there is a significant number that would never have the opportunity of cord blood transplant because of this limited cell dose issue. The fact is that can't be a reason to change the criteria. The fact is that if a dose of 1.7 is inadequate, it's inadequate. It doesn't mean that you can then say it would exclude too many people.

DR. McNIECE: But I wouldn't want to put it in there if it's not available for a lot of products because it would exclude too many patients.

DR. WAGNER: But if it's inadequate then it shouldn't be allowed as a routine thing. So yes, if I cut off a cell dose at a certain level and it means that half the patients are excluded, that's appropriate if the outcome is so poor with that low cell dose.

DR. KURTZBERG: John, I don't think you understand the question. A lot of banks don't have that data available on their cryo-preserved units. You are basing your analysis on a --

DR. WAGNER: That's true.

DR. KURTZBERG: So what Ian is saying, is it realistic to make that a criteria if the data isn't available for more than half of the units or maybe greater than that.

DR. WAGNER: Is that what you are asking?

DR. McNIECE: Yes.

DR. WAGNER: That's different.

CHAIRMAN SALOMON: Then what I would like to do again is say that's a key question that we'll get back to this afternoon.

DR. WAGNER: But I did answer that. That was in my last slide. The last slide says that CD34 should not be used.

CHAIRMAN SALOMON: The last question or comment before the break goes to Dr. Mulligan.

DR. MULLIGAN: I was going to harp on this issue one last time. If you say that the most important thing is the cell dose and if you get up to a certain point that's going to be very important, does the correlation between the CD34 count versus absolute count change as you get up higher? You would think it would as you have more and more cells. I guess I'm asking, would you need to have a CD34 test if indeed the legislation was that you had to have a certain total number? That is once you get to a certain number I would think that the CD34 issue would be less.

DR. WAGNER: So what you are saying that if you exceed a certain cell dose, that we can be pretty assured that the CD34 would have been above the minimum. I would bet that's probably correct. I don't know that answer. Maybe Pablo does. That probably is a good point and maybe that would help us. It's only at the lower cell doses that you would need the CD34 perhaps if you wanted to do it that way.

DR. MULLIGAN: This is important if we are weighing the various kinds of advice that we give if we say the most important issue is to have this ample cell dose. Once you are up at this cell dose, it's not really much of an issue. Another pass which are the units that don't have very much where you have to be more careful, I would think, in terms of holding back what's there.

DR. WAGNER: Sure, that's a good point. I have never evaluated that, but it could be done.

DR. McNIECE: Could I just make a comment? If you look at peripheral blood progenitor cells, the data is exactly the opposite. The total cell numbers don't relate to numbers of CD34 cells. You could have a very high cell dose with a low percentage of CD34 and a low absolute number of CD34. So the two are not connected in any way.

There is a driver because the calculation for absolute CD34s is based upon your total white count. So there is some correlation. However you could have a high white count and very low percentage CD34 level in the cord blood as opposed to a similar cell count with a very high CD34 level. So they are not linked.

CHAIRMAN SALOMON: Okay. We're ending this. We are going to a break. These are really very important questions. I'm not trivializing it, but part of the first part here is to set a database and then discuss it. Everyone wants to discuss it. That's great. Save it for the afternoon.

DR. WANG: I would like to take this opportunity to clarify the questions.

CHAIRMAN SALOMON: We will do that right after the break.

DR. WANG: No, there is no problem of the quality of analysis. The problem is because of the scale of leveling.

CHAIRMAN SALOMON: I want everyone to be able to really pay attention to what you are saying. We will do right after the break. Five minutes. I know if I say five minutes no one is going to be back here in five minutes. Off the record.

(Whereupon, the foregoing matter went off the record at 10:40 a.m. and went back on the record at 11:07 a.m.)

CHAIRMAN SALOMON: Back on the record. So what I promised that we would start immediately after the break is a short response to the questions about data presentation that Butch came up with. Butch, do you want to just make an introductory statement and we'll review it?

DR. TSIATIS: Yes, my understanding is -- the concern I had was apparently a mislabeling of two of the graphs. We have cleared that up.

CHAIRMAN SALOMON: Is that it?

DR. LACHENBRUCH: I think that's all that needs to be said. We have a couple of other things that we'll be in direct communications with Butch on.

CHAIRMAN SALOMON: Good. As long we are giving everyone a hard time about mislabeling, I think John Wagner had a mislabeled slide as well because he was talking about zero to one HLA match on one of his slides and I'm sure that wasn't what you intended. What I would like to do is get started again with introducing Dr. Rubinstein.

DR. RUBINSTEIN: Thank you very much for the invitation and opportunity to share with you the overall data of the National Cord Blood Program. These graphs show the distribution of recipients around the world. Basically it's useful to illustrate that the data that I will be talking to you comes from a group of 1317 patients given cord blood from our program. First the purpose of telling you that some of the reports that have been published by other groups, including the Euro-Cord data, include some of this data, so adding them all up may not be an accurate representation of the number of patients. Next slide please.

These are the patient characteristics. They tell about the U.S. versus non-U.S., the gender and these other percentages of each of the different sub-categories in this analysis. These are risk factors and for these we use IBMTR criteria. Next slide please.

First let me show you very briefly the data for engraftment. We are going to use here the cumulative rates with competing risks. Competing risks are presented in this table over the days after the transplant. They are four types receiving a backup transplant having autologous reconstitution, suffering relapse or death. These are the numbers. Next slide.

These slides presented are cumulative data and here give the Cox regression for the analysis. The effect of the cell dose in the timing and probability of reaching ANC 500 by cell dose is given here and dependent on the cell dose. So these are people who received 100 million cells per kilo. These are better than 50 to 100, 25 to 50 and under 25. The lowest has been seven. That's 10⁶ per kilo. These are all significant, taking as a reference the lower group. All the others are significantly different. Next slide.

Here we have the effect of age. You can see that there is a neat stratification also but it's not nearly as good as the previous one. And using as a reference those older than 16, the 12 to 15 group is barely different and six to 11 group is also not significantly worse and the only one that is not significantly better. The only one that is significantly better is the young kids under five. Next slide.

This is understandable. There is a correlation between cell dose and patient weight and therefore age. For the group between birth and maybe 12 years, there is correlation there. That becomes blurred afterwards. The results show that as you grow older the probability of getting a match with 100 million cells becomes essentially zero and vice versa. You cannot see the red behind here but I assure you that only two patients received less than 25 million cells per kilo and were younger than two. Next slide.

This is a multi-variate analysis taking into account these two factors. We left HLA out of this analysis although statistically it is significant because the significance comes all from one group and that's a six-out-of-six matches. There are only 56 of these. It is significantly better though as a group, looks in a uni-variate and that's a group in the multi-variate. The results using as a reference the lowest of those or the younger group shows very clearly that the TNC is the predominant factor. Next slide.

One of the aspects that I think is important to this Committee is how useful is the pre-freeze TNC content to predict the results in the end. In this slide, we analyze it with respect to what the transplant centers can expect to find.

These are the data obtained at the transplant center plotted against what we found for the same cord blood. This data indicates that the correlation is quite respectable. This is despite the intervention of freezing, storage, transportation, thawing and counting. Counting is also important because the methods vary a lot and particularly how to estimate the dead cells. If you look at these there is an apparent concentration of points that are directly on the diagonal. That was a little surprising. Next slide.

This slide will show why that happened. This is a distribution of the percent recoveries. There is a group here at the 100 percent recovery. Cladd said "They are busted." I added the exclamation point. It is a reasonable relationship between what we found and what the people in the transplant center can expect to find. That was reassuring. Next slide.

Now we will talk about survival. I begin with all the patients in our group. All of these patients were given cord blood from our bank. Next slide please.

This is the overall curve of survival. This curve looks hairy. That's because we marked where the patients are censored. There are a number of patients still on the path right here. That can be seen here. The number of patients that remain at different points in the curve are shown and the decrease is quite rapid. That is not only because some patients died but because of all the other courses of censoring, particularly that they have not been observed for more than that time. Next slide.

Here again is a survival and the TNC, total nucleated cell, dose. We see that for the group as a whole, 100 million cells has an overall survival probability of 51 percent at three years. The important message here is, one, that there is a very nice stratification of results with dose but also, number two, that the lowest dose under 25 million total nucleated cells has a probability of survival of less than half of the 100. Next slide.

Patient age has a similar distribution, but here the age of 12 to 15 or 16 and over are overlapping. These are the ages. Overall these curves are bunched together with respect to those we saw for the cell dose. All of these groups are however significantly different in their Cox regression by age compared to the young patients, six years and under. Next slide.

Now HLA. Here we see the expected stratification and it works exactly as it would be expected. Using six-out-of-six matches as a reference, here note we are only talking about 56 patients in this group. The four-out-of-six is different and the three-out-of-six is different, but the five-out-of-six is nearly the same.

So we made another regression using five-out-of-six as the reference point asking the question is four-out-of-six clearly different, and it is. That is the place at which most of the trouble for a transplant physician lies in the decision to use a product of six with more cells or five-out-of-six cells with fewer cells.

In the analysis of HLA, there is one point that I would like to refer to because it was surprising to us to see the results of it. It's the effect of blanks. Blanks are important because when one of the two participants, the donor or recipient, has a blank then the reaction of donor against patient or patient against unit is different.

If you have a five-out-of-six match, one mismatch, and the mismatch in which there is a blank in one of the two, and if the blank is in the unit, then the patient who has an antigen that is not present in the unit will not be able to see these as different. That's a three. It's most likely a homozygote three. It will not react against this. There is no rejection mismatch.

But there is a mismatch if you look from the point of view of graft-versus-host. This is what some people have referred to as the vectors. The graft-versus-host disease vector would have one mismatch. Next slide will show us the effect of these.

These are all five-out-of-six matches. When the mismatch is p-directional, so both donor and patient are heterozygote for the locus, then you have this effect as we saw before, about 40 percent survival. When the mismatch has a blank and the blank is in the direction against graft-versus-host disease and not in the direction of rejection, the survival is 63 percent which is the same as when you have a perfect match. It's actually a little better. Next slide will show you when you have two of these.

These are patients that are four-out-of-six but have two blanks in the unit. There are only four such patients but all of them are alive. There is only one censor here to use. These are the six-out-of-six matches. These are the five-out-of-six as we just saw in the previous slide. These are not due to a patient with lower risk factors. In fact, two of the four are older than 12, and 37 percent of them have a mean count of 37 million cells per kilo for this group.

Despite that all of them are alive. Of course, there is only four and that doesn't mean very much but it's an interesting phenomena. We have used it in multi-variate analysis. You will see these in a little while. Next slide.

These are other factors that are important. One is where the transplant takes place. Here we compare the U.S. patients. There are almost 1,000 of them, with foreign transplant centers about 200. One thing that worries us is the follow-up of this group is not quite as good as the U.S. patients. You can see this straight line here that immediately sparks suspicion. Next slide.

The risk factors. As I explained, we used for this estimation patients with ALL, AML or CML or patients who don't have these diseases. The known ALL, AML or CML patients are used as a reference. That's the black line here. You can see that the low risk patients are almost the same as the non-malignant results. Then the intermediate patients have a slightly lower survival. The advanced patients have a much lower survival basically parallel to the results that we heard earlier this morning. Next slide.

It has been known for many years that even after you correct for the known factors of HLA, matching and so on, Caucasoids tend to do better than other patients. We have the same here. You can see that this is significant. Next slide.

The survival by the existence of a latent viral infection or at least the presence of CMV antibody in the recipient prior to the transplant is significant. The relative risk is not very large from 0.3 but it's extremely clear that there is a significant difference. Next slide.

This is a very important one and this is the ABO match. These are patients where the unit has not been modified. So they are not washed. The patient is not plasmaphoresed to remove antibody and there is essentially no difference between them. Next slide.

So the multi-variate analysis shows that TNC is important. As you can see, the 50 to 100 million is not a significant difference to the 100 million and over. You can see it for the rest. Age is independently important. The risk groups are significantly different. HLA, the five-out-of-six, in the overall analysis is not significant but four-out-of-six is and so on. Next slide.

It has the same data but here we incorporated in the six-out-of-six those that don't have a rejection direction mismatch. If instead of counting matches as six-out-of-six, we counted mismatches, in the rejection direction this would be still zero. By doing that, we see that the five-out-of-six now becomes significantly different for survival. The overall significance is strengthened. Next slide.

We tried to look very briefly at the survival of the older patients in our subset. We arbitrarily split it at 12 years. Next slide.

This shows the effect of TNC dose. You can see here that it is much less evident than in the data as a whole. It's important to realize that we don't have 100 million per kilo. There are only 14 in the group between 50 and 99 million per kilo. So there is not much and that's why we cannot see the difference. But perhaps with more numbers the difference would become significant. We are at the borderline of significance. The Cox regression however shows that the seven to 24 million group is again at the borderline but is significant. Next slide.

These is HLA. Again six-out-of-six appear to be better but the probability is not definitively there. In the Cox regression, we don't see a significant difference even for the three-out-of -six group which is not difficult to understand because of the numbers. There are only 10 six-out-of-six patients.

The next group shows that even this subset of patients, these five-out-of-six matches with a mismatch in the direction of GVHD and not in the direction of rejection, are also doing well. Again the numbers preclude much speeches. Next slide.

Now the risk factor again is quite evident and the advanced patients are doing poorly, again about 50 percent as well as the early patients. This is significant. As you can see the significance comes from the advanced group. Next slide.

The multi-variate analysis for this subset is again showing significant difference for the lower dose group. There HLA is not significant in this. The other factors are as expected. Next slide.

We changed the HLA, and now incorporating those other five patients we have significance for the five-out-of-six and the four-out-of-six, although you can see the relative risk is almost the same. Next slide.

This is a brief overview of the data of IBMTR analysis of comparison between cord blood and bone marrow transplants in patients with ALL, AML and CML and myelodysplasia in patients under 16 years of age. Next slide.

This briefly describes the patients. The male/female ratio as has been seen is exactly the same. There seems to be a little difference in age. As you can see, there are more younger patients in cord blood. The difference is barely significant. At ethnicity, there is more of a difference. CMV antibody and risk. Although in the risk there is some difference but it's not very important. Next slide.

Here at the end of that comparison the biggest difference, of course, comes in the bone marrow and cord blood for HLA where we have many more mismatched transplants. Only six percent in that comparison are full matches in the cord blood and in the cell dose, which doesn't require any comment. Next slide.

This slide looks at the evolution and survival of these patients. Given bone marrow in the black lines, the six-out-of-six matches and the five- out-of-six matches are very similar. The cord blood, the six-out-of-six matches is only 20, probably to survival rate of 68 percent. These are 49 and 46 percent. The cord blood at four-out-of-six or three-out-of-six, which is the largest group in cord blood, has an overall survival of 30 percent.

So that analysis shows that using bone marrow six-out-of-six as a reference, there is no significant difference for the five-out-of-six and the cord blood has a lower risk but is not quite significant. The five-out-of-six and the three-out-of-six have increasing difference, with the three- and the four-out-of-six being significantly different from the bone marrow. But the five-out-of-six is not either in bone marrow or cord blood transplants. Next slide.

This looks at the same data but this time from point of view of cell dose. Here are the patients who received cord blood under 25 million. These are all in the bone marrow which are used as a reference. This is cord blood over 25 million.

Now you see the numbers are different as the majority are here, and this curve is very close to that of the bone marrow. It's lower. It's not significant now but presumably with many more patients it would be significant. The difference, relative risk, is not a major league difference. Next slide.

At five years survival, the effect of the total nucleated cell dose in a restricted population of 12 to 15 years old, which are the oldest patients here, we again have bone marrow and the cord blood with more than 25 million cells. There is no significant difference here. The cord blood under 25 million is significant. Next slide.

Again this is the same data but this time for patients between six and 11 years old. There is absolutely overlapping in most parts of this curve for bone marrow and cord blood with over 25 million. Next slide.

This is the youngest patients. Here the two curves are different. Bone marrow seems to be better. So if we are going to disallow transplants for some age group, I think this would have to be the one, the youngest of all patients. Next slide.

This makes a somewhat artificial grouping here. These are patients who are in the younger group and have the cell dose over 25 million. It looks at HLA matching. So those that have zero or one HLA mismatches are identical to the bone marrow in most of its extension. Whereas those that have two mismatches are clearly different. This is reflected by the Cox regression. Next slide.

In conclusion of all this data, the effect of histo-compatibility and age on patient survival are similar in cord blood and bone marrow transplants. So the trends that are caused by histo-compatibility and age are similar. The survival of cord blood recipients can be improved by using higher cell doses and particularly by better HLA matches.

If you put this evidence from the point of view of regulating or providing guidance for the selection of tissue for a transplant, the consequence is that we could identify safe levels for the selection of these transplants. Age does not appear to be the factor that needs to be considered. Thank you very much.

CHAIRMAN SALOMON: Thank you very much. This is an expanded dataset that obviously matches nicely with that presented this morning. Any questions? The only thing I'm taking off the table is nucleated cell counts until this afternoon.

DR. RAO: You showed the data when you reanalyzed it with the mismatches in cord blood. It looked like now there was a significance between one and two mismatches. Is that because those mismatches that you included in the expanded set were so much better? Remember in the data that you showed with cord blood transplants that the survival was much better when they had two mismatches but they were in the direction of the graft-versus-host, the blanks.

DR. RUBINSTEIN: That's right.

DR. RAO: Then subsequently they included them in that analysis and then it became significant that the mismatch was better. Do you think that's fair given that it was actually even better than the bone marrow complete match?

DR. RUBINSTEIN: No. When you have mismatches that are in that direction, that makes them better. We like those. I understood your question as asking whether in this comparison those four or five patients are included in the group with two mismatches. The answer to that question is yes.

DR. RAO: So when it's included then you conclude that the one mismatch is now significant as a difference.

DR. RUBINSTEIN: Yes.

DR. RAO: Maybe that's not fair. It may not be different because you?ve changed the bar.

DR. RUBINSTEIN: You are correct but just to be accurate in your presentation of the data, traditionally those are two antigen mismatches. So that's why they are included. But because they are so few, they still cannot compensate and make that subset equal to the one mismatch.

CHAIRMAN SALOMON: How do you do HLA typing in these studies? In organ transplantation now, we pretty much have gone over to molecular typing so that the danger of finding these blanks and not being able to distinguish homozygocity for the allele versus an unusual allele is not as big an issue any longer.

DR. RUBINSTEIN: It is and it isn't. It is not a big issue in the sense that we now know that they are homozygote for sure. When you use serologic data, you cannot tell that. You cannot say that. Or when you use low-resolution DNA data. That is correct.

Now this data is done at the serological level of resolution, which for us is serology with all the known splits. All the accepted splits are in. It is confirmed by medium resolution DNA typing for class I. It is at the sequence level of resolution for class II. It is different for the two classes. We have analyzed the data to see where the increased resolution available from the splits in serology whether that makes the data any better. It doesn't. We are currently analyzing the result of testing all of our five-out-of-six matches by sequence analysis of class I. That data is not available yet but it will be very soon.

CHAIRMAN SALOMON: So is it fair to say then that unlike organ transplantation if you look at splits and you take those out as mismatches, you actually call it split match, although traditionally by serological data a split would be conservatively called a mismatch. This is data from Terasaki.

DR. RUBINSTEIN: It is.

CHAIRMAN SALOMON: It is, yes. But if you take the splits and call them matches, you actually do fine.

DR. RUBINSTEIN: It doesn't make the data any worse.

CHAIRMAN SALOMON: Right. Is the same thing true for cord blood transplantation?

DR. RUBINSTEIN: Yes, that is correct.

CHAIRMAN SALOMON: So the splits act just like an organ. They act as a match.

DR. RUBINSTEIN: Exactly. The increased resolution there does not seem to help you to select a better subset of the match.

CHAIRMAN SALOMON: Dr. Wagner, if you want to step to the mike.

DR. WAGNER: While I can't answer that for cord blood transplants, a split is a mismatch in bone marrow transplants.

DR. RUBINSTEIN: Absolutely.

CHAIRMAN SALOMON: That's what I was asking you. That's not been true for organ.

DR. WAGNER: A split is a mismatch.

DR. KURTZBERG: I think the jury is out on cord blood. I think a split is a mismatch in cord blood as well.

CHAIRMAN SALOMON: Just so we are clear about it when you say a split is a mismatch you are saying you get worse results. Obviously we realize scientifically a split is a split. It's always a mismatch. I didn't mean to confuse anyone but you get worse results with a split.

DR. WAGNER: Yes. Everyone has shown that. It's been shown by multiple people.

DR. KURTZBERG: And it's a moving target in terms of what you called a match in 1995 and what you call a match in 2000 and what you call a match now. It's really difficult and I know Pablo has made that analysis. The definitions change as the technology improves and as the selection of donors change as the technology changes.

CHAIRMAN SALOMON: Just so nobody misunderstands why I brought that up, because in organ transplantation the argument has been on the table that one could improve the distribution of organs because of this data showing that splits can be treated like matches. It is a relevant point. If you don't get away with that, then that's too bad. Then you have to deal that.

DR. RUBINSTEIN: The point I think is that for class II we have clear cut and definitive evidence that in class II, increasing the resolution and getting the subtypes helps. The subtypes contribute to the matching. You should indeed not use low resolution class II data. Our data for class I does not show improvement by going to a more refined typing. The analysis is not complete. There may be some procedure. I would rather suspect that it will not contribute very much because of the very high probabilities that we get with the best matches. If it mattered, I don't know how much better it could be.

DR. CREER: Just one other comment on that. This is Dr. Creer again from St. Louis. The specific set of primers that are often used to PCR amplify these HLA target gene regions that we use for typing sometimes do not include primers that will amplify relatively rare types, particularly those likely to occur in specific ethnic groups. So they will appear as blanks but when studied with sequence based allele-level resolution will actually be found to be not homozygous.

Without actual sequencing of every one of those blanks, you're not sure. You can indirectly perhaps answer the question if you look at the ethnic background of those "blanks" and compare that to the distribution by ethnicity of the total group to see if you are selecting out more rare ethnic groups. If that is the case, then perhaps that might be something contributing to it.

I know your numbers are small and the donor base is such that it's difficult to get these low ethnic representation, but the recipients, when they're blanks, that's when the problem I think becomes even more so and they have a rare ethnic group. They will appear in that particular way, the way you are grading your GVH or rejection vector. You might not have that totally correctly identified.

DR. RUBINSTEIN: Yes, I think that is absolutely correct a priori and there is no argument. It is interesting however that in the data itself the results were so good. If there was a residual heterozygocity, and that procedure heterozygocity was important, it should have been worse, at least a little worse, than the full matches.

DR. HOROWITZ: I think it's an interesting phenomenon that deserves further study, but I think we have to be really careful about going off on biological speculation when you have such small numbers of patients in which the phenomenon is observed. I would say the jury is out on whether those four patients with 4/6 matches having 100 percent survival is a real phenomenon or a random event.

I think Dennis Confer is here from NMDP. I think the most recent data looking at class I and looking at high resolution DNA typing and disparities only detectable by that method, as opposed to serologic or low resolution DNA typing, suggests that class I may be as important as class II in terms of the necessity in matching for bone marrow transplantation.

It takes a lot of numbers to be able to look at those things, but I would say based on the fact that with more and more numbers of cord blood we see that everything that's been true about bone marrow is true about cord blood. That HLA matching is important. All the other prognostic factors are important. I would be surprised if as we get more numbers of patients we didn't see the same phenomenon there.

Then there's splits versus high resolution DNA typing. The typing for these pairs are to the level of splits identifiable with current reagents at the time the transplant was done. It is a moving target. The level of resolution of the donor/ recipient designation as matched or mismatched is different over time in all of these registries' studies. I would be very hesitant to say that class I isn't going to matter in cord blood transplants.

DR. RUBINSTEIN: No, class I does matter, but whether there is additional improvement by matching at the more refined splits and more refined DNA subsets, that is where we are not sure yet. I would expect it as you said. But regarding those unusual one directional mismatches, I must say that there is data in bone marrow. There are 12 patients from Seattle that have been reported only in the book, not in a formal publication, in which the same phenomenon occurs.

DR. HOROWITZ: Yes, we've looked at it in our database which is a lot bigger and we don't see that same phenomenon. I don't know either.

DR. RUBINSTEIN: It probably then is not significant.

CHAIRMAN SALOMON: Dueling numbers, that's a theme here. John.

DR. ALLAN: To change the subject, in one of the latter slides you presented, you said that the zero to five age group, the kids who got the bone marrow did better than did the kids who got the cord blood. Do you know why that might be true?

DR. RUBINSTEIN: No, I don't know the answer to that. I think overall the variability of these numbers is quite substantial. I don't think one can expect them to be always precise just purely on a statistical basis.

DR. ALLAN: Could it possibly be cell dose, just that the kids got more cells from the bone marrow?

DR. RUBINSTEIN: It could be that. It could be also that the patients who came to cord blood were a little later. There are a number of other possibilities but the numbers per se don't tell the whole story.

DR. HOROWITZ: They also tended to be the most mismatched. There is a lot of selection biases in selecting a bone marrow transplant versus a cord blood transplant in clinical practice. If you have an infant, since kids seem to tolerate mismatch, tolerate transplant the younger they are, clinicians are less reluctant to take a more mismatched cord blood in that setting.

If you look at the under fives, there the disparity and mismatch was greatest. There are a lot of two and three antigen mismatch in that very young group of children that got cord blood transplants in that study. So part of it is because that was a more mismatched group.

CHAIRMAN SALOMON: From the FDA, any additional comments, because we're going to go to the public comment here. Given that the intention of this morning was to set a base upon which we will discuss the questions in the afternoon, is there anything that you feel hasn't come out here?

DR. LAZARUS: Right, no, I think that all of these presentations and questions are leading very nicely into the discussion that we're going to have in the afternoon FDA session and that is where we will ask them.

DR. HARLAN: Just a quick comment with regard to the apparent disparity between what you are talking about with splits and solid organ and bone marrow transplant. A major difference I think is the problem with cold ischemia in an organ transplant and that there is such a major effect of how long that organ is cold on its ultimate survival that you might be unable to see a difference in the split in that setting, where you could see it in the bone marrow. So it could explain this apparent disparity. Immunologically speaking, the null hypothesis should be, I think, that the closer the match, of course, the better it will do. It just may be hard to see that in the organ transplant setting.

CHAIRMAN SALOMON: It's taken me several years to get used to the idea of calling a six-out-of -six HLA matches an allo antigenic, but of course it's true but it just drives me nuts. I finally can do it without choking. Definitely what I'm saying is that the rules for bone marrow, the bar is much higher for these guys than for organ transplantation. I just have accepted that.

I had one question. Here again the dumb questions chair's get to ask, but just biologically I think about the known mechanism for graft-versus-host disease. We have been talking about HLA as if it's all the same. Certainly class II antigen mismatches and class I antigen mismatches have a whole lot to do with the CD8 and CD4 cell activation pathways. Can we make some comment? When you say that class II antigen mismatches are really important, I'm fine with that as the data shows it. What's the biology? What is the difference between that and a patient who has a problem with a class I mismatch?

DR. RUBINSTEIN: The first point is that we are talking now specifically about the contribution of these very refined splits. We can see a DF4 in 32 different varieties, for example.

CHAIRMAN SALOMON: I'm not talking about splits now. I'm just talking about class II is relevant, right? So if you had a group of children who received cord blood transplants, or adults for that matter, who received class II mismatched transplants and had an impact of that class II mismatching. Then you compared that to a group that had class I mismatching. I'm done with the splits thing. The question is do you know much about the biology? How different is the biology in a class II mismatch problem and a class I mismatch problem in terms of GVH? Do they have more chronic GVH for example? Do they have more infections? Do their T-cells not come up as quickly? It's just an interesting question.

DR. RUBINSTEIN: Yes, it is a very interesting question and you are right. We know that the biological mechanism of activation of T-cells is different in the two groups. I would like to just answer your question with empirical data. When we compared mismatches at class I or class II as we defined them, which is class I defined by serology and class II at highest resolution, both of them are important. Whether they are equally important or not, I suggested by our data it's still not possible to be sure because the numbers are not sufficient to discriminate the slight effect in which class II could be stronger or class I could be stronger. I think to some extent it is difficult to generalize the answer to that because it will depend on what is the balance of factors that would influence the death or survival of these patients. The empirical data so far indicate that if there is a difference between class I and class II as defined, it's not a major concern.

DR. CREER: One comment. Eppy Petersdorf has actually done a detailed structural characteristics of difference in class I antigen mismatches that are serologic and molecular versus molecular only. The molecular only mismatches tend to map predominantly to the antigen binding group, whereas the combination of serological and molecular mismatches map to both the combination of T-cell receptor contact and antigen binding group and are associated with a higher incidence of rejection of grafts than those that are molecular only. That was a very large number of patients. I can't remember how many that was. That was NMPD data but it was many thousands required to actually map those particular sites.

The general consensus that I get without having a lot of data to show it is that class II antigen mismatches are more predictive of graft-versus-host disease as opposed to class I. The data that I've seen suggests that there is no difference. That still seems to be a bias that we see in transplant centers that search us in trying to match cord blood transplants. They believe that class II antigen mismatches are more predictive for graft-versus-host.

CHAIRMAN SALOMON: Certainly experimental data, unless someone wants to correct me, would not agree with that. Class I mismatching is a powerful driving force in transgenic animals at least for GVH and the immunology of it. That's why I asked if anyone knew anything about the biology. Dr. Chao.

DR. CHAO: I think the data would suggest that there is really no difference whether they are mismatched class I or class II. Both would contribute to graft-versus-host disease. The other part that is different in these patients is a lot of work done by Jim Ferrard demonstrating the importance of cytokines and amplification of that in numerous positive graft-versus-host reaction which always happens. Whether that turns into disease or not depends a lot on the preparatory regimen of the patient because each patient can get total body radiation or very high doses of chemotherapy. A lot of this might be number's game, but I think with larger numbers there is no doubt that class I and class II difference probably contribute equally to GVH.

CHAIRMAN SALOMON: And again I think that's really interesting. I would love to know more about the biology event. Is there any difference? It may have the same negative impact of class II mismatch and class I mismatch. It would really be interesting to know. They can't be the same mechanisms though, right?

One would guess that a class II mismatch would lead to a mechanism predominantly being driven through cytokine activation driving then other HLA minors that aren't in the typical class I mismatch or possibly molecular mismatches that aren't defined by the serology. Whereas class I mismatches might not have that same cytokine drive. You just wonder whether those patients might be more or less susceptible to chronic GVH. Maybe this isn't as big a point.

DR. CHAO: Probably not but I think there are major mismatches. In these patients who are mismatched either class I or class II, it's major HLA mismatch plus minors. That major mismatch I think is a direct antigen presentation versus an indirect antigen presentation. I would suspect that is probably similar.

DR. CONFER: Dennis Confer from National Marrow Donor Program. I don't want to trivialize it, but it's all important. Everything is important. Right now, our data suggests that ABC and DRB1 are all important and they are all related to graft-versus-host disease. They are all related to adverse outcomes and we can't pick out any that are more important than the others.

The thing that I remind you is we'd like to simplify it and be able to come up with simple models. Animal models that are simple don't readily translate into the human situation. But remember HLA identical sibling transplant is a great model here because there you have identical regions of the chromosome no. 6 in the donor and the recipient because they have inherited them from the parents. They are genotypically matched all the way along this and they can still get all the way from no graft-versus-host disease to fatal graft-versus-host disease.

The only situation currently that doesn't cause a risk for fatal graft-versus-host disease is identical twin transplant where literally everything is matched. It is all important. Sorting it out is going to be a monumental task.

CHAIRMAN SALOMON: Your comments are well taken and it actually brings up an interesting thing that typically is a theme in these meetings and hasn't come up as much today. That is the potential use or lack of utility of animal models to study this, because whenever you get into these gray areas of considering biologicals, particularly cell transplantation, to the extent that one has confidence in these surrogate models one can advance significantly ahead of clinical trials and ensure safety, etc. What I'm hearing from you is a certain amount of concern that these animal models are predictive. Any other comments on this area?

DR. HARVATH: I wanted to just make a point about and ask Dr. Rubinstein because he's had such vast experience in cord blood banking in the last decade, is we heard your composite data just now and then we heard the earlier data of your first 530 some odd cord blood unit transplants. Could you help us understand how the product processing advancements may have changed, perhaps becoming better, allowing you to get higher volume product, perhaps better yields and so on from that first experience of the first 530, 540 transplants to now your over 1,000 units?

We know from your papers that you are constantly refining the technology. If you could comment about the importance of the kinds of measures that are needed to get at the quality of that product. So if you could help us understand whether there is something that we should think about when we are addressing these FDA questions regarding the quality of the product and the advances you've made from the beginning of your bank to the current time.

DR. RUBINSTEIN: There have been a number of steps in arriving at the current procedure. We don't consider the current procedure to be optimal. There will be further improvements. Most of the improvements in the banking have been in improving of the processing of the cord blood. We now can routinely obtain 92 percent recovery of all the mononuclear cells in our final product. The average is a little above 95 percent, but we set the limit at 92 percent. That was not always the case.

Another factor that we have seen recently or really documented the importance of this factor recently is the intervention of something we call transient warming events in the management of the cord blood. These events are when you expose liquid nitrogen temperature cord blood, so one that has been frozen as minus 196, and now you take that cord blood from one place and put it in the air and put it in another place. This is unavoidable at the time of shipping because there you must transfer the cord blood to another vessel.

But in our earlier experience when we used the standard doors for liquid nitrogen conservation, every time we took the rack out to select a unit the other two, three or four units that are in the same rack get exposed. So each unit would be exposed as many as four times. That does not happen anymore. We have a system where each unit is stored in a unique destination. So you can access that unit and that unit alone. We don't modify any other.

We have been ordinary about using all the units under the level of nitrogen from the start. But we estimate that the average number of transient warming events for our older units was about four. Now it's only one. The significance of these warming events was demonstrated. We reported these at an ISH meeting a couple of years ago.

CHAIRMAN SALOMON: Okay, last comment.

DR. WAGNER: So now that we are discussing banking issues and the importance of specific banking manipulations on the outcome, one thing as you know that I've tried to address with the banks is really the issue of identity of the unit. How do we verify identity of the unit? As you know over the summer, we had a problem where we had a graft that already went through the entire screening process and yet it was the wrong unit from which they said it was, the unit that we thought.

So in fact what happened is that the patient had undergone a preparative therapy and on the day of the infusion we found that the ABO typing was mismatched from what the bank said the ABO typing was. That then led to HLA typing of that unit. We found out that the entire unit, even though the numbers match and the HLA in the computer matched the HLA of the patient, it was not at all the same unit we believed it to be. How do we verify or how do we improve the safety from a transplanter's perspective of the unit identity?

DR. RUBINSTEIN: I think that's an excellent question and one where everybody who has done banking of any type has had problems. This is a real critical problem. We suggested from the beginning, and we have followed these rules, every unit that is collected has a segment that accompanies that unit throughout its life.

Before a transplant is made, the typing has to be repeated on that segment. This process has been facilitated now immensely by the availability of DNA techniques, because the segments of the cord blood units will not always freeze well at the same rate as the content of the unit itself. So the viability of the cells in the segment for serologic procedures cannot be guaranteed. With the DNA techniques, you can now have an absolutely convincing definitive evidence of identity.

DR. WAGNER: So that leaves us as a transplant center where we have units that are said to be HLA matched or suitably matched with the patient. Then we are faced with a dilemma in those cases where there is not an attached segment: should they be used or should they not be used.

In fact, what we've instituted as a result of this incident, which by the way the patient died -- what happened subsequently is that for those units that do not have an attached segment we will have a back-up unit available that will not be as good, but we'll have a back-up. Then we do a thaw the day of transplant and we will do a rapid HLA typing for class I prior to its infusion. The majority of cases it will be correctly identified, but we have to have that strategy now in place now recognizing that this is potential complication.

One other issue is related to things that have occurred over the past couple of months, and that is as new infectious disease screens come down the pike that are considered to more sensitive, how do we evaluate older cord blood units that were stored with different methods of infectious disease screening?

CHAIRMAN SALOMON: I don't want to say that you can't answer that but not now. That's a very good question. These are all good questions and I would rather focus them all together in one part this afternoon. Thank you very much, Dr. Rubinstein. I would like to go on at this point to the public hearing before lunch.

I have to say this meeting has the largest number of people speaking in the public hearing part, which is fine. It is certainly part of the whole purpose of these public meetings. It's to encourage people and make sure that we demonstrate free access. The only thing I would say is that there are a lot of people so if we could work together to keep these on time and to the point we'll get through them. Otherwise things could get really bogged down.

I would like to start with Mr. Steve Barsh. I know that you through your hard work brought together quite a number of families with some very well-behaved children I must say. Mr. Barsh, if you would make your introductory comments. I'm going to start with Mr. and Mrs. Ackerman after you.

MR. BARSH: I get to say good afternoon. My name is Steve Barsh. I'm a parent of a child who went through an unrelated cord blood transplant recently. I would like to thank the FDA and BRMAC for the opportunity to present today. I would like to thank you for having the courage to allow experimental medical technologies such as cord blood transplantation.

We first learned a little over two years ago that our son, Spencer, had ALD or adrenoleukodystrophy. Fifty percent of ALD children have a deadly cerebral onset of this rare neurodegenerative disease leading to a loss of function and death by the age of ten. The remaining 50 percent have a severe to deadly form as adults. A parents' worse nightmare, this is the disease behind the film "Lorenzo's Oil."

Spencer was one year old when he was diagnosed. Spencer started repeated MRI studies, and on February 14^th MRI studies, just 12 months ago, changes were seen in the brain stem area indicating a deadly cerebral onset had begun. Cord blood and bone marrow transplants are the only accepted therapies, if you call them accepted, not exactly from an insurance point of view, but viable therapies when an cerebral onset of ALD has begun and is caught in time. We had already been looking for a perfect six-out-of-six bone marrow match and we couldn't find one. Plus we were looking at a minimum of three to six months before we could even get to transplant while Spencer would be deteriorating.

Even if you find someone on the donor list, and this has been brought up this morning, it doesn't mean that you can get to them quickly. It's a heartbreaking and gut wrenching process of calling in potential donors, having them screened, confirming they will go through surgery, etc. You are losing precious time waiting.

ALD, like many metabolic diseases, can move very quickly along its destructive path and deficit changes can sometimes be seen weekly and even daily. This is particularly true for rare metabolic diseases such as Hurler's, metachromatic leukodystrophy, Tay Sach's and others. In metabolic diseases, lessening the time to transplant equal less deficits and usually a better transplant outcome.

Time is enemy number one. In cord blood, the unit has already been typed, checked, prepped and is waiting in a freezer right now. It's felt that 95 percent of the U.S. population needing a stem cell transplant already has an acceptable match in a cord blood bank today.

If it's not already obvious, we decided on a cord blood transplant. Spencer's pre-transplant work-up testing started just four weeks after his bad MRI, only five days after my wife and I made the decision that he should be transplanted. He was a little over two and a half years of age at that time. He fully engrafted, by the way since it's a big issue for you, on day plus 17^th. His cord blood transplant was just 10 months ago.

Today is day plus 314, but who's counting? He's doing extremely well. We are seeing improvements in deficit reversal on a weekly and even daily basis. We feel so lucky when we think about those kids with rare metabolic diseases and disorders whose deficits are instead worsening. This is because they are being told that there is no treatment option, it's too late, or they are deteriorating waiting to find a bone marrow match as they mistakenly think it's the only option.

Four months ago at only six months post transplant, physicians began to see improvements in Spencer's MRI as well as his clinical presentation. It appears that Spencer is actually remyelinating as the earlier progenitor stem cells from the cord blood are differentiating into other cell types in his brain. The MRI and clinical observations have been confirmed both by Duke and a diverse team at Children's Hospital of Philadelphia (CHOP).

The CHOP team included Dr. Peter Phillips, Director of Pediatric Neuro-oncology and Dr. Gihan Tennekoon, Acting Chief, Division of Neurology. I believe that the non-technical word that they've used was "amazing," going on to say that they've never seen anything like this. We usually don't see MRI improve.

Cord blood works. These treatment options must be available to children who have these devastating diseases, including malignant diseases. It's imperative that cord blood banks be licensed so they can reimbursed reasonable, break-even fees, not having to live by a set fee that leaves them begging for donations and funds to keep their doors and freezers open.

Finally I would like to raise the point of informed consent, an issue that seems to be important in the past to this Committee. Before Spencer was transplanted, we received a document from our transplant team that clearly delineated the risks we were taking and what was going to be done was experimental. It was given to us days in advance so we could have ample time to review it and digest what it said.

The transplant team scheduled several review meetings with us and was not satisfied until every single question we had was answered and they felt that we had a solid understanding of the process and the risks. The issues we faced were frightening and we knew that this was a life and death decision. The transplant team helped us understand everything, including that we might very well lose our son.

We looked at like any other decision. What was the risk versus what was the reward? We knew the risk of not intervening, certain death. Based on our understanding, we were willing to take the risk. Interestingly, there were families who we became very close with who lost their children during transplantation. One of those families might actually be speaking today. No one expressed the lack of understanding of the risk they had taken. They felt they did the best thing possible for their child given the few medical options they had.

In closing, it's important to remember that you can get to transplant with cord blood extremely quickly which is very critical in many metabolic diseases. In some of the discussion I heard this morning, it seemed to be something that was overlooked. Comparing six-out-of-six and bone marrow versus cord blood, on a metabolic disease these kids deteriorate quickly and time is of the essence.

Also I would like to highlight that the early progenitor stem cells in Spencer now appear to be differentiating into other cell types that are repairing problems in his brain. Again just "amazing" is the term that we hear again and again.

Public cord blood banks must be licensed so they can operate at break-even not a continual deficit potentially having to close. They shouldn't be wasting precious time and energy scrounging for money so they can save lives. We now have a normal three year old son. His name is Spencer Barsh and cord blood saved his life. Thank you for your time.

CHAIRMAN SALOMON: Thank you, Mr. Barsh. I certainly hope that you can come back here in 10 years and tell us about your normal 13 year old son. Mr. and Mrs. Ackerman. We have Connor. That's the young man. Then after that would be Mr. and Mrs. Niakani. Please be ready to go.

MR. ACKERMAN: Good afternoon. My wife couldn't make it today. I'm Glen Ackerman. This is my son, Connor. Three years ago almost to the day, Connor was diagnosed with AML leukemia, subtype M5, and he had the T9 and 11 abnormalities. The doctors gave him a zero chance of survival without a transplant. We were unable to find a suitable bone marrow match, and without going into the long story we ended up selecting a cord blood. We didn't know much about it at the time, but with some help from some friends we ended up at Duke University. We had a four-of-six match. That was in June. He was diagnosed March 3, 2000. He go the cord blood June and he engrafted 37 days after transplant.

Without going into the entire history, he has done well post-transplant. We have had a few bouts in the hospital. For the most part, he has done well. He's only on right now amoxicillin 250 mg twice a day, basically as a prophylactic. He's in the 3^rd grade. He's an average student, plays soccer and swims on a swim team.

I would just like to say I think you should license cord blood to help out other young children. As a career Naval aviator, I just want to put in a plug for the military here. We have numerous folks at remote locations around the world and the speed at which you can get cord blood transplant is very important to them. We also have a significant minority population in the military and my understanding is it's difficult to match bone marrow. That's another consideration for cord blood.

Finally I've worked a lot because we had numerous bone marrow drives for Connor, all of which obviously were negative. However, I know there's a DOD bone marrow organization with C.W. Bill Young I believe. So I know the DOD is involved with bone marrow and with all the military hospitals around the world and with all the military folks, I think you have a ready source of cord blood out there. With the right funding and licensing you could make it work for other folks. Thank you.

MS. NIAKANI: My name is Faraday Niakani and my husband and my son are not with me today. My son, who is nine and a half years, old was diagnosed with ALL at the age of four. He went through treatment for two and a half years. He had relapse. We couldn't find a bone marrow match. Our only option was to have a cord blood transplant. Time was running out and we went ahead with it.

His first transplant did not work. He had to get a second cord blood transplant which the second one took. It was a success. I believe that we should go ahead and license cord blood because it's vital. I have him. We have many complications, but this is a big challenge that I face. I don't know what else. Thank you.

CHAIRMAN SALOMON: Mr. and Mrs. Ruderman.

MR. RUDERMAN: Hello, my name is Joel Ruderman. This is my wife, Lori. This is our little miracle, Hunter. He is almost four years of age. This is our youngest son, Isaac. Hunter, At 10 months of age -- I still remember the date, March 24, 2000 -- was diagnosed with leukemia. He started a protocol here in Washington, D.C. in which he was just on chemotherapy. He relapsed almost within a month. At about 11 months he had already relapsed from chemotherapy. We were looking for a transplant.

We had spoken to the doctors and the things that they had pointed out to us was he needed to have a transplant immediately. So we searched around and we investigated cord blood transplants, which clearly we knew very little about at the time. We decided to go down to North Carolina to have the transplant done on Hunter.

Everything since has been a success. He engrafted at seven days on July 18 and has just continued to improve ever since. His match was only a four-out-of-six, which from my understanding would not have worked for a normal bone marrow transplant. He has just moved forward. He is obviously a success and the reason why we believe licensing should go forward.

MS. RUDERMAN: It saves children. Thank you.

CHAIRMAN SALOMON: Thank you. I wonder if there's an impact on this on behavior of children, because they are so good. Next would be Ms. Kathleen Kim.

MS. KIM: Hi, my name is Kathleen Kim. My daughter, Emily Kim, is currently eight years old. She was diagnosed with ALL when she was three. She went into remission quickly after diagnosis and went through the three and a half years of chemotherapy and was off treatment for about 11 months before she relapsed again. This time it was CNS relapse. We were told that the best option for her was a bone marrow transplant. Being Asian, we also knew that the possibility of finding a perfect match, much less a suitable match, was pretty slim.

As it turned out we found six potential donors in the U.S. or actually in the world but four of them turned out to be four-of-six matches. The fifth was unavailable. The sixth was a five-out-of-six match at the ERB level or mismatch. It was taking a long time to find a donor despite the 17 bone marrow drives that we ran. We obviously did not find the right match.

My husband and I decided rather than sitting back and hoping that eventually someone would come along and be able to help Emily that we did some research and found out about several of the transplant centers across the country that specifically specialized in cord blood transplants. We took Emily down to Duke last August 2001. She got her transplant in October. While it was not a complication-free transplant, I'm happy to say that the complications were not life threatening.

She is a happy, fairly healthy eight year old who just the other day said to me "Mommy, can I do something?" My response was "no." Her response was "But I don't have cancer anymore." I was thrilled when Dr. Kurtzberg contacted me. Actually she didn't contact me; I contacted her with my five million questions and asked me to speak today because I really had a hard time deciding between a bone marrow transplant and a cord blood transplant.

A five-out-of-six mismatch is not a terrible thing with bone marrow as I've seen today from the statistics. However, as Steve Barsh mentioned, time is of the essence. You don't want to wait for your child to relapse and you don't know how long it's going to take for the transplant to take place. There are so many factors that go into bone marrow transplants. You need the donor to agree. It has to work in their time schedule. They have to be healthy enough to be able to do the procedure. We couldn't wait.

In retrospect, to be honest, I wish we had just jumped right into the cord blood transplant because it would have saved us a lot of time and heartache and frustration and anxiety and so many nights of not sleeping. I have no doubt in my mind that the cord blood transplant saved my daughter's life.

As Steve mentioned, I think it's imperative to license the cord blood because there are many kids particularly Asians who don't have a lot of options. A cord blood transplant may be their only option. Thank you for giving us the opportunity to speak today. Hopefully these parents and our words will touch you and you will take these stories into consideration when you think about licensing the cord blood. Thank you.

CHAIRMAN SALOMON: Mr. Herb Lee.

MR. LEE: Hi. My name is Herb Lee. I had not intended to come here today to say anything. I was just going to stop by and see what's going on here. But as I sat back and I saw some of what was being presented, I felt as though I needed to make some comment.

Our son, Andrew, was diagnosed with leukemia when he was three and a half years old. After two years of chemotherapy, he relapsed. We of course looked at the options and initiated a search for a possible bone marrow transplant. As Mrs. Kim mentioned being Asian the chances are very slim of finding a match. From my understanding just a standard match is maybe one out of 30,000. There are only maybe 30,000 Asians registered as far as I know.

We decided that since we could not find a match for Andrew for bone marrow to continue chemotherapy and wait and see if there is something else that could up. After almost another two years of chemotherapy, Andrew relapsed again.

This time we told that it was leukemia but something called mixed lineage leukemia which is a little bit different. The prognosis was as the doctors put it very poor and when I said what does very poor mean in this case they said he was not going to survive. I guess very poor is not good.

Again we initiated a search for a transplant and the results for bone marrow were zero. As I looked at the presentations as I was sitting back, it looks like comparing bone marrow and cord blood and risks and factors and that kind of thing, I am not a medical person, for us it was very simple. The risk was that he was going to die if we didn't do anything.

The difference between the risks of bone marrow and cord blood well there was no choice there. We had zero for bone marrow. There was some chance with cord blood. As a parent of a seriously ill child, the only thing that really matters is having hope when the doctors tell you that the prognosis is not good. Cord blood gave us hope.

Unfortunately Andrew passed away due to a virus infection but he actually went through the transplant very well. He had a good time out there in some ways I'd say. But it gave us hope as a parent. I think it's given life to a lot of children and a lot of families without cord blood. There are some circumstances to me all the data means nothing because there's only one way to give some families hope. Thank you.

CHAIRMAN SALOMON: I believe that the split now is that these were families with children that had undergone cord blood. If I'm wrong I'm sure someone will correct me. I believe the next three people are adults that have had cord blood transplants. I think that's an interesting distinction here. The first would be Mr. Steven Sprague.

MR. SPRAGUE: Good afternoon, ladies and gentlemen. I'm certainly happy to have this opportunity today. My name is Steve Sprague and after listening to your presentation this morning, I'll introduce myself to you as a lucky duck. Any long term cord blood transplant survivor is happy to be anywhere to be able to do anything.

My real purpose here this afternoon is to try to put a smiley face on some of the numbers and slides and statistics that you saw this morning. I just want to take a few minutes to talk to you in patient-speak.

I know you probably can't tell but I'm an aging baby boomer. But I was already a medical veteran before I got my leukemia diagnosis when I was 47. I was a diabetic. I had survived a heart attack. I had had quadruple by-pass surgery in 1993.

Despite of all that, I was totally unprepared for a battle with cancer. That began in November 1995. For those of you who know leukemia, those were the pre-Glivec days for CMLers like me. In those days, chemo only really stalled what was inevitable for a cure and that was the allusive bone marrow transplant. We refer to it as the healing hell those us who have gone through it.

I was resigned when I got my diagnosis to a slowly moving leukemia but for whatever reason 18 months after my diagnosis I went into blast crisis, the end stage of this disease. That was in May 1997. I found myself suddenly in a real pickle. Six weeks after hospitalization, my oncologist got me into first remission while obviously began marrow donor search. I was an only child so of course they had to find an unrelated matching donor if I was even going to have a transplant to consider as an option.

To make a long story short, as unbelievable as it was to me at the time, I was not one of the lucky ones to find a marrow donor match. The doctor looked me square in the eye and said Mr. Sprague, you enjoy your remission for as long as you can. Get your affairs in order. We'll keep looking and we'll try to figure something else out.

At that point in this story, it was a typically sad cancer tale for adult leukemians. Then a series of events that's even hard for me to understand or to describe even now, I'm convinced that those of us in that situation it's really all about fate. It's about the miracles about modern medicine that you folks deal with on a daily basis. It's about faith. It's certainly about being in the right place at the right time.

Here's what I mean. I was beginning to lose my remission just at the same time that my oncologist, Dr. Andy Peccora, up in New Jersey was planning to start a new clinical trial, first one of its kind in the area of end stage adult CMLers using stem cells from umbilical cord blood. Part of that trial in my case would be to try to expand the stem cells ex vivo using one of the experimental medical devices. This was the Aastrom Replicell System.

Now look at me. I was a 239 pounder back then. That was a whole lot of expanding if this thing was going to work. Equally astonishing to me the New York Cord Blood Center found a perfect six-out-of-six cord blood match for me in a matter of days. So I went into the hospital October 30, the day before Halloween. Magic, miracles happened and 40 days later I was discharged certainly by the grace of God but a lot of other things as well. That was in December 1997. A new working immune system, no leukemia and no hair. Of course I didn't have much when I went in.

Fast forward a little bit and here I am now before you five years, two months, 19 days and I count every one of them, still a 100 percent donor cells. All female chromosomes because my donor was a female infant, not a single filly chromosome and still no hair.

I certainly had a lot of time to think about things as a post transplant patient advocate. I learned a lot of things not only about myself but about life and about death and about perspectives, appreciations and priorities and the most important one that I have come to learn about is hope. Several others have spoken about that word. It's an important issue for those in the transplant community.

My point is simply this. Part of that hope for desperate patients that are seeking transplant, patients like I was once, involves options. Heading down the transplant trail as I don't have to tell anybody here sometimes against overwhelming odds is a risky endeavor even in the best of circumstances. But that first critical step can't be taken without some kind of a stem cell match.

I'm not exactly sure what you as a committee have in mind here but I'm obviously convinced that a cord blood transplant is an option that has to be available to any adult willing to undertake the risk and should not be restricted just to children. For me I just walk around as living proof that cord blood can work successfully for adults and even for very large, very ill middle aged adults. Sometimes as I've discovered the gain ends up well worth the taking of the risk.

I would ask you not to deny this life restoring option to adults who may have no other hope. Had that been the case for me five years ago, I think you all know how that story would have ended.

I would remiss if I didn't take this opportunity to thank everybody in this room for what you do and the way you do it so well and certainly to the FDA for giving the public, the survivor community, an opportunity to be heard and to try to convey to you from a patient perspective what going through transplant is all about. Thank you.

CHAIRMAN SALOMON: Thank you and approximately half of us are sympathetic about the hair. The next speaker would be Ms. Heidi Shaw-Tweten.

MS. SHAW-TWETEN: Hello, I'm Heidi. I'm from Rochester, Minnesota. I just wanted to come here and tell you my story. When I was 10 months old, I was diagnosed with Diamond-Blackfan anemia. With that you don't make any red blood cells. So the only treatment for that would be either to have bone marrow transplant or have blood transfusions. Since I had no family members that were a good match for a transplant and the risk of graft-versus-host with a bone marrow transplant was too great, I had monthly transfusions.

By the time I was 20, I had made a lot of antibodies against the blood so the blood was just breaking down. The quality of my life had really diminished. I had to come in about every two weeks for transfusions and I was still tired most of the time. I had to look for other options.

Six years ago I had an unrelated cord blood transplant that matched four-out-of-six. At that time I only had grade one graft-versus-host disease which is totally resolved. Also I'm off all the antibody anti-rejection meds for about five years and all my counts are normal. Since then, my quality of life had greatly increased. A year after transplant, I was already skiing in Colorado. I have graduated from nursing school and I've worked at a major medical center and been married for the last two and a half years. In June, my husband and I are also planning on adopting also a baby boy.

Without the opportunity of having a cord blood transplant, I probably wouldn't be here talking to you today. So I'm very grateful for the opportunity of the cord blood transplant and I hope others have that option. Thanks.

CHAIRMAN SALOMON: And then Ms. Gayle Serls.

MS. SERLS: I'm last.

CHAIRMAN SALOMON: No, you are actually not last. There are two other people after you.

MS. SERLS: A cord blood transplant has extended my life so now that I have menopause and hot flashes.

CHAIRMAN SALOMON: Approximately half of us would sympathize with you.

MS. SERLS: So that's really a good thing now or otherwise I might not have experienced this. In May 1995 I was at my grandmother's funeral when she died at the age of 98 and had been a very active and alert person. I thought that's going to be me. I'm going to live to be at least 98. At that time, I was 45 years old and I was fit and I was strong. I was confident that my body would carry me on for the next 50 some years.

Then two months later I was diagnosed with acute lymphocytic leukemia. Where did that come from? It was out of nowhere. I had been a bit tired but all of a sudden all of my glands had swollen. So here I was. I went into the hospital and three weeks later I found out I had the Philadelphia chromosome which is a nasty little bugger and was set out to kill me. I knew that I was in big, big trouble.

I have two children and they needed a mother. Let me tell you. My insurance company stepped in and they said you could go to Johns Hopkins and talk with them. I thought that's wonderful. That's where the bone marrow transplant program was born I believe and I felt confident that I would get very good treatment.

They told me I was too old. You are too old to have a bone marrow transplant. I said okay what are we going to do. They offered an autologous transplant and then that would be supplemented with something that would boost my immune system and I forget the name of what they were going to do. While I was encouraged, I knew I would accept my own cells and I thought I could live through this and maybe extend my life and if it doesn't work totally something else will come along that will extend my life further.

At any rate, they wanted me to be in remission. The day that I got on the plane to go to Johns Hopkins to start this procedure, I had a lump right here. The next day I reached Baltimore and was in my room. The next morning I woke up and I looked like I had the mumps. I knew it was bad.

I went to see the doctor on my regular appointment and the first thing I said was the leukemia is back. She just started shaking her head. I'll never forget that. I said what else can we do. Is there peripheral stem cell? I was just trying to think of words. She said there's just not enough time. That was the darkest day. I had hope up until that point and then all of a sudden I had no hope.

I got on the plan and went back to Durham and back to Duke. Peter Jennings and my mother helped save my life. I was in Duke. I was on really heavy or big doses of chemotherapy and for the very first time I was sick. I was really sick. I had colitis and something had happened and my eyeballs felt like they were inflamed and I was on morphine.

My mother called me one night when I was in the hospital and she said I just heard on Peter Jennings that cord blood is being used to help patients with leukemia. I immediately called a friend because I was afraid I would forget it because I was on morphine and it was like minute to minute. She called the American Red Cross and the American Red Cross referred her right back to Duke, the hospital I was in.

Before I left the hospital, Dr. Mary Laughlin came to my room. We talked and I thought we were talking if I could be a candidate but she was talking like we're going to do this. I got back into remission. I got out of the hospital. Two weeks later, I was back in the hospital and I had my cord blood transplant on May 1^st.

Now I'm 46 years old. I don't remember what day the cells came in. I had very little graft-versus-host disease. I actually had a good time in the hospital. I was waited on hand and foot. I had very little complications that I remember. I was out of the hospital in about eight to nine weeks and by the middle of August I was off all medications. Since that point, I have had no side effects, no other medications. Nobody told that there might be complications after you get out of the hospital.

If it wasn't for this cord blood transplant which was a four-out-of-six match, as everyone has said before it's time when time is of the essence this is the thing that you can turn to. This is the thing that will give you hope. When I heard in the hospital that I had another option, I had gone from zero to maybe 10 percent of saving my life. That was the biggest 10 percent that I will ever realize. It didn't matter if it was 90 or 100 percent. At that point, my spirit rallied. My body rallied. I was set to go.

It's a wonderful thing. I'm on the opposite end of the scale. I'm six feet tall with 145 pounds. It worked. I hope that it will stay available to help other people like me regain their life. I'm very thankful to be here. Thank you for letting me speak.

CHAIRMAN SALOMON: Again I want to thank all of you. We'll go on with two more presentations but these will gear back to more traditional format of data presentation. I just want to review five points that I wrote down that I thought came out of the aggregate of this public session. I just wanted to put them out there because perhaps these would be things that we would turn to perhaps this afternoon in terms of appropriately contributing their input into our public discussion.

The first was and in no particular order a mention of licensing/reimbursement issues. There were several very pertinent references to racial distributions specifically for Asians but I think that it would be very appropriate to expand those considerations to African-Americans, Hispanics and Native Americans. I'm sure up here I'm missing a major racial distribution and I apologize to them.

A third issue was speed to the procedure, time to the procedure. That was made specifically at first in the case of metabolic diseases but then echoed in several others with respect to these precious gaps between remissions particularly in the adult diseases. Relative to that, my fourth point was also mentioned which was the impact on these families and also on the patients that I just can?t imagine. As a physician, I can get in touch with exactly the stress of running bone marrow drives and not knowing what's going to happen tomorrow as opposed to the whole concept of an efficient banking procedure.

It was interesting that the mention of the military and just take that a little broader and say what sorts of things could enhance the overall process and the diversity of these pools. That might be something to consider. Lastly something came up at the end about high age limitations for transplantation. That was kind of interesting about the idea that you might be too old. Usually the range is somewhere around 55 or 60 for certain transplants even though that's gray. The possibility that with cord blood transplantation older patients might be candidates that wouldn't normally be for bone marrow. That's something I would like to hear a little bit of comment on later because I wouldn't necessarily have guessed that one. It certainly didn't come up in this morning's discussion. Would anyone add anything to that list?

DR. WEISS: I think the other issue was informed consent. That was brought up.

CHAIRMAN SALOMON: Informed consent is such an big issue that I totally agree to the extent that we probably should never have a discussion without taking a moment to do a reality check on informed consent.

DR. KURTZBERG: I think the last issue is that for people who don't have a bone marrow donor this is the only available donor source they may have. A high proportion will find a donor that gives them an option for transplantation.

DR. TSIATIS: Maybe you know more as a chair but I've sat in a number of these committees and I never know from the questions what the FDA is getting at. Maybe when they read the questions this afternoon, they can tell what the background is a little bit. It sounds like there are some issues with licensure and things that I don't know about and I guess we're asked to discuss it. It would be nice to have more background.

DR. HARLAN: The only other point that I heard from the public comment was the concept not proved but the concept that these cells might allow or differentiate into tissues other than bone marrow or support the differentiation of other tissues into good phenotype.

CHAIRMAN SALOMON: Now there are two last speakers. If we can go to Dr. Michael Creer from the St. Louis Cord Blood Bank and trying to keep these presentations around five minutes if we can.

DR. CREER: Thus far the issue of safety of umbilical cord blood for transplantation has been presented primarily from perspective of the transplant center. I would like to address it from the perspective of a cord blood banker.

There are three major issues that we address in the cord blood bank with regard to safety. They relate to the donor, the biological properties of the product and the way we handle it in the lab. With regard to donor suitability and selection, we have adopted the general guidelines that we use for screening whole blood donors with slightly more restrictions applied to that including serological and nucleic acid based testing for pathogenic viruses, viral exposure in the donors and have now achieved unprecedented level of safety in those products. I don't believe that it's a significant problem in any way, shape or form at this particular point in time.

We are now faced with impractical objective of achieving zero risk which we cannot do. I think the main limitation that we have here is the request by many centers to actually screen patients or screen products for inherited metabolic disorders which we cannot possibly achieve practically. There is one issue there that is a whole other subject to address.

With regard to the biologic properties of the product, we know ultimately that what will determine outcome will be the content of a hematopoietic stem cells/progenitor cells (HSC/HPC). I think at this particular point in time we have to acknowledge that we have no practical way of measuring that. The only things we have available are surrogate markers to currently use surrogate markers or total nucleated cell count and CD34 cell count.

We have seen that the one thing that does distinguish umbilical cord blood from all other transplant products at this time is the fact that there is a dose limited effect. The characterizing dose is therefore the single most important task of the cord blood banker. We currently do that based on measurements of total nucleated cell count and CD34. We know biologically that among these two measurements we expect CD34 to be a much stronger correlate of the actual HPC/HSC content. Our inability to demonstrate that may relate not only to the fact that CD34 is maybe not the perfect marker but perhaps the methods we use to measure it are inadequate and that's the main problem.

I would address how we go about characterizing products and what this might potentially impact us in terms of what we think we know about transplants outcomes and their relationship to products. I've shown here the sequence of events that occur between the time of collection and the time of transplantation. Products are characterized in the cord blood bank at two separate stages, immediately after collection and immediately before controlled rate freezing and putting them to sleep in liquid nitrogen.

These measurements preprocessing are primarily used solely for purpose of quality assurance within the laboratory to optimize processing procedures. The post processing information that we obtain on the product is all that's available to the transplant center for search of cord blood products.

We then put the cord blood product to sleep by a process of controlled rate freezing, leaving it in liquid nitrogen which is the most stressful event on the product that could ever possibly occur during the sequence. Product is then removed from that state and thawed. At that point the product is recharacterized by the transplant center and infused into the patient.

It's appropriate to correlate clinical outcomes with what actually gets into the patient so it's very clear that we need to correlate the clinical outcomes with the information that we obtain after thawing. However the information that we obtain after thawing may not be reliably extrapolated and utilized by the transplant center to select products because they could be very different.

To give you some idea of the potential problem this might create the question is could they potentially be different because the technology does not exist to allow us to characterize them or is it because we have not adopted uniform standards for processing. I think that what we can say right now is that adequate technology exists to provide very precise measurements of total nucleated cell count with very good cord blood of variation, a measure of variability in the measurement.

However CD34 cell counts cannot be accurately determined at this point in time with values that range over a fivefold range from 0.4 as you can see to 2.0 at about a 1.4 percent overall cell level which is comparable to what we generally find in stem cell products. When you go down to about a half of a percent, the range is over fourfold different. That's just instrument variability. These are analyses conducted on the same products and the exact same material by different labs.

In order to get a better idea of what the variability might be within stem cell processing labs, I want to clearly acknowledge the people from the American Red Cross particularly Dr. Gary Moroff and Dr. Rebecca Haley for allowing us to participate in the study designed to assess the variability in laboratory cord blood bank laboratory processing and total nucleated cell count and CD34 characterization.

Utilizing a mean plus or minus 25 percent target for characterizing total nucleated cell count in CD34 and in products that have not been cryopreserved but remain in the liquid phase, you can see that with TNC we can get very good precision and nine out of nine labs agree. However, agreement among labs for CD34 is considerably lower.

If you look at the values obtained after thawing the product, TNC values become quite a bit more variable and this is due again to differences in recovery between labs. CD34 looked pretty good although there were two outliers that had to be rejected just to get that kind of agreement. So there seems to be even on the same product thawed the exact same time by each different individual laboratory. You can see after we do the characterization post thaw, there is more variability in results. Next slide please.

In order to determine what the product characteristics might be that we report to the physicians from the cord blood bank side and what's actually going into the patient, I've taken this data from Dr. Wagner's publication regarding the correlation of TNC with CD34. I've compared it with the TNC versus CD34 correlation that we have obtained in the products processed in our lab between January 2001 and January 2003.

The slope of the lines are different. They both show a very strong correlation between CD34 cell count and TNC. What you can see with the individual products that there is a significant difference. What you get by CD34 measurement in a product tells you something totally different than TNC. Even though individual products show an overall correlation what the actual values in any individual product can be quite different.

When you take the products post thaw what you see is that the variation is even greater. The CD34 cell count in some products can be extremely high. Whereas the nucleated cell dose that you would predict from that would be very low. Based on all the information that we've seen so far, you would have two different very disparate predications about outcome in products based on that kind of information. Next slide please.

In order to determine what kind of a problem this might be overall, I have summarized the post thaw recovery results from 60 different transplant centers that we exported products to and compared them with CD34 recovery. In the blue line and I apologize for the contrast here you can see that the TNC recovery varies between a low of zero which one particular institution reported zero recovery while the CD34 recovery is about 55 percent.

Overall the average is around 70 percent with variations extending from a low of 55 to a high of about 98. However if you look at CD34 recovery, it's very variable from a low of 20 to a high of about 90 percent.

You can see that I've ordered the transplant centers on the bottom to show increasing recoveries. Then you will see a break in the line and a solid at the bottom indicating that almost two-thirds of the laboratories do not even measure or report to us CD34 counts post thaw. Next slide please.

To look at some other potential problems we have with these values I took this information, downloaded it from a website, I won't tell what particular cord bank claims this but you can see what we have is misleading information being presented to physicians in a variety of different formats with regard to what they might expect in a cord blood product.

Some people claim 100 percent nucleated cell recovery. At one institution with 25 to 50 percent average over other institutions and the data I have from our 60 sample would indicate that those are clearly unrealistic claims on the front end but the 25 to 50 on the back end. I have never seen one that bad on average. Also viabilities of 99 to 100 percent I think are also clearly not what you would expect. So the information out there for many of the physicians particularly the inexperienced transplant physicians who might be trying to initiate a cord blood search is something that we have to address and maybe establish some standards in terms of recording and advertising. Next slide please.

In terms of what we might have for information with regard to correlation of transplant outcome, we have now exported approximately 520 products from the St. Louis cord blood bank that shows the trend in transplant exports over time. As you can see, virtually over 45 percent nearly 48 percent of all of our products have been in people for one and a half years. So transplant outcomes from our particular bank are going to very limited in time. Next slide please.

To give you some idea of the problems to the particular cord blood bank centers with regard to utilizing transplant outcomes to optimize processing, this just shows some of the responses that we get from our transplant centers with regard to reporting. Only 58 percent of all the transplant centers actually report to us complete post thaw data. Only 50 percent of them actually report to us any kind of outcome information between zero and three months. Only 34 percent give us outcome information extending between three and six months and it drops to 21 and then down to 10 percent for transplant outcome information extending up to one year. As I mentioned earlier, part of that is because a lot of the products we have exported have not been a year out in terms of their transplant but you can see that there is a great deal of loss of information that comes back to the cord blood banks from the transplant centers. So we don't know if we are doing the right thing or not. We have no ability without adequate reporting to help us improve our processing. Next slide please.

In summary I would say that currently there is no standardization of methods used to characterize cord blood, hematopoietic progenitors, cell composition, resulting considerable variability and results were posed by different cord blood banks and transplant centers. There are significant differences in post thaw data, correlating outcomes with cord blood characteristics and post processing data provided to transplant centers to guide product selection. Inexperience transplant centers may be mislead by unsubstantiated claims of cord blood processing efficiency and post thaw recovery. There is a significant number of cord blood transplants for which outcomes have not been reported and we have no idea what's going on with those right now. Finally failure to report clinical outcomes to cord blood bank impedes our progress in optimizing quality of safety in cord blood processing and is a major problem that I would like this committee to try to address. Thank you.

CHAIRMAN SALOMON: Thank you very much. The last speaker of this morning. You really want to ask a question, Alison. Okay.

MS. LAWTON: Just one really quick question because you talked about standardization and obviously the methods used for testing. But one question I have just for my clarification. What consistency is there in the collection and the processing and the storage and shipment for all these different banks?

DR. CREER: The collection model varies. There is a point of contention about whether or not the best way to collect this is to have individual obstetricians collect it or whether it should be done by an ex vivo procedure. We have not seen significant differences in terms of overall yield comparing the two methods between our center and other centers that use two different ways of collecting. The infection rates seems to be the same. Overall sale yield seems to be the same. I don't think the collection component is that critical.

With regard to transplant to the laboratory, we have not seen a significant loss in viability with products received up to 12 to 18 hours. The components of collection with regard to how you do and how fast you get it to the lab seem to be problems that we can overcome and that don't seem to have a real significant impact on the product quality.

The issue with regard to the processing we all utilize pretty much the same standard approach to red cell depletion and plasma depletion of our products. The approaches that we use have been published in the literature. I think efficiencies for banks with experience are quite comparable. So what we actually get back from with the products that we begin with is very comparable across labs.

Contamination rates are very similar. I think the real problem that results in the variability is in the laboratory assessment of what the progenitor cell content is in the product which is the ultimate determinant of outcome. We have no practical way to assess that.

The surrogate markers that we are currently using are imprecise and being utilized at different points in time and are giving us misleading information about how to correlate that with transplant outcomes. Until we can get around that, our ability to show statistical significance with regard to transplant outcomes for these product characteristics is minimal. That's what I think we need to address. We have got to get more tight data or variability will never allow us to show what the real dose effect is and give us the ability to give guidelines to other transplant centers so that they know what they should get and to be able to predict that from what we've already published. That's critical. If we try to extend this or license this, I don't think we can do it until we can give firm guidelines.

CHAIRMAN SALOMON: I think that there are a lot of complicated issues here. I'm not sure they are all just based on typing things. We'll get into the details of that this afternoon.

DR. KURTZBERG: One sentence. There is an NIH sponsored on-going current study called COBLT which is collecting data prospectively and which has standardized approaches that will answer some of this question.

CHAIRMAN SALOMON: The last speaker of this morning is Mary Laughlin who is at Case Western Reserve and is going to talk about a retrospective analysis based on data from the International Bone Marrow registry comparing bone marrow to cord blood.

DR. LAUGHLIN: Good morning. Thanks for the opportunity to present this data. This is a study that is coordinated by the IBMTR to analyze alternative donor transplants for adults with hematologic disorders. This is a comparison of unrelated donor bone marrow or peripheral blood and unrelated donor cord blood. The study chairs including clinicians Dick Champlin and John Wagner and myself are collaborating with Mary Horowitz and her group at IBMTR as well as Pablo Rubinstein and Cladd Stevens at the New York Blood Center. Next slide please.

The primary objective of this study is to determine one and two year survival after allogeneic transplant in adults with hematologic disorders lacking in HLA identical sibling donor comparing these transplants using bone marrow or peripheral blood from HLA matched or mismatched adult donors with that from umbilical cord blood. Secondary objectives of this trial are to monitor the incidence and kinetics of hematopoietic recovery, the incidence of severity of graft versus host disease, the transplant related mortality, disease of recurrence and disease of free survival. This study has been on-going since March of last year. Next slide please.

This slide outlines the patient demographics of the current study cohort. I'd like to begin with the caveat that this is a preliminary data analysis and that as such has not yet been subjected to peer review. The current study cohort includes 186 adults over age 16 who have undergone myeloablative conditioning and infusion of minimally manipulated single unit cord blood. These patients are compared with 451 adults who have also received full myeloablative conditioning and infusion of neither six to six or one antigen mismatched adult derived bone marrow or peripheral blood graft.

There are no differences as far as demographics of sex. There are differences in demographics of age with a higher proportion of the cord blood patients being a younger age group compared to the bone marrow group.

There are also differences in weight. The cord blood patients are slightly lower weight than the bone marrow cohort. Further differences in race with a higher proportion of minority groups represented in the core blood study group. Disease entities, there's a higher representation of CML in the bone marrow group. Next slide.

Further demographics to share with you before we go into the analysis includes disease status. There were differences in that. In the cord blood patients, there was a higher proportion of patients with advanced disease as defined by relapse, primary induction failure, blast crisis or advanced myelodysplasia. A higher proportion overall of the cord blood patients received TBI based conditioning compared to the bone marrow group. GVHD prophylaxis also differed.

The majority of the cord blood patients received a combination of cyclosporine and steroids due to concerns of administration of methotrexate and its potential impact on engraftment. The nucleated cell dose in these patients was fully a log less than that of the adult derived grafts. HLA compatibility was one antigen mismatch in 21 percent of the patients. The majority of the patients, 69 percent, received two antigen mismatch grafts and ten percent of the patients received three antigen mismatched grafts. There were four patients who had six-out-of-six matched grafts. They were not included in the analysis due to the very small numbers. Next slide please.

This is a uni-variate analysis of outcomes after transplant. Treatment related mortality was lower in adults transplanted with fully matched adult derived grafts. Importantly in comparing bone marrow from adult donors that is mismatched at one antigen which is considered a current standard of care in multiple transplant centers, treatment related mortality was similar in comparing adults transplanted with HLA mismatched cord blood.

Other parameters, including leukemia free and overall survival, again was better in the patients receiving fully matched bone marrow grafts. However importantly equivalent in comparing these two patient populations, those receiving bone marrow with one antigen mismatch from adult donors and HLA mismatch. Next slide please.

Cox proportional analysis was performed and the model was to test the main effect of the graft source. Potential confounding variables that were built into the model included patient age, sex, disease status, the graft cell dose and HLA compatibility and TBI versus non-TBI conditioning. Next slide please.

This Cox proportional analysis in the analysis of relative risk of treatment related mortality using matched bone marrow from adult donors as a reference point demonstrates that if you received a graft from an adult donor that was mismatched at one antigen or various levels of mismatch cord blood grafts you demonstrated an equivalent likelihood of treatment related mortality which was all higher than that of the reference group.

Age was also a factor in predicting treatment related mortality as a continuous variable and disease stage in the aspect patients with more advanced disease demonstrated a higher treatment related mortality not unexpectedly. Next slide.

The relative risk of relapse again using patients infused with matched bone marrow as a reference point was lower in all patients receiving HLA mismatched grafts whether they were from an adult donor or derived from cord blood. There seemed to be a lowering risk of relapse as increasing HLA disparity was observed.

Further disease stage not unexpectedly predicted a higher relative risk of relapse. Here patients with more advanced disease had a 6.4 fold higher risk of relapse. Next slide.

The relative risk of overall mortality again using matched bone marrow as reference point was notably higher in patients receiving HLA mismatch grafts either from adult donors or from cord blood. Again age broke out in the Cox proportional analysis as a predictor of overall mortality, as a continuous variable and patients with more advanced disease were more likely to die of their procedure. Next slide.

In an analysis where now the reference group, the patients receiving fully matched bone marrow from adult donors, is excluded. We see no differences in comparing cord blood of various HLA disparity with single antigen mismatched grafts infused from adult donors. Next slide.

So in summary we find in our retrospective analysis and again I want to emphasize that this is a preliminary data analysis that cord blood is equivalent to one antigen HLA mismatched grafts from adult unrelated donors in all transplant outcomes including treatment related mortality, relapse, leukemia free survival and overall survival. We also found that HLA matched grafts from adult donors is associated with improved transplant outcomes compared with one antigen mismatched adult derived grafts and cord blood.

The Cox regression analysis thus far does not identify HLA mismatched cord blood to be associated with poorer transplant outcomes as observed in recipients of adult derived grafts. In addition, cord blood cell dose in this dataset was not associated with improved transplant outcomes in these adult recipients but I wish to emphasize the narrow dose range that may be contributing to this observation. Thanks for your attention.

CHAIRMAN SALOMON: Thank you, Mary. I don't know about you but I'm hungry. How about we go to lunch and try and get back here in an half hour and I'll start the meeting as soon as we have a quorum. See you in a little bit. Off the record.

(Whereupon, at 1:21 p.m., the above-entitled matter recessed to reconvene at 2:13 p.m. the same day.)

A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N

2:13 p.m.

CHAIRMAN SALOMON: Welcome back from a fast lunch to the afternoon meeting where we are going to initiate a discussion of the four questions from the FDA. Before we get started particularly in responding to David's questions about just wanting a little more context on the questions which is fine and that should be an on-going process as we go along, I would like to start off with Phil Noguchi. Phil, do you want to make some comments?

DR. NOGUCHI: Yes. Thank you, Dan, and thanks everyone for getting a little heartburn over the quick lunch. What we've seen today is we came to this meeting with a certain set of data that has been evaluated. We certainly have seen a lot more data being presented. We have had an extraordinary public meeting where we are seeing it's not just the numbers. It's not just the curves but at the end of those curves, there are people. There are people who have benefitted from the emergence of this technology.

You have also heard from Dr. Rubinstein and Dr. Creer that it isn't just the outcomes and it isn't just the patient and it isn't just the source of the blood and the blood banks that produce them and the centers but it's all of them together. What we have seen and are seeing that the datasets are very interesting. They are very promising in some aspects but they are incomplete and we are asking for advice on how we can go further and begin to identify those things that are critical versus those things that are yet to be developed in the future.

As we step through the questions, those concepts of what do we now need will become very clear. Thank you, Dan, for allowing some exploration of many of the concepts within the questions during the morning session. With that, let's get stared.

CHAIRMAN SALOMON: The first question I have taken the liberty of modifying it slightly with permission of the FDA. The question 1(a) is gone. It will be discussed under two with a slight modification to the way we will phrase two. For right now, it goes as follows. Does everyone have the questions?

What factors should the Agency consider in determining the safety and effectiveness for the use of placental umbilical cord blood for hematopoietic reconstitution? That's obviously front and center today. With revision, the first thing is specific disease indications including specific types of cancers and severity of diseases for which there are data on the use of cord blood, time to engraftment, HLA matching issues and the incidence of graft versus host disease.

I must say in the context of this morning's discussions I feel that we certainly aired enough data. I'm not sure I've had time to digest all that data. It was definitely put there. We will need some help perhaps in synthesizing it down which is part of it. What kind of things do we think in discussion points?

Let me underline one thing though. In a way, you can all relax. None of this is going to lead to any kind of vote. What we want to provide the FDA is a high quality discussion that's balanced and shows what things we feel comfortable as a group saying and others that the field feels is just not known yet and in so perhaps establish priorities for future questions in consideration. Really everyone is encouraged to participate and feel that in the end there has to be some big pressure to come to final conclusion. Does anyone want to start with 1(b) specific disease indications? Mary.

DR. HOROWITZ: The thing that struck me in listening to the presentations this morning and preparing for coming to this meeting is that the results of cord blood transplantation are very similar to the results of bone marrow peripheral blood transplantation. All the things that effect bone marrow transplantation, and I'll use bone marrow to refer to both bone marrow and peripheral blood, will effect cord blood transplantation.

Some patients who get bone marrow transplants have a lower probability of a good outcome than others but still it's an acceptable approach in a variety of diseases. I haven't heard anything or seen anything that would make me think that the spectrum diseases for which cord blood transplantation should and can be used as any different from the spectrum of diseases for which bone marrow transplantation should be used.

What I'm struggling with is where are the differences in cord blood transplantation versus bone marrow transplantation that require some special consideration. We are not talking about the general safety, effectiveness of hematopoietic stem cell transplantation in general. We are talking about cord blood transplantation and trying to find out what's a little different here and there's not much I think actually that's very different in cord blood transplantation than the general field of allogeneic hematopoietic stem cell transplantation.

The differences have to do with we are operating at the limits of cell dose. There are special consideration for stem cell product where we are operating at the shoulder of the curve where we have special area actually includes inadequate for cell dose where we don't have to worry about that too much in bone marrow transplantation. We are in the adequate range by in large.

The other thing is that we have an experience with HLA mismatching in the unrelated donor setting that we actually don't have in bone marrow transplantation. We have a fairly extensive experience with multiply mismatched cord blood transplants that gives us actually a more precise indication of what the outcome is with this stem cell product than we do with bone marrow transplantation. The data that was presented today says it makes a difference but actually the data with two antigen mismatched cord blood transplants looks pretty good when you compare it one antigen bone marrow transplants.

I would say the specs on the disease issue should be no different than our usual specs on disease for bone marrow transplantation but we probably need to have some special considerations about cell dose.

CHAIRMAN SALOMON: That's an excellent place to start. There are a lot of implications that flow from the first statement. Bone marrow transplantation, mobilized peripheral blood stem cell transplantation and placental umbilical cord blood transplantation are essentially all the same.

When we started there certainly is a scientific premise that would suggest that they shouldn't be the same. In other words, if you had no data, the hypothesis would be that these would be distinct. Certainly the umbilical cord from the others because you have these immature T-cells so you should have less graft versus host disease. I would like some comment on that.

There are some biological reasons to suggest that there may be differences. There's the issue in the so-called SCID repopulating populations that suggest that these proliferate better, that they are more primitive cells. Dr. Wagner picked up on that in a little bit of a different way in human patients.

Given that as where we started, I agree with you that the data we saw today when you get down to clinical outcome parameters which is what the data was that overall the things that are going to drive therapy with these tools are not distinguishing. How about some discussion along that spectrum of things? Richard and then David.

DR. MULLIGAN: I'm very impressed by those last comments. That succinctly really summed up much of what we did here. That's the crux of the entire issue we have here. Two things. Is there equivalence or not and data? If we can come to some consensus that there is, that's very important. Very important to separate this dosage. It sounds like all the safety issues, the uncertainty of the questions really have to do with this issue of the limitation that you have per patient. I would propose that we leave out that, the limitation, to try wean out all those issues and deal with the other part first and directly deal with the issue of the dosage.

CHAIRMAN SALOMON: Precisely, yes. David.

DR. HARLAN: I'm just going to agree but I propose a way that would codify all of this pretty quickly is just to say that the indications for umbilical cord transplant are identical to anything that exists for HSC or peripheral blood transplantation except... The only except I heard was if you have an HLA identical bone marrow donor immediately available because there the dosage is better and I thought the results were better in that one exception. Correct me if I'm wrong.

CHAIRMAN SALOMON: That's fair. Then what we are really saying though is if you had an HLA identical cord blood donor it would be the same too.

DR. HARLAN: There's a difference. There you have the dose problem. The one advantage that you have in bone marrow and peripheral blood is you don't have the limitation of dose.

DR. HOROWITZ: Although I should point out that in the pediatric study where dose is not as limiting there was some data suggesting that perhaps if you had HLA identical cord blood transplant your outcome may be better than with a HLA identical bone marrow transplant. Dose has to figure into that. There are not sufficient data with HLA identical cord blood transplants in adults to be able to make a statement one way or the other in my opinion.

CHAIRMAN SALOMON: Don't forget to remind me that I have this one point I have to make.

DR. WAGNER: I'll try to remember that. In terms of addressing this point, are you saying that if there is an HLA matched unrelated marrow donor then that should be chosen in preference to a cord blood donor except maybe with the caveat of six-out-of-six cord blood donor?

DR. HARLAN: Based upon the data I heard ?-

DR. WAGNER: But that's not exactly what I said though. Basically what the data suggests is that umbilical cord blood at least if it's matched in one antigen mismatched that appears to be equivalent to a six-out-of-six marrow.

It wasn't saying one was better than the other. In fact, the cord blood six-out-of-six looked preferable to all other stem cell sources. However the one antigen mismatched cord blood appeared to be equivalent the six-of-six marrow. Whether you look at the IBMTR New York Blood Center dataset or whether you look at the Minnesota dataset, that seemed to be the case. You can't say you should select one over the other because they were quite comparable in outcome.

Now you could argue that if you have your choices of six-of-six marrow donor versus a four-out-of-six cord blood donor particularly if the cell dose is low then you can argue perhaps in that setting that you should choose the marrow donor in preference. What makes this all difficult is the fact that it's this issue of timing of transplant. Therefore to come up with a strategy that says if you have a potential six-out-six you must wait for that. Practically speaking that can't happen. That's where the dilemma comes up. Sometimes you just have to go ahead and use the more mismatched graft because you simply can't wait. That's something to keep in mind.

DR. KURTZBERG: I agree with John said. I can't stress that enough. I do a lot of work with the children with inborn errors in metabolism. We can go forward in two weeks with the cord blood and we can't do that with a bone marrow donor unless it's in the family.

I don't think you can tell a family with two matched siblings there is no regulation that says which one you are going to choose. If you have equivalent donors then you as the transplanter have the authority given the context of the entire situation to pick the best donor. So then I wouldn?t, when you have equivalency, put a restriction on which goes first. I think that's really important.

DR. HARLAN: As the subject matter expert, the one that has done a lot of these, you should be the one that suggests. Are there any limitations that should be placed on umbilical cord blood in your expertise?

DR. KURTZBERG: I think cell dose is the thing that should be defined as safe or unsafe. It should be targeted to a certain minimal level. I think we have the data to make those recommendations.

DR. HOROWITZ: Go back to the statement. The indications which I think of as the underlying disease for which the transplant is going to be done are the same for bone marrow, peripheral blood and cord blood transplantation. In the absence of any compelling data which I didn't hear today that one thing is really better than another because of all the difficulties in these kinds of analyses that have to do with why one went to a cord blood rather than a bone marrow transplant which have to do with the issue that Joanne brings up which is you can't wait all the time, CMV positivity of the donor, it would be hazardous to try and regulate that decision making process.

CHAIRMAN SALOMON: I didn't take David's comments as he was suggesting that an outcome of this was that we said you can do this or this but you couldn't not that. Number one we all know that no one is going to allow that sort of advice to be given to the FDA. You don't want that in any medicine. We all realize that there are so many different things that impact on the decision processes that we want to just come up with some general views.

What I thought David was challenging was just is there any situation in which one could argue that a bone marrow or peripheral blood was different. That was the premise. So far, no one has come up with that. In the absence of that, we can make a very clear statement.

I have one. That was the point I wanted to be reminded of. It would be perhaps in some of these diseases where we are playing out this new hypothesis that these stem cells might hone to disease tissues such as occurs in the ALD children that maybe in that instance there might actually be a rationale for using an umbilical cord blood. Again I know there was no data shown today that will allow a decision on that. I'm just throwing that out as the one thing I can even think of now right now.

DR. KURTZBERG: It's a very exciting field. There is a lot of early information that is suggestive that cord blood might be a good stem cell source but there are again so many complex decisions that go on in making a decision to treat those children that it might get in the way of their care if that kind of marrow restriction was placed on their care. I believe they should have cord blood but I believe more importantly they should have a transplant. I don't think we are at the point where we have enough information to say it has to be this kind or that kind.

CHAIRMAN SALOMON: And I wasn't going anywhere that. I was just saying that it was the only thing I could think of that might eventually be a distinguishing feature in terms of the diseases that might be treated because that's what we are trying to swear. Mary.

DR. HOROWITZ: It should be noted that differentiation into non-hematologic tissues has been demonstrated for bone marrow and for blood stem cell transplants not just cord blood transplants without a lot more data of one being better than the other.

CHAIRMAN SALOMON: Of course, do you believe that data too?

DR. HOROWITZ: It depends. We can talk about that later.

DR. KURTZBERG: Frankly, I don't think there will ever be a study done where you analyze bone marrow versus cord blood. The other logistics that go into that are just too daunting.

DR. RAO: Everybody said that bone marrow and cord blood are effectively equivalent. Is the data quite clear for this one mismatch in bone marrow with one mismatch in cord blood or is that still just a matter of timing and availability that one would say there is no choice? In other words, preference of one or the other?

DR. KURTZBERG: I'll start. There are so many biases. I've done 500 transplants myself but I'm biased. I accept more mismatch. I use a certain type of typing. I have a certain timeframe. I only use a certain bank. So whatever comes out of my center is biased by those decisions that I've made for better or worse. In my data, it is not clear that the one-antigen or two-antigen mismatch make a difference.

Again you have to realize that the majority of patients who come to this are people without donors. They are not people who are deciding between a bone marrow donor and cord blood donor. They are people either who don't have a bone marrow donor or because of time, don't have time to wait for a bone marrow donor.

So you are talking about availability of a therapy versus no therapy. What we have to do is decide if this is safe, if it's equivalent enough to the other stem cell sources that might be available to allow these people access to a similar therapy that they otherwise could not have.

DR. RAO: Are you saying something like it's a choice after when compatible bone marrow is not available? That's why I asked this question about the one mismatch because then the possibility of bone marrow becoming available would be that much less.

DR. KURTZBERG: In the majority of patients that come to cord blood, they have looked for a bone marrow donor and have not found them at least in a timely fashion. But yes there will be some who have a five-out-of-six match and make a choice between one or the other.

DR. HARLAN: Joanne, you said something that tickled a thought in my mind too and that is all of the cases that we heard in the public comment were people who really had no other choice. In the 500 that you transplanted, are these otherwise fatal diseases for which there is really no other choice and should that enter in our recommendations at all?

DR. KURTZBERG: I think in general any people who go to any sort of stem cell transplantation don't have a casual disease.

DR. HARLAN: I mean not a casual disease but these are rapidly terminal diseases that we heard about today. The area where I could see this becoming a slippery slope is if whatever recommendations that we make mean to people okay we start using this for something that's maybe not quite so terminal. Then the risk benefit equation changes considerably.

DR. KURTZBERG: You can go first.

DR. HOROWITZ: I think that's unlikely. Whether you are talking about a bone marrow transplant or a cord blood transplant, as long as you have informed consent I know the patients and their families won't take it likely. The relative efficacy of stem cell transplant versus non-transplant therapy is beyond the scope of this conversation and probably unresolvable in a lot of instances.

Most of the patients who go to unrelated donor transplantation regardless of graft source have diseases for which you can get reasonable agreement that there aren't a lot of other options. These are not transposes that they can be done easily. They tend to be done in larger tertiary care centers. I don't think we have to worry about a slippery slope here.

DR. HARLAN: I can tell you from my experience that we do. I deal with diabetes all the time and patients say I don't care what I have to do I want to get rid of this disease. They will do anything even if the risk benefit ratio isn't supported.

DR. HOROWITZ: I doubt that there would a transplanter that would do it for diabetes.

DR. KURTZBERG: This is not different whether you use cord blood or you use bone marrow or you use peripheral blood. This issue is generic to stem cell transplantation not different for cord blood.

CHAIRMAN SALOMON: That was the point that I wanted to bring us back to and that is notwithstanding that I totally agree with David in that this is a classic where in your field that it's probably not a very slippery slope. But outside of you and stuff you haven't thought about yet, it's real slippery. The key thing here is if there is no critical difference between hematopoietic stem cell of any of these sources at least as far as the data shows today if that's going to be the conclusion that we are rapidly getting to and then we're going to talk about the matching and then we're going to talk about the dose then risk benefit issues for those on the slippery slope, and I agree it's there, it won't matter what they are arguing, the safety and efficacy issues are still going to be the same for as long as we are using a stem cell. David.

DR. STRONCEK: I do worry about the slippery slope too. You have to remember with unrelated bone marrow donors there are criteria that centers have to meet. It's fairly liberal but there still is a criteria and the centers have to be qualified. Right not, there is no such criteria for cord blood transplant centers. I agree that right now the centers doing it are very good, very academic centers. It doesn't mean that as this becomes more routine that other centers that aren't qualified won't do it and they won't get as good outcomes.

DR. HOROWITZ: Related to that, I find it very concerning, the data of the St. Louis Cord Blood Bank that they are not getting outcome data on half of the transplants that they are facilitating. There?s a lot of ways to get outcome data. You can't get an unrelated donor marrow from NMDP unless you agree to provide outcome data. I don't think that the criteria for obtaining a cord blood should be any less stringent than that or else how can you evaluate the field and how can you evaluate the outcome from that cord blood bank if half the data aren't there and you don't know whether that's a representative omission or not?

CHAIRMAN SALOMON: I think that's a very important point. Let's hold that for question three which is outcome parameters for clinical data which I think is a nice segue to your question of how you can get the outcome data. I think that's really important.

The slippery slope thing is worth discussing just a little bit. Two things in the field right now that we heard about today relate to this as well. One is a non-myeloablative preparative regimen and two, less HLA matching. One thing that could happen here is that centers that can do non-myeloablative regimen which are much less taxing than fully myeloablative regimen, number one then would feel emboldened to go in directions that would be defined and then if you open up and say one good thing about cord blood stem cells is that you can break down some of these HLA match criteria then again the pressure is going to be I can go to this cord blood registry and I can find a two to three antigen mismatch cord blood for me and cure my fill in the blank, diabetes, lupus, rheumatoid arthritis, allergy. I do think we need to consider that a little bit.

DR. KURTZBERG: Why do people have an M.D. degree if people are going to be that cavalier about performing a transplant? We're in a lot of trouble if that's true. I don't know how anybody can regulate that. I think transplanters are generally very responsible and take what they do very seriously and understand that the risks are quite high regardless of which one of themes you take to an ending. This is not done lightly. (1) I just can't imagine that you could control it. (2) I can't believe that the practice of medicine would go in that direction.

DR. HOROWITZ: Could we ask John Wagner to give his opinion on that? Except for a small number of transplants that were presented this morning, almost all the data that is in the literature or was presented this morning has to do with myeloablative preparative regimen. So I don't know how you can say much about the others right now.

DR. WAGNER: Right. There are two points. First, I think there is a risk of slippery slope but we need to go back and say would you do a non-myeloablative prep using bone marrow cells. If you are willing to do that without any oversight, then yes you are willing to do unrelated cord blood without any oversight as well. It may or may not stop if someone really wanted to do it.

On the other hand in terms of specific disease indications and I know this is going to be controversial but the only disease that makes me nervous only because I don't know the data is CML. CML is a disease which in chronic phase where the results of unrelated marrow transplants if you have a six-of-six , the results are very good as has been reported by a number of institutions.

Now it's not to say that cord blood is any different. The question is whether or not we have the data to be able to say that it is the same. Maybe you do at the large repositories of data could address that issue. But because of the Seattle data first reported by them and then by other institutions, if you have a six-of-six unrelated marrow donor and you have CML in chronic phase and you have the right placement of the moon in the sky, you can actually have the outcomes in excess of 70 percent success. That's the only indication that is not known to me. I don't know that it would be any different if we have cord blood. It's just that I don't know the answer.

CHAIRMAN SALOMON: So that's the conservative/gold standard argument. Phil.

DR. NOGUCHI: Before we leave this topic, what I do want to say is Dr. Kurtzberg is absolutely right. Responsible transplanters have ethic and a standard that is will established. However cord blood in its own right is a limited study. That is you don't get more than one cord blood sample per person normally per donor. It is being banked under various conditions. If you use one and it happens to be a match for somebody that's one less person that may be able to get a use of it for a well established procedure for which bone marrow transplantation is useable.

I think it is not for an in-depth discussion today but unfortunately FDA does have to consider all these other things. We think the data is certainly very encouraging for the outcome for the kinds of transplantations that are typically done by transplanters. But as this data becomes more widely appreciated, there is going to be the push for what other diseases can we do that for.

That's where we really do need to be very clear about what data exists for what is available as well as recognizing that the difference between cord blood and other bone marrow peripheral blood stem cell transplantation is. It limited in terms of the numbers. It is limited in terms of the dosing and everything else. The others are at least theoretically you might be able to do it more than once from the same donor.

CHAIRMAN SALOMON: Mary.

DR. LAUGHLIN: I had a couple of comments. One was I wanted to add a caveat to Dr. Wagner's comments of the aspect of CML.

CHAIRMAN SALOMON: This is Mary Laughlin from Case Western.

DR. LAUGHLIN: The data is antedating the initiation of oral Glivec and certainly the demographics of CML patients that are stepping forward into unrelated transplant are not in chronic phase generally. The role of cord blood may be even more preeminent now in the treatment of CML in the Glivec phase because we are now more often seeing patients with blast crisis or accelerated phase having grown through oral Glivec. That was one comment I wanted to make.

The other from the standpoint of a leukemia physician in the trench is this aspect of somewhat doing the math of 10 leukemics who needed transplant to having a donor sister that matches them. The remaining eight turning to unrelated sources of grafts whether they are from adults donors or cord blood and then the kinetics of their disease determining which of those grafts is optimal for their disease.

It would seem to me from the standpoint of a licensure mindset that those diseases where stem cell transplant is a standard indication doesn't require stringent mechanisms such as IMD but that perhaps by these newer non-myeloablative approaches to extend into renal cell cancers, autoimmune processes that the restrictions might be to incur greater stringency in those less well defined indications for stem cell transplant in general.

DR. KURTZBERG: Just relating to CML, this was our data for CML in kids. These are kids in accelerated phase or blast crisis and two in chronic phase. There were only 11 patients but this is mismatched cord blood in terms of what we have with the median follow-up of 4.7 months but actually I think that's probably about 19 months. This is not data that would suggest to me that cord blood wouldn't be a reasonable source for kids with CML who come to us.

CHAIRMAN SALOMON: Can you just educate me? This would be very good for adults with blast phase CML. Is that also true for children?

DR. KURTZBERG: Yes, this is a mixture of accelerated phase blast crisis and there are a couple in there with first or second chronic phase.

CHAIRMAN SALOMON: Are children less prone to problems?

DR. KURTZBERG: I don't think so.

CHAIRMAN SALOMON: So it's just as bad for them as it is for adults.

DR. KURTZBERG: Yes.

DR. WAGNER: The only point to make however is that in contrast to what Mary Laughlin indicated, it's a very small subset in the narrow transplant setting that fits all those perfect criteria. When you look at that perfect criterion that's the only group that makes me somewhat nervous. It's not to say that I know what's right or wrong. It's just that I don't know what's right or wrong. That's my only point.

CHAIRMAN SALOMON: I think again to reassure you that's not what the FDA is worried about. The FDA is not going to try and tell you that you shouldn't do this or that in that specific group. I think you agree, right, Phil? That's not ever going to come out in an FDA reg.

DR. NOGUCHI: No, FDA always recognizes the autonomy of the physician to within his or her best judgment to make the best decision for the patient absolutely.

CHAIRMAN SALOMON: Unless somebody has something to change, then I think there?s a consensus here that based on clinical data so far regardless of whatever we postulate from theoretical points of the biology of these types of preparation, there doesn't seem to be any reason for the FDA to consider in at least the spectrum of diseases a difference in any sort of hematopoietic stem cell source. We are going to get now to a couple of the key issues. Does everyone agree with that as a first answer?

DR. NOGUCHI: That's with what the underlining assumption in terms of standard bone marrow and peripheral blood stem cell experience with the diseases for which it's being done.

CHAIRMAN SALOMON: That's right. I still think it's important that you realize that there are these other indications. There are various levels of concern. My bone marrow transplant guys don't realize that if I get it in my mind to do this particularly the non-myeloablative I'm going to put them in the hospital and just do it. I don't need you.

I'm just teasing, but if someone feels strongly about it, that's what they are going to do. We need to consider a very separate group of things. Again I don't know that it needs to be considered from any hematopoietic stem cell which is the fundamental reason we're here today.

As Rich directed us after Mary's comments the two key issues seem to be the HLA typing and the dose. Let's talk about the HLA typing first. All right. Let's talk about the dose.

DR. ALLAN: Actually I have a question that I was a little bit confused about this morning. I don't remember who presented the data but it was that there was less relapse in those patients that had greater mismatch. Is that what it was?

CHAIRMAN SALOMON: That's classic in the GVL literature. As you increase dose, you get more GVL but you also get graft versus leukemia and graft versus host disease. That's the dark and light side of this.

DR. ALLAN: But in terms of survival, would you want to have more mismatches because that way you get less relapse and might still have survival?

DR. HOROWITZ: No. That analysis was the analysis of the adult bone marrow versus cord blood transplant where there was a suggestion of less relapse with increasing mismatch. I have to go back to the multiple comparisons issue that was pointed out earlier quite rightfully. The overall p-value for that whole category of variable was not significant. The most significant p-value was 0.05. The number of comparisons was in excess of 20 so I wouldn't make much out of that in this analysis. GVL is nice. Survival is better. It's quite clear that more mismatch gives you more transplant related mortality.

CHAIRMAN SALOMON: The only thing one could go based on today's discussions would be if the premise could be put out for discussion that if the cord blood allows you to get away with a more mismatch with equivalent or less GVHD which is what the data showed us today, would you potentially get a GVL benefit in the process? That might at some point argue that cord blood would be a better source. I'm just trying to get these points into the discussion. I know we are back to HLA and we said we were going to talk about dose.

DR. KURTZBERG: If you look at the composite groups that have presented, there are a higher proportion of high risk patients but there is no comparison group. You would have expected a higher relapse rate with a standard matched sibling transplant in those groups than you see in the cord blood patients. Both of them have alternative donor sources that are not related matched have a lower relapse rate than a matched sibling transplant. I don't know that you have the data to say more than that but your job was in the data where the relapse rates were in the 20 percent range.

In our group and Pablo's group where the portion of the patients that are high risk are already 80 percent you would have expected a 60 to 80 percent relapse rate. There is no comparison or randomized trial that actually answers the question directly.

CHAIRMAN SALOMON: That could be one of the things that we can't really comment on because the data is not out there. That might be worth looking into. Of course, the null hypothesis would be that the same biological mechanism gives you less GVHD, would give less GVL. It's hard to know.

DR. WAGNER: In terms of this question about HLA and the importance of HLA, I agree it's probably the easier of the two issues to talk about. Maybe the way to get to it is actually to find what is unacceptable. Go the other way around. Most of us here would agree that some degree of matching is important. The question is how much matching is important. From my point of view, what I would at least propose to you is that the data that we've seen for all intents and purposes has really been focused on one and two antigen mismatches with some zero mismatches and some three and beyond.

Just based on that alone, I think that we have to focus on saying probably we don't have enough data yet unless Pablo would say differently to say that a three antigen mismatch transplant probably should not yet be done routinely until we have more data to say that it's okay. Then everything beyond that in terms of more mismatches probably there is really no data to comment on. At least three and beyond should not be routine use.

CHAIRMAN SALOMON: Does everyone agree with that statement that three and beyond is in a different class than two and less?

DR. HARLAN: Couldn't we even make it simpler? Isn't it the same criteria that's currently used for bone marrow and peripheral blood? If there is no data to suggest what criteria should be used, why not just leave it alone?

CHAIRMAN SALOMON: I'm glad you brought that up. Do the bone marrow transplanters want to respond to that?

DR. KURTZBERG: No one would use the four-of-six antigen without some additional method to prevent graft versus host disease. So it's not the same.

DR. HOROWITZ: There's actually not very much experience in that. There's enough experience to know that two antigen mismatching the bone marrow setting is probably not a good thing. John points out that the real difference in cord blood transplants between the two and the three is not that we know that the three's are worse but there is ample experience with the two antigen mismatch to say that has an acceptable outcome to feel comfortable with it. There's just not enough experience with the three and more antigen mismatch to feel comfortable with it.

DR. STRONCEK: John, what do you mean by antigen? Yes, as defined here. For bone marrow and unrelated donors, people will do basically sequencing so they get allele specific matches and then they start talking about eight antigen and 10 antigen matches. Now with cord blood, it would be theoretically nice to have everyone sequenced but that's so costly. I don't think that happens when you match up your cord bloods, does it?

DR. WAGNER: It's being done now. We treat cord bloods the same as we treat unrelated marrow donors except that that's relatively new and only in those cases where we can get DNA to be able to do the high resolution 10 antigen testing. What we are doing is we are comparing class I lower intermediate resolution, A and B and DRB1. That's really the only dataset that's comparable to bone marrow dataset.

When we talk about what level of matching is the minimum, we're talking about A, B and DRB1. The others are interesting but we don't even know what the impact is for sure yet in bone marrow transplants much less in cord blood. So what I would suggest to you that clearly with cord blood, zero and one antigen mismatches at that level of typing is certainly adequate. We have substantial data in the two antigen mismatched cord blood transplant to at least be able to adequately inform our patients as to what the outcomes should be. I would suggest that zero, one or two at that level of typing should be considered adequate.

DR. STRONCEK: This is too technical. I'm confused. When you do your actual transplants, you get DNA from that cord and then you do really high resolution typing but then when you do this group data you are looking at lower resolution typing for A and B antigens and high resolution for DRB-1. Is that what you said?

DR. WAGNER: It depends which analysis you are asking to be done. All this data that's been presented was dataset that was several years old and it's at the level of typing that was available then. We are just now beginning to do a higher resolution typing prospectively for future patients but there is no data analysis that we have at our institution to be able to ask that question. As we pointed out in our discussions, it would require a huge dataset before we can make any statement whether it be bone marrow or cord blood as to what degree of mismatch should be allowed.

DR. KURTZBERG: The standard of practice now for cord blood is not to do high resolution DNA typing at A, B, C, D, RD, PD, Q. The standard of practice is A and B at medium resolution at the serologic level. DRbeta1 at high resolution DNA.

CHAIRMAN SALOMON: That is what I was asking Dr. Rubinstein during his presentation exactly what he was talking about in the typing. What David is getting at is really critical and that is if we are going toward just a recommendation to the FDA that two or less mismatches seems to be equivalent that's a very important one. Then we have to be clear that we are defining to them what does two mismatch mean. I think that's where David was coming from. Dr. Rubinstein.

DR. RUBINSTEIN: With respect to the high resolution typing, the data that I presented this morning has been typed for DR at the highest resolution possible. Even the old transplants we have the DNA of the donor and recipient in essentially all cases and we have retrospectively typed them so we know where they are for class II at the sequence level.

For class I we are at the moment finishing the sequencing of A, B and C for the patients who have one mismatch. For those patients we will have an answer about whether we can increase discrimination possible with better matches in parallel to an improvement of the each situation. It will be retrospective but that's one point.

The second point is that all our cord bloods are now tested from the start with sequence for DR. We do that because it turns out to be cheaper than to do the sequence of an intermediate DNA level and then prior to the transplant to come back and do the sequence.

CHAIRMAN SALOMON: I think that what I'm hearing here is still what we are talking about in this first recommendation which is relatively low level typing not sequence based typing. But if the recommendation was going to be that we should go on and get data on that, we don't really have to make that recommendation because you are already doing that. That means it's underway and we can look forward to that sort of information.

DR. WAGNER: That's my point though. So you can still make that recommendation as appropriate because one bank doesn't mean that everyone's doing the same. Also with being very specific, it's low resolution or low to intermediate at A and B. DRB is high.

CHAIRMAN SALOMON: Is everyone comfortable with our recommendation on HLA matching? My understanding is that we are saying that for cord blood two or less mismatches as defined currently by the field which is relatively low level resolution is acceptable. And that, we acknowledge, is distinct from current practice today for bone marrow and peripheral blood allogeneic transplants. However as was pointed out, I don't know if everybody got it, there are people doing that now coupling in strategies to do aggressive T-cell depletion or to use antibodies and other strategies. It's not that there isn't research going on to even make that equivalent to hematopoietic cells.

DR. WAGNER: I may have missed it. Can you repeat what you said for DR?

CHAIRMAN SALOMON: I tried to get at this DR, A, B, C difference and I thought that what I heard overwhelmingly is not make a distinction. It was two mismatches.

DR. HOROWITZ: How we define them is different. We define DR with high resolution. We define A and B with low to intermediate resolution.

DR. WAGNER: Right. That's the distinction I wanted to make sure was made.

CHAIRMAN SALOMON: I have to say that was where I was going earlier. I am surprised that you can't go further and distinguish the difference between DR and A and B matching, not resolution but just in the effects but you can't. I don't know I don't believe that but I'm sure that's true from the data.

DR. ALLAN: I just wanted a clarification. When you say two mismatches or less and then John said that there is very little data on three mismatches or more. If there is more data in the future, you could change. But you are not going to have more data if you say two or less. You are not going to have any three mismatches so you will never know whether or not three would actually be useful.

CHAIRMAN SALOMON: I think the answer there, Jon, is that no one is trying to recommend regulations on anybody. What we are saying as a committee is that if someone wanted to do three or more, that would be in a group that would have to be taken differently than if they wanted to do two or less.

DR. ALLAN: You know how things operate. It gets around pretty quick. Everybody goes you shouldn't do three so it's not going to get done.

DR. WAGNER: That certainly is possible. I'm thinking from a different point of view. What we are saying is not that you can't do these things. It's just that in terms of what is acceptable based on what we know today the category has data that seems to support it being acceptable. This may be but whatever mechanism they can still study it. It's just a matter of how you define what's routine versus what's not routine.

DR. NOGUCHI: Perhaps just a better way of putting it from the way we typically look at things is you can make statements about two or fewer mismatches. You really can't make a statement three or more. We try to avoid such things as an acceptable versus not acceptable because it depends a lot on the clinical situation as Joanne and other rightfully point out.

MS. LAWTON: I just wanted to add to that discussion if we are recommending two or less and clearly you don't want to stop the research to understand whether three could be acceptable. Presumably the FDA will look on a case by case basis and proposal, regimen, etc. as to whether that's appropriate research to continue to try and find out because we want to make sure that we don't stifle that research. That's all.

DR. NOGUCHI: Again, to say that it's acceptable versus not acceptable I think we would be more in the category of saying we hope that you will be able to make a statement. For two or less there is a set of outcomes for which we can make some predictions that will be useable in the clinic. For three or more you can't make any statement. Acceptable versus not acceptable is really not quite what the FDA is looking for here.

CHAIRMAN SALOMON: I think that's what we've told you.

DR. LAUGHLIN: I guess I would also reemphasize John's data that as you increase HLA disparity you can somewhat overcome or you can rescue the less optimal transplant outcomes by increasing cell dose. It's certainly been in our experience in the adult recipients within the constraints of what you can attain cell dose that our outcomes in three to six mismatches has been below the threshold that we consider acceptable. There are some differences in the pediatric data versus the adult data in that in the pediatric data you have a greater wobble cell dose-wise that you don't have that luxury in the adult recipients.

CHAIRMAN SALOMON: Before we go on to focus on dose a little bit more, incidence of graft versus host disease. I'm not sure what to say about the data I saw today. Part of it showed that there would be less graft versus host disease with CD34 enriched populations from cord blood. Also the statements that we made that you can get away with two mismatches or less which would certainly argue strongly for that premise.

DR. WAGNER: The data that I've seen and published has only demonstrated that if you put all patients together that cord blood is associated with less graft versus host disease. Data presented today which is the cleanest dataset that Mary and I and others worked on was looking at HLA identical sibling transplants with cord blood versus marrow. There you were able to more easily tease out what was the contribution to graft versus host disease. There clearly acute graft versus host disease was significantly less in the setting with cord blood. So there is something biologically different.

With unrelated transplants, there it's a little bit more complicated because there is also other level of disparity which we can't account for because we don't test for them. Nonetheless the data goes flows in the same direction that would appear that there is a less risk of severe graft versus host disease with unrelated cord blood as compared to marrow.

The one exception of that was my dataset at the University of Minnesota which is a matched pair of analysis. The surprise was that in children when you compare mismatched cord blood with matched marrow in that very specific setting there the results were equivalent. So it's at least as equivalent if not better.

DR. KURTZBERG: Again that's a mismatched cord blood with matched marrow. If you are thinking about availability, the people getting the mismatched cord blood are not the people who have matched donor. The other thing which John presented in his talk was the chronic GVH which is really a marked difference between bone marrow as far as I am concerned. It's true in children as well as in adults. Chronic GVH rates and the severity of chronic GVH is lower with cord blood than with bone marrow.

CHAIRMAN SALOMON: I have to say that so far the committee is saying in concert that GVHD is less with cord blood. Acute GVHD. We?ve segued into chronic. I would like to come back to chronic for a second. I think we can agree that the weight of the data is pretty good that acute GVHD is less with cord blood. That's a good thing. Dr. Rubinstein.

DR. RUBINSTEIN: When we look at what the transplant center tells us about GVH, the picture is not so transparent. The centers that have the most transplants who all agree with what we have heard that the frequency of acute, severe GVHD is lower in cord blood. But the other centers report higher frequencies. I'm not certain what this means. I didn't show this data in the morning. I don't think we are comfortable with just giving frequency data. I think there is a problem in the evaluation of these parameters.

CHAIRMAN SALOMON: Certainly if I was running a multi-center clinical trial specifically with acute graft versus host disease, I would be much more inclined to trust the GVHD evaluations in an experienced center than a smaller center that didn't have that kind of experiences. There is a lot of subjective, experiential based judgments on judging acute GVHD. That may be very powerful. Mary.

DR. HOROWITZ: Even in big centers I might add. What you can be confident of is that it's not worse than bone marrow transplants. I think that's the only issue that this committee has to deal with. I feel much more comfortable when we look at things like survival because there is no inter-center variability with that one. Everything that we've said up to now can be held up by looking at the survival data which is the data that we can be really sure of. When we start looking at the intermediate outcomes of acute GVHD, chronic GVHD engraftment, we are bound to leave out some that are important.

I go back to the conversation earlier about the T-cell depletion trail where everyone assumed even though there was no survival difference since there was less GVHD and the T-cell depletion arm, the quality of life was going to better. It doesn't look like that's the case because we can't think of every surrogate, of every cumulative outcome that adds up to a person who is alive and feeling good. So we pick out some of them. GVHD is of course a traditional one in transplantation. We can feel really comfortable that it's not worse and I don't think we should talk about it anymore. Then the recommendations because they are held up with the survival data are stronger on HLA and then we'll get to cell dose eventually and indications.

CHAIRMAN SALOMON: Yes, we'll get to cell dose eventually. I don't totally agree with what you said in that survival data of course is very comforting and you're right, it's clear. You're alive or dead. It's kind of binary. But the problem is when you start comparing using survival data as your only parameter then you do get into trouble because there are differences in diseases.

There are differences in some of the therapies. A certain survival rate would be very appropriate for one disease and totally unacceptable for another. So I don't think we should get too much comfort from a single variable.

DR. HOROWITZ: What you just said about differences in therapies and differences in diseases is true about all the intermediate outcomes as well. One of the things in evaluating GVHD is that GVHD prophylaxis regimens may very much influence the maximum grade of GVHD when used by different centers. It may decrease GVHD as T-cell depletion did with bone marrow transplants and increase an adverse outcome that you don't happen to be measuring.

CHAIRMAN SALOMON: All I'm saying is that we just should realize that there are multiple parameters and not just survival. That's the only point I was trying to make.

DR. WAGNER: My only comment about graft versus host disease is the fact that we can't throw away the sibling data which was the best way of evaluating graft versus host disease and in the sibling data, works the cleanest for both acute and chronic. It was less for cord blood. However for the unrelated transplant patients as Pablo pointed out, there is a marked discrepancy between institutions and by reports both for acute and chronic graft versus host disease. As you have pointed out, Mary, never have we seen a report where it says it's higher with cord blood. I think that it's safe to say it's at least equivalent despite the level of mismatching with cord blood in contrast to marrow.

CHAIRMAN SALOMON: I think that's clear. What I would like to put out is that I'm not comfortable at the moment saying anything about chronic GVHD and this is for discussion. I really don't think there was any data presented today that made me comfortable that there is enough known about chronic GVHD in any of these. I think we should tell the FDA that there isn't enough data to make any comment on that right now.

DR. WAGNER: There is plenty of data. We just didn't present it. The only data that was presented today was my data which showed at risk of 10 percent. I would tell you that you can look in any transplant literature and you will never see another report that says 10 percent for bone marrow transplants for any disease because bone marrow transplants is higher. It's always higher than 10 percent.

DR. KURTZBERG: The data for unrelated marrow ranges from 40 to 70 percent depending on the population. The data that John and I put together and I can pull those slides up from an analysis of about 300 patients between our combined institutions held up the 10 percent figure.

DR. WAGNER: So even if we graded differently, at least within our own institution where we graded it and despite the fact that you might say that the way we graded is different, we have two investigators, Stella Davies and Dan Weisdorf, who blindly graded all the graft versus host disease between the unrelated marrows and the cord blood. That's how we've presented it. At least it applied to the same criteria to both arms.

DR. HOROWITZ: Did you see a difference in chronic GVHD in the same subset, where you did see a difference in acute GVHD?

DR. WAGNER: Yes, we did except unfortunately I didn't bring the slide to present to you. We did see a difference.

DR. HOROWITZ: In the pediatric study with the New York Blood Center versus the IBMTR data, we did not see a difference. In those children in that analysis, there was not a significant difference in chronic GVHD between graft sources.

DR. WAGNER: And that's what the Europeans have presented.

DR. HOROWITZ: So I think it's inconclusive.

DR. WAGNER: But where we do agree is that it's at least not higher.

DR. HOROWITZ: It is not higher, right. I can agree with that.

DR. RUBINSTEIN: I could give you a few. In the aggregate data, the chronic GVHD has gotten up to 30 percent. But if you make subsets in the data, it's very clear that the results are very different. We can make some extrapolations that people with more experience will grade these patients lower.

DR. HARLAN: Dr. Horowitz moved about five minutes ago to table this discussion. I second that motion because I think everybody is saying that there is no?evidence that it's worse. Incomplete data. I just think end of discussion.

CHAIRMAN SALOMON: We don't need to be that formal, guys. Fine, we'll go on. I'm just trying to make sure we answer everything here. Time to engraftment. I tried to go there when I started talking about immune reconstitution but just like allogeneic, you have your own term for immune reconstitution too. Immune reconstitution now is T-cell function. According to that hypothesis, what I'm hearing from you is the difference of reconstitution of the granulocytes is not really that critical a determinate of outcomes. Now the question is is there a difference in time to engraftment for immune reconstitution that distinguishes cord blood from other hematopoietic stem cell sources.

DR. WAGNER: Now I have to sit down.

DR. KURTZBERG: I don't think that's ever been studied. I think there is more data about immune reconstitution after alternative donor either T-cell depleted marrow or cord blood than there is after standard marrow transplantation with matched donors. There's never been any head-to-head comparison. From what I know there is not a distinct difference between the cord blood and the marrow.

CHAIRMAN SALOMON: One way we would look at it is that there doesn't seem to be a significant difference at least from what I saw today in infection related deaths when you compare cord blood to bone marrow which would support a contention that there's no difference in functional immune reconstitution as an impact on the patient. So that would be supporting what you are saying.

On the other hand, Dr. Chao showed data where the TRECs appeared in the subset of patients at one year and not until three years in a lot of them and those were non-myeloablative. That may be changing the rule a little bit.

DR. KURTZBERG: No, actually what he showed was cord blood children and adults after ablated transplant where the children came in much sooner than the adults. Then he had those dots on top of it which was the non-myeloablative showing that in the non-myeloablative cord blood the adults came in sooner.

DR. HOROWITZ: That's similar in bone marrow.

DR. KURTZBERG: That relates to the injury to the host as much as it does to stem cells.

CHAIRMAN SALOMON: Again in terms of engraftment, what I'm trying to get at this morning and now is that I don't think that's an issue here. I don't that engraftment is a big problem in thinking about hematopoietic stem cells of cord blood, bone marrow or peripheral blood derivation.

DR. WAGNER: I don't know if I'm comfortable with that statement.

DR. HOROWITZ: Hematopoietic recovery is slower. I don't think that there's anyone in this room that would say that.

DR. WAGNER: It's unequivocal.

DR. HOROWITZ: Unequivocally slower. Hematopoietic recovery is slower but long term immune reconstitution probably isn't different but that's outside my area of expertise.

DR. WAGNER: Based on the data that I showed, if you look at those patients with an adequate cell dose, the transplant related mortality or the non-relapse mortality was low. It was only 20 or 25 percent as long as you had an adequate cell dose.

DR. HIGH: That's the point I wanted clarification on. If you correct for the dose, is there a difference in time to hematopoietic reconstitution? There is.

DR. KURTZBERG: Yes, it's longer to neutrophils and it's longer to platelets. Once it gets there, it's durable. The differences are in weeks not in months and years.

DR. HIGH: We're edging in on cell dose.

DR. KURTZBERG: There is a relationship to cell dose. But regardless of that, it's still delayed.

DR. TSIATIS: Was there an analysis shown that showed the relationship of survival to time from engraftment?

CHAIRMAN SALOMON: There was actually the reverse. It didn't come out in the multi-variate analysis.

DR. KURTZBERG: I don't think anybody has ever done it quite that way. Maybe Pablo has. There's definitely a relationship. I think Gluckman's looked at 60 days to an ANC of 500. Other groups or programs have looked at 42 days to an ANC of 500. Although engraftment occurs past that point for some patients, the mortality is higher in that group.

DR. WAGNER: I've never really analyzed it that way so I can't tell you. But we could.

CHAIRMAN SALOMON: Dr. Rubinstein, do you have any comment on that?

DR. RUBINSTEIN: I think Joanne's statement is perfect. I will sit down.

DR. NOGUCHI: I think our experience with peripheral blood stem cells and mobilization would suggest that the data really are inadequate to try to get at any meaningful differences in overall survival but that with peripheral mobilization and selection you do get delayed engraftment. It may or may not have any major consequence on the long term outcomes. I think that this data does appear to be along the same line. So I guess the question is, at least from what we are hearing, is that there is nothing major that jumps out at you that would suggest that delay of engraftment is some reason to steer away from cord blood versus bone marrow at any point.

CHAIRMAN SALOMON: Right, but I would say, again could be argued with, from my impression here and also just based on my experience as a physician that you have to face the fact that to anyone undergoing a cord blood transplant is going to engraft slower, not talking about anything specific but the ANC, by platelet count, etc. That's going to probably contribute some morbidity and possibly some mortality.

It doesn't come out in the big statistics and that's important but it's going to mean that they are going to get some more platelet transfusions maybe. It may mean that they are going to longer antibiotic prophylaxis maybe. It may be more than one readmission or a longer primary admission. There will be price to pay, not a monetary price, possibly, but --.

DR. TSIATIS: I just feel that's a real jump without actually looking at and doing a formal analysis of that relationship.

CHAIRMAN SALOMON: I don't think it's a big jump based on the fact that that is what happens in hospital care. I do this all the time with organ transplant patients. I'm going on 20 years of experience here.

DR. TSIATIS: I don't know the data and I don't know the area. I can see that time to engraftment is going to be important but when the relative relationship of time to engraftment for cord blood is the same as the relative relationship with bone marrow, again that's not clear. To make a jump that one is slower than the other means that it's going to have some bad effects again is a jump.

DR. KURTZBERG: It does have a monetary effect. The first in-patient stay is longer. It costs more. In our setting, our average length of stay for an autologous transplant in the pediatric population is 34 days, and that includes the chemotherapy. For a matched sibling it's usually 39 or 40 days and for cord blood it's 48 days. You used to be as high as 60 to 70 days. We've brought it down with an active outpatient clinic.

CHAIRMAN SALOMON: Remember, Butch, being the hospital is bad.

DR. KURTZBERG: It would cost more money too.

CHAIRMAN SALOMON: Anything that keeps you in the hospital longer is bad.

DR. WAGNER: Just two points. Remember this is still cell dose issue. Those with the lowest cell doses have the very prolonged recovery periods in terms of neutrophil recovery and they're the ones that do badly. I'm sure that somehow this very slow rate of recovery in that cohort that's at the lowest cell doses, cell recovery is playing an impact upon survival. So I bet you're right. There's an impact that's predominantly affected by the lowest cell doses. The other thing is if you look at the impact of cell dose in those that engrafted, the cell dose effect is still there.

CHAIRMAN SALOMON: So cell dose.

DR. HARLAN: Just one comment. I'm still not exactly clear despite what you've said and what Phil Noguchi said as to what the product of this is going to be but I'm envisioning now some document that could say something like starting from where we started today that umbilical cord blood is equivalent in all ways to traditional bone marrow transplant with these exceptions: where we can have a bullet that says for instance it looks like you can relax stringency slightly on HLA match; it looks like graft versus host disease incidence is no greater and maybe less, you have to have that proviso; and the third one is this last issue of that time to engraftment appears to be consistently longer, whether it has any relevance or not we can't say but these are observations.

CHAIRMAN SALOMON: That's about where we are right now. Yes.

DR. RUBINSTEIN: This is just about the last point that Dr. Wagner made. In the overall data, when we do a multi-variate analysis after engraftment, cell dose is no longer a significant predictor of survival. If you take the patients not after the transplant but after they engraft and then you ask the question is cell dose important, in the overall data it isn't.

CHAIRMAN SALOMON: We are now talking about cell dose. This is what everyone has been waiting for. The curves that show these family of curves with improved survival, Dr. Rubinstein, you are saying is all engraftment driven. Comments?

DR. KURTZBERG: I think there is no question that cell dose influences engraftment but cell dose also influences survival.

CHAIRMAN SALOMON: So why do you say that and he says that?

DR. KURTZBERG: I think he would say that too but I'll let him talk for himself.

DR. RUBINSTEIN: If you do the multi-variate analysis starting from the transplant and then of course cord blood the survival is very dependent on the cell dose because you must engraft if you are going to survive. Immediately there is some difference in pieces of the data in both directions. But once you engraft, the cell dose is no longer a predictor.

CHAIRMAN SALOMON: That is not what you said, Joanne, if I heard you right.

DR. KURTZBERG: To me doing that kind of analysis is the same as saying you are only going to now analyze the leukemic survival after achievement of remission. And I don't think that's a fair analysis. I think you have to analyze from the day you start the therapy which is before you start the transplant.

From that point, cell dose influences both engraftment and survival and even after engraftment. In our data if engraftment is delayed, survival is poorer even if that patient engrafted. I don't think Pablo's data shows that. But I don't think it's a valid analysis either way. I think the analysis has to be from the start of the therapy and take in all competing risks. If you go from start of therapy then cell dose impacts both engraftment and survival. That's the more important analysis to do.

CHAIRMAN SALOMON: Mary, did you want to comment on that?

DR. HOROWITZ: We haven't looked at the data from time to engraftment. John, I thought you said that even among patients who engrafted, cell dose made the difference.

DR. WAGNER: When we looked at our dataset as Joanne just said, we both looked at Duke and Minnesota but also more recently just looking at the Minnesota dataset and the impact of cell dose occurs both from the very beginning as well as from those that had already engrafted. If you start the time clock after engraftment, there is still an impact of cell dose. My guess is those with the very delayed recoveries even though they recover they still are negatively impacted by cell dose. My interpretation is still pretty much the same.

DR. KURTZBERG: We have analyzed immune reconstitution as it relates to cell dose using one year as the endpoint and cell dose impacted immune reconstitution as well.

DR. CREER: If you look at the survival curves too, you never really see that they separate out based on cell dose until at least maybe 60 to 90 days which is the time beyond engraftment. They really don't begin to disperse until after. I would say that even though there is delay in engraftment, that delay is not the factor that is contributing to the death. The problem determining survival is occurring after that 60 to 90 day period.

CHAIRMAN SALOMON: Let's take a different view of the cell dose because I don't think this is the most critical thing. The part of the cell dose is are there recommendations we want to make in this discussion to the FDA that gives them some sort of a sense of what is an appropriate cell dose and then think about the pros and cons of that sort of a thing because it has to do with the size of the patient that one could do and the availability of the numbers of units. If you set a really high cell dose, you decrease the number of units that are going to acceptable. This is an awfully critical piece of the conversation. I would like to focus on that rather on this rather tenuous engraftment issue.

DR. LITWIN: Could we perhaps as we are entering this discussion talk about numbers in terms of cell dose? That would be useful to see where people are.

CHAIRMAN SALOMON: That's where I'm going. I hope that's what I meant. That didn't come out right. That is what I meant. Let's have some discussion about numbers here with the realization that you are going to pay a price when you make the numbers. The good news is you will get better survival to a certain point but the bad news you will decrease the number of useable units. Mahendra is saying let's start with Dr. Wagner's number. You guys messed around with me a little bit because some of you were talking about per kilo and then Dr. Rubinstein was showing total. Your best data was 100 million total.

DR. WAGNER: That's per kilo.

CHAIRMAN SALOMON: One hundred million per kilo?

DR. KURTZBERG: In 10 kilo kids you can get 100 million.

DR. WAGNER: Yes. Basically again what I would say is we should step them back and say what's unacceptable and work our way that way. Therefore I would say that certainly anything less than a dose of 1.5 times 10⁷ per kilo is unacceptable. Now remember I'm talking about nucleated cells now, not CD34, because as I said in my talk CD34 in principle is a better measure. On the other hand, it's not practical yet.

CHAIRMAN SALOMON: We'll get back to the CD34 in question four. So 15 million or more is what John is putting on the table per kilo.

DR. WAGNER: Per kilo, yes.

DR. KURTZBERG: I would actually say for adults that's reasonable but for children I would make the bar higher.

DR. HOROWITZ: I would make it higher for both.

DR. KURTZBERG: For children, I think it should be 25 or even 30 and for adults I would say 15.

DR. WAGNER: That doesn't make any sense. If a cell dose is acceptable, it's not going to be different only because otherwise you could limit the use of cord blood for adults but that may be appropriate. I mean the fact is why would it biologically have to be higher for children than adults. If anything it would be the other way around.

CHAIRMAN SALOMON: I can answer that but do you want to make a comment?

DR. McNIECE: Ian McNiece, Johns Hopkins. I would really warn against going to some of these minimum doses. We have learned a lot through bone marrow and peripheral blood stem cell transplants. With PPCs we have optimal doses. We don't have minimal doses because the number of the grafts even with lower doses still engraft. If you look at the Kaplan-Meier probabilities, a number of those grafts still reach neutrophil engraftment within 20 to 30 days. We ran a study with expansion at the University of Colorado. Our median dose was 0.85 times 10⁷ per kilogram. We had engraftment in all patients. We gave as low as 0.28 times 10⁷ per kilogram. We still got neutrophil engraftment. These were expanded products. The quality was somewhat different. We are talking about different things.

I would like to see a recommendation for an optimal dose but I would certainly warn against giving a minimum dose. As we have heard, a lot of times you don't have an option in terms of the graft. So you are going to prevent people going forward with getting a transplant because of your cell dose. What is their option?

CHAIRMAN SALOMON: Again, this comes back to what Phil said. We're not trying to say acceptable and unacceptable. We're just talking about optimal.

DR. McNIECE: Minimum is different than optimal.

CHAIRMAN SALOMON: That's right. You're right. We should talk about optimal, not necessarily use the word minimal. That implies an unacceptable dose. Rich.

DR. MULLIGAN: How about if we just first build a non-controversial dose without using CD34 that everyone agrees on? What would that be?

DR. KURTZBERG: That's the discussion we are trying to have.

CHAIRMAN SALOMON: That's what the people are saying. We have 15 million per kilo on the table as an optimal dose possibly. Then Joanne and Mary raised the bar on that thinking it ought to be higher and John saying he didn't think that you ought to play with the biology like that. If it's optimal, it's optimal.

DR. KURTZBERG: But 15 million is not optimal. I have trouble with that. Fifteen million may be the lower threshold but it is not optimal.

DR. WAGNER: You know that's not what I'm saying.

DR. KURTZBERG: I know but we just have to be careful how we choose our words.

DR. WAGNER: What I said was specifically that working our way from the bottom up I think that everyone would agree that 1.5 or below is unacceptable. I'm not saying that above 1.5 that there might not be a better level. I'm just working my way up to figuring out what that threshold should be. The data that I presented would suggest to you that a graft failure rate of 30 percent and a time to recovery of 32 to 35 days is pretty poor. Therefore routine use of cord blood -- that cell dose should not be considered for routine use.

DR. KURTZBERG: I agree with that. I also think you have to realize that people have options sometimes where you might have a four-of-six unit that gives 50 million and a six-of-six unit that gives 10 million. I'm not saying we have all these answers but a traditional transplanter who has no knowledge of the impact of cell dose is going to pick the six-of-six unit. That to me is substandard because of that cell dose.

CHAIRMAN SALOMON: That's excellent. We do have to realize that now bringing in the idea that there is more than parameter to deal with is critical. Alison.

MS. LAWTON: I have a couple of comments. I thought that I wrote down here and I wanted to check if this is correct that, Dr. Rubinstein, you presented data that showed at below 25 million it was statistically significant that you did get a drop-off below that dose. Then the second comment I would just make which is slightly different is this is obviously a very difficult discussion on the dose. Maybe one way to go is whatever dose or whatever discussion we come up with is that again it becomes part of informed consent for the patient because if that patient still doesn't have a choice they may still have a better chance with a low dose than having no dose whatsoever. As long as they are informed of that, maybe that's the way to go.

DR. RUBINSTEIN: I think those points are excellent but I wanted to say that in discussing the dose, it's very important to make a precision whether we are talking of the dose as seen in the bank before it is frozen or in the clinic when it is thawed. You should anticipate some loss there. In our experience and I showed the curve that represents all of the replies to us from the transplant centers, the data shows that apex of that distribution is about 70 to 75 percent, somewhere in that area. The estimate of 15 million that John is making maybe more or less equivalent to the threshold of 25 million that I was talking about.

CHAIRMAN SALOMON: So for clarification, I'm glad you brought that up. I was thinking that when you talk about 25 million, are you standardized to that's the time of infusion after thawing? That's what I was thinking.

DR. WAGNER: But remember you don't that up front.

DR. KURTZBERG: Let him answer that. Your numbers are at banking not at thawing.

DR. WAGNER: No, it's infused.

DR. KURTZBERG: I'm asking Pablo.

DR. RUBINSTEIN: When we represent to you the numbers in a unit, we tell you the numbers that are pre-freeze.

DR. KURTZBERG: Right. As far as I'm concerned, that's the only numbers we can work with. I know what you are saying and you're right. But when you select a unit for your patient, you don't know what it's fly out at. You have to select it by what's there.

CHAIRMAN SALOMON: I'm making a point here kind of dumbly. I want to get this clear. When John presented us data about 15 million per kilo, he was talking about thawed. When Dr. Rubinstein presented us data that showed 100 million had the best results, he was talking about banked, pre-thaw.

DR. WAGNER: No, this is actually a critical point that I was misleading you to think that this is the cryo-preserved cell dose. It is in fact the infused cell dose. So in fact it may then turn out to be closer to the 2.0 to 2.5.

DR. KURTZBERG: This is a very slippery slope because transplant centers get different recoveries. People with expertise in thawing may get a better recovery. Different banks have different ways of counting and the correlation. I don't think we should go there.

CHAIRMAN SALOMON: I think we have to. I don't think we can not go there. But what I'm hearing you say is that it should be based on the bank number.

DR. KURTZBERG: Right, that's what I'm saying.

CHAIRMAN SALOMON: That's what we should stay with is the bank numbers. That's what I hear you saying.

DR. KURTZBERG: That's what I'm saying.

DR. WAGNER: And I am saying the same thing. I agree with that. I wasn't thinking about that when I told you the 1.5. I was giving you data that was on infused but I agree with you entirely. It should be based on the bank number and not the infused number because you don't have that up front.

DR. CREER: This gives you an idea of the difference between the two. If you look up there between all the units banks between 0.6 billion and over 3.0 billion in our bank, the inventory distribution and the selection of units based on their total cell count in each product and if you look at the actual mean TNC dose, it's pretty much the same.

The products are being selected by the transplant centers to achieve a specific TNC dose. That median dose in the range between 6.8 and a low actually of 4.8 to a high of about 7.3. Those are much higher numbers than you saw in any studies presented today because this is post processing, not post thaw. That's where the big difference has been. We have a median CD34 dose that we send out that based on post-processing data that varies from a low of 3.7 to 6.8 which is well above all the thresholds that you saw.

What I'm saying is that the data you've been seeing for outcomes correlation is based on post-thaw information. That difference is very significant with what we show them and what they are selecting post-processing. And that's the problem.

DR. KURTZBERG: Now Pablo's data is all post-processing, what went in the freezer.

DR. CREER: Right.

DR. KURTZBERG: John is saying that his data was post-thaw which was only 100 patients. All the other data you saw today was post-processing.

DR. CREER: And you get consistent recoveries on TNC but not on CD34.

CHAIRMAN SALOMON: We'll get back to the CD34 thing.

DR. KURTZBERG: Actually on TNC in the 500 units we thawed, we've recovered TNC from a range of 65 percent to 98 percent. It varies somewhat on banks, somewhat on transportation and somewhat on I don't know what but it's the same staff doing the same procedure.

DR. CREER: Single institution experience. When you look at every institution, it varies from a low of 20 to a high of 90. What I'm saying is in a single institution you have very precise recoveries, very precise calculations of CD34 and TNC. When you tease out the data and look at correlation of outcomes by single institutional experience, you get a different view than you do when you look at multi-institutional studies where the variability and recovery changes things. That's the difference.

Dr. Wagner sees a very clear correlation with CD34 because he gets consistent recoveries at his institution. When you put multiple centers in there who calculate CD34 using different techniques, get variable recovery post-thaw, you can?t find a correlation because the variability in the data is so great. That's the problem.

CHAIRMAN SALOMON: Now we're talking about CD34 again which we will get to in a minute.

DR. CREER: Well, it's the same for TNC but it's a little more consistent.

CHAIRMAN SALOMON: Alison.

MS. LAWTON: I agree with you. I think we need to have a separate discussion on CD34 but just on the TNC and just given the discussion, it seems an obvious thing, and I know it's one of the things that we'll get on to discuss, some standard data that we want to collect across all centers and across everything that's done so that we start to see that pre-freeze and post-thaw information and start to try to understand so that people going to those centers and getting the units will understand what they are getting in the future.

DR. McNIECE: I just come back to trying to put any minimum level. I think some of the comments made earlier were we are going to try to inhibit research. Coming back to the study that we did at the University of Colorado with expansion, one of the problems we had with the FDA -- we had to go through extension to get discussions with them was that those very concerns about the minimum doses going back. Our studies clearly demonstrate that expanded cells do behave differently than unexpanded at least at cell doses. If you start to put down minimum doses again, you are going to inhibit research. I think you need to be careful.

CHAIRMAN SALOMON: Nobody is talking about minimum doses and no one is talking about expanded cells which is not on the table today. That's going to be a separate discussion.

DR. McNIECE: I understand that but I've heard 15 million. I've 25 million being raised as minimum levels for safety. I'm just saying if you start to put those doses down that's going to inhibit what research projects can be done to look at improving engraftment and other parameters because it's going to be impossible to do other things. We don't know what cells in those grafts truly give you engraftment. That's the whole problem now. We're dealing with gray matter that we can't define.

CHAIRMAN SALOMON: If what you are saying is I don't want to talk about cell doses because it might limit future research, then that point is made. That's not what we are doing right now. We're discussing cell doses. But you can go on record as saying that you don't want to talk about cell doses because it makes you nervous that it might affect research that you want to do and that's fine. I heard it.

DR. McNIECE: Okay.

DR. LITWIN: The FDA in the past in connection with approving cell separation, cell selection devices has used the term and it's used the term "target." The target dose is not a minimal dose so it does not preclude using lower doses. For those in the field who know those doses of 1.4 million bone marrow per kilo, 2.0 million for peripheral blood, I think it's well appreciated that these are certainly not maximal doses and that you probably could do better. We, faced with the issue of having to license a device, use the approach which may be somewhat similar to what John was recommending that is if we looked at the data which was limited but enough, we saw that below 2.0 million per kilo for peripheral blood there were many bad results. There was a clear difference.

Since the FDA is safety oriented in almost all circumstances we felt that we made a reasonable choice as it's held up reasonably well in the field. Although most peripheral blood transplants now are probably three to five million, that two million per kilogram figure I think did protect or at least give good advice. I would recommend that we use the term "target" dose which would not constrain anybody without trying to imply that we're giving the best possible dose under the best possible circumstances.

CHAIRMAN SALOMON: Rich.

DR. MULLIGAN: Another way of looking at it is that maybe we just give the information that we heard here which is essentially here's the results. You give this amount and you get this survival percentage. You use this amount and you have less. I don't think anyone would be trying to say that you wouldn't do those transplants that were already done for which we have statistics. I'm not sure we have to go past that. We now have a body of information that says you get this level of survival, this dose and you get this level here. Based on that, you make the decision.

CHAIRMAN SALOMON: Let me try and synthesize what we've talked about so far and then we can tear that apart. I think what we all agree on is that dose is a critical thing. Whenever something is going to develop in this field, dose is going to get put front and center and it's appropriate. That's one thing I think we all agree on.

Second, clearly no one in this field is ready to say that there is a given dose below which you shouldn't do a transplant. Everyone has given a different version of why you can't do that but the bottom line is that.

Your version was that it might inhibit research into expanded cells. That's a very good point because expanded cells could have totally different dose rule.

Another reason might be that adults would be dying and their only option would be a transplant. It might very well violate an optimal dose range or a target range to save a life and that would be very appropriate. That's a good reason too.

Certainly size, the gentleman who was 239 pounds, he is standing there as a slimmed down version of him but congratulations on that too. Certainly none of us want to restrict a transplant on that basis. I think there are a lot of reasons not to put any fixed number here. I think we can tell the FDA that number is really important and that to the extent that one can do it in the situations choosing a higher number would be really of value except where the choice becomes more complicated like a higher number but a worse HLA match and that it's not going to be a single parameter decision but one that might need to be integrated with more than one parameter. Is there anything else to add to that?

DR. RUBINSTEIN: Really to support that since dose and HLA are completely independent estimators and they involve different biological reasons or mechanisms on how they determine, their effect is independent. We have however tried to look at some sort of a equivalence to understand when should we choose a higher dose and when we should choose a better HLA match. If you compare the group that had the five-out-of-six, the group that had the four-out-of-six match which are the two numerous groups in this analysis and then you sort each one out by dose, it is remarkable that at least in the data that we now have which doesn't mean that it will be the same in another setting but for this data 25 million cells make the difference and four-out-of-six becomes the same or gives the same survival curve to five-out-of-six. If you have 25 million then you are equivalent in terms of the overall survival.

CHAIRMAN SALOMON: From what I've heard, I'm comfortable with adding to what I said that 25 million integrator per kilo would be a good target dose.

DR. HOROWITZ: And that's only thing that I wanted to add. I don't think we should be afraid to say that 25 million is a target dose. We have seen a lot of data to suggest that there is some kind of threshold effect somewhere in that range and we shouldn't be afraid to put it on paper as a target.

CHAIRMAN SALOMON: So there's two things. One is 25 million as a target dose. Twenty-five million as a target dose translates to 15 million after you thaw it out. Then maybe I would be happier with 50 million as a target dose and everything else would be covered by informed consent.

DR. NOGUCHI: I think you are hitting on upon some of the nuances of what we really needed to set in the future. If what you are saying is that's based on the post-processing but pre-thaw, then I think that should be explicitly stated. It appears that the dataset in terms of post-thawing is incomplete.

CHAIRMAN SALOMON: Remember I said there were two things that were really bothering me. We'll get to the post-thaw in a second.

DR. NOGUCHI: Make sure of that.

CHAIRMAN SALOMON: What John is saying was that his was 15 million per kilo post-thaw. We made a rough calculation that 25 million post-processing was probably reasonable to say in a big center trial was 15 million post-thaw. If that's true, I didn't love his data in the 15 million post-thaw group. It was okay but it was much better in what I would consider the 50 million post-processing group or the 25 to 30 million post-thaw. I'm just trying to get this out. I'm talking about target dose. I'm not trying to set any lower limit.

DR. KURTZBERG: But in this world that you live in, you are not going to know your post-thaw recovery. There is some variability for a variety of reasons that we don't have to go into. So you have to deal with what went in the bank and use that to base your decision.

CHAIRMAN SALOMON: I'm just saying I would rather see us set a target dose a little higher than 25 million.

DR. KURTZBERG: But I think the worse case average number for post-thaw recovery is 70 percent of the cryo-preserved count and that's being very conservative. I think the math that you are doing is a little too conservative.

DR. HARLAN: Could I try to elevate what Richard Mulligan said again? Would anybody have a problem with saying below this number the survival is this poor and just leave it at that? That way you say below it, it's really bad. You should as well as you can above that. I like that.

CHAIRMAN SALOMON: I'm okay with stopping here and just saying it's all on the table to the FDA. It's really not important at this point for us whether it's 30 million. I agree with that.

DR. MULLIGAN: Maybe just to preempt the CD34 discussion if we agree on this, what do we have to talk about CD34 about?

CHAIRMAN SALOMON: We'll get back to CD34. So we are sticking with the 25 million post-processing. Now I said there were two things that bother me and Phil picked up on it too. I don't think that we can dismiss what the gentleman was saying from the St. Louis Blood Bank that there really are in his experience very large variations in cells count from post-processing to post-thaw.

What I would add now to this almost complete piece is that going forward we ought to really be paying attention to the reproducibility of the post-thaw. Joanne, what you are saying is that you have experience so you are comfortable with this and that's great. I think as an overall thing to the field there has to be some assurance that this is really true. If it's really true that their center is getting 20 percent yield I would be very uncomfortable with them selecting and then I'm not sure what to tell the FDA as to why are they getting 20 percent yield. Is it something to do with the processing, the shipment, partial thawing as Dr. Rubinstein mentioned or is it problems in these institutions?

DR. KURTZBERG: I know that Pablo from the start went to each institution for their first thaw and probably still does that because you can read about that but when you do it, it's different. COBLT actually adopted a protocol for training for thawing and it's a certification protocol where three units have to be done that are not going to be transplanted to patients. As a person who lives on both sides of the fence, I strongly recommend a thawing training certification requirement for any center that's going to transplant.

CHAIRMAN SALOMON: I totally agree and that's where I was going. That's probably one of the most important things we are telling the FDA right now. If they are going to go forward with this, they are going to have to in some sense certify not only the processing end of this but the post-thaw end of it in the institutions. That's reality for cell transplants. That's fine. Rich.

DR. MULLIGAN: May I add one thing and that is I would go with where the data is. I mean the data that we are basing everything on comes from a certain place. Then the data depends presumably upon the method that was used in that case so that would be a benchmark as opposed to just some other methods that has not been attached to some any statistics.

CHAIRMAN SALOMON: I agree with that too in the sense that we want to consider. Yes, I remember that data. I think we have this one. Does anyone want to add anything to it? Have we missed something? Mary, are you comfortable now? Do you think that we de-emphasized at the end the number?

DR. HOROWITZ: That's okay. That's fine as long as we make it clear that there is an optimal dose.

CHAIRMAN SALOMON: There's a target dose.

DR. HOROWITZ: There's a target dose. There may be an optimal dose but we don't know what it is. Most of the data that are available in terms of numbers of patients studied suggest that 25 million is a reasonable target dose. The only thing I want to make sure that we address is that if we are going to be able to refine optimal dose, we are going to need to have outcome data. So I hope that we will make that recommendation in addition just certifying that transplant centers know how to file the product, we need to be able to collect and analyze the outcome data for their patients served.

That's another reason why we can't find correlations. If you only have outcome data on half of the transplants you've facilitated and you don't really know whether that's a representative sample or not, then you may find correlations that are there.

DR. KURTZBERG: I totally support that but I think the FDA could do something helpful in this arena which is that there is no mechanism within the health care system to support the collection of that data. If it is required as part of standard patient care which I think it should be for transplants, then the FDA can say this is part of the standard of medicine in transplantation. Then the insurers may in the long run or the hospitals may in the long run realize that they have an obligation to support the collection of this data. Right now that doesn't happen. Everybody scrambles in one way or another to make it happen. That's one of the reasons that feedback is not as good as it should be.

CHAIRMAN SALOMON: The comment I have and then maybe Phil or somebody else from the FDA wants to comment. Often times as you know Joanne, we come to these situations in which we say we need a registry. We need database. We have to be really careful that our recommendations unless you really want them to be are that the FDA should have a database because I don't think that's what you were saying. They don't have the funding to do that. These are what we call unfunded mandates and they are scary.

DR. KURTZBERG: The IBMTR has a database.

CHAIRMAN SALOMON: Right. The FDA can insist however -- Well the FDA should comment for themselves.

DR. KURTZBERG: But for me as a transplanter to provide that data back to whoever there is no programmatic support to do that unless it's a research project supported by a grant. But as part of routine day-to-day hospital care, the collection of data is not part of what they fund. They should.

CHAIRMAN SALOMON: We agree with that. The question now is -- that's the right thing to say. What that has to do with the FDA though is what I'm trying to put into context. The FDA is not going to maintain the database and they are not going to fund the database and they are not going to fund the collection of the data. That much I can say after five years of doing this. The question is could the FDA somehow say that if you want to participate in, down the road of course, a regulated or licensed process involving cord blood, could you stipulate that it would have to be shared with fill in the blank database which might be the IMBTR?

DR. NOGUCHI: I think all things are possible. Part of the gestalt that we're trying to develop here is what the impediments to getting the data for which it appears that there is pretty good support within the community to provide but the mechanisms are somewhat vague.

Sometimes we can't even find the wherewithal to pay our own employees so yes, it's difficult for us to go beyond even our own basal activities. However, the point is well taken. Our actions do have effects, they can change attitudes and what we want to do is to do this in judicious fashion.

If we can just go back to number because every time you talk about 25 million, Dan, you neglect to add that that's post-thaw that comes from the bank.

DR. WAGNER: Post-processing.

DR. NOGUCHI: Excuse me. Post-processing coming from the bank. Dr. Kurtzberg is correct. Her institution is very good but the goal is to provide this at many different institutions, not all of which are going to be at the same level of confidence. I just have to say from our point of view we would be thinking about making sure and in terms of the certification that's certainly a part of it but data is also a part of it too. How can you begin as a physician to analyze where the deficits in a cord blood comes from unless you actually know how good your thawing procedure is? I do want to make that statement up front that we think a part of this issue is the responsibility of a transplanting center as well.

CHAIRMAN SALOMON: In fact, let me go back and say it's my fault then if it didn't come across clearly. We were saying specifically to you that this is important that you have to have a way of certifying that the transplant center is getting at least 70 percent yield on thawing. If not then there are systematic problems on the transplant center side. We are specifically recommending to you that you see the problems both in the processing and the thawing of the transplant side. Yes, we totally agree with that.

DR. MCMANUS: John McManus, from M.D. Anderson. Related to that, now we get a number of different cord blood from many of the different banks. Many of them tell us how they want us to thaw the product. So will we have to get certified in each one of those procedures or do we just go with one standard one and say we no longer will do St. Louis's procedure on how to thaw bank products?

CHAIRMAN SALOMON: That's a great point. There probably ought to be said at the beginning of something like this at the time when the FDA is really going to step up to regulate something like this, you are going to have to set a gold standard and then you can use whatever thawing procedure you want as long as you adequately reassure the authorities that it fits the gold standard.

That's what we are doing in islet transplantation right now where you have to show that you can get the right number of islets and they are functional and cure a couple of patients. Then after that you can argue to deviate from that in your IND.

DR. NOGUCHI: The idea is to identify those specifications that are meaningful in terms of outcomes. We have heard some numbers thrown around and we're not asking for what those numbers should be. I think that the goal of post-thaw, you should get as high a viability as you possibly can just as a general principle. Those are the principles we would suggest you start from. If it's 20 percent I would say post-thaw, likely everybody would be really concerned about what processes are being done.

CHAIRMAN SALOMON: Unless the committee wants to disagree, we're telling you that it should be around 70 percent or better if you want to talk about a 25 million post-processing number as a target.

DR. NOGUCHI: I think if you are willing to do that, that's great.

CHAIRMAN SALOMON: I think that's what we're telling you. I'm a little bit cognizant now that I'm under the pressure to get this done here which is good. I love this. That's what I live for.

Question number two deals with age. We've had discussion but we haven't yet talked about age. There are two ways to go with this to set it up. One way is to say that there's no evidence here that age is the determining factor as long as you get adequate dose. If that's true, we're done.

DR. KURTZBERG: Second.

CHAIRMAN SALOMON: I love it. That's good. Does anyone want to take that alternatively? Is there a minority opinion on this?

DR. HARLAN: I agree actually. The only caveat is there was this question about the increased incidence of GVHD and should we mention that. That the older the patient, the higher the incidence of GVHD.

DR. HOROWITZ: You can say age does not appear to differentially impact on cord blood transplantation and bone marrow transplantation as long as cell dose is adequate.

CHAIRMAN SALOMON: All right. Question number three, this is really important. What do you want to see for outcome parameters for trials? This gets back to the argument Mary and I had about she likes binary outcome parameters, live and dead. I understand. If I was running something like the IBMTR, I would probably be happiest with that. So I understand. What kind of outcome parameters besides survival? What have we talked about? Disease-free survival/survival I guess is the same thing. Time of engraftment.

DR. HOROWITZ: Time to hematopoietic recovery.

CHAIRMAN SALOMON: Time to hematopoietic recovery as opposed to ANC, platelet and lymphocyte counts.

DR. HOROWITZ: Graft versus host disease, acute and chronic. Immune reconstitution.

CHAIRMAN SALOMON: Well immune reconstitution you don't want to use that term because you don't have a number for that. You don't do MLRs or antigen stimulation tests. You have lymphocyte counts, leukocyte counts and platelet counts.

DR. HOROWITZ: If we're talking about clinical trials. If we're talking about clinical research that's different from all the other things that we talked about. I think that it is hard to just say a point estimate for immune reconstitution. There are a variety of things that came be looked at in immune reconstitution. But I think if someone is really going to do clinical research in this area, that group has to consider some parameters of immune reconstitution. What's hard to say is what they should be in every single trial.

TREC are one but they are not going to be done in every center. I don't think you can ignore the immune reconstitution issue. I would say that all of these outcomes are outcomes that have to be considered in all trials of allogeneic hematopoietic stem cell transplant not just cord blood transplantation.

CHAIRMAN SALOMON: That brings actually another view of this and that is there is nothing in an outcome parameter that we currently have missed in the field. There isn't any unique outcome parameter for cord blood transplant. I'm just putting that out for argument.

DR. HOROWITZ: Except for those processing things.

CHAIRMAN SALOMON: That's good. Let's say we're going to come back and talk about the processing things because that's really important. How about outcome parameters for the clinical trials not related to processing? Mahendra.

DR. RAO: One thing that seemed to be important given what has been reported as this whole issue of transdifferentiation if it's going to be studied by a prospectively collecting data, should you be considering looking at this liver contributions or looking at some other tissue?

CHAIRMAN SALOMON: That's a liver biopsy. Is that what you are suggesting?

DR. KURTZBERG: I don't know any IRB that's going to approve screening liver biopsies to look for hematopoietic cell differentiation in that kid who's doing well. You could say that in livers that are biopsied for medical indications you could also study it. But you are not going to get that as a standard follow-up procedure.

CHAIRMAN SALOMON: I actually want a piece of that guy's coronary artery if he's still here but I don't think he's going to let me do that either.

DR. HOROWITZ: Those studies are hard to do trying to figure out what's going in a recipient in a methodologically sound way. I don't think that's something you can say across the board.

CHAIRMAN SALOMON: But his donor cells were actually X-chromosome. His are Y.

DR. KURTZBERG: I have pictures of autopsy brains showing donor cells in the brain with X, Y if you want to see them.

DR. HOROWITZ: But do you know whether they are from fusion or from real transdifferentiation?

DR. KURTZBERG: They are confocal microscopy and they look like they are not fusion. I don't think we should go into that here.

CHAIRMAN SALOMON: So what I'm hearing is that even though we can all think of some of these cutting edge sorts of things, those are going to be very difficult and we're not suggesting that the FDA start sticking those in. That would be going too far. Alison.

MS. LAWTON: One question I have and I'm certainly not an expert in this area but I know it's been discussed in detail is would you want to make sure that you collect the information on the high resolution typing to try and understand that further in your outcome data?

DR. RUBINSTEIN: I think it will be difficult to do but I would like to have you consider the possibility of asking for data on donor type of the reconstituted hematopoietic system. In many places, they just see the reconstitution and consider it a success. It is not necessarily so.

CHAIRMAN SALOMON: So you are talking about percent. I know this is another word that you like. Percent chimerism when it's 100 percent one. That's what you want. I agree with that.

DR. MULLIGAN: That's a very important thing. In a mouse system, there is definitely a well known thing where if you leave the irradiated recipient and you don't give back any bone marrow, you die. When you give back a low dose of bone marrow, you will get radio-protection but you'll get largely host bone marrow reconstitution and that was exactly the question I was next going to ask. Have people looked in detail to see in fact whether long term reconstitution dose effects the amount of actually donor reconstitution?

DR. KURTZBERG: Most of the centers doing this get chimerism on a regular basis. I can only speak to my own data but we get chimerism every three months the first year and yearly thereafter. Patients who have 100 percent donor chimerism maintain 100 percent donor chimerism.

DR. MULLIGAN: And do the survival statistics that we saw depend upon the percent of donor reconstitution?

DR. KURTZBERG: In my data, people either engraft or they don't engraft. We're not seeing mixed recovery except if you look two weeks post-transplant you may see a mixture of patient and host. But everybody is achieving 100 percent. So I can't really say that it impacts a lot of them.

CHAIRMAN SALOMON: Everybody is achieving 100 percent. That's what Rich is asking.

DR. KURTZBERG: Yes.

CHAIRMAN SALOMON: Therefore in a sense 100 percent engraftment is an outcome variable.

DR. KURTZBERG: It is in the mini transplants and occasionally it's possible that I have a host recovery. That's 100 percent too. It's not usually a mix when you are using ablative regimen.

DR. MULLIGAN: So are you saying even at low doses when they seem to get their neutrophil back it's always 100 percent or not?

DR. KURTZBERG: We are talking about ablative.

DR. MULLIGAN: Yes, but low dose.

DR. KURTZBERG: The lowest cell dose ablative patients who are grafted and graft with donor cells. They are 100 percent donor cell.

CHAIRMAN SALOMON: Interestingly enough even in the experience now with non-myeloablative stem cell transplants and I don't have data on the cord blood but on peripheral blood and bone marrow, it's almost always 100 percent chimerism. Not always. I'm talking about at a year or more.

DR. HOROWITZ: But non-myeloablative peripheral blood transplants contrary to what people expected we don't see many patients with durable mix chimerism. They revert to host. You may have transient mix chimerism. Most of them revert to host or become 100 percent donor in the long term.

CHAIRMAN SALOMON: This isn't published yet. We did 30 -- a unique genotyping method for doing it and it was 100 percent at nine months in 30 non-myeloablative allos. If you did 1,000, I'm sure you'd get some but it's really unusual.

DR. KURTZBERG: One of the things Pablo is raising has to do with more standards of practice in transplantation medicine again disregarding stem cell source. It should be to me a no-brainer to check your cells once your patient engrafts to make sure that they're donor but he's right. Not everyone does that.

These are some minimal standards that might be address in another meeting which don't just relate to cord.

DR. RUBINSTEIN: We could propose to the centers to do those studies for them if they have difficulty in doing it. I wanted to mention that the mouse model with cord blood transplants that is allogeneic mouse donor of cord blood, they can achieve permanent chimerism in the sense that if you do it from two donors, you will get a chimera that will be permanent for the life of the animal. In humans, that doesn't seem to happen. There may be chimerism but it eventually resolves for one or the other.

CHAIRMAN SALOMON: The other thing that I didn't see in enough data would be some standardization of marking acute and chronic graft versus host disease as outcome parameters. Again I don't mean to insult the field and say they obviously know that they need to that. It was interesting that you presented a lot of acute GVHD data but didn't really step up with your chronic GVHD data. That has to be a part of it.

DR. KURTZBERG: Mary can probably speak to this better than I can but there are committees as part of some of the national trials now trying to define the universal criteria for grading and scoring chronic GVHD. That's in progress but again it's related to all stem cell transplantation. It's very hard.

CHAIRMAN SALOMON: In general, what we've said now is we're going to deal with this processing issue now. For clinical trials that essentially I get the feeling that pretty much we are all comfortable with the general outcome parameters for other hematopoietic stem cell transplant trials and for that matter, donor chimerism is a relatively important measurement for any hematopoietic stem cell trial.

DR. HOROWITZ: No, I'm going to have to say something because the clinical trials network and the transplant community have said that chimerism studies for a standard myeloablative allogeneic transplant are not necessary because they are a waste of money.

CHAIRMAN SALOMON: We're talking about clinical trials.

DR. HOROWITZ: In clinical trials, they are not always required. It depends on the end point being looked at. If you talk about bone marrow peripheral blood stem cell transplants, in myeloablative regimen, the incidence of finding recipient cells is so low as to make it not worth looking for. We could say that chimerism studies in cord blood transplants might be something that should be considered as routine that might not always be considered routine in bone marrow transplants.

CHAIRMAN SALOMON: Perfect. Processing.

DR. MULLIGAN: Just on that point, even though we're not going to talk about the manipulation cell in in vitro amplification, I think that's the area where this is going to be very important. When you begin to change the stem cells or you begin purifying different parts of the stem cell population then all bets are off.

CHAIRMAN SALOMON: I totally agree. I'm thinking that this committee will get those kind of protocols in the not too distant future. Kathy.

DR. HIGH: Are we on processing now?

CHAIRMAN SALOMON: We're on processing.

DR. HIGH: I just wanted to get one clarification. In your data about in vitro functionality of ten year cryo-preserved umbilical cord blood and 15 year, is that a settled issue so now if it's 20 years or 25 years it's okay? Do we know?

CHAIRMAN SALOMON: Remember that's based on SCID repopulating efficiencies. It's not in vitro. It's in vivo. But I don't know. That's really a good question. For the stem cell transplanting clinicians, what do you guys think? If you had a routine patient and you had offer of a three year old versus a 25 year old unit or a 15 year old unit. What would you pick?

DR. KURTZBERG: I'd pick the one with the highest cell dose.

DR. STRONCEK: From a laboratory medicine and transfusion medicine point of view, if you haven't tested it, you don't know. When people have looked at it, things seem to cryo-preserve pretty well. So it's likely they'll be viable but you really don't know. You don't know what the fall off is.

DR. RUBINSTEIN: We have some empirical data. Because we started cryo-preserving and freezing this annulated cord blood in 1993 every year we do give a few transplants from units that were collected in the first year. I can tell you that the last three units that we transplanted at the end of last year and beginning of this year engrafted with the same kinetics that a newer unit would provide.

CHAIRMAN SALOMON: How old were those?

DR. RUBINSTEIN: Eight and a half years. One of them was nine years.

DR. CREER: We confirm that over five years, the same thing. No change in rate of engraftment when matched for cell dose and no effect on outcome over five years when matched for cell dose.

DR. RUBINSTEIN: So if this has not changed after nine years then the likelihood that it would suffer an abrupt change some time in the near future is not very much.

CHAIRMAN SALOMON: Liane.

DR. HARVATH: I would like to say that in general I would agree with the caveat that we know that not all cryo-preservation facilities are created equal. Both Pablo and Phil Coelho commented about transient warming events. I think they have some pretty compelling data that if you freeze your product in such a way that it has gone through a number of cycles of transient warming events, you take a hit in the cell viability.

We don't really have the in vivo data to show what that impact would be on the outcome of engraftment. If a facility is incredibly rigorous and they have well documented records of how they cryo-preserve and monitor the status of their products than these data will probably show promise in terms of cryo-preserving for a decade or perhaps more. Many of us know of places where there have been freezer meltdowns and very good facilities. Those accidents happen. We have to be really cautious about the kind of message we send out as a committee making a recommendation that we are very confident that this is always the case.

DR. HOROWITZ: I agree and I also agree with the fact that you don't know until you know. There has to be some kind of program of periodically testing a hypothesis.

CHAIRMAN SALOMON: Personally as a recommendation, I think David said it well. In laboratory science, you don't know until you test it. That probably covers your concerns as well.

DR. KURTZBERG: I agree but testing here is biologic. It's dictated by the practice so it means collecting information.

CHAIRMAN SALOMON: Exactly. In terms of processing, we really would all agree that if you are going to get into this in terms of regulating licensing, INDing you are going to want to know this whole pathway. That's going to be the processing step, the freezing step, the quality of the long term freezing, some of the points of view you made. I see my post-docs come in to our liquid nitrogen farm and pull the whole rack out and set it up there and talking on the phone, 200,000 cells go into that. There are a lot of issues here.

Another issue would be shipping. I've done a lot of cell shipping and we've all had disasters doing that. That's a really big issue that in the clinical trials there should be some standardization for how you ship. Shipping dry ice, shipping liquid nitrogen, reassurance that you are not losing temperature in the shipping process, time shipping, etc. Then the thaw and what kind of yield you're getting. Does that cover it? We've really had a pretty good discussion of that today.

All right. We get to finish on the CD34 question. Pablo's number one is if you use the total nucleated count and you give enough as defined empirically by the data and we've had that discussion already then even though we may admit that there could be differences in CD34, and there might be some who think that's really important and some who think that's less important, that it will all come out in the wash because you gave enough. That might be the recommendation.

Just give enough by total nucleated count and make sure that the rest of the processes are done well and you will be okay. Then you avoid all these downstream issues with CD34 that have confounded this field for 10 years. I remember ISH obsessing about how to count CD34 cells. Here we are ten years later and still these guys aren't sure how to count CD34 cells.

DR. HARVATH: Dr. Salomon, may I ask a question? I would like to ask a question of my former colleagues at FDA. If you were to recommend quantification or even make it a requirement of CD34 enumeration, are there enumeration assays that have been licensed by FDA for the quantification of CD34 in cord blood? If so, would that be something that would need to done using licensed kits or ones that are alternative methods?

DR. NOGUCHI: To my knowledge, we do not have such a kit licensed for cord blood. I hope I'm stating that correctly. It is of interest that much of the discussion has been about the variability or non-variability of CD34 which suggests to us that more work needs to be emphasized in this area. As far as I know, we don't have something we could mandate as a requirement which is why I think the extensive discussion here is very helpful to us to enable us to really think about what are the future needs.

DR. CREER: I think the main distinction that people have in the CD34 field is whether or not you actually gate out non-viable cells. The ISH protocol requires that the immuno-phenotype labs that usually do most of the CD34 counts. Transplant centers don't so they are not using a gating strategy that eliminates non-viable cells. That's very key. That's the most important part of it. You can just ask the transplant center who is doing your CD34 and how they are doing it. If it's not done by a processing lab but a path lab, the focus is entirely different.

DR. HARLAN: I have a specific question and then a general statement. The specific question is for the experts in the field. How well established is it that the CD34 population is the best? I have heard about these side population cells and CD34 positive and CD38 negative. Then the general statement is maybe we should just suggest in our language that assays to better characterize what the true stem cell population is and just leave it very vague like that.

DR. KURTZBERG: I don't think it's practical at all to start requiring banks or transplant centers in the clinical practice to look at side populations, CD34 and all of that negative populations because those are research questions. But even looking at 34 isn't standardized as you are hearing. It's harder to do on the thawed cells than it is on the fresh cells.

There are not established agreed upon protocols amongst different transplant centers or banks. That's not to say that's not possible but it's also a moving target. The technology is changing. The assays are changing. The companies bring out new methods that sometimes can be adapted for cord blood.

I think the real practical issue is that many of the units that are in banks right now that may be good units for patients do not have pre-cryo-preservation 34 numbers. So it's not practical at this point to require that. A statement like this is of interest and it's been shown in other stem cell transplants to be a reliable predictor of engraftment and should be looked at or something like that but I don't think we can go further than that.

DR. CREER: One important point is the CD34 count measured in peripheral blood, bone marrow and umbilical cord blood are totally different measures of stem cell activity. You can't take CD34 counts from one and extrapolate it to the other. Peripheral blood has too many CD34 cells that are already committed. It's much different than cord and bone marrow.

CHAIRMAN SALOMON: Basically what I'm hearing is that it's an interesting scientific question but not one that steps up to the point of displacing as a fundamental measurement than nucleated counts. Again it comes to numbers. The numbers have been fine with total nucleated counts. In a clinical trial, Mahendra whispered to me I think you would want some data on this. It should be one of the questions you want to solve for this field down the road but it's scientific one not a clinical one right now.

DR. RUBINSTEIN: One issue that has been raised is the effect of nucleated red cells in the total count and whether that is relevant to the relationship between the cell count and the speed of engraftment or other issues.

CHAIRMAN SALOMON: Thank you for bringing that up because that came up in our conference call to ask that question. So what do you think? What are these numbers when people are counting them? Are they counting nucleated red cells? I think they are, right?

DR. RUBINSTEIN: Yes, most of the automated instruments at the moment will count nucleated red cells into the TNC. But we studied the relevance specifically of nucleated red cells on engraftment. We found to our surprise that there is a very good correlation. It's almost as good as the TNC as a whole. There is no problem in having people just count TNC.

CHAIRMAN SALOMON: Just to be clear, are you saying that if you counted the total nuclear red cells, the more nucleated red cells in a cord blood unit the better the reconstitution?

DR. RUBINSTEIN: Yes.

DR. KURTZBERG: Pablo, I'm aware of that data but every so often, we have a unit that we bank that has more than 50 percent nucleated red cells. I've always wondered if those really are giving the same cell dose so that the functional leukocyte dose is actually 50 percent of what you are measuring. I wonder how much experience you might have with those units.

DR. RUBINSTEIN: They have typically extremely high colony counts, much higher than would be predicted by the CD34 counts. In our experience, they engraft as if they had a lot more white cells than we find in those units.

DR. CREER: One thing we've noticed is red cell depletion removes nucleated red cells. So pre-processing and post-processing differ a lot. You almost never recover after thaw nucleated red cells. You have to be very careful when you make that correlation to state specifically at what point in time your correlated nucleated red cell count with engraftment potential because I think it should be at a pre-processing stage and not later after you have selectively removed them or lost them. That's very important. I agree with you. I think nucleated red cells are a marker for erythroid potential. There are erythroid colonies that you get when you have a higher nucleated red cell count.

DR. RUBINSTEIN: I am always talking about these numbers pre-freezing. It's an interesting observation. The reason we looked into that is there's a German group that raised the alarm about these nucleated red cells. In their opinion, when you thaw these units, those nucleated red cells lyse together with the rest of the red cells. We have done those experiments and we find that these cells don't lyse like the others. That's one of the problems with nucleated red cells. They still maintain some metabolic ability to resist stochatic lysis so those cells can be counted after thawing.

DR. HALEY: Rebecca Haley from the American Red Cross. In our cord blood banks by agreement among the investigators, what we do with nucleated red cells is we make a notation on the cord blood information record when we have more than 20 per hundred white cells.

Whether they survive freezing or not, we are worried that the transplanter may get the wrong idea about how many white cells there are in the unit. Whether they survive freezing or not, they probably are not going to contribute to overall engraftment. They may be an indicator of engraftment as Dr. Rubinstein has said but we think in the interest of full disclosure the transplanter needs to know what's there. We have made it a practice to say what's there.

DR. RUBINSTEIN: I think that should be required. In our information, I guess this is what we would call a label. In the label there is complete description of the cell count including the nucleated red cell counts.

CHAIRMAN SALOMON: Then I think we've answered the CD34 question.

DR. HARLAN: One other thought occurred to me. Dr. Wagner raised it about the typing immediately pre-transplant because of the mistake that they had uncovered. Should we include that?

CHAIRMAN SALOMON: That's a good point. Obviously that's been in the news recently, the possibility of errors in typing and John had experience that quite pre-dated that. It's easy for all of us to agree we don't want that to happen. Joanne.

DR. KURTZBERG: Pablo has stated that they have a contiguous segment on their bag that they type before they release the unit from the bank which is one practice. There are some transplant centers that would also like to retype themselves. That's practical if there are enough segments on the bag but not if there aren't enough segments on the bag and that varies from product to product and from bank to bank.

The actual typing though is not fast turn around typing. It's not available at every transplant center. You don't want to keep the product in the thawed state while you are waiting to check. John did ABO which just by chance happened to be different in that patient. That's not going to be something that you can depend on to be 100 percent accurate. There are issues that relate to that. The requirement of the unit that it has at least a single segment that is still attached and can be used for final confirmation of identity is really important.

CHAIRMAN SALOMON: When it gets to the host institution.

DR. KURTZBERG: I don't know because that's not always going to be possible. Let's say that the bank has done some other testing on that unit. Let's say I asked them for a segment to test for Hurler enzyme. So they used one up already and maybe they have their internal QC that uses the second and then they only have one left. Then the bank and/or the transplant institution have to compete for that last segment. I'm not sure I know the right answer but if the programs are reputable that one should trust the other. Some confirmation has to be done either at the point it's shipped or at that point it's received.

CHAIRMAN SALOMON: Let me explore that for a second. If you call me up and I'm a bank. I say I have confirmed it. You say when. I say I confirmed it on a separate segment within a week of freezing it which was three years ago and I confirmed it before offering it to you.

DR. KURTZBERG: I think the confirmation has to be done at the time it's being either offered or received not at the time it's being put in the freezer.

CHAIRMAN SALOMON: That's what I was asking.

DR. KURTZBERG: Most places have something that they do at the time it is put in the freezer but that isn't confirmation. You need confirmation at the time received. If you say it has to be done by both, the bank and the transplant center, there aren't always going to be enough segments to do that.

CHAIRMAN SALOMON: I was trying to follow you through. One of the ways to deal with the time issue is that you initiate your confirmation during the shipping which gives you a little time. Then by the time it arrives at the host institution I can confirm from the bank that it was what I sent you.

DR. NOGUCHI: Dan, this topic itself is a very generic one for a whole lot of things for all cellular therapies. It's important to discuss it but I suggest this is not the time to do it in detail. A statement or consensus that even simple things about is it what you think it is and do you know what you think you did it's an important one to say but to put a bookmark and say in the general area that errors will be made and we need to mitigate them as much as possible is really where we want to go. For the purposes of today, that's not one that we need immediate feedback on other than to say yes, be careful.

CHAIRMAN SALOMON: Dr. Rubinstein.

DR. RUBINSTEIN: Maybe this is beating a live horse but I have a question and a suggestion on this issue. Sometimes we have done the confirmation from the segment and for a number of reasons the transplant center chooses not to use that unit. We don't ship it to them. So we store it and we note in the computer what was found and we have all the data available. I would suggest that those units should be available to other patients.

CHAIRMAN SALOMON: What Phil is telling us is you're right. We're not going to solve these. These are really important details but if the FDA is comfortable that we've gotten to the level of detail they are comfortable with then we've done our job today.

DR. RUBINSTEIN: The suggestion is that perhaps when we do the confirmatory typing on the segment we might provide the transplant center with a little bit of that DNA for their confirmation if they want it.

CHAIRMAN SALOMON: Anyone feel like we missed something important today? Joanne? You have your light on. You're okay?

DR. KURTZBERG: No.

CHAIRMAN SALOMON: Ellen, Stephen, Ruth, are we okay? Phil.

DR. NOGUCHI: Just a very quick and succinct thank you for the Committee for dealing with these extraordinarily complex and really perplexing issues. Given the advice, we're going to study this very carefully and review the transcripts and move on from here. As we always say, we can't do this without all your input. I would also say the public hearing today was immensely helpful in making sure that all voices have been heard including those who have benefit from it and including the presentation on new outcome data.

CHAIRMAN SALOMON: Joanne. Then from my point of view, I would like to thank all of you for all the hard work today. I always try to end a session by thanking Gail Dapolito and Rosanna Harvey and Bill Fress and the rest of the FDA staff that worked so hard and then sit quietly making sure I don't screw up. These things don't happen without them. I really appreciate that.

(Applause.)

CHAIRMAN SALOMON: We are now going to take a break for ten minutes. When we come back, we're going to have in open session presentations from investigators involved in the most recent site visit. I encourage all of you to stay since these are always very succinct and informative about the research going on intramurally in the FDA. Then we will go to a closed session. Anyone on the committee who is going to be here tomorrow for the gene therapy discussions we will start on time. Some people have arrived at 3:00 p.m. today. Otherwise I'll give the talks and we'll get this done. Thanks. Ten minute break. Off the record.

(Whereupon, the foregoing matter went off the record at 4:45 p.m. and went back on the record at 4:57 p.m.)

CHAIRMAN SALOMON: Back on the record. What we are going to do in this first part of the open session is review the SIF of the Division of Monoclonal Antibodies. This was chaired by Dr. Mahendra Rao. We can do this one of two ways. Mahendra, do you want to introduce these people and take over this piece of it?

DR. RAO: You do it.

CHAIRMAN SALOMON: Just do it? Okay. Then if you will allow me to continue in chair's role then I would like to introduce Keith Webber who will provide an overview of the Division. Then we'll go on to presentations from the Laboratory of Cell Biology, the Laboratory of Immunobiology and the Laboratory of Molecular and Developmental Immunology. Keith.

DR. WEBBER: As introduction, I'm Keith Webber. I'm the Director of the Division of Monoclonal Antibodies. The Division is to review at the site visit. First I'd like to thank everybody who is here and let you know that we certainly appreciate your participation in the site visit review. Without a doubt, it's one of the most important aspects of our program to keep the quality and integrity of our research programs as high as I believe they are. We definitely all appreciate your help in that manner. I'm going to quickly give you a thumbnail sketch of the Division and what we do. Then I'll introduce the folks who are under review within our Division.

Essentially Division has about 17 research reviewers who are as the name implies do both research and review. They spend about half their time on each of those aspects. They split 50-50 on doing research and regulatory review work on product files. We also have six full time reviewers in the Division who as their name implies are 100 percent regulatory review personnel. We also have about 20 additional staff that includes post-docs in laboratories, laboratory technicians and secretaries who are very important to the function of our Division.

With regard to our regulatory load, the Division currently has about 20 licensed products. Those are naked antibodies, radio-labeled antibodies. We have an antibody drug conjugate that we consult with Center for Drugs on. Also some of the fusion proteins like Enbrel are within our Division and then the latest one, Alefacept. We also have about 460 prelicensing files. Those are the ones that are under clinical trial or in INDs or they are master files that are supporting either license applications or other INDs. Next slide please.

This is a very abbreviated diagram of the layout of the Division. Within the Director's office, we have Patrick Swann who is the supervisor of the full time reviewers in the Regulatory Science and Policy branch. Dr. Ezio Bonvini is the acting deputy director of the Division. His responsibilities are primarily for oversight of the research activities of the Division.

The research activities in our Division are distributed into three laboratories. The Laboratory of Cell Biology which is led by Dr. Kathleen Clouse. This laboratory focuses primarily on studying the modulation of cytokine networks within monocytes, macrophages and other immune cells. The Laboratory of Immunobiology is headed by Dr. Ezio Bonvini. They study primarily the signaling activities within immune cells that control their growth and differentiation as well as neoplastic progression tumor cells.

We also have a Laboratory of Molecular and Developmental Immunology which is alternatively headed by Steve Kozlowski and/or Margie Shapiro as acting lab chiefs of that laboratory. This laboratory focuses primarily on studying the humoral and cellular immune response and control of those processes. Next slide.

The folks who are being reviewed currently are Dr. Kathleen Clouse here which I've already introduced. Her research is primarily focused on studying the interactions between HIV and the cytokine network of macrophages. She will briefly tell you about that today. Next slide.

Ingrid Markovic here who is studying the mechanism of entry of HIV into immune cells particularly T-cells and macrophages. Next slide.

Dr. David Frucht who is a tenure track investigator in the Laboratory of Cell Biology. He is studying the role of T-bet and Stat4 in antigen-presenting cells. He is also one of the folks in our division who is involved in doing counter bioterrorism work. Next slide.

Dr. Jerry Feldman is another tenure track investigator who is in the Laboratory of Immunobiology. He's studying the effect of immune complexes on interferon-gamma signaling pathways mostly in dendritic cells. He is also the Office of Therapeutic point person for issues regarding transmissible spongiform encephalopathy agents and he is doing some research as well with regard to prion protein effects on interferon-gamma expression.

Finally last but not least I should say that Dr. Kozlowski who is in the Laboratory of Molecular and Developmental Immunology and is studying the effects of T-cell activation. Since the hour is late, I'm not going to go further unless there are any questions for me specifically. We will move on directly to Dr. Clouse's presentation.

DR. CLOUSE: I'd like to thank you for the opportunity to give you a summary of our research program concerning cytokines network and HIV pathogenesis. The focus of our research program is actually to investigate the modulation of cytokine networks in HIV disease and to determine the impact of these changes on the disease process and to use this information to evaluate immune-based therapies for HIV as well as other infectious diseases.

The research can be covered under three specific aims. The first is to determine the effect of HIV on cytokine production. The second is to determine the effect of cytokines on HIV replication. The last is to identify novel agents that block HIV fusion and entry and therefore prevent a variate cytokine production.

I'll focus initially on the first specific aim to determine the effect of HIV on cytokine focusing on human macrophages. The macrophages play a pivotal role in HIV pathogenesis. They are the first cell type infected with HIV in vivo. They serve as a vector for the spread of infection not just to other macrophages but also to T-cells. They function as regulatory cells that control the pace and intensity of disease progression. They do this by the production of cytokines. Furthermore, they serve as a reservoir for HIV which is quite resistant to the effects of current highly acted antiretroviral therapy.

Our initial work to determine what cytokine genes were turned on by HIV-I infected macrophages showed that they did not produce pro-inflammatory cytokines or GM-CSF or other cytokines and chemokines induced by GP-120 alone. They do however produce macrophage colony stimulating factor which is important because M-CSF functions as a survival factor for the macrophage and infected macrophages are known to survive longer than their uninfected counterparts. M-CSF also facilitates infection of macrophages in part via the up regulation of the HIV receptors CD4 and the chemokine co-receptor CCR5.

We then went on to show M-CSF antagonists can actually inhibit HIV-I replication in macrophages. Shown here as a typical infection of macrophages with a CCR-5 M-tropic strain of HIV-I, you can see substantial virus replication over the course of 21 days. We monitor replication by measuring reverse transcriptase activity. However in the presence of either polyclonal or monoclonal antibodies to M-CSF or soluble M-CDF receptors, there is a dramatic inhibition of HIV-I replication.

We then went on to do studies to look at the specificity of the M-CSF induction by viruses. We used a panel of single-stranded RNA viruses. We found viruses such as measles virus and respiratory syncytial virus all though they replicated macrophages and they are functional in that they produced pro-inflammatory cytokines and beta chemokines, they do induce M-CSF production.

In contrast, M-tropic strands of HIV-I, HIV-II and also SIV replicate and induce M-CSF production in infected human macrophages. Furthermore, exogenous M-CSF enhances replication of these lentiviruses. The replication of both HIV-I and HIV-II in macrophages can be inhibited using M-CSF antagonists. This has led us to propose the following cytokine network for sustaining a macrophage HIV reservoir where you have an HIV infected macrophage which produces M-CSF but can function in an autocrine and paracrine manner.

Beta chemokines are produced as well as infectious HIV particles. Beta chemokines such as mip-1 alpha and beta attract uninfected cells to the area. They are rendered more susceptible to infection by M-CSF. They are infected by the released virus particles and the network continues. Pivotal for maintaining this network is the production of M-CSF.

The second area of research dealt with the effect of cytokines on HIV replication. For this we focused on the effect of interferon-alpha on HIV replication and human macrophages and T-lymphocytes. We specifically addressed the question of whether interferon-alpha therapy is beneficial or detrimental for HIV disease.

It's pretty well known that interferon-alpha therapy is effective early in the course of HIV disease but as the disease progresses it becomes less effective and may actually hinder HIV patients. There are many reasons that are thought to be the cause of this. We looked at a number of isolated interferon-alpha components as well as the number of hybrid interferon-alpha molecules and looked at their effect on virus replication in both macrophages and T-cells. We have come up with the following few effected macrophages with an M-tropic strain of HIV-I that uses the CCR-5 chemokine co-receptor. You see again substantial virus replication. But in the presence of interferon-alpha, there's a dose dependent inhibition of virus replication. This does not represent a block in virus release but an actual block in virus replication.

In contrast if you infect T-cells with a T-tropic strain of HIV-I that uses the CXCR-4 receptor again you see substantial virus replication. This is accompanied by death of the CD4 cells. But when you add an interferon-alpha here, you actually see an enhancement of HIV replication. This enhancement does not correlate with an increase in the proliferation of the T-cells or an up-regulation of the receptor or the co-receptor for HIV. We feel we now have a model that we can actually use to determine one of the mechanism that may contribute to the exacerbation of HIV-I disease mediated by interferon-alpha.

I would like to acknowledge the people involved in these first studies. The people in my laboratory. Dr. Judy Beeler's lab collaborated on the measles and RSV studies. People in Indira Hewlett's lab on the HIV-II. Vanessa Hirsch's lab contributed SIV islets. All of the interferon-alpha are being done in collaboration with the laboratory of Dr. Kathy Zoon. The final area of research concerning the identification of novel agents that block HIV fusion and entry will be presented by Dr. Ingrid Markovic in my lab.

DR. MARKOVIC: As Dr. Clouse mentioned part of our laboratory's efforts are dedicated to try to understand the mechanism of HIV entry and eliciting the way how this process can be blocked. Shown here is the schematic representation of HIV entry, this here being the viral particle with the envelope surviving the virus.

Embedded in the envelope is gp120 and gp41 which is the protein which is responsible for fusion between viral envelope and the target membrane. When the virus approaches the target membrane it binds to CD4 which is its primary receptor. Following that binding, there is a confirmation on change in gp120 which induces the exposure of the co-receptor binding site allowing gp120 to now bind to the co-receptor which is CXCR4 or CCR5 depending on the type of the virus. That would finally lead into the final confirmation on changes, this being the six helix bundle stage which brings the viral and the target membrane in the close up position and allows the fusion to take place.

Now previously there were a number of ways which people had thought about in order to block the viral entry. Primarily those were trying to block gp120 and binding site or blocking using a receptor ligands trying to block the CD4 binding site and CXCR4 binding site. Another way was trying to interfere with the confirmation on change so the six helix bundle stage does not form.

There were problems associated with these different approaches whether there are the formation of escaped mutants like here or the fact that chemokine ligands can actually induce replication of the HIV-I, the viruses inside the cell. Our approach was trying to actually use inhibitors of cell by cell like interchange of gp120, gp41 and try to understand the mechanism first whether this is possible and second what would the mechanism of this action be.

The hypothetical role that we had in mind is how far this thiodisulfide interchange could be applied in HIV entry, this here being the gp120, gp41 in its initial confirmation with disulfide bonds in the original form. In the presence of CD4 and CXCR4 in the target membrane in a critical thiodisulfide oxidoreductase, gp120 would undergo unfolding and reduction of critical disulfides opening now the free-style groups. We actually have the data to show the distance remedied indeed does actually. As a logical consequence of this intermediates we think that protein now undergo final refolding to achieve the fusion competent confirmation. The disulfides will be restructured in a different orientation where the previously unpaired systems can now be bonded.

We decided to focus our interest on protein disulfide isomerase as a punitive agent which would be this thiodisulfide isomerase. The reason being that it's the most widely published enzyme because it is present in many different cell types. The function of this enzyme is to characterize reduction, formation, and isomerization of disulfide bonds. It does that by a two active catalytic site which is shown here in green which are basically two free cell groups. It localized in the ER as well as the cell surface. It's been shown by our lab as well confirmed by other labs that it can be inhibited by DTNB, Bacitracin and anti-PDI monoclonal as well as polyclonal antibodies.

The first question that we had we wanted to confirm whether PDI is indeed present at the cell surface and if so does it actually co-localize with the CD4 and the CXCR4. Our approach was to use co-focal microscopy. Here shown is the triple labeling where the green indicates PDI, the red region indicates CXCR4 and the blue is CD4.

We did this in two different conditions. In the absence of soluble gp120, we can pretty much see that PDI is isolated in distinct regions on the membrane and it does not co-localize with either the receptor or the co-receptor. The magenta color up here indicates co-localization of CD4 and CXCR4 which is occurs obviously in the absence of gp120.

However when we added gp120 to those cells we now see that there is a majority of the PDI is now co-localized in this white region which indicate triple co-localization of PDI with CD4 and CXCR4. This indicated to us that probably gp120 acts as some sort of an assembly mechanism bringing PDI into conflict with CD4 and CXCR4.

Next we wanted to see using cell-cell fusion assay which was our assay to estimate where the viral entry takes place. I should say to substitute for viral entry assay. We wanted to see whether the inhibitors of PDI such as antibody DTNB and Bacitracin can actually block this process. Shown here are just the most effective doses of each inhibitor namely antibody DTNB and Bacitracin which have been very effective in blocking fusion in comparison to the fusion of cells which were in control or that were treated with isotype antibody control. I should say that we have a dose response for each of these inhibitors. This is just the most effective dose of each of those inhibitors.

We next proceeded to look at the infection assays again trying to estimate the potential of these inhibitors to inhibit infection. On the X axis it stays Post-Infection and the Y axis is the Reverse Transciptase Activity which are indicative of the level of the infection. The three bell-shaped curves are control to antibody control namely polyclonal and monoclonal antibody controls. This is a typical bell-shaped curve that you get in control with HIV infection. However in presence antibodies, both poly- and monoclonal we see a significant delay in the onset of infection whereas DTNP and Bacitracin completed abolished infection.

Next we decided to form a working model. Actually this is a working model that we have now in our minds when we are thinking about this problem. Namely this is gp120 and gp41 bound to CD4 and the target membrane. This induces conformational change in the protein and exposes a certain high affinity site which allows PDI to bind to gp120 which in turn induces subsequent confirmation on changes allow it to bind to the co-receptor and finally the fusion can take place.

I would just like to acknowledge that all this work was done in laboratory headed by Dr. Kathleen Clouse. These are our collaborators from CBER, USUHS, and NIAID. Thank you.

DR. FRUCHT: I'm David Frucht, one of the newest investigators in the DMA. The specific aims of my program are to divide expertise in evaluating the safety and efficacy of new biologics that effect cytokine networks. Specifically I've been concentrating on doing reviews on hematological and infectious disease related biologics.

As part of this Aim 1, Aims 2 and 3 also fit under Aim 1 which is to establish research programs to establish the role of key transcription factors and the function of antigen presenting cells (APCs) and to determine the role of cytokine signaling and host responses to anthrax-lethal toxins. That's a new project I just started in the last couple of months.

Perhaps the most important process that occurs during immune responses is the differentiation of T-helper cells into either TH1 or TH2 effector cells. TH1 effector cells make interferon-gamma and are important for cellular immune responses, control of intracellular infections like Tuberculosis and Anthrax, Toxoplasma, Salmonella and also are important for tumor surveillance. TH2 effector cells on the other hand make IL-4 and are important for humoral immunity, control of helminthic infections and regulation of allergic responses. We now know that this process is regulated by cytokines. IL-12 signals via Stat4 to induce TH1 differentiation whereas IL-4 signals via Stat4 signal transduction molecule to induce TH2 differentiation.

Two other transcription factors are also very important. T-bet in TH1 cells which is important for programing TH1 cells to make interferon gamma and GATA-3 which programs TH2 cells to make IL-4. My program mainly deals with TH1 differentiation specifically with Stat4 and T-bet.

Instead of concentrating on T-cells, I've concentrated on antigen-presenting cells. Antigen-presenting cells are absolutely required for TH differentiation. In fact if you take a T-cell and you stimulate it through the T-cell receptor in absence of antigen presenting cells you don't get cytokines, production or proliferation, in physiological conditions. So antigen presenting cells are very important.

We also know evolutionarily that they are very important because the bugs actually target antigen-presenting cells like Tuberculosis and Salmonella. Also the toxins made by bacteria also specifically target antigen-presenting cells such as Anthrax-lethal toxins.

Our first major finding was that Stat4 is produced by antigen-presenting cells and is required for IL-12 dependent signaling and interferon gamma production. In this experiment using antigen presenting cells for wild-type mice and Stat4 knockout mice -- and Stat4 knockout mice have the genes deleted to they don't express Stat4 messenger protein -- you can see clearly that in the wild-type mice following stimulation with IL-12 or Il-12 in combination with Il-18 you get high levels of interferon gamma in the wild-type samples. However you see no induction of interferon gamma in the Stat4 knockout samples.

We also went on to show that Il-12 dependent nitric oxide production depends on Stat4 as well. This is important because it allows control of intracellular infections such as Toxoplasma.

Our second major finding was that T-bet message RNA is induced by interferon gamma in antigen-presenting cells. As you can see in this Northern blot looking at messenger RNA, high levels of T-bet message RNA are induced following interferon gamma stimulation, higher levels than we see even in our NK cell line which was our positive control.

Now you might recall from one of my first slides that T-bet regulates interferon gamma production. What we've shown is that T-bet message is present in antigen-presenting cells. It's regulated by interferon gamma. We have very good evidence that this positive feedback sustains interferon gamma production by antigen-presenting cells. What wasn't previously known was that interferon gamma regulates T-bet message in T-cells as well. This positive feedback loop is absolutely critical for TH1 differentiation in the absence of IL-12. These are experiments that we did in collaboration with Bill Paul.

I would also like to mention that we've been limited by the fact that we could only look at message levels of T-bet because a good antibody didn't exist. But now there are some commercial antibodies that we are testing and looking at right that might let us make more definitive statements about these positive feedback groups. Also we're going to be getting the T-bet knockout mice by June. We are very happy about that.

In the very basic model of TH or cellular immunity, antigen-presenting cells make IL-12 which signals to Stat4 to promote TH1 differentiation. TH1 cells also express high levels of T-bet and T-bet allows the programing of TH1 cells to make interferon gamma. What we showed is that APCs express T-bet. It's regulated by interferon gamma and this positive feedback loop likely sustains a long term interferon gamma production by APCs. It's more important probably in T-cells, where T-bet is also regulated by interferon gamma. Again this wasn't recognized before and is absolutely essential for TH1 differentiation in circumstances where IL-12 is not present.

We also showed that Stat4 is expressed by antigen presenting cells and allows signal transduction response to IL-12. One of the genes that's turned on is interferon gamma. Interferon gamma present at that time of antigen presentation can be very important in promoting cellular immunity at a very early stage. We call this the "jump start" model. This model appeared in "Trends in Immunology" in a paper we wrote recently. I'm happy to say that it seems to reaching some level of acceptance because in Laurie Glincher's recent review and annual reviews of immunology, she highlighted some of our data in this regard.

I would like to conclude by talking about our last project which just started a few months ago looking at the role of cytokine signaling and host responses to anthrax-lethal toxin. Anthrax-lethal toxin specifically targets macrophages. It kills them through apoptosis by leading MAP kinase kinase. This is pro-survival pathway. We have developed toxin potency assays for rapid screening of potential therapeutics. We are very interested in looking at particular cytokine pathways; pro-inflammatory cytokine pathways like IL-1, TNF and IL-12; suppressive pathways like IL-4 and Il-10 and pro- survival pathways like GN-TSF and M-CSF and IL-7.

Finally I would like to acknowledge that this represents the work of a lot of people especially in my lab. I would like to make a special point that we just found out today by letter that Carla Lankford won an award from the American Federation for Medical Research, Henry Christian Award, for her work on the positive feedback model for T-bet during the short time she was in my lab. Thank you very much.

DR. BONVINI: Mr. Chairman, Mr. Rao, distinguished members of the Advisory Committee. My name is Ezio Bonvini. I'm here to represent Dr. Gerald Feldman who could not be here because he's attending a regulatory conference in San Diego.

My task before you today is to briefly summarize the work of Gerry in the lab. I'll start by mentioning the two projects he's involved with. The major focus is represented by the PRIF program. That deals with the signaling interference or crosstalk we will call it between immunocomplex interferon gamma in monocytes. Monocytes are an important first line of defense against pathogens and the critical antigen processing and presenting cells such as we heard in the regulation of the immune response.

I will only briefly touch the second aspect which is something that Dr. Feldman started recently in our laboratory and that is the effect of prion proteins on monocyte reliability cells. This is dictated by the fact that studying immunocomplexes and their role in regulating immune responses is quite an important area of interest within the FDA. That's because not only are autoimmune diseases directly linked to immunocomplexes but also because the use of antibody therapeutics has forced us scientists of the FDA and the field in general to be more and more aware of the IC receptor or the C portion of the antibody. The ?- IC receptor pertain to both the effector function of antibody and the generation of adverse effects.

As I stated earlier, monocytes are critical regulatory cells in a disease and they are activated by a number of cytokines including interferon gamma. The path of activation by interferon gamma has been fairly well illustrated by work of numerous investigators, including Dr. Feldman when he worked with Dr. Finland in the laboratory at CBER. What this pattern shows is that interferon gamma hits a specific receptor on the surface of monocytes and triggers the activation of a Janus Tyrosine kinase family, JAK-1, JAK-2 which in turn activates a transcription factor, Stat1, which is translocated to the nucleus where induces a number of activation of a number of genes.

Now in the work by other scientists and in particular Emil Unanue and Bob Schreiber showed that immunocomplexes interfere with the response to interferon gamma in a negative fashion. With the knowledge of how the interferon gamma signal perceives, Dr. Feldman set up to understand where the effect was and by using a retrograde approach he went on to show that immunocomplex pretreatment will down-regulate the entire JAK-STAT activation pathway but he did so without affecting interferon gamma binding, receptor expression or the forceful relation of the alpha chain of the interferon gamma receptor. However the inhibition correlated well with the inhibition or the reduction in activation of the JAK-STAT pathway.

There are several IC receptors on the surface of the monocyte. We now can put some names on these receptors. By the process of exclusion, Dr. Feldman showed that this receptor, the CD64, FC gamma receptor 1 is the one important for inducing inhibitory response. Surprisingly the CD32 which is the FC gamma receptor 2(b) which is the prototypical inhibitory receptor was not involved. The FC gamma receptor 1 hits the Stat kinase which also went on to activate a tyrosene phosphatase, SHIP-1, which presumably serves as an inhibitory effect early on upstream with the activation of JAK kinase.

We think that this is the first mechanistic evidence of how the CD64 exerts an inhibitory crosstalk on the interferon gamma receptor. I think that this is an important first step to appreciate the potential impact of immunocomplices and therapeutic antibodies on FCS.

I mentioned that Dr. Feldman got involved in a second project and that stemmed from his leading role within the Office of Therapeutics in the issue pertaining to the disease CJD. Because of his interest in monocytes, he went on to study monocyte-derived dendritic cells. The characteristics of dendritic cells are actually quite important because they are found in the plaque associated with CJD disease, were they producing inflammatory cytokines and they probably have a role in the transport of prion protein from gut to the lesions.

He went on to show that dendritic cells are activated by a prion-derived peptide in the treatment of dendritic cells for this prion derived peptide-induced deactivation which leads to the induction of pro-inflammatory cytokines which leads to the further activation and maturation of these cells. This is fairly similar to what happened in monocyte induction by him and by Murphy's group recently about these cells responding to chemotaxis.

However there are quite important differences. For instance, the dendritic cells show no constant response whatever in response to this peptide, whereas monocytes do. The two cell types show differential sensitivity to cartoxis toxin treatment which is a tool that we use to determine the involvement of G-protein in this signaling. In the case of monocytes, this prion peptide was involved in the activation of a chemotaxis receptor which was not involved in the dendritic cells. That led us to conclude that dendritic cells have a functionally different response to prion peptide as do monocytes. Perhaps more importantly, the prion peptide can use a differential signaling mechanism possibly which is as specific.

I would like to acknowledge the people in this project, the laboratory of Michelle Jessen. Dr. Feldman would like to encourage a couple of collaborators, Petra Lenz from DNCI and John O'Shea from NINLS. Thank you very much for your attention.

DR. KOZLOWSKI: Hi. I'm Steve Kozlowski in the Division of Monoclonal Antibodies. I'm also involved in the regulation of these products and I would like to talk about the activation of T-cells, a critically important issue both in immuno-protection and immunopathology.

Clearly to activate a T-cell, you need to have an antigen reach it. That's generally done through an antigen-presenting cell. This doesn't happen though in isolation. It happens in the context of many other cells in the tissue in which it occurs. It occurs also after an activation and signaling of both these cells through their passageway to get to each other and through soluble mediators that are present when they interact; and finally, last but not least, through the direct interaction between the antigen-presenting cell and the T-cell.

We approached studying this through two opposite ends. We looked at isolated T-cell ligands in a very minimalist in vitro activation system and we looked at the opposite side, in vivo outcome of T-cell activation. The way we linked these two different ends was to use a very similar system. We used a transgenic T-cell mouse which had CD8 cells primarily all sharing the same receptor specificity which is allo-reactivity against the H2LODMHC molecule in the context of endogenous peptide.

In moving to this first project that isolated T-cell ligands, T-cells are activated through presentation of antigen to the specific T-cell receptor. This has been called signal 1. It's been called signal 1 because it is the critical attribute of the immune cells that allows them to have both specificity and memory for specific antigens, the antigen-specific derived signal.

However important that signal is though, it's been found that it's often not enough and there are other additional signals necessary for full T-cell activation. In a model that's many decades old now, we would say that the antigen specific signal plus another signal gives you activation, whereas the antigen specific signal alone can be even negative much less not necessarily positive.

The providing of the signal 2 has been termed co-stimulation. A classic molecule for this is CD8 on the T-cells and its ligand, B7 molecule that allow full activation of the T-cells. However although this is a prime example of this signal, there are a myriad of other molecules and I'm showing you a small subset of them which also can function as a signal too in some way or form. There are also negative second signals which can actually down regulate T-cell activation.

All of these molecules are potential sites for therapeutic intervention. In fact I would say the vast majority of these molecules have antibodies to block them under IND or being licensed. So it's critical to understand the role of these.

One of the questions we had from this is although there's a lot of data to say signal 1 alone is not enough is that all was true. In other words, is it absolutely true that you always need a second signal to activate a T-cell?

Using a cell-free system, T-cell contained, but antigen-presenting, cell-free system where you activate it with purified MHC peptide in the absence of any presenting cells, we asked the question could we activate T-cells and was the second signal absolutely necessary? We found for naive CD8 T-cells, and they were sorted to be a naive phenotype, they could be activated in the absence of antigen-presenting cells with purified ligand.

Now activated may not be fully activated but it occurs in a large fraction of cells when you look by flow cytometry for markers of activation. It's associated with high levels of IL-2 production. These cells mature to become lytic effectors. We would define this as full activation.

Furthermore if we add blocking antibodies to LFA-1 or B7 molecules, which both are forms of co-stimulation, we don't alter the activation. Using the system though required using a high dose of extremely potent peptide.

Under physiological conditions, second signals may be necessary but in some aberrant situations which may lead to autoimmunity or to failure of immunosuppression in some circumstances, this isolated signal, when alone, can operate. By using weaker signals we can add back other molecules. We have already added back I-chem 1 and we are looking to add back other molecules to test in this isolated system the interaction between these second signals.

I want to make a point though here. With all the things that we added back including IL-4, we have never seen a TH2 response. We have never seen these cells make IL4 in vitro. They will make interferon gamma but not IL4.

Although data with isolated ligands has its advantages, nonetheless the question is can you go back and say these interactions are relevant. So despite the fact that antigen MHC, co-stimulators, soluble mediators, even changes in cell surface typology are good readouts and markers of T-cells, nonetheless, what happens in vivo? To look in vivo, we used an adoptive transfer model. We look though at our outcome as cytokine production and C2.

I want to talk a little bit about what David Frucht mentioned before but type I versus type II cytokines. There is type I cytokines, T-cytokines that family would be interferon gamma and type II IL4. They have very different responses on the overall immune phenotype and they tend to polarize. In fact, David Frucht gave you an example of an interferon group. Cytokines tend to feedback positively and that allows a polarization effect to one type or another. This has tremendous impact because it alters disease outcomes both in disease models and in some examples of human disease.

There are a number of factors that can affect these things including cytokine, antigen dose, co-stimulatory molecules, chemokines and antigen presenting cells can all affect this polarization which affects disease outcome. As a broader range of what's regulated all of these things are potential therapies under development in some way or another. All of them may have impact on diseases. Again in vivo it's not any of these in isolation. It's all of them together in concert that have an effect.

To look at this we adoptively transferred the same T-cells, purified the same way that we used for our in vitro experiments. We generated a graft-versus-host disease model. We took these cells. We isolated the spleens. We purified the T-cells. We added 2x10⁶ purified T-cells. I can tell you these were counted after processing. We never froze them.

These were then transferred into an F1 mouse which was selected so that it would not reject the cells that were transplanted and so that it would allow the allo response to be generated against them. Then we immunized these mice with antigenic peptide that it sees endogenously but at much higher levels and much stronger response.

Then we looked at tissues, generally at five days later, but sometimes at different time points. The first thing we looked at we looked at was whether these cells were migrating to these tissues so we stained with an antibody specific for clonotypic T-cell receptor and looked in these tissues. We found that primarily we saw two CT-cells in spleen, some in mesenteric lymph nodes and in testes. We didn't see them in any of the other tissues.

Then to look at the immune phenotype as I mentioned before we looked at cytokine families. We first looked at IL4. We stained spleen and testes for IL4-producing cells. Strikingly, and very unexpectedly, because in vitro we could not make these cells make IL4, we found large numbers of IL4-positive cells in the spleen. However, in the testes we found none. We looked at interferon gamma, the type 1 T-cytokine, and we found staining of interferon gamma positive cells in the testes but none in the spleen. We saw a complete and total opposite polarization depending on the tissue that you looked at.

We further went on to do double staining and proved that although not all these cytokine producing T-cells are the adoptively transferred T-cells, many of them are. The exact same clonotypic T-cell is making the opposite cytokine in the different tissues.

This has a number of implications. Tissue specific immune responses to identical epitopes in which we have clonotypic T-cells exactly the same receptor making opposite cytokines, depending on the organ they were in. It's dependent on peptide. If we do this without adding the antigen, it doesn't happen. It occurs with peptides of a 30-fold difference in avidity so it's not some narrow range of dose and not only have we shown it by immunohistochemistry but also by ELISA.

I think the take-home from this is modulating the immune response phenotype, even though it may ameliorate pathology in one organ, it might exacerbate it in another. You really need to know what's going on in all the tissues. In fact for toxicology issues for the FDA it may be very important to look in models that not only evaluate infiltrates in tissues but look at specific immunophenotypes of the responses in these tissues.

Ultimately the goal of understanding both individual ligand-receptor interactions and in vivo responses is trying to integrate all this together. The immune system is a myriad of interactions. Many cytokines were defined by the term pleiotropies and redundances.

Many of them do the same things but many of them do slightly overlapping differences in terms of exactly what they do. If you start thinking about in vitro figuring this out, I did a little back of the envelope calculation here. There are about 80 cytokines and say 15 of them are involved in immune responses. There are 40 chemokines that are known and say 10 of them are involved in immune response. And say out of a dozen or so co-stimulatory molecules, five of them. That's 10²⁰ possibilities and that's not even looking at doses. So you clearly can't mimic this entirely in vitro.

The question is you need to be able to make correlations. I think our strategy is to use an identical system in vivo and in vitro in terms of the T-cell and the T-cell receptor to look at it at difference scales of activation, both from individual ligands to whole tissue responses and to try and assess the contribution of different ligands and compare them to what we see in the microenvironment.

In my lab, Julia Goldstein, Trina Chen, Elena Govina have done most of the work with the in vitro system. Lee Zhang has done all these adoptive transfer experiments. Thank you.

CHAIRMAN SALOMON: Thank you, Steve. Just to show you how young I am, we were post-docs together at the NLI ten years ago. That's true. I think now we go to the closed session, right? I want to thank everyone who has presented this afternoon. I've said it again and again but it can't be said enough. The tremendous respect all of us have for the high quality intramural work that's being done at the FDA. I know you don't get enough appreciation. I know it's a battle.

Appreciation equals money. I get it. The fact that you get so much good quality work done while there is so much competing for your time and you get really relative to what we have in Research Institute like at Scripps or Harvard is just fantastic. You do have our respect. It's really an important part of the FDA mission. Off the record.

(Whereupon, at 5:49 p.m., the above-entitled matter concluded.)