Bringing the Promise of Molecular Medicine to Oral Cancer Screening

March 2006  

Oral and pharyngeal cancer remains common and deadly in the United States.  In 2005, according to the American Cancer Society (ACS), an estimated 29,370 new cases of oral and pharyngeal cancer were diagnosed, while an estimated 7,320 Americans died from these diseases.  The ongoing tragedy is with early detection and timely treatment, most of these deaths could have been prevented.  The NIDCR has made a concerted effort to help lay the scientific groundwork for a new generation of screening tools for oral cancer.  Recently, the Inside Scoop spoke with NIDCR grantee Dr. Miriam Rosin, a senior staff scientist at the British Columbia Cancer Agency in Canada.  She and her colleagues in British Columbia are now developing a novel, province-wide oral cancer screening program that integrates for the first time telltale molecular features of a developing tumor with more traditional cancer screening tools.  This project offers a glimpse of the much-touted promise of molecular medicine and serves as a template for future molecular-based cancer screening programs elsewhere in the world, including the United States.

When people talk about oral cancer screening, the message is early detection is the key.  Is this still the right message?

Absolutely.  The earlier that you catch a suspicious sore, or lesion, in the mouth, the better the outcome will be for the patient.  The problem is the morbidity and mortality rates for oral cancer have remained fairly static for several decades.  That might seem surprising.  These lesions, after all, are often readily visible in the mouth and thus are more amenable to excision than, say, a lesion in the liver or kidney.  But I think the lack of progress tells us three things:  One, we need better screening tools that more sensitively and more specifically determine whether a suspicious lesion is precancerous and should be removed.  Two, we need better educational outreach efforts.  Those at risk for oral cancer need to know that they must act to safeguard their health.  And three, within the dental and medical communities, we need to develop a more seamless system to manage patients.  We can’t allow anyone to get lost in the system.

Let's take the first issue of better screening tools.  What's needed?

What’s needed is that scientific buzzword called translation.  That means we need to take the most promising molecular discoveries in the laboratory and translate them into more informative clinical screening tools.

Why's this so critical?

Let me first give you a little background.  When you think about oral cancer screening, there are really two barriers to success.  The first is the visualization of the lesion in the mouth.  You need to determine whether it stems from infection or trauma, in which case they will heal on their own, or whether the irregularity is a mass of cells of abnormal shapes or sizes called a dysplasia and should be biopsied.   If you don’t make the right call at the front end, the whole system falls apart.  Everyone loses.

What's the second barrier?

The second barrier is, assuming the lesion is a dysplasia, how do you then take the next step and predict risk?  If it’s a severe lesion, it’s easier to make the determination.  But, if it’s a mild or moderate lesion, the lines become blurred.

You mean looks can be deceiving?

Exactly.  Looks can be deceiving in the visualization process, and they also can be deceiving in the pathology lab.  You can’t just look at these lesions under a microscope and know definitively how they will behave in the future.  To better characterize a mild or moderate dysplasia, you need access to molecular information.  You need to know whether abnormal molecules are present in the cells that drive the abnormal growth.

And you and your collaborators are now folding in--or, as you said, translating--this molecular information into a new generation of oral cancer screening tools?

We are.  The whole focus of our group is to take promising molecular information, such as a relatively common chromosomal abnormality associated with oral cancer, validate its diagnostic benefits, then integrate it into our oral cancer detection program here in British Columbia.  That doesn’t mean that we intend to discard our existing screening tools and start anew.  Not at all.  It means we need to integrate the best of the old with the promise of the new.

How do you do this?

Let me give you one of our preliminary success stories.   A few years ago, we took a set of 116 samples of oral epithelium that were classified as low grade hyperplasia.  These are the early lesions for which it is hard to predict their cancer risk.  Because these were stored samples, we had the luxury of knowing the outcomes for each.  Some progressed to cancer, some didn’t.  But what’s especially important is the samples looked the same under the microscope.  You couldn’t differentiate them by any of the traditional features that pathologists evaluate.

Then why look at them?

Because we wanted to go beyond appearances under a microscope.  We wanted to know whether these samples contained telltale chromosomal changes that predisposed them to progression.  Sure enough, those that advanced to cancer had characteristic changes on chromosomes 3 and 9 that seemed to be essential to their progression.   But these changes weren’t present in the lesions that didn’t progress.  We concluded that screening for these chromosomal changes in the mild to moderate lesions might be helpful, but we also knew that dentists neither had the time nor the tools to test for chromosomal abnormalities.  So, we got to thinking, what if the early lesions that contained these chromosomal changes could be visualized in their offices with relatively little time or effort?

And that's where the best of the old came in, right?

Precisely.  We thought of toluidine blue, a staining dye that has been used for decades to identify various tumors including oral cancer.  Dentists are very familiar with it, and preliminary work from our laboratory and David Sidransky’s lab at Johns Hopkins showed that toluidine blue appears to differentially stain early precancerous lesions. 

That’s absolutely correct.  People were worried because toluidine blue flagged only a portion of mild to moderate lesions as positive.  But this raises a fundamental question:  Do we really need to identify every low grade lesion?  To my thinking, detection just for the sake of detection is an extremely high and artificial standard.  At the end of the day, what we really want to know is outcome.  We want screening tools that catch just those low-grade lesions that will become cancerous.

Exactly.  Last year we published the results of a study in which we monitored 100 patients with oral premalignant lesions for about three and a half years.  We found toluidine blue detected 16 of 17 cases of high-grade dysplasia, and it preferentially stained the oral premalignant lesions with minimal or no dysplasia that had high-risk clinical and molecular characteristics.  That’s precisely the information that clinicians and patients need to make informed decisions.

By reaching out to high risk populations.  Our research team just completed a pilot study in Vancouver ’s Downtown Eastside, one of the city’s poorest neighborhoods where drinking and smoking, two well known risk factors for oral cancer, are widespread.  We screened about 250 Eastside residents and identified two cancers and seven precancerous growths.  Now, we’ve launched a more extensive screening program in this neighborhood.  But I should add that outreach works best when the third element of progress - a seamless management system - is in place, and we’re working very hard in British Columbia to create this infrastructure.

We’ve created the British Columbia Oral Cancer Prevention Program.  It’s a partnership of the BC Cancer Agency, University of British Columbia,SimonFrasier University, and the Portland Dental Clinic here in Vancouver.  The NIDCR, of course, is a major supporter of the program.  We’re establishing lifelong management of oral cancer for the four-million people who live in the province.  The program builds on the idea of taking biological discoveries into the community as soon as they are validated.   

Right.  We’ll integrate informative visualization devices and molecular probes to complement existing screening tools.   The program will involve all 2,900 dentists in the province.  They will be trained through continuing education programs in how to use the various screening tools as they emerge.  We also will create a referral pathway, meaning if dentists don’t know how to perform a certain screening test, they can find someone who does.  We’re trying very hard to create a cohort of dentists that will serve as the leaders of the effort, mastering the available screening technologies and engaging their colleagues to optimize the screen.

Dentists will continue to send their biopsies to the British Columbia oral biopsy service.  It’s directed by my colleague Dr. Lewei Zheng and is a central component of the system.  Right now, about one in five dentists across the province submits biopsies every year to the service, totaling around 4,000 biopsies per year.  Also of interest, Dr. Zheng maintains archival samples that date back about 25 years.  What we’re working on now is to transform the biopsy service into a provincial registry that trains and interacts with dentists, registers patients, and monitors their progress via a computer tracking system.  Our overriding goal is to ensure that patients receive treatment.  If they get hung up in the system, everyone loses.