Experts on pancreatic cancer met last week to exchange ideas about how to make progress against this deadly disease. Much of the discussion focused on how to identify the most promising potential therapies and test them efficiently in the clinic. Many participants said that combinations of targeted therapies may prove effective against the disease, even though few patients have benefited from the targeted drugs tested to date.

More than 90 researchers, clinicians, patient advocates, and pharmaceutical industry representatives discussed topics ranging from mouse models and cancer stem cells to the pitfalls of designing and interpreting clinical trials. The meeting was organized by the NCI Gastrointestinal Cancer Steering Committee and was the first in a series of state-of-the-science conferences sponsored by NCI's Coordinating Center for Clinical Trials.

"We are facing a bit of a crisis in pancreatic cancer clinical research," said Dr. Margaret Tempero of the University of California, San Francisco. "We seem to make progress in identifying potential therapies in isolated institutional clinical trials, yet when we test these agents in a broader population, they fail."

Failure may occur for several reasons, and they are not always clear. Patients in early-stage trials could be less sick than patients in the broader population and therefore fare better, for instance.

Many participants said that a single drug is unlikely to be effective against this genetically complex disease. Pancreatic tumors are notorious for resisting the antitumor effects of targeted drugs by activating alternative pathways.

Some of the molecules involved in transmitting cancer-promoting messages have been identified, but the most important players in these complex networks are not known. "The essential problem with pancreatic cancer is that we're trying to make progress in the face of biological uncertainties," said Dr. James Abbruzzese of University of Texas M.D. Anderson Cancer Center.

A recent study suggesting that multitargeted therapies may be required to treat glioblastoma brain cancer could apply to pancreatic cancer, said Dr. Abbruzzese. Both cancers are genetically and clinically diverse and resistant to therapies.

Pancreatic cancer is the fourth most deadly cancer in the U.S., though it is not among the 10 most common. The cancer is frequently diagnosed only after it has spread beyond the pancreas, and treatment options are limited. Surgery is the best hope for a cure among some patients with cancer localized to the pancreas.

The primary chemotherapy drug used to treat pancreatic cancer, gemcitabine (Gemzar), helps a minority of patients. With the exception of erlotinib (Tarceva), few of the targeted drugs tested to date have shown a benefit for patients.

Still, critical genes have been identified, and there was reason for optimism at the meeting. Mouse models have generated interest because the mice, like their human counterparts, develop tumors that spread easily and resist most treatments. In addition, recent studies suggest that some patients may benefit from adjuvant therapies following surgery.

"We have much more knowledge about the basic science of pancreatic cancer today than we did several years ago," said Dr. Philip A. Philip of Karmanos Cancer Center, Wayne State University. "We also have inventories of targeted drugs that we could use if only we can figure out how to test these drugs in patients in ways that provide us with informative results."

The group plans to report its findings and recommendations in 2008. In the meantime, there was a consensus to do more pilot clinical trials with smaller numbers of patients before embarking on large randomized trials, said Dr. Philip. This would ensure that the most promising treatments are taken to final-stage trials with a much better chance of success.

The need to develop experimental tools such as mouse models was also stressed. "There was general agreement on the importance of these tools in helping us to design trials and in choosing drugs and selecting patients to get the maximum benefit," said Dr. Philip.

—Edward R. Winstead