Speakers call for advances in breast cancer biology to inform clinical research

December 11, 2008

Emerging knowledge about breast cancer biology must be integrated into clinical trials in order to personalize treatment in patients with early breast cancer. Two speakers at an educational session on clinical trial design discussed our advancing understanding of basic science and the ways in which that can change and refine the design of clinical trials.

SAN ANTONIO-Emerging knowledge about breast cancer biology must be integrated into clinical trials in order to personalize treatment in patients with early breast cancer. The goal is to avoid overtreatment or undertreatment, said Martine Piccart, MD, PhD, chair of the Breast International Group. Dr. Piccart spoke at an SABCS 2008 educational session on clinical trial design.

"Future clinical trials should have a strong translational component and a multidisciplinary research team that includes basic scientists. If not, they will be considered outdated and perhaps even unethical," said Dr. Piccart, who is also director of medicine at the Jules Bordet Institute in Brussels, Belgium.

Thanks to the work of basic scientists, knowledge of the key cancer cell pathways has expanded tremendously, but its effect on patient care is far from realized, she said.

For example, 96,329 breast cancer patients have been studied in clinical trials of taxanes, solidly establishing the risks and benefits of paclitaxel (Taxol) and docetaxel (Taxotere) with a 5% absolute reduction in mortality versus anthracyclines alone.

"But in spite of an astronomical number of patients, we have made minimal progress in understanding which individual patient will benefit from these drugs," she said.

Personalized cancer treatment is still just a dream, in spite of huge progress in the field of basic science.

"Why? Because the basic scientist is not talking to the clinical scientist," Dr. Piccart said.

She said the work of basic scientists is needed to "identify the Achilles heel of the tumor," which will help identify which patients can safely avoid treatment, which will need aggressive treatment, and which drug will be most effective in that individual.

"We know that newer drugs are more effective than older drugs, but some patients may get as much benefit from older -- and much less expensive -- chemotherapy agents. We still have no valid tools to select these patients," she said.

The way to accelerate progress is to boost translational research. Without robust diagnostics, Dr. Piccart predicted that personalized medicine will remain out of reach for the next two decades.

In another lecture delivered as part of the educational session, Susan Hilsenbeck, PhD, a professor of medicine at Baylor College of Medicine in Houston, discusses some of the hurdles that researchers will face. One is to develop better tests for molecularly targeted compounds in trials.

For a number of reasons, the current clinical trial approach may not be applicable, Dr. Hilsenbeck noted. For drugs requiring a molecular target, treatment effects may be masked when the population includes nonexpressors of the target as well as expressors. Furthermore, targeted agents may be cytostatic rather than cytotoxic, and cannot be judged by the traditional measure of tumor response. Even phase I studies may need to be rethought, since maximum tolerated doses may not be the optimal doses for targeted agents.

Finally, the development of biomarkers that serve as surrogates for response is critical in the assessment of the newer agents, Dr. Hilsenbeck said.

"Changes in biomarker expression with treatment will give us a hint of the biological activity of the drug," she said.

Reliable assays for these biomarkers must be developed and independently validated.

Dr. Piccart added that for a biomarker to be clinically useful, it must not only separate responders from nonresponders, but must fulfill additional criteria.

"There have been over 35,000 peer-reviewed articles on candidate biomarkers, and very few have been useful in the clinic," she pointed out.

A useful biomarker will also be prognostic for "good outcomes versus bad outcomes," she said.

She is optimistic that two gene expression assays, the 21-gene recurrence score (OncotypeDX) and the 70-gene signature (Mammaprint), will prove capable of clearly doing so in two large clinical trials now under way.