In a recent Breast Cancer Action Newsletter (February/March 2004), Musa Mayer expresses her doubts about new genetic research presented at the 26th San Antonio Breast Cancer Symposium. [Abstracts are available at www.sabcs.org, by clicking on ‘Abstract Online 2003′ and logging in as a guest.] Mayer says that the hope that genetic analysis of a person’s cancer would lead to individualized prognosis and treatment recommendations is not living up to expectations. Researchers are developing new drugs that target specific proteins in a cancer cell, but these targeted therapies only work in specific cancers. For example, Herceptin[R], which received FDA approval five years ago, is a monoclonal antibody that targets cancer cells that make too much HER-2, a protein found on the surface of about 25-30% of breast cancer cells. Genentech, the drug’s manufacturer, developed a test to identify which tumors might respond to Herceptin. Without targeting a specific population, the drug would not show a benefit in breast cancer trials. Genentech has developed another drug called Avastin[R] that targets VEGF, a gene responsible for angiogenesis, but has not yet found a way to identify patients who might benefit from it. Like many of these drugs, it does not show any significant effect among the general population in clinical trials.
The test was credited with predicting outcome better than any other single prognostic factor except for tumor grade. The test was not 100% accurate. One observer at the San Antonio Symposium noted that 10% of the patients in the low-risk group were misclassified. Dr. Soonmyung Paik, director of pathology, the National Surgical Adjuvant Breast and Bowel Project, worked with Genomic Health on this study. He admitted that it may never be possible to predict whether cancer will recur in an individual patient. Statistician Donald Berry of the MD Anderson Cancer Center concurred, noting that the complexity of genetic analysis, difficulties with uniform data collection and pathology, and other methodological and statistical issues make interpretation of genetic tests problematic.