Over the past few decades, intense efforts in breast cancer research, coupled with advancing technologies, have generated an overabundance of biologic information to be exploited for diagnosis and treatment. At the annual meeting of the American Society of Clinical Oncology, the session "Integrating Newer Science into Breast Cancer Prognosis and Treatment: A Review of Current Molecular Predictors and Profiles" focused on one particular aspect of these efforts: the prediction of response to primary chemotherapy.

The session was led by Daniel F. Hayes, MD, of the University of Michigan Comprehensive Cancer Center, and Gabriel N. Hortobagyi, MD, of the University of Texas, M. D. Anderson Cancer Center.

Two new studies were discussed during the session. Each study was based on the hypothesis that molecular markers could be used to determine whether a patient will respond to therapy. This possibility has considerable implications for patient care and quality of life. By identifying biologic predictive factors of response to neoadjuvant therapy, clinicians may be better able to accurately identify patients who will likely benefit from the treatment and to spare unnecessary toxicity for patients who may not benefit.

However, studies of predictive and prognostic markers need to be critically analyzed for scientific rigor. Both Hayes and Hortobagyi emphasized that a major problem with not only these studies but also throughout the entire biomarker field is the variability of study design that can reduce the clinical relevance of study results.

Thierry Petit, MD, of Centre Paul Strauss in France, presented a study in which he and his coauthors compared the value of five markers (SBR grade, hormonal receptors, KI67, HER2, and topoisomerase II alpha status) in predicting response to neoadjuvant anthracycline-based chemotherapy. The results indicate that high tumor grade (SRB grade), absence of hormonal receptors, and high tumor cell proliferation (KI67) were predictive of complete clinical and pathologic response. HER2 and topoisomerase II had no predictive value in this study.

Furthermore, clinical complete response was superior to radiographic complete response in predicting a pathologic complete response. According to Petit, these results are not surprising, as the first three markers are well established, whereas the other two are still controversial.

The results of a similar study focused solely on HER2 expression as a marker of response to neoadjuvant chemotherapy with fluorouracil, doxorubicin, and cyclophosphamide (FAC). Again, no correlation between clinical response or pathologic response was found; however, a correlation between overexpression of HER2 and radiographic response was indicated.

The results suggest HER2 is a weak predictor of response to doxorubicin, but tumors with low HER2 expression also respond. Moreover, HER2 has limited clinical utility as a predictive marker for anthracycline-based chemotherapy.

In their commentary on these two studies, Hayes and Hortobagyi noted several confounding factors that prevent investigators from drawing definitive conclusions. These factors are apparent in many response predictor studies partly because the field has grown so rapidly. They said that the myriad new technologies have contributed an element of chaos to the field of tumor marker analysis. As investigations proceed, they cautioned, new methodologic parameters are needed to improve the identification of predictors of treatment response.

SOURCE:
Annual Meeting of the American Society of Clinical Oncology, May 21, 2002, Orlando, Florida