Bevacizumab treatment for solid tumors: boon or bust?

IN 1971, FOLKMAN 1 PROPOSED THAT MALIGNANCIES COULD only grow to a significant size, or metastasize to other organs, by stimulating new blood vessel growth. This process, designated angiogenesis or neovascularization, is now accepted as necessary for cancers to extend beyond the in situ state and metastasize to distant sites. Tumorinduced angiogenesis requires a complex interplay between the cancer cell and surrounding stroma at the primary cancer site and at the metastasis site, including recruitment of bone marrow–derived, circulating endothelial cell precursors. Folkman’s proposed mechanism led to development of therapeutic agents that target tumor vasculature, including antibodies against endothelial cell growth factor receptors and inhibitors of tyrosine kinases that are presumed to contribute to neovascularization. Arguably the most successful strategy to target cancer-associated neovascularization consists of a monoclonal antibody that binds to vascular endothelial growth factor (VEGF) A. This antibody, bevacizumab, was shown in early phase 1 studies to induce occasional responses as a single agent and to enhance chemotherapy-induced tumor shrinkage. The subsequent clinical development of bevacizumab has included successes and disappointing failures. Following promising phase 2 trials, a randomized controlled trial (RCT) of capecitabine with or without bevacizumab demonstrated a doubling of response rates in participants receiving bevacizumab but no difference in progressionfree or overall survival among women with metastatic breast cancer whose disease had previously progressed during standard anthracyclineand taxane-based therapies. Favorable results for patients with colorectal cancer, non–small cell lung cancer (NSCLC), and renal cell cancer suggested that bevacizumab had a role in treatment of patients with these malignancies, whereas negative RCT results for patients with pancreatic and prostate cancers tempered enthusiasm. Perhaps the most promising findings for bevacizumab were observed by investigators from the Eastern Cooperative Oncology Group (ECOG), who reported that, when administered with paclitaxel as first-line chemotherapy for patients with measurable metastatic breast cancer, the antibody not only increased response rates but appeared to double the time to progression. Curiously, few of these studies demonstrated a survival benefit from the addition of bevacizumab to chemotherapy. Explanations for this observation included early follow-up, insufficient statistical power, and the confounding effects of subsequent postprogression treatments. Nonetheless, the US Food and Drug Administration (FDA) approved bevacizumab for treatment of colorectal cancer, NSCLC, and renal cell cancer. The FDA approval for breast cancer was controversial, including an initial rejection based on concern that the ECOG study was not placebo-controlled. After an objective, third-party review of the radiographs used for this determination provided similarly favorable findings, the FDA provided accelerated approval of bevacizumab for metastatic breast cancer, overriding the recommendation against approval made by the Oncology Drug Advisory Committee (ODAC). The ODAC expressed particular concern about the lack of overall survival benefit and the rare but occasional life-threatening toxic effects reported with bevacizumab. Recently, the FDA has rescinded its approval of bevacizumab for metastatic breast cancer, based on the results of 2 other RCTs that were less favorable and suggested uncommon but occasional lethal adverse events. The findings for metastatic disease stimulated the conduct of several RCTs of bevacizumab in the adjuvant setting. Again, results have been mixed. Results were favorable for progression-free survival, but not overall survival, for gynecologic cancers. However, results were negative for all end points in colorectal cancers. These inconsistent results raise important questions: Does bevacizumab have cancer-specific benefits? Is it more effective when combined with a chemotherapeutic partner, and if so, is this benefit specific for different agents? Why, despite the impressively solid preclinical data and the promising early clinical results, has bevacizumab not been more successful in improving overall survival?