Most trials in metastatic breast cancer do not use overall survival as the primary endpoint, according to Dr. Robert. "It's a very uncommon endpoint," he said. "It is always nice to see [an improvement in overall survival] achieved, but I don't think it should be the sole metric or a necessary metric if you see a significant progression-free survival with acceptable toxicity."

Using overall survival as an endpoint "is a very big challenge for a number of reasons," he noted. While a crossover design "is a bit of a carrot to encourage people to participate in a clinical trial," it also means "you are reducing the chance of showing an impact on overall survival, because there may be some benefit in second-line treatment." That was actually demonstrated in RIBBON-2, where everybody gets first-line chemotherapy, and then is randomly assigned to chemotherapy or chemotherapy plus bevacizumab. Thus, some patients are getting bevacizumab in the second-line setting, where there was a progression-free survival of about 2 months.

Another reason overall survival is challenging as an endpoint is because breast cancer is quite heterogeneous, Dr. Robert said. Although median survival is 2 to 3 years, many patients are living even longer. ■