*2.4. Statistical Considerations*

The primary end point of the phase II study was the rate of response to combination chemotherapy. The study was powered to detect a significant improvement of 18% relative to the 5% estimated from previous studies [1,3]. Assuming a one-sided a = 0.05% and 80% power, sample size was calculated to be 39 patients, with 6 patients at the recommended dose level in the phase I study. PFS and OS were assessed by the Kaplan-Meier method and log-rank test.
