Impact of phase II trials with progression-free survival as end-points on survival-based phase III studies in patients with anaplastic gliomas

BACKGROUND: To assess progression-free survival (PFS) as the appropriate end-point for phase II trials for anaplastic gliomas (AGs) and to determine the impact of PFS on survival-based phase III trials. METHODS: Combined data from 16 phase II studies (N = 529 patients) were analyzed to determine progression-free survival (PFS) at 6, 9, and 12 months and the impact of age, Karnofsky performance score (KPS), number of prior chemotherapies, and response to treatment on PFS. RESULTS: The specific chemotherapy used was the major effector of PFS at 6, 9, and 12 months. Age, KPS, treatment response rate, and number of prior chemotherapies did not affect PFS to the same extent. Hierarchical cluster analyses and linear least squares fitting of PFS(9 )v PFS(12 )demonstrated the existence of three therapeutic efficacy groups with PFS rates at 6, 9, and 12 months ranging from lowest (A) to highest (C). The PFS(6 )was 15% in group A and 41% in group C (p < .0001); the PFS(12 )was 9% in group A and 33% in group C (p < .0001). Further, 80% of patients at recurrence had a 23% likelihood that each chemotherapy would provide > 1 year of additional life. CONCLUSION: Based on PFS rates at 6, 9, and 12 months for AG patients, a differential of 1.5 to 2 years is the norm and could invalidate overall survival as an end-point for phase III studies in patients with AG. PFS is a more reliable end-point because it reflects the true antitumor benefit of the chemotherapy.