Evaluating Traditional Phase 2 Trial Design Of Systemic Therapies In Recurrent/Advanced Cervical Cancer: A Nrg Oncology/Gynecologic Oncology Group (Gog) Ancillary Data Study.

e16592 Background: The identification of effective systemic agents for recurrent cervical cancer has been illusive despite multiple phase 2 studies. This analysis of ancillary data from phase 2 GOG studies was conducted to explore whether traditional endpoints of tumor response, progression-free survival (PFS) and overall survival (OS), were reflected in subsequent phase 3 trials. The aim was to generate hypotheses for refining future trial designs and to identify prognostic factors for response to treatment. Methods: From 1995 to 2012, all 35 phase 2 trials in recurrent cervix cancer were included. Patient, treatment, and outcome data were collected from a database and individual chart review. Generalized estimating equations estimated overall response rates; multivariate analysis determined factors associated with tumor response. PFS and OS were compared based on agents that went forward to phase 3, and by clusters of prognostic factors as defined by the Moore criteria. Results: 1237 eligible patients were included. Only 11% (95% CL 8-15%) had a complete or partial response. Factors associated with better tumor response included performance status of 0, time to first relapse > 1 year, no prior platinum chemotherapy, and non-African-American race. The odds for low risk patients (0-1 factors) to respond to therapy was 6.03 times (95% CL 1.69 – 21.59, p = 0.0057) the odds for high risk patients (4-5 factors). The number of risk factors influences overall survival (high risk vs low risk patients HR 1.9 [1.42-2.55], p < 0.0001). There were no statistically significant differences in PFS (p = 0.09) or OS (p = 0.20) for patients on studies of drugs rejected vs accepted for phase 3 study. Conclusions: Factors prognostic for response to treatment include performance status, time to first relapse, prior platinum chemotherapy, and race. Stratifying risk groups should be strongly considered during the design of future trials. It is unclear what criteria should be used to select agents for subsequent phase 3 studies, as current criteria are not robust. This information drives the need for refining clinical trial design and selection criteria for effectiveness.