Impact of the algorithm for declaring exact progression date on progression-free survival (PFS) estimates in advanced lung cancer clinical trials.

Abstract

7629 Background: PFS based endpoints are controversial due to ascertainment bias, missing assessments, and symptomatic (nonradiographic) progression. We evaluated the impact of progression date determination on PFS estimates in advanced non-small cell lung cancer (NSCLC), and extensive stage small cell lung cancer (SCLC). Methods: Individual patient (pt) data from 8 consecutive NSCLC (n=485) and 4 SCLC (n=116) trials conducted through NCCTG were pooled. The disease assessment schedule was every 6 or 8 weeks (6 trials) during treatment, and every 3 or 6 months (6 trials) post treatment. The reported progression date (RPD) was defined as either the scan date for radiographic progression or the clinical deterioration date determined by the treating physician for nonradiographic progression. Disease progression date was determined by using: RPD (method 1), midpoint between last progression-free scan date and RPD (method 2), or one day post last progression-free scan date (method 3). PFS was defined as the earlier of time from registration to progression or death from any cause, using the 3 methods, and estimated using the method of Kaplan-Meier. Results: Within NSCLC (SCLC) 53% (58%) of pts reported progression during treatment, and 12% (7%) of pts died without progression. The median time from the last progression-free scan to progression for NSCLC (SCLC) during and post treatment was 1.4 (1.4) months (mos), and 2.8 (3.0) mos. 51% of pts did not have a progression-free scan post-baseline in SCLC. The median PFS ranged from 2.3-5.6 mos (method 1), 1.6-4.3 mos (method 2), and 0.03-2.9 mos (method 3) within NSCLC. The results were similar for SCLC trials. Conclusions: PFS estimates differed substantially across the 3 methods, with the magnitude of difference large enough to conclude efficacy (observed range: 30% to >100%). A standard for progression date determination should be adopted to ensure accurate comparisons between trials and to guide trial design. Progression date determination method Trial type NSCLC Median (95% CI) months SCLC Median (95% CI) months 1 3.9 (3.6-4.5) 2.7 (1.6–4.2) 2 3.0 (2.6–3.4) 1.8 (0.8–3.3) 3 1.8 (1.6-1.9) 0.03 (0.03–1.4) No significant financial relationships to disclose.