Efficacy translation between phase II and phase III trials in gynecologic oncology: a cross-phase meta-comparative...

The reliability of phase II results informing phase III outcomes has been inconsistent across multiple tumor treatment settings. We aim to quantify the magnitude of treatment effect attenuation between phase II and III gynecologic oncology trials.

We systematically identified phase III gynecologic cancer trials published between 2010 and 2025 through ClinicalTrials.gov. Each phase III study was paired with its preceding phase II trial evaluating the same intervention. Study demographics, design, and efficacy outcomes were extracted. Correlations between phase II and phase III progression-free survival and overall survival were assessed using Spearman's rank coefficients, and directional differences were compared with paired and non-parametric tests.

Of 221 screened phase III trials, 25 matched phase II-phase III pairs were included. The median progression-free survival data were 11.3 months (inter-quartile range; 6.1-16.8) in phase III versus 7.2 months (inter-quartile range; 4.1-13.2) in phase II (p =.11). Progression-free survival data values were correlated (ρ = 0.59, p =.005). Phase III superior pairs (n = 13) had a median gain of 4.6 months (inter-quartile range; 2.2-9.4) versus 2.3 months (inter-quartile range; 0.5-9.8) for phase II superior pairs. The median overall survival was 20.4 months (inter-quartile range; 13.2-38.2) versus 14.9 months (inter-quartile range; 10.8-21.3) (p =.06); overall survival values were correlated (ρ = 0.57, p =.002). The median overall survival gain was 5.9 months (inter-quartile range; 2.7-17.0) for phase III superior versus 3.2 months (inter-quartile range; 2.8-3.5) for phase II superior pairs (p =.022). The hazard ratios for progression-free survival and overall survival did not differ significantly between phases and were not correlated.

Phase II and phase III gynecologic oncology trials demonstrate moderate concordance of progression-free and overall survival, without systematic inflation of early-phase effect estimates. Greater use of randomized phase II designs and improved demographic reporting may enhance translational validity.