MATERIALS AND METHODS: Using APOLLO individual patient data (IPD) and OPTIMISMM aggregate covariate data plus pseudo-IPD for outcomes, the APOLLO population was re-weighted to match OPTIMISMM aggregate baseline characteristics. Bayesian posterior distributions of DPd versus PVd for progression-free survival (PFS) and overall survival (OS) were estimated using a likelihood-weighted Bayesian Cox model with fixed weights.

RESULTS: At baseline, APOLLO included a higher proportion of patients who received ≥ 2 prior lines of therapy, were refractory to prior therapies, and had advanced International Staging System stage versus OPTIMISMM, which would otherwise disadvantage APOLLO versus OPTIMISMM. The PFS hazard ratio (HR) favored DPd over PVd at 0.59 (95% credible interval [CrI]: 0.36, 0.89) with 99% probability of DPd superiority versus PVd. The OS HR appeared to favor DPd over PVd at 0.80 (95% CrI: 0.45, 1.30), with 83% probability of DPd superiority versus PVd; however, the estimated OS benefit was not conclusive.