Durvalumab (PD-L1) + platinum/pemetrexed: biomarker signals (BAP1/DDR, LOH) emerge despite DREAM3R OS miss in 1L pleural mesothelioma

Durvalumab is an anti–PD-L1 monoclonal antibody combined here with first-line platinum/pemetrexed chemotherapy in unresectable pleural mesothelioma (DREAM3R, randomized Phase 3; NCT04334759) that stopped early for slow accrual (n=173; 94% epithelioid) and did not meet its primary OS endpoint. In translational analyses, PD-L1 TPS ≥1% trended to higher ORR vs <1% (65% [31/48] vs 49% [45/92]; p=0.077), but PD-L1 and overall TMB were not associated with PFS or OS; higher clonal multicopy mutations correlated with longer OS (24.44 vs 17.61 months; p=0.0082), while higher genome-wide LOH predicted shorter OS within the D+Ch arm (15.77 vs 24.57 months; p=0.0042). Potential predictive pockets for D+Ch vs Ch on PFS were hinted in tumors with inactivating BAP1 (9.89 vs 6.21 months; p=0.064), homozygous DDR gene deletions (6.14 vs 5.75 months; p=0.069), and LATS2 wild-type (7.79 vs 5.9 months; p=0.027)—but these are subset, underpowered, multiple-comparison–prone findings; clinically, current 1L SOC often uses nivolumab + ipilimumab (CheckMate 743: mOS 18.1 vs 14.1 months vs chemo; HR 0.74), so any chemo-IO strategy needs clearer OS value or a crisp biomarker-defined niche.