Atezolizumab (PD-L1) + MAPK-targeted therapy links ATC long survival to “inflamed + differentiated” tumors, not BRAF/RAS status

Atezolizumab is an anti–PD-L1 checkpoint inhibitor combined here with MAPK-pathway targeted therapy—BRAF V600E patients received vemurafenib (BRAF inhibitor) + cobimetinib (MEK inhibitor), while RAS/NF-mutant patients received cobimetinib alone—in a Phase 2 ATC trial (NCT03181100). Median OS numerically favored the BRAF V600E cohort (43 months; 95% CI 16–NE) vs the RAS/NF-mutant cohort (8.7 months; 95% CI 5.1–37.0), but with striking overlap (BRAF V600E OS as short as 81 days; RAS/NF OS up to 1702 days); correlative WES/transcriptomics (pretreatment n=14–16; OS groups: <2 years n=9 vs >2 years n=7) found no OS-linked co-mutations, but did find that longer OS tracked with epithelial/differentiation programs (e.g., PAX8) plus immune activation/antigen presentation (MHC-II) signatures, while shorter OS tracked with EMT, hypoxia, and oncofetal programs—patterns that held “within oncogene” subsets and showed therapy-consistent nuance (e.g., MAPK/ERK activity associated with better OS in BRAF V600E, worse OS in RAS). Clinically, this is directionally consistent with what we see in ATC benchmarks: BRAF V600E-targeted therapy dabrafenib + trametinib has reported ORR ~69% and median OS ~86 weeks in prospective data, while single-agent PD-1 blockade (spartalizumab) has shown ORR ~17–19% in ATC—i.e., this abstract’s key contribution is less “new best OS number” than a plausible baseline biology readout to explain why outcomes can diverge dramatically even within BRAF or RAS groups.