Biomarkers of treatment benefit with atezolizumab plus vemurafenib plus cobimetinib in BRAF

The phase III IMspire150 study (NCT02908672) demonstrated significantly improved progression-free survival (PFS) with atezolizumab, vemurafenib, and cobimetinib (atezolizumab group) versus placebo, vemurafenib, and cobimetinib (control group) in patients with BRAFFive hundred fourteen patients were randomized to atezolizumab (n = 256) or control (n = 258). Outcomes were evaluated in subgroups defined by key biomarkers, including programmed death-ligand 1 (PD-L1) expression, lactate dehydrogenase (LDH) level, tumor mutational burden (TMB), and interferon-γ (IFN-γ) gene signature. Exploratory recursive partitioning analysis was then used to model associations between PFS and baseline covariates, including key biomarkers.PFS benefit for atezolizumab versus control was greater in patients with high TMB [≥10 mutations/Mb; hazard ratio (HR) 0.73; 95% confidence interval (CI) 0.52-1.02; P = 0.067] versus low TMB (10 mutations/Mb; HR 0.92; 95% CI 0.65-1.30; P = 0.64) and similar between patients with strong IFN-γ (≥median; HR 0.76; 95% CI 0.54-1.06) versus weak IFN-γ (median; HR 0.79; 95% CI 0.58-1.08). In patients with elevated LDH, PFS benefit for atezolizumab versus control was greater in the PD-L1- subgroup (HR 0.53; 95% CI 0.29-0.95; P = 0.032) than in the PD-L1+ subgroup (HR 1.16; 95% CI 0.75-1.80; P = 0.51). Recursive partitioning analysis showed that IFN-γ discriminated PFS outcomes in patients with normal LDH, whereas TMB discriminated outcomes in patients with elevated LDH in the atezolizumab group. Neither IFN-γ nor TMB discriminated PFS outcomes in the control group.Treatment benefits in the atezolizumab group seemed to be most evident in patients with elevated LDH and PD-L1- tumors. LDH remains the primary predictor of outcomes regardless of treatment. IFN-γ and TMB further differentiate outcomes for patients treated with atezolizumab, vemurafenib, and cobimetinib.