Biomarkers associated with efficacy of first-line nivolumab vs. sorafenib in advanced hepatocellular carcinoma: results from the CheckMate 459 phase III trial.

BACKGROUND & AIMS: First-line nivolumab demonstrated a higher objective response rate (ORR) and favorable safety profile vs. sorafenib in advanced HCC in the phase 3 CheckMate 459 trial; however, no significant improvement in overall survival (OS) was observed. Here, we report exploratory biomarker analyses from CheckMate 459. METHODS: Archival or fresh tumor samples were collected before treatment and subjected to whole-exome sequencing, whole-transcriptome RNA sequencing, immunohistochemistry, and serological testing. Biomarkers were examined for association with clinical outcomes (ORR by blinded independent central review [BICR] per RECIST v1.1, progression-free survival [PFS] by BICR, and OS). RESULTS: The minimum follow-up was 33.6 months. In the nivolumab arm, 201, 230, 343, and ≥ 293 patients were evaluable by whole-exome sequencing, RNA sequencing, immunohistochemistry, and serology; 199, 239, 334, and ≥ 279 patients were evaluable in the sorafenib arm, respectively. Baseline tumor mutational burden was not predictive of survival with nivolumab. OS outcomes with sorafenib were improved in patients with mutated vs. wild-type Wnt/β-catenin pathway; consequently, OS was improved with nivolumab vs. sorafenib in patients with unaltered Wnt/β-catenin pathway (HR 0.66; 95% CI 0.47-0.91). High T-cell inflammation gene signature score and tumors with high programmed death ligand 1 levels (combined positive score ≥10) had improved OS with nivolumab vs. sorafenib (HR 0.68; 95% CI 0.47-0.98 and HR 0.55; 95% CI 0.33-0.93, respectively); PFS and ORR were also improved. Among circulating blood biomarkers, high interferon-γ, low pro-collagen 11, and low frequency of granulocytic myeloid-derived suppressor cells were associated with improved nivolumab efficacy vs. sorafenib. CONCLUSIONS: These exploratory analyses from CheckMate 459 identified several candidate biomarkers that may be associated with efficacy of first-line nivolumab vs. sorafenib and support further studies to identify patients with advanced HCC who may benefit from programmed death-1 inhibition. IMPACT AND IMPLICATIONS: Nivolumab demonstrated a higher objective response rate and favorable safety profile as first-line treatment compared with sorafenib for patients with advanced HCC in the CheckMate 459 trial, yet did not show a significant improvement in overall survival. The exploratory analyses presented here have identified several candidate biomarkers that may be associated with efficacy of nivolumab in the first-line setting vs. sorafenib. These findings support additional studies to identify patients who may benefit from programmed death-1 inhibition treatment. CLINICAL TRIAL NUMBER: NCT02576509.