Predictive value of genomic mutation signatures and tumor mutational burden for induction immunochemotherapy in locally advanced squamous cell carcinoma of the oropharynx, larynx, and hypopharynx.

Abstract

e14592 Background: The predictive role of programmed death-ligand 1 (PD-L1) expression in patients (pts) with head and neck squamous cell carcinoma (HNSCC) is imperfect: clinically meaningful responses are observed in a subset of pts with low PD-L1 expression, highlighting the need to identify additional biomarkers. This study evaluated the association of genomic mutations and tumor mutational burden (TMB) with the efficacy of induction immunochemotherapy (ICT) in HNSCC. Methods: Molecular genetic analysis was performed on tumor biopsies from pts enrolled in a prospective phase II study of ICT with pembrolizumab plus platinum-based chemotherapy for locally advanced HNSCC between 2022 and 2024 (NCT05551767). The clinical relevance of genomic alterations was classified according to the ESMO Scale for Clinical Actionability of molecular Targets (ESCAT). The association of the spectrum of somatic mutations and copy number alterations with objective response rate (ORR) and progression-free survival (PFS) on ICT was analyzed. TMB (mutations per megabase, mut/Mb) and microsatellite status were assessed using next-generation sequencing. Results: A total of 64 tumor samples were analyzed. The most frequently mutated genes were TP53 (n = 18, 36%), SDHA (n = 11, 22%), CDKN2A (n = 6, 12%), PIK3CA (n = 6, 12%), NOTCH1 (n = 6, 12%), and NTRK3 (n = 6, 12%). Copy number alterations were dominated by amplifications of FGF19 (n = 16, 32%), FGF4 (n = 16, 32%), FGF3 (n = 16, 32%), and CCND1 (n = 15, 30%). Four tumors (6%) had TMB ≥10 mut/Mb; no microsatellite instability was detected. TMB ≥10 mut/Mb was the only biomarker meeting ESCAT tier I evidence. All 4 pts with TMB ≥10 mut/Mb achieved a partial response to ICT; only 1 pt experienced disease progression within 12 months after treatment initiation, while 3 pts remained progression-free with follow-up of 20–27 months. Among 58 evaluable pts with TMB < 10 mut/Mb, 37 (63.8%) achieved an objective response after 3 cycles of ICT. There was no statistically significant difference in ORR between TMB ≥10 mut/Mb and < 10 mut/Mb cohorts (p = 0.35). High TMB also did not significantly impact PFS (hazard ratio 0.46, 95% confidence interval 0.06–3.47, p = 0.45). Conclusions: The high clinical activity of ICT observed in this cohort cannot be explained by TMB as a predictive biomarker. These findings suggest that alternative biomarkers are required to better predict benefit from induction ICT in locally advanced HNSCC.