Indication-specific tumor evolution and its impact on neoantigen targeting and biomarkers for individualized cancer immunotherapies

Abstract

Individualized neoantigen specific immunotherapy (iNeST) requires robustly expressed clonal neoantigens for efficacy, but tumor mutational heterogeneity, loss of neoantigen expression, and variable tissue sampling present challenges. To characterize these potential obstacles, we combined multi-region sequencing (MR-seq) analysis of five untreated, synchronously sampled metastatic solid tumors with re-analysis of published MR-seq data from 103 patients. Branching evolution in colorectal cancer and renal cell carcinoma led to fewer clonal neoantigens and to clade-specific neoantigens (those shared across a subset of tumor regions but not fully clonal), with the latter not being readily distinguishable in single tumor samples. Prioritizing mutations with higher purity- and ploidy-adjusted variant allele frequency enriched for globally clonal neoantigens (those found in all tumor regions), whereas estimated cancer cell fraction derived from clustering-based tools, surprisingly, did not. Neoantigen quality was associated with loss of neoantigen expression in the bladder cancer case, and HLA-allele loss was observed in the renal and non-small cell lung cancer cases. Our results show that indication type, multi-lesion sampling, neoantigen expression, and HLA allele retention are important factors for iNeST targeting and patient selection.