Clinical interpretation of variants in CHIP-associated genes detected using tumor-only comprehensive genomic profiling in NSCLC

Introduction The interpretation of variants in genes associated with clonal hematopoiesis of indeterminate potential (CHIP) detected by tumor-only comprehensive genomic profiling (CGP) remains challenging. We investigated the prognostic significance of variants in CHIP-associated genes and evaluated whether the normal cell fraction (NCF) framework could be extrapolated to tumor-only CGP to provide clinically meaningful stratification in NSCLC. Methods We analyzed 3294 patients with NSCLC who underwent tumor-only CGP via FoundationOne CDx and were registered in the Center for Cancer Genomics and Advanced Therapeutics. Forty-six CHIP-associated genes were evaluated. Associations between mutations and post-CGP overall survival (OS) were assessed using the Cox proportional hazards model. NCF-based classification used an exploratory threshold of NCF ≤ 1. The associations of NCF and variant allele frequency (VAF) with OS were evaluated. Results TP53 and SETD2 mutations remained significantly associated with OS after multivariable adjustment, with TP53 associated with shorter OS and SETD2 associated with longer OS. In TP53 -mutated cases, NCF-based classification stratified survival outcomes; however, NCF was strongly correlated with VAF (Spearman's ρ = 0.91; P < 0.0001). Restricted cubic spline analyses showed no clear VAF threshold, with no significant association after multivariable adjustment ( P = 0.24). In univariable continuous-variable analyses, VAF, but not NCF, was associated with OS, although both showed limited prognostic discrimination (C-indices, 0.54 and 0.53, respectively). Within SETD2 -mutated cases, NCF-based classification did not provide meaningful prognostic stratification. Conclusions Although NCF-based classification stratified the prognosis of TP53 -mutated NSCLC, this separation largely reflected TP53 VAF distribution rather than biologically distinct subgroups.