Background: Heterogeneity among cancer cells within a tumor significantly impacts cancer development and resistance to treatment. Currently, no standard measure exists for assessing ITH in routine clinical settings. In this study, we applied a quantitative metric to assess intratumor genetic diversity in NSCLC, using a large, multicenter database.

Methods: We accessed the Tempus database for next-generation sequencing results from patients with NSCLC who underwent immune checkpoint inhibitor (ICIs) monotherapy as their first-line treatment. We calculated the mutant-allele tumor heterogeneity (MATH), defined as the ratio of the distribution’s range to the median of mutant-allele fractions across tumor-specific mutated loci. The primary outcomes were progression-free survival (PFS) and overall survival (OS).

Results: A total of 160 pembrolizumab Significant differences in PFS between high and low MATH groups (MATH-H and MATH-L, respectively), using the 50th percentile as the cutoff, were observed among patients with stage 4 adenocarcinoma who had TMB-H above the 25th percentile (Table). In this subgroup, with 27 patients in the MATH-H/TMB-H group and 22 in the MATH-L/TMB-H group, the MATH-H/TMB-H group showed improved PFS (median PFS 5.19 vs. 1.61 months; p = 0.02). Differences in PFS were also observed among patients with TP53, KEAP1, and MET mutations (p < 0.001, p = 0.02, and p = 0.004, respectively). TP53 mutations were found in 27.6% of the MATH-H group and 27.8% of the MATH-L group, KEAP1 mutations in 8.2% and 10.1%, respectively, and MET mutations in 5.1% and 8.9%, respectively. No significant PFS differences were found in groups with low TMB, other stages, or squamous histology.

Conclusions: Our findings suggest that higher genetic diversity within stage 4 TMB-H lung adenocarcinoma, as quantified by MATH, might be linked to improved clinical outcomes among those who receive pembrolizumab as first-line therapy and supports further exploration of MATH as a predictive/prognostic biomarker.

VIEW THE PUBLICATION