Sequencing cohorts link genomic features to survival in early-stage lung cancer

A clinical sequencing analysis of 4,097 patients with primary non–small cell lung cancer found that higher tumor mutation burden, whole-genome doubling and mutations in TP53, KEAP1, SMARCA4 and STK11 were associated with worse overall survival, while EGFR mutations were associated with better survival. The study used the MSK-IMPACT 50 K cohort for discovery and included 1,176 patients from the ALCHEMIST adjuvant chemotherapy trial as an independent comparison cohort. Because the analysis lacked stage and treatment data, its findings represent prognostic associations rather than causal effects or evidence that a genomic feature predicts benefit from a particular therapy.

Early-stage non–small cell lung cancer has an unresolved survival risk despite surgery performed with curative intent: approximately 35% of patients die within five years. The study examined whether tumor genomic characteristics could help explain differences in overall survival among patients whose cancers were subjected to large-scale clinical sequencing. The investigators evaluated tumor mutation burden, whole-genome doubling, individual gene mutations and mutational signatures using multivariable Cox regression adjusted for age, sex and tumor mutation burden.

Co-mutation patterns were assessed in 3,064 patients whose tumors had been sequenced on uniform gene panels. Tumor mutation burden was independently associated with poorer overall survival in the discovery cohort. For each log10 unit increase in tumor mutation burden, the adjusted hazard ratio for death was 1.34, with a 95% confidence interval of 1.22 to 1.49 and P = 8.2 × 10⁻⁹.

Whole-genome doubling was also independently associated with worse survival, with an adjusted hazard ratio of 1.35, a 95% confidence interval of 1.21 to 1.51 and P = 8.9 × 10⁻⁸. The analysis detected no significant interaction between the two measures, for which P was 0.25. Gene-level analysis identified four mutations associated with adverse outcomes.

Mutations in SMARCA4 carried an adjusted hazard ratio of 1.45 and P = 4.7 × 10⁻⁵; TP53 mutations had a hazard ratio of 1.42 and P = 1.7 × 10⁻⁹; KEAP1 mutations had a hazard ratio of 1.41 and P = 2.6 × 10⁻⁶; and STK11 mutations had a hazard ratio of 1.25 and P = 0.002. By contrast, EGFR mutations were associated with lower mortality risk, with a hazard ratio of 0.72 and P = 9.4 × 10⁻⁷. These estimates describe adjusted associations with overall survival in the analyzed population; they do not establish that the mutations caused the outcomes or that they modified the effectiveness of surgery, chemotherapy or another treatment.

The ALCHEMIST cohort supplied an independent clinical-trial population for comparison. The reported cross-cohort result concerned the prevalence of STK11 mutations rather than every survival estimate from the discovery analysis. STK11 mutations were substantially underrepresented in the trial cohort compared with the real-world MSK-IMPACT cohort: 6.9% versus 13.7%, respectively, yielding an enrichment ratio of 0.51.

This difference shows that the two populations did not have the same distribution of at least one adverse-prognosis feature. The authors argue that the depletion of adverse-prognosis mutations in the adjuvant chemotherapy trial population warrants attention in biomarker-driven trial design. The comparison should be interpreted in light of the study’s observational framework and missing clinical information.

The absence of stage and treatment data limits the analysis’s ability to separate genomic associations from possible confounding by disease severity, management or other clinical factors. It also prevents the survival estimates from being interpreted as treatment-predictive effects. Within these limits, the analysis linked several tumor genomic features with markedly different overall-survival risks among patients with early-stage non–small cell lung cancer.

Each log10-unit increase in tumor mutation burden corresponded to an adjusted hazard ratio of 1.34, while whole-genome doubling corresponded to an adjusted hazard ratio of 1.35. At the gene level, the largest reported hazard ratio among the four adverse markers was for SMARCA4, at 1.45. TP53, KEAP1 and STK11 mutations were also associated with higher hazards, whereas EGFR mutations were associated with a lower hazard of death.

The authors concluded that tumor mutation burden, whole-genome doubling and mutations in TP53, KEAP1, SMARCA4 and STK11 were associated with worse overall survival in this setting, while EGFR mutation was associated with better survival. They also identified the underrepresentation of STK11 mutations in the adjuvant chemotherapy trial population as an issue for biomarker-driven trial design. The results define candidate prognostic patterns and a measured difference between the analyzed cohorts, but the available evidence does not establish their clinical utility or show that any of these genomic features should guide adjuvant treatment selection.

Subject of Research: Prognostic genomic markers in early-stage non–small cell lung cancer

Article Title: Genomic determinants of overall survival in early-stage non–small cell lung cancer: a large-scale clinical sequencing cohort analysis with independent cohort comparison

Article References: Chen, S., Lin, X., Zhang, W., He, R., Cai, W., & Li, K. (2026). Genomic determinants of overall survival in early-stage non–small cell lung cancer: a large-scale clinical sequencing cohort analysis with independent cohort comparison. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09098-7

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09098-7

Keywords: non–small cell lung cancer, overall survival, tumor mutation burden, whole-genome doubling, SMARCA4, TP53, KEAP1, STK11, EGFR, clinical sequencing, Genomic, determinants