8042 Background: Patients with stage I-III non-small cell lung cancer (NSCLC) often have very different outcomes, even when they share the same anatomic stage. This suggests that tumor biology plays an important role in prognosis. Fraction genome altered (FGA), a measure of chromosomal instability derived from next-generation sequencing (NGS), is already used in clinical practice but its prognostic value in early-stage NSCLC is not well established. Therefore, we evaluated whether FGA is associated with overall survival (OS) in early-stage NSCLC and whether it provides prognostic information beyond standard clinical factors and tumor mutation burden (TMB). Methods: We performed a retrospective analysis of 7,722 patients with NSCLC using clinical-grade targeted NGS with available copy-number data (MSK CHORD 2024). Disease stage was categorized as stage I-III versus stage IV. FGA was analyzed in quartiles, with high FGA defined as the top quartile. OS was assessed using Kaplan-Meier and multivariable Cox proportional hazards models adjusting for age, sex, smoking history, histology group, sequencing panel, stage, and log-transformed TMB. Incremental prognostic value was evaluated using concordance (C-index) and likelihood ratio testing (LRT). The results were validated using several sensitivity/statistical analysis. Results: After quality control, 7,722 patients were included (3,934 deaths), including 4,343 with stage I-III disease (1,559 deaths). In early-stage NSCLC, increasing FGA was associated with progressively worse OS across quartiles (log-rank χ² = 199.6, p < 0.0001). 5-year OS declined from 65.6% in the lowest FGA quartile to 39.6% in the highest quartile. Patients with high FGA had significantly inferior survival compared with low-FGA patients (5-year OS 39.6% vs 63.0%, p < 0.0001). In multivariable models restricted to stage I-III, high FGA remained independently associated with mortality (Hazard-Ratio = 1.60, 95% CI: 1.42-1.82, p < 0.0001). Adding FGA to clinical models improved survival prediction (C-index = ~0.70) and overall model performance (LRT = 125.9, p < 0.0001), and this improvement was greater than that seen with TMB. When both biomarkers were included, FGA continued to provide strong prognostic information, whereas TMB did not improve prediction once FGA was already in the model. We found our results to be consistent across internally validated analysis. Conclusions: FGA was strongly associated with OS in stage I-III NSCLC. FGA identified a subgroup of early-stage patients with substantially worse outcomes despite similar anatomic stage, indicating important underlying biologic differences. Since, FGA can be obtained from routine targeted NGS already used in clinical practice, it presents a practical biomarker that may improve risk stratification beyond standard clinical factors and TMB, and help guide surveillance and perioperative trial selection.
Shrestha et al. (Thu,) studied this question.