Background

Unsuccessful KRAS-specific treatment approaches in non-small-cell lung cancer (NSCLC) might reflect underlying disease heterogeneity. We sought to define clinical "syndromes" within advanced KRAS mutant NSCLC to improve future clinical trials and create a clinical framework for future molecular development.

Patients and methods

To test a series of a priori hypotheses regarding KRAS-mutant NSCLC clinical syndromes, we conducted a multi-institutional retrospective medical record review. Survival probabilities were estimated using the Kaplan-Meier model. Between-group differences were assessed using the log-rank test. Multivariate Cox regression analyses and Wilcoxon rank sum testing were used to assess progression-free survival and overall survival (OS) differences.

Results

Among 218 patients with advanced KRAS-mutant NSCLC, OS and progression-free survival with first-line chemotherapy did not differ by intrathoracic versus extrathoracic spread, smoking intensity, or the specific KRAS mutation. Metastatic disease at diagnosis resulted in significantly worse OS than recurrent, unresectable disease (median OS, 14.6 vs. 40.9 months; P = .001). Among the patients with metastatic disease at diagnosis, nonscalp, soft tissue metastases (syndrome X; 6% of cases; 95% confidence interval [CI], 2.5%-10.1%) signified a poor prognosis (median OS, 7.5 vs. 15.9 months for the controls; P = .021). The response to first-line chemotherapy (syndrome Y; 41% of cases; 95% CI, 32.3%-50.6%) signified a good prognosis (median OS, 26.7 vs. 11.9 months; P = .002). The overlap between these 2 syndromes was minimal (2 of 111). Multivariate analysis confirmed these observations. The hazard ratio for death for syndromes X and Y was 2.64 (95% CI, 1.13-6.14) and 0.45 (95% CI, 0.28-0.76), respectively.

Conclusion

Chemotherapy-responsive disease and nonscalp, soft tissue spread might represent distinct clinical syndromes within KRAS-mutant NSCLC. The molecular biology underlying this heterogeneity warrants future studies.