Genetic signature for lung cancer brain metastases. Use of Comprehensive Genomic Profiling for Biomarker Discovery for the Management of Non-Small Cell Lung Cancer Brain Metastases

Background Clinical biomarkers for brain metastases remain elusive. Increased availability of genomic profiling has brought discovery of these biomarkers to the forefront of research interests. Method In this single institution retrospective series, 130 patients presenting with brain metastasis secondary to Non-Small Cell Lung Cancer (NSCLC) underwent comprehensive genomic profiling conducted using next generation circulating tumor DNA (Guardant Health, Redwood City, CA). A total of 77 genetic mutation identified and correlated with nine clinical outcomes using appropriate statistical tests (general linear models, Mantel-Haenzel Chi Square test, and Cox proportional hazard regression models). For each outcome, a genetic signature composite score was created by summing the total genes wherein genes predictive of a clinically unfavorable outcome assigned a positive score, and genes with favorable clinical outcome assigned negative score. Results Seventy-two genes appeared in at least one gene signature including: 14 genes had only unfavorable associations, 36 genes had only favorable associations, and 22 genes had mixed effects. Statistically significant associated signatures were found for the clinical endpoints of brain metastasis velocity, time to distant brain failure, lowest radiosurgery dose, extent of extracranial metastatic disease, concurrent diagnosis of brain metastasis and NSCLC, number of brain metastases at diagnosis as well as distant brain failure. Some genes were solely associated with multiple favorable or unfavorable outcomes. Conclusion Genetic signatures were derived that showed strong associations with different clinical outcomes in NSCLC brain metastases patients. While these data remain to be validated, they may have prognostic and/or therapeutic impact in the future.