REPRODUCIBILITY OF A 12-GENE PANEL FOR PREDICTING BENEFIT WITH ADJUVANT CHEMOTHERAPY IN SURGICALLY RESECTED NON-SMALL CELL LUNG CANCER PATIENTS

Introduction: Lung cancer is the deadliest cancer worldwide with very low survival rates. Curative treatments are available for early-stage patients with non-small cell lung cancer (NSCLC), including both surgery and adjuvant chemotherapy. However, regardless of adjuvant chemotherapy (ACT), most NSCLC patients treated with curative approaches will experience relapses and disease progression. Currently, specific criteria lack to define patients who will be benefited from ACT. Aim: To validate a 12-gene panel for predicting outcome and ACT benefit in NSCLC patients. Methods: Surgically resected NSCLC patients was retrospectively selected (n=119) and tumor tissue samples (FFPE) were used for RNA extraction (RNeasy FFPE Mini Kit, Qiagen). The expression of 12 genes (ATP8A1, AURKA, C1orf116, COL4A3, DOCK9, HOPX, HSD17B6, IFT57, MBIP, NKX2-1, RRM2, TTC37) was evaluated by nCounter ElementsXT (NanoString) and normalized by housekeepings genes (package NAPPA_R). The risk score was determined from a cutoff value (R mclust package), which stratifies two groups for evaluating predictive performance: high-risk patients (ACT benefit) and low-risk patients (ACT non-benefit). The Kaplan-Meier and Log-rank tests (univariate) and Cox Regression and Hazard ratio (multivariate) were used for survival analysis and for evaluating prognostic performance of the 12-gene panel. Results: In a subset of cases (n=93), 46 patients were classified as high-risk group and 47 patients were classified as low-risk group. The high-risk group presented lower survival time compared with the low-risk group (p=0.0011, HR=3.49) regardless of treatment. The low-risk group was further stratified into ACT-treated (n=19) and untreated (n=23) patients and overall survival was similar in both treated and untreated groups (p=0.7). The high-risk group was also stratified into ACT-treated (n=22) and untreated (n=18) patients, overall survival was slightly different in both treated and untreated groups (p=0.1). Taking into account likely confounders for disease outcome, patients classified into the high-risk group presented decreased risk for unfavorable outcome (HR=0.19; p=0.01) when ACT treatment is intended (HR=0.38; p=0.05). Histology and disease stage at the diagnosis did not influence outcome. Conclusion: The 12-gene panel stratified resected NSCLC patients into high- and low-risk groups showing a prognostic value and potential predictive performance for predicting ACT benefit for high-risk patients.