7023 Background: Adjuvant treatment of patients with early-stage lung adenocarcinoma is based on post-surgical pathological staging and patient performance status. Disparate outcomes within each staging group suggest that additional prognostic markers could improve our understanding of risk-benefit and potentially lead to better treatment decisions. A proliferation-based, mRNA expression profile was applied to public microarray data of surgically treated lung adenocarcinomas and a cohort of FFPE samples to test its potential prognostic utility. Methods: Public expression data (Director’s Consortium, DC) were derived from Affymetrix HG-U133A arrays. Clinical FFPE samples were assayed by quantitative PCR. A cell cycle progression (CCP) score was calculated from the expression average of 31 cell cycle genes normalized by 15 housekeeper genes. The prognostic value of the CCP score to predict stage I and II patient outcomes was evaluated by Cox proportional hazards analysis with disease-related death as the primary outcome measure. Results: In 256 DC cases, the CCP score was a significant predictor of death in univariate (p=0.0001) and multivariate analysis (p=0.001, HR 1.57, 95%CI 1.20-2.05) using age, stage, gender, smoking status and treatment as covariates. Similarly, in a second data set (GSE31210, n=204) the CCP score was highly associated with death (univariate, p=0.001; multivariate analysis, p=0.003, HR 1.81, 95% CI 1.24-2.66). Using quantitative PCR, the signature was applied to 381 FFPE samples with a median follow-up of 5 years collected at the MD Anderson Cancer Center and the European Institute for Oncology. In the presence of clinical covariates (as above and tumor size and pleural invasion), the CCP score remained the most significant predictor of death in univariate (p=0.0003) and multivariate analysis (p=0.007, HR 1.50, 95% CI 1.11-2.02). Conclusions: A 46 gene mRNA signature is a significant predictor of disease-related death in early-stage lung adenocarcinoma, providing independent prognostic value in the presence of clinical variables. This molecular predictor of cancer survival will be studied in additional cohorts for its ability to impact clinical treatment decisions.
Read full abstract