Gene expression-based survival prediction in lung adenocarcinoma: a multi-site, blinded validation study

Abstract

Although prognostic gene expression signatures for survival in early-stage lung cancer have been proposed, for clinical application, it is critical to establish their performance across different subject populations and in different laboratories. Here we report a large, training-testing, multi-site, blinded validation study to characterize the performance of several prognostic models based on gene expression for 442 lung adenocarcinomas. The hypotheses proposed examined whether microarray measurements of gene expression either alone or combined with basic clinical covariates (stage, age, sex) could be used to predict overall survival in lung cancer subjects. Several models examined produced risk scores that substantially correlated with actual subject outcome. Most methods performed better with clinical data, supporting the combined use of clinical and molecular information when building prognostic models for early-stage lung cancer. This study also provides the largest available set of microarray data with extensive pathological and clinical annotation for lung adenocarcinomas.

Figure 1

Classifier Performance: Hazard Ratios of methods A–I on validation datasets for four hypotheses, along with 95% confidence intervals.

Figure 2

Kaplan-Meier estimates of the survivor function for method A on each validation dataset for the 4 hypotheses. a: MSK test set all stages, b: MSK test set with covariates all stages, c: MSK test set stage 1 only, d: MSK test set stage 1 only with covariates, e: CAN/DF test set all stages, f: CAN/DF test set with covariates all stages, g: CAN/DF test set stage 1 only, h: CAN/DF test set stage 1 only with covariates.

Figure 3

Kaplan-Meier estimates of the survivor function for method A (cross-validated) on training sets UM and MSK. a: all stage, b: all stages with covariates, c: stage 1 only, d: stage 1 only with covariates.