Multiple gene expression-based signatures have been identified in diffuse large B-cell lymphoma that are predictive for survival outcomes. Most studies assess predictive significance based on P values from multivariable Cox regression. Few investigations have evaluated the incremental usefulness of these signatures. Recent developments in statistical methodology extend the use of concordance measures on censored survival data. We applied these methods to evaluate the added value in survival risk prediction from 3 published gene-based signatures on 2 sets of patients with diffuse large B-cell lymphoma treated with CHOP or R-CHOP. Our results indicate these gene-based signatures are inferior to clinical factors and provide little added value in risk assessment. To develop highly discriminating risk prediction models, we need to use appropriate approaches and consider more than gene expression. However, the study of gene expression and clinical outcomes retains considerable potential to enhance understanding of disease mechanisms and uncover new therapeutic targets.