Abstract P3-05-05: Classification using dynamic re-discovery of the strongest prognostic features in each analysis outperforms static gene expression signatures for prognostic prediction in breast cancer

Abstract

Abstract Introduction: To overcome limitations of previous multigene prognostic classifiers, we propose a new dynamic predictor. Our classifier does not use a single universal training cohort and an identical list of informative genes to predict the prognosis of new cases but a case-specific predictor is developed for each test case. Methods: We analyzed gene expression data from 3,534 breast cancers with clinical annotation including survival. For each test case we select a case-specific training subset including only molecularly similar cases and a case-specific predictor is generated. This method yields different training sets and different predictors for each new patient but preserves the independence of model building and validation. The model performance was assessed in leave-one-out validation and also in 325 independent cases. Results: Prognostic discrimination was high for all cases (n = 3,534, HR = 3.68, p = 1.67E-56). The dynamic predictor showed higher overall accuracy (0.68) than genomic surrogates for the 21-gene predictor (0.64), the 97-gene genomic grade index (0.55) or the 70-gene prognostic signature (0.41). The dynamic predictor was also effective in triple-negative cancers (n = 427, HR = 3.08, p = 0.0093) where the above classifiers all failed. Validation in independent patients yielded similar classification power (HR = 3.57). Discussion: We developed a new method to make personalized prognostic prediction using case-specific training cohorts. The dynamic predictors outperform static models developed from single historical training cohorts and they also predict well in triple-negative cancers. The dynamic classifier is available on line at http://www.recurrenceonline.com/?q = Re_training. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P3-05-05.