Disseminated high-grade serous ovarian cancer (HGS-OvCa) is an aggressive disease treated with platinum and taxane combination therapy. While initial response can be favorable, the disease typically relapses and becomes resistant to treatment. As genomic alterations in HGS-OvCa are heterogeneous, identification of clinically meaningful molecular markers for outcome prediction is challenging. We developed a novel computational approach (PSFinder) that fuses transcriptomics and clinical data to identify HGS-OvCa prognostic subgroups for targeted treatment. Application of PSFinder to transcriptomics data from 180 HGS-OvCa patients treated with platinum-taxane therapy revealed 61 transcript isoforms that characterize two poor and one good survival-associated groups (P = 0.007). These groups were validated in eight independent data sets, including a prospectively collected ovarian cancer cohort. Two poor prognostic groups have distinct expression profiles and are characteristic by increased hypermethylation and stroma-related genes. Integration of the PSFinder signature and BRCA1/2 mutation status allowed even better stratification of HGS-OvCa patients' prognosis. The herein introduced novel and generally applicable computational approach can identify outcome-related subgroups and facilitate the development of precision medicine to overcome drug resistance. A limited set of biomarkers divides HGS-OvCa into three prognostic groups and predicts patients in need of targeted therapies. (C) 2015 AACR.