Identification of multigene predictors of prognosis in patients with ovarian cancer

The high recurrence and mortality of ovarian cancer (OC) necessitate reliable prognostic tools. As multigene models outperform single-gene markers, they are crucial for advancing clinical decision-making. From 4 microarray datasets (N = 859) and a large OC cohort (N = 1793), we identified 87 overall survival (OS)- and/or progression-free survival (PFS)-associated genes. Multigene models revealed that 25 combinations and 5 combinations predicted at least 5-year OS and 5-year PFS, respectively. Notably, compared with all combinations of 4 randomly selected genes, combinations involving ALDH1A2, DCN, GATA6, and PDGFRA were frequent among the top 20 OS models; 6 combinations predicted at least 50-month OS, and tissue microarray (155 OC samples) analyses confirmed this. Particularly, DCN + GATA6 predicted the longest OS (90 months), with a survival difference of 3 years. Most genes correlated with immune cell abundance, especially macrophage abundance. Overall, these genes/combinations serve as valuable biomarkers to optimize OC clinical management.