Fig. 6

Construction and validation of HRDscore system. A Differentially expressed genes between pairs in HRD clusters. B Risk plot. The red dot meant sample with high risk, and the blue dot meant sample with low risk. C Sankey plot shows patients’ distribution in HRD clusters and HRDscore risk groups. D The HRDscore in different subgroups was stratified by age, Gleason score, PFI status, pathologic T and N stage, as well as clinical M stage. E The survival differences regarding DSS, DFI, PFI, and OS between high HRDscore and low HRDscore groups. F The survival differences regarding PFI between high HRDscore and low HRDscore groups in subgroups stratified by age (left) and Gleason score (right). H, I The results of multivariate Cox analyses for H DFI and I PFI. P value was set in bold if it was less than 0.05