Personalized Optimization of Systematic Prostate Biopsy Core Number Based on mpMRI Radiomics Features

Abstract

Abstract Background Prostate cancer (PCa) is definitively diagnosed by systematic prostate biopsy (SBx) with 13 cores. This method, however, can increase the risk of urinary retention, infection and bleeding due to the excessive number of biopsy cores. Methods We retrospectively analyzed 622 patients who underwent SBx with prostate multiparametric MRI (mpMRI) from two centers between January 2014 to June 2022. The MRI data were collected to manually segment Regions of Interest (ROI) of the tumor layer by-layer. ROI reconstructions were fused to form VOIs, which were exported and applied to subsequent extraction of radiomics features. The t-tests, Mann-Whitney U-tests and chi-squared tests were performed to evaluated the significance of features. The logistic regression was used for calculating the PCa risk score(PCS). The PCS model was trained to optimize the SBx core number, utilizing both mpMRI radiomics and clinical features. Results The predicted number of SBx cores were determined by PCS model. Optimal core numbers of SBx for PCS subgroups 1–5 were calculated as 13, 10, 8, 6, and 6, respectively. Accuracies of predicted core numbers were high: 100%, 95.8%, 91.7%, 90.6%, and 92.7% for PCS subgroups 1–5. Optimized SBx reduced core rate by 41.9%. Leakage rates for PCa and clinically significant PCa were 8.2% and 3.4%, respectively. The optimized SBx also demonstrated high accuracies on the validation set. Conclusion The optimization PCS model described in this study could therefore effectively reduce the number of systematic biopsy cores obtained from patients with high PCS. This method can enhance patient experiences without reducing tumor detection rate.