IntroductionFocal therapy for localized prostate cancer has the potential for oncological control without the side effects of radical therapies. However, there is currently no validated method for monitoring treatment success. We assessed the diagnostic performance of prostate-specific antigen (PSA) parameters and MRI compared to histological outcomes following focal therapy. Patients and methodsPatients from 3 Ethics Review Board approved prospective studies of focal high intensity–focused ultrasound (HIFU) (Sonablate 500) for localized prostate cancer (T1c–T3a, Gleason grade≤4+3, and PSA≤20). Post-HIFU PSA nadir, 6-month PSA, PSA density, and early (<3wk) and late (6mo) MRI (T2-weighted, dynamic contrast-enhanced±diffusion-weighted) was assessed for predictive accuracy of cancer on postoperative biopsy, using receiver operating characteristic (ROC) analysis and sensitivity, specificity, and positive and negative predictive estimates. ROC areas for MRI and PSA were compared. Calculations for statistical significance (P≤0.05) were obtained in a subset of patients comparing area under ROC for 6-month MRI and PSA criteria, across 4 different histological definitions of disease significance. ResultsOf 118 men, 111 underwent at least 1 postoperative biopsy (median 6 cores), with an overall positive biopsy rate of 37% (41/118), over a mean follow-up period of 716 days post-HIFU. Areas under ROC for early and late MRI were (depending on definition of significant disease) 0.65 to 0.76 and 0.77 to 0.85, respectively, with sensitivity, specificity, and negative predictive values of 68% to 91%, 52% to 55%, and 85% to 98% (early MRI), and 63% to 80%, 67% to 73%, and 86% to 97% (late MRI). The area under the ROC curve was statistically significantly higher for late MRI than 6 months and nadir PSA for residual disease >3mm or any Gleason 4 tumor. ConclusionsEarly and late MRI performed better than PSA measurements in the detection of residual tumor after focal therapy.