OBJECTIVE To evaluate the diagnostic performance and complication rate of the in-bore magnetic resonance imaging-guided transrectal targeted prostate biopsy (MRGB) in a repeat biopsy population on the basis of a nearly 4-year learning curve (2014-2017). MATERIALS AND METHODS A total of 142 consecutive males with previous biopsies and persistent suspicion of prostate cancer (PCa) due to high prostate-specific antigen level initially underwent MRGB in the case of prostate imaging reporting and data system (PI-RADS) 3-5 lesions. Cancer detection rate (CDR), number and length of cores, biopsy time, operator experience, complications, and prediction of clinically significant (cs) PCa (Gleason score >= 7) were investigated. RESULTS PCa was found in 57% of patients. CDR in PI-RADS 3, 4, and 5 lesions were 46%, 52%, and 74%, respectively. csPCa was found in 9%, 25%, and 48% of patients. In univariate analysis the PI-RADS score (P = .0067) was a significant predictor of csPCa. In the multivariate logistic regression, age (P = .0007), number of previous biopsies (P = .0236), and prostate-specific antigen density (P = .0250) were significant predictors of csPCa. Location and size of the index lesion, number and length of cores obtained, and operator experience did not affect CDR. Concerning learning curve, biopsy time and number of cores obtained improved significantly after 10 procedures. Complications requiring medical intervention were seen in 6% (infections 2%). CONCLUSION In a re-biopsy setting the MRGB showed sufficient diagnostic performance in detecting csPCa in PI-RADS 3-5 lesions, with low complication rate. The skill of performing biopsy is quickly acquired, and location of index lesion did not have an impact on CDR. (C) 2018 Elsevier Inc.