Purpose: To use meta-analysis to assess multiparametric magnetic resonance (MP MR) imaging for detection and staging of untreated men with localized prostate cancer.Materials and Methods: We searched for original cohort studies and randomized trials in PUBMED,EMBRASE, and ScienceDirect up to June 15 2015.We selected only studies that used histopathotogy after radical prostatectomy as reference standard.We undertook random effects meta-analysis,subgroup analyses, and meta-regression.Results: We included 135 studies comprising 13182 patients.A combination of T2-weighted imaging and diffusion weighted imaging gave higher diagnostic test accuracy than T2-weighted imaging alone (83% versus 75%) for a difference of 8% (95% confidence interval (CI) 6%, 10%).Controlled trials of biopsy na(i)ve patients showed that a targeted biopsy guided by MP MR imaging detected more significant prostate cancer lesions than a transrectal ultrasound (TRUS) biopsy (77% against 53%).To detect extracapsular extension of tumor lesions, MP MR imaging had pooled sensitivity of 59% (95% CI 56%, 62%) and pooled specificity 85% (95% CI 83%, 85%).MR-estimated tumor volume was linearly associated with pathological tumor volume.Median Spearman rank correlation coefficient was 0.56.To delineate DIL in prostate segments, MP MR imaging had a sensitivity of 58% (95% CI 58%, 61%) and specificity of 88% (95% CI 87%, 88%).MP MR imaging may support selection of candidates for active surveillance and focal therapy, and changed treatment for some patients given definitive treatment for the whole prostate.Conclusion: MP MR has value for detection and staging of patients with localized prostate cancer.