OBJECTIVETo investigate the accuracy and use of body‐coil magnetic resonance imaging (MRI) in the local staging of prostate cancer before radical prostatectomy (RP).PATIENTS AND METHODSFifty‐six patients undergoing RP were staged before surgery using body‐coil MRI; none was denied surgery on the basis of their scan results. All scans were reported before RP by one of three consultant radiologists and afterward by a colleague with a special interest in prostate MRI, unaware of the patients’ clinical details.RESULTSThe overall sensitivity of MRI at detecting extracapsular extension was 50% on general reporting and 72% when reported by the specialist radiologist; the respective specificities were 84% and 86%. Of the 55 patients included in the study, 18 (33%) had extracapsular disease on histological analysis. MRI was most accurate in the 17 patients at high‐risk (prostate‐specific antigen, PSA, >10 ng/mL and Gleason score ≥ 8) and eight at intermediate risk (PSA < 10 ng/mL and Gleason score 7). In the former group with specialist analysis, the sensitivity was 100%, although this decreased to 67% with general reporting. Both gave a specificity of 82%. Intermediate risk disease gave a sensitivity and specificity of 75%, irrespective of reporting method. The ability of MRI to detect extraprostatic tumour in the 30 low‐risk patients (PSA < 10 ng/mL and Gleason score 2–6) was poor; the sensitivity was 25% with general and 50% on specialist review, although both methods gave a specificity of >90%.CONCLUSIONBody‐coil MRI is sensitive and specific for identifying extracapsular extension of prostate cancer in patients with high‐ or intermediate‐risk disease. Patients at low risk frequently have microscopic extension which is not detected. Opinion from a radiologist with a special interest in prostate MRI can increase the reporting accuracy even when unaware of the patients’ clinical details.