Use of the Prostate Imaging Reporting and Data System (PI-RADS) for Prostate Cancer Detection with Multiparametric Magnetic Resonance Imaging: A Diagnostic Meta-analysis

CONTEXT: In 2012, an expert panel of the European Society of Urogenital Radiology (ESUR) published the Prostate Imaging Reporting and Data System (PI-RADS) for prostate cancer (PC) detection with multiparametric magnetic resonance imaging (mp-MRI). Since then, many centers have reported their experiences. PURPOSE: To review the diagnostic accuracy of PI-RADS for PC detection with mp-MRI. EVIDENCE ACQUISITION: We searched Medline and Embase up to March 20, 2014. We included diagnostic accuracy studies since 2012 that used PI-RADS with mp-MRI for PC detection in men, using prostatectomy or biopsy as the reference standard. The methodological quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool by two independent reviewers. Data necessary to complete 2x2 contingency tables were obtained from the included studies, and test characteristics including sensitivity and specificity were calculated. Results were pooled and plotted in a summary receiver operating characteristics plot. EVIDENCE SYNTHESIS: Fourteen studies (1785 patients) could be analyzed. The pooled data showed sensitivity of 0.78 (95% confidence interval [CI] 0.70-0.84) and specificity of 0.79 (95% CI 0.68-0.86) for PC detection, with negative predictive values ranging from 0.58 to 0.95. Sensitivity analysis revealed pooled sensitivity of 0.82 (95% CI 0.72-0.89) and specificity of 0.82 (95% CI 0.67-0.92) in studies with correct use of PI-RADS (ie, clear description in the methodology and no adjustment of criteria). For studies with a less strict or adjusted use of PI-RADS criteria, or unclear description of the methodology, had pooled sensitivity of 0.73 (95% CI 0.62-0.82) and specificity of 0.75 (95% CI 0.61-0.84). CONCLUSIONS: In patients for whom PC is suspected, PI-RADS appears to have good diagnostic accuracy in PC detection, but no recommendation regarding the best threshold can be provided because of heterogeneity. PATIENT SUMMARY: Pooling of results from all previous studies that used a relatively new 5-point scoring system for prostate magnetic resonance imaging showed that this scoring system appears to be able to detect prostate cancer accurately