Interobserver agreement in prostate cancer detection using multiparametric MRI.

PURPOSE

The purpose of this prospective observational study was to evaluate the diagnostic performance of multiparametric (mp) magnetic resonance imaging (MRI) for prostate cancer detection and to assess the interobserver variability, using the Prostate Imaging Reporting and Data Systems (PI-RADS).

METHODS

50 patients (mean age 68.42±6.58 years) with suspected prostate cancer fulfilling the inclusion criteria and without any exclusion criteria were enrolled. All patients were examined with mp-MRI protocol, as per European Society of Urogenital Radiology (ESUR) guidelines, before systematic transrectal ultrasound (TRUS)-guided biopsy. All examinations were read by three independent radiologists with 3-year experience in prostate MRI. Sensitivity (Se), specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV) were calculated. Interobserver agreement was evaluated using Kappa Cohen coefficient of agreement.

RESULTS

mp-MRI and histopathological results of TRUSguided biopsy showed a very good agreement in prostate cancer detection. The overall Se, Sp, PPV and NPV ranged between 93.3-96.7%, 55.0-80.0%, 76.3-87.9% and 88.2-94.1%, respectively. The Kappa Cohen coefficient of interobserver agreement was 0.643 between Readers 1 and 2, 0.664 between Readers 1 and Reader 3 and 0.568 between Readers 2 and 3.

CONCLUSIONS

Our results showed a high Se for the detection of prostate cancer with mp-MRI and a high NPV to rule out prostate malignancy. PI-RADS version 2 provides an adequate standardization of mp-MRI, allowing a good level of interobserver agreement.