A novel approach to differentiate prostate cancer from prostatitis in the peripheral zone

Abstract Objectives To analyse the different imaging manifestations of prostatitis and prostate cancer (PCa), and propose systematic differentiation and validate its diagnostic performance. Methods Men with histologically proven prostatitis or PCa who had pretreatment multiparametric MRI and lesions in the peripheral zone were retrospectively identified from January 2018 to May 2022. The images were evaluated by 2 experienced radiologists, and key imaging features for differential diagnosis were summarized and analysed via univariable and multivariable logistic regression. The best differentiating model was generated via refining the original Prostate Imaging Reporting and Data System (PI-RADS) standard (rPI-RADS). Two inexperienced radiologists reevaluated the images according to the PI-RADS and rPI-RADS criteria, then compared. Results A total of 315 patients were enrolled in this study, including 181 patients with clinically significant PCa (67.3 ± 7.7 years) and 134 patients with prostatitis (64.8 ± 9.6 years). The differentiated model was based on diffuse distribution, wedge shape, T1 iso-hyperintensity, rim enhancement, and delayed periprostatic enhancement, thus forming the rPI-RADS standard according to these features. Paired comparison study showed that 22 out of 121 (18.2%, P < .001) inflammatory lesions were corrected from PI-RADS category 4-5 to rPI-RADS category 2, and 20 out of 121 (16.5%, P < .001) were corrected from PI-RADS category 4-5 to rPI-RADS category 1-3 by 2 inexperienced radiologists. Conclusions The supplementary descriptions of wedge-shaped and diffuse distributed lesions were helpful for inexperienced radiologists. The refined principle of score reduction may reduce false-positive PI-RADS scores for prostatitis. Advances in knowledge A systematic differentiation of PCa from prostatitis was proposed, and validated by inexperienced radiologists.