Added value of amide proton transfer-weighted magnetic resonance imaging to Prostate Imaging Reporting and Data System version 2.1 in differentiating clinically significant prostate cancer.

Background: Amide proton transfer-weighted magnetic resonance imaging (APTw-MRI) and apparent diffusion coefficient (ADC) values can effectively differentiate clinically significant prostate cancer (csPCa). However, their added value in Prostate Imaging Reporting and Data System version 2.1 (PI-RADS v. 2.1) remains uncertain. This study aimed to investigate the added value of APTw-MRI and ADC to PI-RADS v. 2.1 in differentiating csPCa.

Methods: This study included 161 participants who underwent prostate biopsy or radical prostatectomy within 1 month following MRI. Six magnetic resonance (MR) parameters [minimum APT value (APTmin), maximum APT value (APTmax), mean APT value (APTmean), minimum ADC value (ADCmin), maximum ADC value (ADCmax), and mean ADC value (ADCmean)] and the PI-RADS score were acquired by two radiologists. Receiver operating characteristic (ROC) curve analysis was performed to evaluate and select parameters with the greatest performance from the APT and ADC values. Multivariate logistic regression analysis was conducted to construct models based on selected MR parameters and the PI-RADS score. The Delong test was applied to compare the area under the ROC curve (AUC) between models.

Results: The lesions of csPCa exhibited higher APT values and lower ADC values than not clinically significant prostate cancer (ncsPCa). The greatest comprehensive performance in differentiating csPCa was achieved by APTmean (AUC: 0.723; cutoff value: >1.644) and ADCmean (AUC: 0.759; cutoff value: <0.656). The addition of APTmean significantly improved the performance of the model based on the PI-RADS score (AUC: APT PI-RADS 0.867 PI-RADS 0.813; P=0.002). The introduction of ADCmean further improved the performance of the model based on APTmean and PI-RADS (AUC: combined model 0.875 PI-RADS 0.867, P=0.69), and it retained high diagnostic efficacy (AUC: 0.835) in the validation cohort. Further verification through calibration curve and decision curve analysis demonstrated that the adjusted PI-RADS model exhibited remarkable precision and practicality.

Conclusions: APTw-MRI and ADC are valuable parameters for differentiating csPCa and can provide additional value for improving the efficiency and clinical benefit of PI-RADS v. 2.1.