Abstract
BACKGROUND: To measure the diagnostic value of T2WI and DWI radiomics model, combined with advanced biomarkers, in distinguishing benign prostatic hyperplasia (BPH), prostate cancer (PCa) and prostatitis. METHODS: A total of 90 patients with prostate diseases were selected from our hospital from January 2022 to January 2024. All patients underwent T2WI and DWI MRI examinations. Regions of interest (ROI) were delineated, and imaging features were extracted using radiomics analysis. In addition, novel biomarkers, including Prostate Cancer Antigen 3 (PCA3), Sarcosine, Glypican-1 (GPC1), Urokinase Plasminogen Activator Receptor (uPAR), and Thymidine Kinase 1 (TK1), were analysed for their diagnostic significance. Feature selection was performed using LASSO regression, and a random forest model was established for classification. The receiver operating characteristic (ROC) curve was used to evaluate the diagnostic performance of the T2WI and DWI radiomics model with and without biomarker integration. RESULTS: Among the 90 patients with prostate diseases, 50 cases were PCa, 20 cases of prostatitis and 20 cases of BPH were detected by biopsy. The PI-RADS v2 score in the PCa group presented elevation relative to those in the BPH and prostatitis groups (P<0.01). The ADC values in the PCa group were reduced relative to those in the BPH and prostatitis groups (P<0.01). The integration of biomarkers with radiomics analysis led to improved diagnostic performance. The AUC value, sensitivity, and specificity of the T2WI and DWI radiomics model were higher relative to those of PI-RADS V2. CONCLUSIONS: The T2WI and DWI radiomics model, when combined with novel biomarkers, enhances the accuracy of distinguishing PCa, BPH, and prostatitis. This approach may provide an advanced diagnostic tool for personalised prostate disease management.