Abstract： Objective:To establish and evaluate a predictive model of bi-parameter magnetic resonance imaging(Bp-MRI)radiomics for positive surgical margin(PSM)after radical prostatectomy for prostate cancer(PCa).Materials and Methods:The imaging and clinical data of 105 patients who underwent laparoscopic radical prostatectomy via extraperitoneal route in Gansu Provincial People's Hospital were retrospectively analyzed,and they were classified into 40 cases with positive postoperative margins and 65 cases with negative postoperative margins by postoperative pathological findings.The dataset was partitioned into a training set(n=73)and a test set(n=32)in a 7∶3 ratio.Subgroup analysis was performed within both the training and test sets,comparing clinical and MRI data.Region of interest(ROI)was delineated using the ITK-SNAP software from T2WI,diffusion-weighted imaging(DWI),and apparent diffusion coefficient(ADC)sequences.The"Pyradiomics"package was utilized to extract a total of 312 features from these ROIs.The features were then subjected to dimensionality reduction and model construction using the least absolute shrinkage and selection operator(LASSO)algorithm.A predictive model was built based on a logistic regression(LR)classifier.The efficacy of the imaging model in predicting PSM after radical prostatectomy for PCa was evaluated using the area under the curve(AUC)of receiver operating characteristic(ROC).Additionally,decision curve analysis(DCA)was employed to assess the clinical net benefit of the model.Results:Ten radiomic features closely related to PSMs were ultimately selected.The LR model achieved an AUC of 0.869(95%CI:0.786-0.952)in the training set and an AUC of 0.858(95%CI:0.726-0.991)in the test set.The DCA indicated that the model offers a significant clinical net benefit.Conclusions:The predictive assessment of positive margins after radical prostatectomy for PCa based on Bp-MRI radiomics is informative and helpful for clinical preoperative risk stratification and postoperative treatment.