Objective:To investigate the predictive value of a radiomics model based on biparametric magnetic resonance imaging(bpMRI)for biochemical recurrence(BCR)after radical prostatectomy(RP)in elderly prostate cancer patients(≥60 years old).Methods:A retrospective analysis was conducted on data from 175 patients treated at Beijing Hospital from August 2017 to December 2021.Based on pathological results, image segmentation was performed on preoperative bpMRI T2, diffusion weighted imaging(DWI), and apparent diffusion coefficient(ADC)sequences.Pyradiomics was utilized to extract radiomic features, and Cox regression, Spearman correlation coefficient, and LASSO regression were employed for feature dimensionality reduction, leading to the construction of radiomic labels.Clinical models and image-clinical combined models were developed using multifactorial Cox regression analysis, and the performance of these models in predicting BCR was evaluated using the concordance index(C-index).Results:The 175 patients were randomly divided into a training set(122 cases)and a test set(53 cases)at a ratio of 7∶3, with 24 cases(19.7%, 24/122)and 11 cases(20.8%, 11/53)experiencing BCR, respectively.A total of 5 775 radiomic features were extracted from the three sequences, and after dimensionality reduction, 5 features were selected to construct the radiomic labels.The radiomics model exhibited C-index values of 0.764(95% CI: 0.655-0.872)and 0.769(95% CI: 0.632-0.906)in the training and test sets, respectively.Multifactorial Cox regression analysis revealed serum prostate-specific antigen(PSA)( HR=1.032, 95% CI: 1.010-1.054), postoperative pathology International Society of Urological Pathology(ISUP)grade grouping( HR=1.682, 95% CI: 1.039-2.722), and positive surgical margins( HR=2.513, 95% CI: 1.094-5.774)as independent predictors of BCR.The clinical model exhibited C-index values of 0.751(95% CI: 0.655-0.846)and 0.753(95% CI: 0.630-0.877)in the training and test sets, respectively.Following combined modeling of clinical factors and radiomic labels, the image-clinical combined model demonstrated the highest C-index values, namely 0.782(95% CI: 0.679-0.874)and 0.801(95% CI: 0.677-0.915)in the training and test sets, respectively. Conclusions:The radiomics model based on bpMRI can predict the occurrence of BCR after RP in elderly prostate cancer patients.Combined modeling of clinical factors and radiomic labels can enhance predictive efficiency.