Objective:To investigate the value of the combined ultrasound imaging radiomics model for predicting progression-free survival in endocrine therapy for prostate cancer.Methods:A total of 283 prostate cancer patients who received endocrine treatment at the Inner Mongolia Autonomous Region People′s Hospital and the First Hospital of Hohhot from July 2018 to January 2023 were retrospectively collected, of which 198 patients from the Inner Mongolia Autonomous Region People′s Hospital were randomly divided into the training set and the validation set according to the ratio of 7∶3, and 85 patients from the First Hospital of Hohhot served as an independent external test set. They were classified into a progression group and a non-progression group based on whether the patients progressed to desmoplasia-resistant prostate cancer 18 months after the start of endocrine treatment.Based on the two-dimensional ultrasound images, the imaging radiomics features were extracted and the imaging radiomics score (Rad-score) were constructed, the immunopathology and other clinical data were analysed, and three prediction models were constructed using logistic regression: the clinical model, the ultrasonography model, and the ultrasonography-clinical combined model, respectively. The predictive efficacy and clinical utility of the models were assessed by the ROC curves and clinical decision curves.Results:Five ultrasonographic features were included in the ultrasound model; the prostate-specific antigen nadir, the neutrophil-to-lymphocyte ratio before treatment, and the expression level of tumour proliferating cell nuclear antigen 67 (Ki-67) were incorporated into the clinical model; and the Rad score computed from the output of the ultrasound model for the screening features, together with the prostate-specific antigen nadir (PSA nadir), the neutrophil to lymphocyte ratio (NLR) before treatment, and the expression level of Ki-67 were used to construct the ultrasound-clinical joint model. The joint model achieved the highest predictive performance in both the training and validation sets of the three groups of models, with the area under the curve of 0.85 and 0.84, and the clinical decision curve showed good clinical benefit.Conclusions:The combined ultrasound-clinical model constructed in this study based on two-dimensional ultrasound images of prostate cancer before endocrine therapy can predict progression-free survival of endocrine therapy for prostate cancer more accurately.