Predictive risk model for bone metastasis in intermediate-risk prostate cancer: a single-center retrospective analysis.

To develop and validate a nomogram to identify patients with intermediate-risk prostate cancer who do not require a bone scan. Data from 305 patients with intermediate-risk prostate cancer who were hospitalized at the Department of Urology of the First People's Hospital of Nantong between September 2023 and December 2025 and who underwent pretreatment bone imaging (99mTc-MDP scintigraphy) were included in this retrospective analysis. Patients were randomly assigned to training and internal validation groups at a ratio of 7:3. After the optimal predictive features were selected via minimum absolute shrinkage and selection operator (LASSO) regression, the predictive model was constructed using multiple logistic regression, and internal validation was performed with split-sample and bootstrap calibration. A total of 305 people with moderate-risk prostate cancer were included in this study. Of these, 54 (17.7%) had bone metastases, and 251 (82.3%) did not. Prostate-specific antigen (PSA), fibrinogen (FIB), hemoglobin, and the systemic immune-inflammatory index (SII) were incorporated into the final model. We thoroughly evaluated the performance of the model according to clinical utility (decision curve analysis), calibration (calibration curve, Hosmer-Lemeshow test), and discrimination (ROC analysis, AUC). In both the training (AUC 0.880, 95% CI 0.812-0.947) and validation (AUC 0.877, 95% CI 0.782-0.971) groups, the discrimination, calibration, and clinical utility of the model were excellent. The PSA, SII, FIB, and hemoglobin levels are independent predictive factors for bone metastasis in patients with intermediate-risk prostate cancer, and the constructed nomogram has good predictive value for identifying patients who do not need bone scans. A limitation of this study is its retrospective single-center design, which may lead to selection bias and limit the generalizability of the model.