Development and validation of a non-invasive nomogram for predicting bone metastasis in newly diagnosed breast cancer.

Background: Bone metastasis (BM) is a major factor contributing to reduced quality of life and poor prognosis in breast cancer patients, occurring in approximately 65-75% of advanced cases. Early identification of high-risk individuals for BM at the time of initial breast cancer diagnosis is crucial for improving outcomes. However, current imaging and biopsy methods have limitations including low sensitivity, high cost, and invasiveness. This study aimed to develop a noninvasive nomogram model using routine clinical indicators to predict BM risk in newly diagnosed breast cancer patients. Methods: A retrospective single-center case-control study was conducted with 376 patients (134 with BM, 242 without) from January 2010 to January 2024. Data on demographics, laboratory indicators [e.g., alkaline phosphatase (ALP), cancer antigen 15-3 (CA15-3), albumin (ALB)], and imaging characteristics (e.g., tumor size) were collected. Patients were randomly divided into training (n=263) and validation cohorts (n=113) in a 7:3 ratio. Independent predictors were identified using univariate and multivariate logistic regression analyses, and a nomogram model was constructed. Model performance was evaluated using the concordance index (C-index), receiver operating characteristic (ROC), area under the curve (AUC), calibration curves, and decision curve analysis (DCA) to assess discrimination, calibration, and clinical utility. Results: Multivariate analysis identified four independent predictors: low ALB [odds ratio (OR) =0.89, 95% confidence interval (CI): 0.83-0.95], elevated ALP (OR =1.01, 95% CI: 1.01-1.03), elevated CA15-3 (OR =1.02, 95% CI: 1.01-1.02), and large primary tumor size (OR =1.02, 95% CI: 1.01-1.05). The nomogram model showed good performance: Training AUC =0.85, Validation AUC =0.79; calibration curves approximated the ideal line, and DCA demonstrated clinical utility. Conclusions: We successfully developed and validated a nomogram model based on routine indicators (ALB, ALP, CA15-3, tumor size) for noninvasive and accurate prediction of BM risk in newly diagnosed breast cancer patients. This model exhibits good predictive performance and clinical utility, providing a practical tool for early screening of high-risk patients and guiding intervention decisions.