Abstract / Summary
Accurate identification of bone metastasis at initial evaluation is essential for staging and treatment planning in newly diagnosed prostate cancer (PCa). We retrospectively analyzed a prospectively maintained database of 355 patients from the Jixi Road Campus as the development cohort and 203 patients from the High-Tech Campus as the external validation cohort. Associations between five inflammatory-immune indices and bone metastasis were examined in the development cohort using multivariable logistic regression and restricted cubic splines. Predictors were selected using repeated least absolute shrinkage and selection operator regression followed by stepwise logistic regression. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), calibration plots, the Brier score, and decision curve analysis. Bootstrap resampling and the separate-campus cohort were used for internal and external validation, respectively. Bone metastasis was identified in 67 patients (18.9%) in the development cohort and 48 patients (23.6%) in the external validation cohort. The neutrophil-to-lymphocyte ratio (NLR) and systemic immune-inflammation index remained associated with bone metastasis after full adjustment. Their odds ratios (ORs) per one-standard-deviation increase in the transformed index were 1.58 (95% confidence interval [CI] 1.12–2.22) and 1.44 (95% CI 1.03–2.03), respectively. Repeated predictor selection retained log-transformed prostate-specific antigen, clinical T stage, biopsy Grade Group, and NLR. The model yielded an AUC of 0.905 (95% CI 0.869–0.941) in the development cohort. The optimism-corrected AUC was 0.898 (95% CI 0.864–0.934), and the external validation AUC was 0.884 (95% CI 0.828–0.941). Calibration and clinical net benefit remained consistent during internal and external validation. Adding NLR modestly increased the AUC from 0.887 to 0.905 (change in AUC [ΔAUC] = 0.018; P = 0.034). We developed a model combining clinicopathological factors with NLR to estimate the individual probability of bone metastasis. NLR provided modest complementary information, and the model may support decisions about further imaging in newly diagnosed PCa.