Abstract
BACKGROUND: Elderly patients (≥70 years) with bladder cancer undergoing radical cystectomy (RC) represent a vulnerable population with heterogeneous outcomes. Traditional pathological lymph node (pN) staging has limitations. This study evaluated the prognostic value of lymph node ratio (LNR) and log odds of positive lymph nodes (LODDS) compared to pN and developed a novel nomogram for this demographic. METHODS: Using data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database [2004-2015], 1,018 elderly bladder cancer patients post-RC were identified and randomly split into training (n=712) and internal validation (n=306) cohorts. An independent external cohort (n=260) was included. Predictive performance of pN, LNR, and LODDS was assessed using time-dependent area under the curve (AUC) and concordance index (C-index). Prognostic factors were identified via least absolute shrinkage and selection operator (LASSO) and multivariable Cox regression. A nomogram predicting 1-, 3-, and 5-year overall survival (OS) was constructed and validated. RESULTS: LODDS demonstrated superior prognostic discrimination compared to pN and LNR across all cohorts (training C-index: LODDS 0.602 vs. pN 0.573 vs. LNR 0.579). The final nomogram incorporated race, tumor stage (T stage), metastasis stage (M stage), chemotherapy status, and LODDS. It showed robust performance: training C-index =0.647 [95% confidence interval (CI): 0.622-0.672], internal validation C-index 0.650 (95% CI: 0.611-0.690), and external validation C-index =0.729 (95% CI: 0.687-0.770). Calibration curves indicated strong agreement between predicted and observed survival. LODDS maintained superior stratification within the ≥80-year subgroup. Risk stratification based on the nomogram significantly differentiated survival outcomes (log-rank P<0.001). CONCLUSIONS: LODDS provides enhanced prognostic stratification over pN and LNR in elderly bladder cancer patients post-RC. The developed and validated LODDS-based nomogram offers a practical tool for individualized survival prediction, aiding clinical decision-making in this growing population.