Analysis of risk factors for postoperative bladder cancer in patients with upper tract urothelial carcinoma and construction of nomogram prediction model.

Abstract

Objective: To explore the risk factors for bladder cancer following upper tract urothelial carcinoma (UTUC) surgery and to construct a nomogram prediction model.

Methods: A retrospective cohort of 292 patients with UTUC treated at our hospital from February 2010 to April 2020 was enrolled and divided into a training set (204 cases) and an internal validation set (88 cases). Based on the postoperative occurrence of bladder cancer, the training set was subdivided into a bladder cancer group and a non-bladder cancer group. An additional 268 UTUC patients from other hospitals during the same period were selected as an external validation group. Logistic regression analysis was used to determine the influencing factors. The nomogram risk prediction model for postoperative bladder cancer was developed. The area under the curve (AUC) of the receiver operating characteristic (ROC) curve, calibration curve, Hosmer-Lemeshow (H-L) test, and decision curve analysis (DCA) were used to evaluate the predictive performance of the nomogram model.

Results: Tumor stage T3-T4, tumor grade G3, history of bladder cancer, and preoperative ureteroscopy examination were independent risk factors for postoperative bladder cancer in UTUC patients, and preventive bladder perfusion was an independent protective factor (P < 0.05). Internal verification: The AUC values of the ROC curves for the training and validation sets were 0.864 and 0.831, respectively. The calibration curves showed good agreement between the nomogram's predictions and actual observations. The H-L test revealed that χ² = 7.555 and 7.365, P = 0.478 and 0.392, respectively. The clinical practicality of the nomogram in DCA was relatively high. External validation: The AUC was 0.847. The calibration curve showed good agreement, and the H-L test resulted in χ2 = 6.175, P = 0.538. DCA confirmed the model's high clinical utility.

Conclusion: The nomogram, which integrates tumor stage, tumor grade, history of bladder cancer, preoperative ureteroscopy, and preventive intravesical instillation, demonstrates a high predictive ability for postoperative bladder cancer.