Objective To investigate the application value of a competing risk prediction model for postoperative prognosis in patients with bladder cancer constructed using ultrasound blood perfusion parameters.

Methods A total of 420 patients with bladder cancer who underwent treatment at our hospital between January 2018 and December 2023 were retrospectively enrolled. The participants were randomly assigned to a training group (n = 294) and a validation group (n = 126) at a ratio of 7 to 3. Contrast-enhanced ultrasound was performed preoperatively on all participants to assess blood perfusion parameters, and clinicopathology data were collected. The Fine-Gray method was used for multivariate analysis to determine the factors affecting bladder cancer-specific mortality and to construct a nomogram. The performance of the model was evaluated using the validation group.

Results The median follow-up period was 15 months. Comparisons of tumor-related parameters, including tumor number, diameter, pathological grade, pathological stage, muscle layer invasion, peak intensity (PI), mean transit time (MTT), and time from peak to half (TPH) among the groups with different outcomes showed statistically significant differences (P < 0.05). Cox regression analysis indicated that PI (hazard ratio HR = 2.081, 95% CI: 1.069-4.053), MTT (HR = 1.570, 95% CI: 1.243-1.982), TPH (HR = 1.657, 95% CI: 1.341-2.048), pathological grade (HR = 1.829, 95% CI: 1.210-2.766), and pathological stage (HR = 1.788, 95% CI: 1.213-2.636) were all factors significantly associated with bladder cancer-specific mortality (P < 0.05). Bootstrap internal validation showed that the C-index for 3- and 5-year cumulative incidence of bladder cancer–specific mortality was 0.887 and 0.834, respectively. Time-dependent receiver operating characteristic analysis showed areas under the curve (AUCs) of 0.816 and 0.869 at 3 and 5 years, respectively. Calibration curves demonstrated high agreement between the predicted and observed probabilities.

Conclusion The competing risk nomogram constructed based on ultrasound blood flow perfusion parameters show potential application value for predicting bladder cancer-specific mortality.