(1) Background: This research aims to identify candidates for trimodality therapy (TMT) or radical cystectomy (RC) by using a predictive model. (2) Methods: Patients with nonmetastatic muscle-invasive bladder cancer (MIBC) in the Surveillance, Epidemiology, and End Results (SEER) database were enrolled. The clinical data of 2174 eligible patients were extracted and separated into RC and TMT groups. To control for confounding bias, propensity score matching (PSM) was carried out. A nomogram was established via multivariable logistic regression. The area under the receiver operating characteristic curve (AUC) and calibration curves were used to assess the nomogram's prediction capacity. Decision curve analysis (DCA) was carried out to determine the nomogram's clinical applicability. (3) Results: After being processed with PSM, the OS of the RC group was significantly longer compared with the TMT group (p < 0.001). This remarkable capacity for discrimination was exhibited in the training (AUC: 0.717) and validation (AUC: 0.774) sets. The calibration curves suggested acceptable uniformity. Excellent clinical utility was shown in the DCA curve. The RC and RC-Beneficial group survived significantly longer than the RC and TMT-Beneficial group (p < 0.001) or the TMT group (p < 0.001). However, no significant difference was found between the RC and TMT-Beneficial group and the TMT group (p = 0.321). (4) Conclusions: A predictive model with excellent discrimination and clinical application value was established to identify the optimal patients for TMT among nonmetastatic MIBC patients.
Keywords: SEER database; nomogram; nonmetastatic muscle-invasive bladder cancer; predictive model; trimodality therapy.