Development of a Preoperative Nomogram Incorporating Biomarkers of Systemic Inflammatory Response to Predict Nonorgan-confined Urothelial Carcinoma of the Bladder at Radical Cystectomy

Objective To develop a preoperative multivariable decision-making tool to predict nonorgan-confined urothelial carcinoma of the bladder (NOC-UCB) using standard clinical and pathological factors as well as biomarkers of systemic inflammatory response. Materials and Methods We retrospectively analyzed a prospectively maintained single-institutional database comprising 310 patients with clinically N0 M0 UCB who underwent radical cystectomy (RC) with pelvic lymph node dissection without neoadjuvant cisplatin-based chemotherapy (NAC). NOC-UCB was defined as pT3-4/Nany or pTany/N + disease. A predictive nomogram was built based on significant variables in a bootstrap-corrected multivariable logistic regression model. The accuracy was measured by the area under the curve. Decision-curve analysis was used to evaluate the clinical net benefit. Results NOC-UCB was found in 147 (47%) of the 310 patients. On multivariable analysis, T stage at transurethral resection of the bladder, lymphovascular invasion, abnormal imaging, and Glasgow prognostic score (GPS) were all independent predictors of NOC-UCB and formed the basis of the nomogram. By adding the GPS, the accuracy of the nomogram improved by 4.7% to 81.7%. The decision curve analysis showed a net benefit of this model compared with the Green model and the strategies of treating all patients or no patient with NAC. Limitations include the retrospective design and the lack of a validation cohort. Conclusion NOC-UCB at radical cystectomy can be accurately predicted. The accuracy of preoperative models can be improved by adding biomarkers of systemic inflammatory response, such as the GPS. The use of this nomogram may help physicians to accurately identify patients with NOC-UCB who may benefit from NAC.