Prognostic nutritional index combined with NLR to construct a survival prediction model and decision analysis of patients with muscle‐invasive bladder cancer after surgery

Abstract

AbstractObjectivesTo build a nomogram prediction model, assess its predictive ability, and perform a survival decision analysis on patients with muscle‐invasive bladder cancer (MIBC) to study risk factors affecting overall survival (OS).MethodsA retrospective analysis was performed on the clinical information of 262 patients with MIBC who underwent radical cystectomy (RC) at the Urology Department of the Second Affiliated Hospital of Kunming Medical University between July 2015 and August 2021. The final model variables that were included were chosen using single‐factor stepwise Cox regression, optimal subset regression, and LASSO regression + cross‐validation with the minimum AIC value. The next step was to do a multivariate Cox regression analysis. The establishment of a nomogram model by fitting and the screening out of independent risk factors impacting the survival of patients with MIBC having radical resection. Receiver Activity Characteristic curves, C‐index, and a calibration plot evaluated the prediction accuracy, validity, and clinical benefit of the model. The 1‐, 3‐, and 5‐year survival rates were then computed for each risk factor using a Kaplan–Meier survival analysis.Results262 eligible patients in total were enrolled. With a median follow‐up of 32 months, the follow‐up period ranged from 2 to 83 months. 171 cases (65.27%) survived while 91 cases (34.73%) perished. Age (HR = 1.06 [1.04; 1.08], p = 0.001), preoperative hydronephrosis (HR = 0.69 [0.46, 1.05], p = 0.087), T stage (HR = 2.06 [1.09, 3.93], p = 0.027), lymphovascular invasion (LVI, HR = 1.73 [1.12, 2.67], p = 0.013), prognostic nutritional index (PNI, HR = 1.70 [1.09, 2.63], p = 0.018), and neutrophil‐to‐lymphocyte ratio (NLR, HR = 0.52 [0.29, 0.93)], p = 0.026) were independent risk factor for the survival of bladder cancer patients. Create a nomogram based on the aforementioned findings, and then draw the 1‐year, 3‐year, and 5‐year OS receiver operating characteristic curves by the nomogram. The AUC values were 0.811 (95% CI [0.752, 0.869]), 0.814 (95% CI [0.755, 0.873]), and 0.787 (95% CI [0.708, 0.865]), respectively, and the calibration plot matched the predicted value well. The 1‐year, 3‐year, and 5‐year decision curve analyses were higher than the ALL line and None line at threshold values of >5%, 5%–70%, and 20%–70% indicating that the model has good clinical applicability. The calibration plot for the Bootstrap 1000‐time resampled validation model was similar to the actual value. Patients with preoperative combination hydronephrosis, higher T‐stage, combined LVI, low PNI, and high NLR had worse survival, according to Kaplan–Meier survival analysis for each variable.ConclusionsThis study might conclude that PNI and NLR were separate risk factors that affect a patient's OS after RC for MIBC. The prognosis of bladder cancer may be predicted by PNI and NLR, but additional confirmation in randomized controlled trials is required.