Enhanced Bellmunt Risk Score for Survival Prediction in Urothelial Carcinoma Treated With Immunotherapy

Abstract

The discrimination performance of Bellmunt risk score for immune checkpoint inhibitor (ICI) therapy is largely unknown. This study aimed to validate and enhance discrimination of the Bellmunt score in patients with urothelial carcinoma treated with ICIs. Cox proportional hazard analysis were used to validate overall survival (OS) discrimination performance of the Bellmunt score in patients with urothelial carcinoma treated with atezolizumab in IMvigor210. The c-statistic (c) was used to evaluate the ability of C-reactive protein (CRP), neutrophil-to-lymphocyte ratio (NLR), lactate dehydrogenase (LDH), PD-L1 gene expression level on immune cells (PD-L1 ICs), albumin, time from prior chemotherapy, and tumor site count to enhance the Bellmunt score. External validation of an enhanced Bellmunt score utilized the independent atezolizumab arm of IMvigor211. In IMvigor210, Bellmunt score displayed moderate OS discrimination (c=0.66). Addition of CRP (one point for CRP>30 mg/L) to the Bellmunt score resulted in greatest improvement in performance (c=0.70), followed by NLR (c=0.69). On external validation, CRP-Bellmunt score had superior performance (OS c=0.67, PFS c=0.60) than original Bellmunt score (OS c=0.64, PFS c=0.59) with 30% of patients reclassified into a higher risk group. Patients with CRP-Bellmunt score of 0, 1, 2, or 3-plus had 1-year OS probabilities of 63%, 44%, 21%, and 15%, respectively. CRP inclusion within the Bellmunt score enhanced the ability to discriminate high risk patients misclassified using the original model. We propose that the CRP-Bellmunt score may enable improved patient stratification in ICI clinical trials and provide more accurate prognostic information for patients with urothelial carcinoma initiating ICIs.