Prognostic factors in patients with metastatic urothelial carcinoma who have been treated with atezolizumab

Background: In the current study, we evaluated pretreatment prognostic factors foroverall survival (OS) in patients with metastatic urothelial carcinoma who have progressed after first-line chemotherapy in the Expanded-Access Program ofatezolizumab.Methods: In this study, we present the retrospective analysis of 113 patients withurothelial cancer treated with ATZ after progression on first-line chemotherapy.Eligible patients included metastatic urothelial carcinoma patients treated with at leastone course of ATZ. Univariate analysis was used to identify clinical and laboratoryfactors that significantly impact OS. Variables were retained for multivariate analysis ifthey had a statistical relationship with OS (P?0.1) and were then included the finalmodel if P?0.05.Results: In univariate analysis, primary tumour location in the upper tract, increasedabsolute neutrophil count (ANC), increased absolute lymphocyte count, neutrophil-tolymphocyte ratio (NLR)>3, liver metastases, baseline creatinine clearance (GFR) < 60ml/min, Eastern Cooperative Oncology Group (ECOG) performance status (1) andhemoglobin<10 mg/dl were all significantly associated with OS. Three of the fiveadverse prognostic factors according to the Bellmunt criteria were independent ofshort survival: liver metastases (HR¼ 0.323; 95% CI 0.174-0.60; P <0.001), ECOG PS1(HR¼ 0.459; 95% CI 0.236-0.895; p¼0.022), and haemoglobin level <10 mg/dl (HR¼0.373; 95% CI 0.217-0.642; P <0.001). In addition, NLR>3 (HR¼ 0.474; 95% CI 0.234-0.962; P ¼0.039) and GFR <60 ml/min (HR¼ 0.546; 95% CI 0.328-0.907; P ¼ 0.019),maintained a significant association with OS in multivariate analysis. Patients weredivided into three risk categories: the favourable risk group (0-1 prognostic factor;median OS¼20.1 mo.), the intermediate-risk group (2 prognostic factors; medianOS¼10.08 mo.) and the poor-risk group (3 prognostic risk groups; median OS¼ 2.2mo.) (log-rank P< 0.001).Conclusions: This model confirms the Bellmunt model with the addition of NLR>3 andGFR <60 ml/min. Taken together, these factors can be used for prognostic parametersin clinical trials that use immunotherapy in patients with bladder cancer who haveprogressed after first-line chemotherapy.