RESUMO
BACKGROUND: The current prognostic models for patients with advanced urinary tract cancers were developed and validated in the chemotherapy setting. As immunotherapy has become the backbone of novel treatments, updated prognostic scores are needed. METHODS: A comprehensive analysis of inflammatory indexes from peripheral blood and clinical factors was planned on the entire real-world cohort of pretreated patients with advanced urinary tract carcinoma receiving atezolizumab in the prospective, single-arm, phase IIIb SAUL study. Univariable and multivariable analyses with overall survival as the primary endpoint, bootstrap internal validation, Schneeweiss scoring system and calibration test were performed to develop a novel immunotherapy prognostic score. RESULTS: Thirteen clinical variables from 1001 patients were analyzed. The following eight prognostic factors were included in a model: ECOG PS, liver and bone metastases, histology, pre-treatment steroids, systemic immune-inflammatory index (i.e., neutrophils-to-lymphocytes ratio times platelets count), hemoglobin and lactate dehydrogenase. The prognostic model was able to stratify patients into five risk groups with significantly different (P<0.001) median overall survival of NR, 18.0, 8.7, 4.6 and 2.4 months, respectively. The c-index for OS was higher than the Bellmunt Score one (0.702 vs. 0.672). CONCLUSIONS: A novel 5-class prognostic model contemporary to immunotherapy provides robust prognostic discrimination of patients with advanced urinary tract carcinoma homogeneously treated with immunotherapy through baseline affordable and reproducible clinical and laboratory factors. It could be quickly adopted in clinical practice to inform patients about prognosis with immunotherapy and assess the benefit of novel immunotherapy combinations in clinical trials.