Using real-world data to define a validated nomogram for advanced bladder cancer patients who respond to immunotherapy

Abstract

Aims Immune checkpoint inhibitors (ICIs) are used in incurable urothelial cancers, both in chemo-naïve and platinum-refractory patients. Efficacy and toxicity data published outside controlled clinical trials are limited. We report overall survival, progression-free survival and toxicities of ICIs in locally advanced (LABC) or metastatic bladder cancer (MBC). We aimed to develop and validate a prognostic model for these patients. Materials and methods A multicentre real-world individual patient-level data study (n = 272) evaluating ICIs in the first-line platinum-ineligible or platinum-refractory setting for LABC/MBC between March 2017 and February 2020 was undertaken. Cox regression analyses evaluated the association of prognostic factors with overall survival. Data were split to create a training (n = 208) and validation (n = 64) cohort. The backward elimination method with a P-value cut-off of 0.05 was used to develop a reduced prognostic model using the training data set. The concordance index and assessment of observed versus predicted survival probabilities were used to evaluate the final model. Results The median follow-up was 18.9 (15.8–21.5) months. The median overall survival and progression-free survival in the training cohort were 9.2 (95% confidence interval 7.4–10.5) and 4.5 months (3.5–5.7), respectively. The most common grade 1/2 adverse events recorded were fatigue (47.8%) and infection (19.9%). Five key prognostic factors found in the training set were low haemoglobin, high neutrophil count, choice of immunotherapy favouring pembrolizumab, presence of liver metastasis and steroid use within 30 days of treatment. The concordance index for the training and validation cohorts was 0.66 (standard error = 0.05) and 0.64 (standard error = 0.04), respectively, for the final model. A nomogram was developed to calculate the expected survival probabilities based on risk factors. Conclusions Real-world data were used to produce a validated prognostic model for overall survival in LABC/MBC treated with ICIs. This model could assist in patient stratification, interpreting and framing future trials incorporating PD-1/PD-L1 inhibitors in LABC/MBC.