Distinct immune and inflammatory response patterns contribute to the identification of poor prognosis and advanced clinical characters in bladder cancer patients.

Due to the molecular heterogeneity, most bladder cancer (BLCA) patients show no pathological responses to immunotherapy and chemotherapy yet suffer from their toxicity. This study identified and validated three distinct and stable molecular clusters of BLCA in cross-platform databases based on personalized immune and inflammatory characteristics. H&E-stained histopathology images confirmed the distinct infiltration of immune and inflammatory cells among clusters. Cluster-A was characterized by a favorable prognosis and low immune and inflammatory infiltration but showed the highest abundance of prognosis-related favorable immune cell and inflammatory activity. Cluster-B featured the worst prognosis and high immune infiltration, but numerous unfavorable immune cells exist. Cluster-C had a favorable prognosis and the highest immune and inflammatory infiltration. Based on machine learning, a highly precise predictive model (immune and inflammatory responses signature, IIRS), including FN1, IL10, MYC, CD247, and TLR2, was developed and validated to identify the high IIRS-score group that had a poor prognosis and advanced clinical characteristics. Compared to other published models, IIRS showed the highest AUC in 5 years of overall survival (OS) and a favorable predictive value in predicting 1- and 3- year OS. Moreover, IIRS showed an excellent performance in predicting immunotherapy and chemotherapy's response. According to immunohistochemistry and qRT-PCR, IIRS genes were differentially expressed between tumor tissues with corresponding normal or adjacent tissues. Finally, immunohistochemical and H&E-stained analyses were performed on the bladder tissues of 13 BLCA patients to further demonstrate that the IIRS score is a valid substitute for IIR patterns and can contribute to identifying patients with poor clinical and histopathology characteristics. In conclusion, we established a novel IIRS depicting an IIR pattern that could independently predict OS and acts as a highly precise predictive biomarker for advanced clinical characters and the responses to immunotherapy and chemotherapy.