Transcriptomic classification of non-muscle invasive bladder cancer and its clinical and prognostic implication

About This Grant

This study addresses the unmet clinical needs for management of non-muscle invasive bladder cancer (NMIBC). NMIBC represents ~75% of bladder cancer cases and has a favorable five-year survival rate, but typically recurs (50-70% in 5 years) and progresses to muscle-invasive disease (MIBC, 10-30%). Intravesical Bacillus Calmette-Guerin (BCG) is the most effective therapy to prevent NMIBC recurrence and progression, yet BCG has a 30% failure rate, with various adverse effects leading to intolerance. Two critical needs exist: 1) a risk stratification tool to identify patients at high risk of recurrence and progression for early and aggressive treatment, and 2) a response prediction tool to identify patients who are unresponsive or intolerant to BCG for other treatment options. Our goal is to develop and validate risk-stratification and BCG-response tools by incorporating molecular signatures into the current pathological system for optimal clinical care of NMIBC. Transcriptome analysis is powerful to identify genes, and importantly, to define cancer molecular subtypes associated with different therapeutic and prognostic outcomes. This approach has been applied in bladder cancer but primarily focused on MIBC. Research on NMIBC is an unmet need. Building on the Bladder Cancer Epidemiology, Wellness, and Lifestyle Study (Be-Well), one of the largest prospective cohorts of NMIBC patients, we propose to conduct a comprehensive transcriptomic analysis of NMIBC and examine the utility of molecular subtypes with additional prognostic genes in predicting NMIBC treatment response and prognostic outcomes. Our central hypothesis is that molecular signatures (i.e., molecular subtypes and/or other genes) could unveil NMIBC heterogeneity and thus improve the current pathological classification system for tailored NMIBC care. We will: Aim 1. Develop a risk stratification tool for NMIBC prognostic outcomes by incorporating molecular subtypes, genes, clinicopathological and demographic factors. Primary outcomes will be disease recurrence and progression, with survival explored. We will define molecular subtypes and identify genes by NMIBC prognostic outcomes (1a), build risk prediction models in Be-Well (n=928) (1b), validate molecular signatures and the risk-stratification tool in an independent validation cohort (n=959) (1c), and explore the tool in sex and race specific groups in the pooled cohorts (1d). Aim 2. Develop a response prediction tool for BCG outcomes by incorporating molecular subtypes, immune signatures, genes, clinicopathological and demographic factors. Primary outcomes will be BCG unresponsive with BCG intolerant explored. Given the high immunogenic nature of bladder cancer and BCG therapy, we will develop the BCG- response prediction tool with further consideration of immune signatures in patients who received BCG in Be- Well (n=426) and a validation cohort (n=691). We will explore characterization of molecular subtypes by sex, race/ethnicity, and etiological risk factors. Clinical translation will be accelerated given the generation of NMIBC-specific molecular signatures using NanoString in a large health care delivery system setting.