Impact of primary tumor location, histology, and host factors on objective response to immune checkpoint inhibitors in metastatic urothelial carcinoma.

Abstract

493 Background: Factors affecting response to immune checkpoint inhibitor (ICI) are poorly understood in metastatic urothelial carcinoma (mUC). While tumor PD-L1 status is often used as a biomarker, it is not always predictive and ICI also benefits patients (pts) with PD-L1 negative tumors. Therefore, we sought to study the effect of some host and disease-related variables like gender, ethnicity, body mass index (BMI), platelet to lymphocyte ratio (PLR), and neutrophil to lymphocyte ration (NLR) on objective responses in pts with mUC treated with ICI. Methods: We performed a retrospective analysis of adult pts with mUC who received ≥2 cycles of ICI (pembrolizumab or atezolizumab) at the Cleveland Clinic from 2015 to 2020. Tumor and host-related factors evaluated are listed in the table below. We focused on meaningful treatment response, so only partial response (PR) and complete response (CR) were included as responders, while stable disease (SD) and progressive disease (PD) were counted as non-responders. Analysis was carried out with Fisher’s exact test and Wilcoxon rank sum test as applicable. Results: A total of 124 pts with mUC that received ICI were included. Gender did not correlate with response (p>0.99) or duration of response (p=0.37). Ethnicity did not correlation with response (p=0.78) or duration of response (p=0.24). Histology (UC, mixed variant histology or non UC) did not correlate with response (p=0.13) or duration of response (p=0.87). Location of primary malignancy (upper tract versus lower tract) did not correlate with response (p>0.99) or duration of response (p=0.36). BMI (p=0.23), PLR (p=0.9), and NLR (p=0.9) did not correlate with objective response. Conclusions: In our single center experience of pts with mUC treated with ICI, host factors (gender, ethnicity, histology, BMI, NLR, PLR) and location of primary tumor did not correlate with treatment response or duration of response. Although there were few African Americans represented in this study as commonly seen for minority representation, it is encouraging that no significant differences in responses were observed. The role of BMI and gender in response to ICI treatment in mUC was not observed. While there are limitations of a retrospective analysis, our study warrants investigation into predictive factors of response to ICI in mUC. Ongoing work integrating radiomics and pathomics will further our understanding and develop potential predictive biomarkers of response to ICI in mUC. [Table: see text]