Characterizing immune-genomic signatures associated with response to immune checkpoint inhibitors (ICI) in patients with metastatic urothelial carcinoma (mUC).

Abstract

e16582 Background: Response rates with ICI are modest at around 20% in mUC and there is a lack of validated biomarkers to select pts who are likely to derive benefit from ICI. There is an unmet need to develop biomarkers of response and resistance to avoid unnecessary treatment (Tx) and spare physical and financial toxicity. Here we report tumor immune-genomic characteristics associated with response and resistance to ICI in mUC. Methods: We selected 58 mUC pts who received at least 2 cycles of ICI (atezolizumab or pembrolizumab) between 2015-2020 and had pre-Tx archival formalin fixed paraffin-embedded (FFPE) tissues available. Pts were categorized by their response to ICI as responders (R) (Complete Response (CR), Partial Response (PR), and Stable Disease (SD)) (N = 30) or non-responders (NR) (Progressive Disease (PD)) (N = 28). Tumor was macrodissected from FFPE slides and RNA extraction was done using Qiagen Rneasy FFPE kit. 100ng of RNA was used for transcriptomic analysis using the Nanostring PanCancer IO 360 panel. Data normalization and analysis was carried out using nSolver 4.0. Results: Differential expression analysis comparing (R) vs (NR) displayed upregulation of gene signatures associated with Immunoproteasome, Antigen presenting machinery (APM), IFN Gamma, PD-L1, and IFN downstream. Additionally, the tumor inflammation signature (TIS), which is trained to predict response to ICI, was upregulated in (R) vs (NR) (Table 1). Single gene analysis showed trend towards upregulation of ISG15 (logFC = 0.7821, p = 0.00337), TAP1(logFC = 0.7649, p = 0.00422), and STAT1 (logFC = 0.6053, p = 0.00510) in (R), P-values, were above significance threshold after adjustment for multiple comparisons. Receiver operator characteristic (ROC) analysis with area-under the curve (AUC) scores assess capacity of gene signatures to predict response to ICI (Table 1). In our cohort, the median overall survival for (R) was 44.1 months compared to 24.5 months in NR (p = 0.0100). Conclusions: In our cohort of mUC pts treated with ICI, we found an association between signatures of tumor immunogenicity (APM and Immunoproteosome) and anti-tumor immune activity (IFN Gamma, IFN Downstream, TIS) in pre-Tx tissue with favorable response to therapy (including SD). These findings are hypothesis generating and need to be further validated in prospective trials. Ongoing genomic and immunologic correlative studies in our cohort will further help understand comprehensive biomarkers of response and resistance to ICI in pts with mUC. [Table: see text]