Abstract 1389: Defining immune phenotypes of checkpoint inhibitor response in human live tumor fragments preserving the native tumor microenvironment.

Abstract Background: Immune checkpoint inhibitors (ICIs) provide durable benefit for a subset of patients, yet current FDA-approved biomarkers (eg, PD-L1, MMR/MSI, TMB) inaccurately predict clinical response. To address this gap, we have developed an ex vivo platform using live tumor fragments (LTFs) and multiplex cytokine profiling to accurately predict ICI response in patients (P0.05). Ex vivo cytokine profiling of LTFs produced from clinically relevant tumor specimens provides a unique opportunity to elucidate the biology and mechanisms of action of ICI response in the context of the native tumor microenvironment (TME). Methods: Fresh core needle or forceps biopsy specimens collected as part of three ongoing observational clinical trials (NCT05478538, NCT05520099, NCT0634962) and a biobank biopsy collection study were profiled on the platform. Biopsy specimens were cut into LTFs, encapsulated in hydrogel and treated with ICI (αPD-1, αPD-L1, αPD-1+αCTLA-4 or αPD-L1+αCTLA-4) ex vivo. Longitudinal profiling of 46 cytokines was performed on 193 patient samples representing multiple solid tumor types. Unsupervised hierarchical clustering with Wards method was used to identify cytokine-based response patterns and classify specimens into primary responder and non-responder clusters. Upregulated cytokines within the response cluster were defined using Mann-Whitney-U tests with Bonferroni correction for multiple comparisons, annotated for their function, and subclusters were classified into immune phenotypes. Results: Comparison of response and non-response clusters identified 21 significantly upregulated cytokines (P0.05) known to be associated with immune response to ICI. Additionally, response subcluster analysis identified 4 distinct tumor-type agnostic immune phenotypes: (1) T-cell inflamed with regulatory counterbalance, (2) mixed regulatory T-reg/myeloid (3) cytotoxic inflamed with T-reg-VEGF counterbalance, and (4) polyfunctional activated with myeloid angiogenic remodeling. Furthermore, subcluster assignment was consistent in 6 of 7 (85%) patient specimens with multiple replicates, demonstrating that profiles of response, when present, are reproducible between replicates from the same tumor. To date, patients with available clinical response data to an ICI therapy were identified across different subclusters. Conclusions: The ex vivo platform identifies distinct and expected immune phenotypes in response to ICI treatments. These data underscore the platform’s potential to accurately predict clinical response to ICI. Future studies will help to understand the relationship between these immune phenotypes and the depth and durability of ICI response in patients. Citation Format: Erika von Euw, Julie Zweng, Nicholas Dana, Chetan Sood, Hilary R. Hernan, Laura C. Hrycyniak, Payton McDonnell, Pichet Adstamongkonkul, Christin Johnson, Nathan Marhefke, Amreen Nasreen, T S. Ramasubramanian, Katherine Rexroad, Sidney Schneider, Christina Scribano, Aishwarya Sunil, Ellen Wargowski, Christina Vivelo, Lindsey Vedder, Sean Caenepeel, David Alexander Braun, Hinco J. Gierman, Hilario Ramos. Defining immune phenotypes of checkpoint inhibitor response in human live tumor fragments preserving the native tumor microenvironment abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1389.