Abstract 3745: Allele-specific HLA LOH frequencies and survival outcomes in cancer: a real-world analysis

Abstract While immunotherapies have revolutionized cancer treatment, the unpredictability and the limited durability of clinical responses continue to pose challenges to achieving broader success in treating solid tumors. Therapies involving T cell receptor (TCR)-HLA:peptide interactions have shown promise, including the first-in-class FDA approval of afami-cel in 2024, but concerns have been raised whether their response rate will be negatively affected by HLA loss of heterozygosity (LOH) as a mechanism of resistance. Therefore, it is important to understand the frequencies and the potential impact which HLA LOH has on overall survival. Given the heterogeneity of HLA genes as well as multiple modalities relying on a specific HLA allele for their mechanism of action, allele-specific, in addition to allele-agnostic, analyses may help explain variance in treatment outcomes. We analyzed a real-world genomic dataset of 78, 418 cases (Montesion et al.), that were heterozygous for one or more HLA class I loci, to assess HLA LOH frequencies in paired groups of frequent HLA-A alleles and hotspot mutations. We also used an orthogonal dataset from MSK IMPACT and AACR Genie BPC to examine the impact of LOH on survival in colorectal cancer (CRC; n=579) and lung adenocarcinoma (LUAD; n=610) patients. Allele-agnostic HLA LOH was detected in 16.6% of all Montesion et al. cases. Rates varied by cancer type but had no impact on the survival from diagnosis in CRC or LUAD patients treated with any therapy. Allele-specific LOH was less frequent: A*01:01 (5.3%), A*02:01 (6.3%), A*03:01 (6.6%), A*11:01 (6.3%), and A*24:02 (6.5%). The frequency of A*02:01, A*03:01, and A*11:01 allele-specific LOH remained low, even when analyzed by driver mutations, ranging between 5.9% - 9.7% for KRAS mutations (G12C, G12D, G12V) and 5.9% - 6.2% for wild-type (WT) KRAS. The TP53 R175H mutation was associated with a limited increase in A*02:01 LOH (from 4.4% in WT to 8.1% in TP53 R175H). Notably, survival outcomes analyzed in the context of disease stage and allele-specific LOH indicated divergent trends between CRC and LUAD, suggesting that the prognostic influence of specific HLA alleles may be cancer-type, stage, and/or treatment dependent. For instance, A*02:01-specific LOH was associated with worse survival in CRC (HR 0.5355, 95% CI 0.2991-0.9589), but showed a trend toward improved survival in LUAD (HR 1.342, 95% CI 0.7713-2.336), in all patients treated with any therapy. Our findings highlight that over 90% of cancer cases found in the database are genomically intact at the HLA class I locus, preserving the molecular complex needed for TCR-HLA:driver mutation-based therapies. However, the association of A*02:01-specific LOH with poorer survival in CRC and the opposite trend in LUAD underscores the need for further investigation into allele-specific factors that may influence outcomes across various cancer types, treatment lines, and therapeutic modalities. Citation Format: Tomasz Sewastianik, Christian Roy, Michael Gormally, Meagan Montesion, Patrick Halvey, Aastha Jindal, Christopher A. Klebanoff, Gregory J. Opiteck, Dirk Nagorsen. Allele-specific HLA LOH frequencies and survival outcomes in cancer: a real-world analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3745.