HLA discordance between blood and tumor: A treatment conundrum?

Abstract

e21577 Background: Tebentafusp is a bispecific gp100 peptide-HLA-directed CD3 T cell engager that is now FDA approved for HLA-A*02:01 patients with unresectable or metastatic uveal melanoma (UM). Tebentafusp requires the targeted gp100 peptide to be presented by the major histocompatibility complex Class I complex HLA-A*02:01 allele for optimal cell killing. HLA-A*02:01 has been reported to be carried by approximately 50% of the North American and Western European population. The clinical trials for tebentafusp all required central blood tests to determine HLA status. In March of 2021, Caris Life Sciences began the launch of HLA Genotype as a standard reporting element for patient tumors profiled by Caris. Somatic HLA mutations have been described in different tumor types and has strong implications for a possible mechanism of immune evasion. Tumor heterogeneity between tumors and within tumors may also impact HLA tumor testing. As more novel drugs that rely on HLA genotyping are developed, understanding the relationship between blood and tumor HLA testing has become crucial. Methods: A single institution retrospective analysis was performed of patients who had tumor HLA genotyping performed as part of their tumor profiling by Caris Life Sciences. HLA Genotyping by Caris was performed using Whole Exome Sequencing of the tumor. Of the patients who had tumor HLA genotyping results, patients who also had blood HLA typing were identified. The concordance rate of HLA-A status between blood testing and tumor testing was determined. The prevalence of HLA-A*02:01 in our patients was also determined (either blood or tumor). Results: Forty-seven patients with both tumor and blood HLA genotyping were identified (37 UM, 8 cutaneous melanoma, 1 mucosal melanoma, 1 renal cell carcinoma). The majority of the patients were Caucasian (n = 44), with only two Asian patients and one African patient. There were two UM patients whose tumor and blood HLA genotyping were discordant (concordance rate of 95.7%). One patient was found to be HLA-A*01:01, *026:01 by blood but tumor genotyping revealed HLA- A*01:01, A*02:01. The other patient was found to be HLA-A*02:01, A*01 by blood but tumor genotyping revealed HLA-A*03:01, A*33:03. The prevalence of HLA-A*02:01 in the entire group was 34%, while the prevalence of HLA-A*02:01 in the Caucasian patients was 36.4%. None of the non-Caucasian patients were HLA-A*02:01 by blood or tumor genotyping. Conclusions: Among our 47 patients with both tumor and blood HLA genotyping, we found two patients with discordant HLA results. Neither of the discordant UM patients had received tebentafusp yet, but it brings up a crucial question about how these patients may respond to a therapy that relies on the HLA-A*02:01 complex. Our study highlights the importance of continuing to use blood HLA genotyping as the standard test to determine eligibility for a treatment such as tebentafusp, as the concordance rate with tumor genotyping is high but not 100%.