Next-Generation Sequencing for HLA Genotype Screening and Matching to HLA-Restricted Therapies

Abstract

This case series evaluates whether human leukocyte antigen (HLA) genotype could be inferred from standard next-generation sequencing, comparing concordance with confirmatory whole-exome sequencing and assessing if this assisted in matching patients to clinical trials.

In 2022, the first human leukocyte antigen (HLA)–restricted therapy, tebentafusp, gained US Food and Drug Administration approval.¹ A second, afamitresgene autoleucel, was recently approved, and clinical trials with HLA restrictions are increasing rapidly (Figure 1).² However, HLA evaluation is often not part of routine next-generation sequencing (NGS) ordered to evaluate newly diagnosed cancers for actionable alterations. Validated methods exist to infer HLA genotypes from NGS results but are not broadly adopted.^3,4,5 In this case series, we investigated whether we could infer HLA genotype from standard NGS, compared concordance with confirmatory whole-exome sequencing (WES) and approved tests, and assessed if this assisted in matching patients to clinical trials.

Figure 1. Growth of Human Leukocyte Antigen (HLA)–Restricted Therapeutic Trials.

This bar chart reports the annual cumulative number of HLA-restricted therapeutic clinical trials worldwide for any malignant indication. Their incidence has been increasing over the past 2 decades. Each annual bar is shaded to convey the discrete contribution from each HLA-restricted therapeutic class to the cumulative count. Classes include T-cell receptor (TCR) engineered adoptive cell therapy (ACT), TCR fusion protein, and vaccine. The total annual number of new trial registrations across all therapeutic classes is provided in the balloon plot below each bar with larger circles indicating larger numbers. The search criteria is provided in the eMethods in Supplement 1.

Methods

From 2014 to 2023, specimens from 58 236 patients underwent MSK-IMPACT (Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets), a hybridization capture-based tumor-normal NGS platform (Figure 1).⁶ Polysolver, version 3.0, derived HLA class 1 calls for HLA-A, HLA-B, and HLA-C using MSK-IMPACT data in a research setting.³ HLA genotype calls from MSK-IMPACT were compared to recaptured WES and CLIA-certified clinical HLA genotyping for patients referred to clinical trials requiring it (eMethods in Supplement 1). Data on race and ethnicity were collected from electronic health records. Ethics approval and informed consent were not required by Memorial Sloan Kettering because data were deidentified. Significance was determined as 2-tailed P < .05.

Results

For 3600 patients, we compared investigational HLA class 1 genotypes (the 6 HLA-A, HLA-B, and HLA-C alleles) to WES recapture results from the same DNA libraries and found high concordance across all versions of MSK-IMPACT used for investigational HLA genotyping (Figure 2); 95% to 98% of HLA genotypes matched those derived by WES recapture. Additionally, 380 of 672 screening referrals for HLA-A*02-restricted trials at MSK from 2013 to 2024 also received MSK-IMPACT testing. HLA eligible status by investigational and CLIA-certified testing was highly similar (98.9% agreement for 376/380; P < .001). Three of 4 discordances were between alleles within the same HLA-A*02 supertype having more than 99.99% sequence homology.

Figure 2. Human Leukocyte Antigen (HLA) Genotyping From MSK-IMPACT Improves Clinical Trial Screening Efficiency for HLA-Restricted Therapeutics.

These bar charts report the allele frequencies for the 20 most abundant (A) HLA-A, (B) HLA-B, and (C) HLA-C alleles across the MSK-IMPACT institutional cohort, with data available for investigational HLA genotyping in 58 236 patients. D, This bar chart shows the percentages of patients with potentially actionable alterations before implementing investigational HLA genotyping (29%, defined as tumor mutational burden > 10, microsatellite instability–high, alteration with OncoKB Therapeutic level 1 or level 2) and after (91%; additionally inclusive of patients with an HLA-eligible genotype for a HLA-restricted trial). The percentages of patients with potentially actionable alterations including HLA-eligible genotypes by racial category including Asian (n = 4486), Black or African American (n = 4034), and White patients (n = 44 434), from a total 58 236 patients. HLA alleles used as criteria for an HLA-restricted phase 1 trial include A*02:01, A*01:01, A*02:05, A*02:06, A*11:01, A*24:02, A*68:01, B*58:01, C*01:02, C*07:02, C*08:01, and C*08:02. The process for investigational HLA genotyping and actionability analysis is detailed in the eMethods in Supplement 1.

Investigational HLA genotyping substantially increased the number of patients with potentially actionable profiles (Figure 2). However, Black patients were less likely to be HLA-eligible for an HLA-restricted trial compared to Asian and White patients (63%, 95%, and 90%, respectively.)

When patients’ investigational HLA genotype indicated a potential match for an ongoing trial, we provided real-time automated notifications to the primary oncologist and trial investigators. One trial (NCT03709706) screened 31 patients, with only 4 who were HLA eligible after clinical testing. After midstudy implementation of this tool for prescreening, 14 patients with investigational HLA genotypes available were screened. All were confirmed HLA eligible using clinical testing, decreasing HLA ineligibility from 87% to 0%.

Discussion

Our clinical dataset demonstrates that inferring HLA genotypes is highly concordant to dedicated HLA testing. Timely testing is critical to match patients to the best therapies and facilitate trial participation. Appropriate comprehensive NGS should profile the full landscape of clinically actionable somatic and germline variants. Our approach can be applied broadly to in-house panels and several commercial NGS companies have incorporated HLA reporting in their standard or investigational panels.

One study limitation is that the current MSK-IMPACT panel cannot similarly resolve class 2 alleles. Although their clinical significance is presently limited, a growing number of investigational trials aim to elicit class 2 responses and knowledge of class 2 HLA haplotypes may become crucial for future targeting of tumor-restricted antigens.

HLA A*02:01 is the most common allele, a frequent subtype for HLA-restricted trial eligibility, and higher in White patients compared to Asian and Black patients. Trials are expanding beyond HLA A*02:01 with 18 active trials covering 8 non–HLA A*02:01 genotypes, but disparities persist particularly for Black patients. Enriching research datasets with HLA genotyping and incorporating this into standard NGS can support equitable development of therapies targeting HLA alleles and ensure universal access to the latest advancements in cancer treatment.

Supplement 1.

eMethods. Supplemental Search Criteria and Processes

Supplement 2.

Data Sharing Statement