Background: Genomic loss of the mismatched HLA haplotype (HLA-Loss) is a frequent mechanism of immune evasion and relapse following haploidentical hematopoietic stem cell transplantation (haplo-HSCT). Prognostic data in the era of novel immunotherapies (e.g., CAR-T, targeted agents) remain limited.

Methods: We collected clinical data from patients with hematologic malignancies who underwent haplo-HSCT and tested positive for HLA loss at Hebei Yanda Lu Daopei Hospital and Beijing Lu Daopei Hospital between November 2020 and September 2024. HLA loss and specific lost loci were identified using next-generation sequencing (NGS) and HLA-KMR analysis. NGS is used for detection when tumor cell content exceeds 10% in haploidentical transplants and exceeds 50% in fully matched transplants. This method can detect specific genotypes at five HLA loci: HLA-A, B, C, DRB1, and DQB1. KMR, while more sensitive for tumor cell content below 10%, cannot cover all genotypes due to probe limitations. Consequently, when KMR detects HLA gene loss, it can only identify the presence of loss but cannot distinguish whether the loss occurs at a single locus or involves an entire haplotype. Clinical characteristics and outcomes of patients with HLA loss were assessed. The last follow-up date was July 1, 2025.

Results: Among 45 included patients, HLA-Loss was detected via NGS in 35 and KMR in 10. Of the NGS-confirmed cases (n=35), haplotype loss occurred in 16 (45.7%), while locus-specific losses included: 4 loci (20.0%, n=7), 3 loci (25.7%, n=9), and 2 loci (8.6%, n=3). Median HLA loss proportion was 100% (range: 64.7–100%).

HLA-loss positive patients included 19 females (42.2%) and 26 males (57.8%). Median age at transplant was 26 years (range: 3-56 years). Diagnoses included acute myeloid leukemia (60.0%, n=27), acute lymphoblastic leukemia (20.0%, n=9), myelodysplastic syndrome / MDS transformed AML (13.4%, n=6), non-Hodgkin's lymphoma (4.4%, n=2), and mixed-phenotype acute leukemia (2.2%, n=1). All patients received haplo-HSCT; pre-transplant disease status was CR (80.0%, n=36) or NR (20.0%, n=9). Conditioning regimens were busulfan (Bu)-cyclophosphamide (Cy) (66.7%), total body Irradiation (TBI)-Cy (17.8%), Bu-fludarabine (flu) (13.3%), or TBI-flu (2.2%). ATG-fresenius (ATG-F) or ATG-thymoglobuline (ATG-T) were used for GVHD prophylaxis. Median time to relapse/MRD positivity post-transplant was 344 days (range: 65-1979 days). Post-relapse treatments included chemotherapy (80.0%), targeted agents (55.6%), DLI (26.7%), and CAR-T (13.3%). Twenty-two patients (48.9%) underwent second HSCT with donor change.

The 5-year overall survival (OS) was 46.2% (95% CI: 30.0–62.5) for all patients. Second HSCT significantly improved 5-year OS (55.2% [30.1–80.2] vs. 36.4% [15.6–57.3]; p=0.021). No survival difference was observed between haplotype loss vs. non-haplotype loss (41.0% [15.6–66.4] vs. 53.2% [28.7–77.7]; p=0.596) or by number of loci lost (p=0.760). Pre-transplant CR status predicted better 5-year OS vs. NR (51.9% [33.3–70.6] vs. 22.2% [0.0–49.4]; p=0.013). TBI-based regimens showed inferior 20-month survival vs. Bu-based regimens (33.3% [2.5–64.1] vs. 75.0% [60.9–89.1]; p=0.010). A total of 25 deaths occurred. The causes of death were: relapse(12 cases, 48%), infection (8 cases, 32%), graft-versus-host disease (GVHD) (2 cases, 8%), cytokine release syndrome (CRS) (1 case, 4%), cardiac arrest (1 case, 4%), and liver failure (1 case, 4%).

Conclusion: Second transplantation improves survival in patients experiencing relapse following haplo-HSCT who demonstrate HLA loss. Survival outcomes were independent of HLA loss pattern (haplotype vs. locus-specific) or extent (number of loci lost). Poor prognostic factors were pre-transplant NR status and the use of TBI-based conditioning.

This content is only available as a PDF.
2025
Sign in via your Institution