CHIME: A Novel HLA Eplet Model Predicting Non-Relapse Mortality in Haploidentical Hematopoietic Cell Transplantation

Highlight

This study proposes the CHIME score, a novel molecular mismatch model based on HLA eplet mismatches, which better predicts non-relapse mortality and severe acute graft-versus-host disease (aGvHD) in haploidentical hematopoietic cell transplantation with post-transplant cyclophosphamide (PTCy) compared to traditional HLA allele mismatch models. CHIME also correlates with cytokine release syndrome when combined with CD3 cell dose, suggesting improved risk stratification for transplant outcomes.

Study Background

Haploidentical hematopoietic cell transplantation (HaploHCT) has become an increasingly utilized option to treat hematologic malignancies when matched donors are unavailable. While post-transplant cyclophosphamide (PTCy) effectively reduces graft-versus-host disease (GvHD) and improves outcomes, the role of human leukocyte antigen (HLA) mismatch in this setting remains contentious. Traditional assessments focus on antigen or allele mismatches at the HLA locus, but these may insufficiently capture immunogenic disparities that drive transplant complications.

The concept of eplet mismatch, which quantifies mismatches at a molecular level based on epitope differences on HLA molecules, offers a refined approach to assessing donor-recipient compatibility. A molecular model could enhance prediction of transplant risks such as GvHD, cytokine release syndrome (CRS), and non-relapse mortality (NRM), thereby informing donor selection and conditioning strategies more precisely.

Study Design

This retrospective cohort study analyzed 265 patient-donor pairs undergoing HaploHCT with PTCy for hematologic malignancies. Eplet mismatches were quantified across HLA class I and II molecules, and the interaction amongst eplet mismatch load, HLA class, and mismatch direction vector was modeled. This interaction was termed the Class-wise HLA Imbalance of Mismatched Eplets (CHIME) score, stratifying patients into three risk groups (0, 1, 2).

Clinical endpoints included non-relapse mortality (NRM), severity of acute GvHD (grades III-IV), and incidence of cytokine release syndrome (CRS). Adjusted hazard ratios (HRs) were calculated controlling for known transplant-related confounders. Additionally, combinations of CHIME score with CD3+ T cell dose were explored for associations with CRS.

Key Findings

The CHIME model demonstrated a significant association with clinical outcomes. Patients with a CHIME risk score of 1 versus 0 showed nearly a threefold increased risk of NRM (HR 2.85; 95% CI, 1.10-7.37), whereas those with a CHIME score of 2 had an even higher risk (HR 3.63; 95% CI, 1.36-9.68). These associations remained statistically significant after adjusting for multiple transplant variables (adjusted p = 0.046).

Regarding severe acute GvHD, CHIME scores of 1 and 2 correlated with substantially increased hazards: HR 7.82 (95% CI, 1.03-59.22) and HR 8.08 (95% CI, 1.01-64.63), respectively, with similar statistical significance after adjustment (adjusted p = 0.046). This suggests CHIME effectively stratifies risk for life-threatening GvHD complications.

Moreover, the combination of CHIME with the CD3 cell dose yielded a correlation with higher CRS incidence, indicating that both molecular mismatch and the T cell dose contribute synergistically to inflammatory complications post-transplant. This highlights the potential of CHIME as part of a composite risk assessment tool for transplant-related toxicities.

Notably, the CHIME model outperformed traditional allele-level mismatch assessments in predicting NRM and severe aGvHD, underscoring the value of high-resolution molecular mismatch analyses to better capture immunological disparities relevant to transplant outcome.

Expert Commentary

The introduction of CHIME is a significant advancement in tailoring risk prediction for HaploHCT recipients, especially in a setting reliant on partially mismatched donors. By accounting for differential eplet mismatches in class I and II HLA molecules and the mismatch vector, CHIME captures a nuanced immunogenetic interaction that translates into clinically meaningful risk stratification.

Expert opinion suggests that this refined model could guide donor selection beyond conventional allele typing and potentially inform immunosuppressive strategies and prophylaxis customization. However, prospective validation in larger multicenter cohorts is necessary to establish generalizability and integration into clinical algorithms.

Biologically, eplets represent functional antigenic determinants recognized by recipient antibodies and T cells. The link of CHIME to both severe GvHD and cytokine release syndromes reinforces the mechanistic relevance of molecular mismatches in triggering potent alloimmune responses despite PTCy immunomodulation.

Current limitations include the retrospective design and the single-center nature of the cohort, which may restrict applicability to broader populations. The predictive capacity for relapse outcomes was not detailed, which would be valuable to assess potential graft-versus-leukemia benefits.

Conclusion

CHIME is a novel, evidence-based HLA eplet mismatch scoring system that significantly associates with increased non-relapse mortality and severe acute GvHD in HaploHCT with PTCy. It represents an important step forward in molecularly informed donor compatibility assessment, improving outcome prediction beyond traditional HLA antigen or allele mismatch approaches. Integration of CHIME with cellular dosing further enriches risk stratification, potentially guiding personalized transplant management.

Future research should focus on prospective validation, exploration of CHIME’s predictive value for relapse outcomes, and incorporation into clinical donor selection algorithms to optimize HaploHCT outcomes.

Funding and ClinicalTrials.gov

The original study by Carter et al. was supported by institutional funding sources not explicitly detailed in the abstract. Clinical trial registration information was not provided in the source citation.

References

1. Carter M, Park S, Taniguchi M, et al. CHIME: a novel HLA eplet model associates with non-relapse mortality in haploidentical hematopoietic cell transplant with post-transplant cyclophosphamide. Bone Marrow Transplant. 2026 Jul 4. PMID: 42401707.

2. Hall EM, Petersdorf EW. Mismatching of HLA epitopes: implications for hematopoietic cell transplantation. Blood Adv. 2019;3(21):3296-3301.

3. Fleischhauer K. Human leukocyte antigen mismatching and the risk of graft-versus-host disease. Semin Hematol. 2014;51(4):266–273.