Dynamic ΔEASIX score is a robust predictor of survival and relapse after allogeneic hematopoietic cell transplantation in acute leukemia

Abstract Allogeneic hematopoietic cell transplantation (allo-HCT) remains a curative therapy for acute leukemia in transplant-eligible patients. However, post-transplant complications, including relapse and non-relapse mortality (NRM), contribute to significant transplant failure. Traditional prognostic models, which rely on static pretransplant variables, often fail to capture the dynamic physiological stress of the post-transplant course. The Endothelial Activation and Stress Index (EASIX) score, reflecting endothelial injury, has emerged as a promising prognostic marker. While recent studies have highlighted the prognostic value of EASIX at specific timepoints or its trend over time in predicting NRM (Nawas et al. Blood Adv 2022), the prognostic significance of its change (ΔEASIX) in predicting both overall survival (OS) and relapse has not been thoroughly evaluated. We investigated the prognostic value of ΔEASIX (post-transplant EASIX minus baseline; pretransplant day –10) over multiple timepoints in predicting relapse and survival following allo-HCT in patients with acute leukemia.This retrospective single-center study included 137 adult patients with acute leukemia, 65% with acute myeloid leukemia (AML) and 35% with acute lymphoblastic leukemia (ALL), who underwent their first allo-HCT from a matched sibling donor between 2011 and 2024. The median age was 41 years, and 84% of patients had a Hematopoietic Cell Transplantation–Comorbidity Index (HCT-CI) score of 0. The majority of patients (90.5%) received a myeloablative conditioning regimen. All patients received a calcineurin inhibitor plus methotrexate for graft-versus-host disease (GVHD) prophylaxis. The incidence of grade ≥2 acute GVHD was 19% for the gastrointestinal tract, 22% for the skin, and 15% for the liver. Chronic GVHD requiring systemic treatment occurred in 18% of patients. Relapse occurred in 27% of patients, with a median relapse-free survival (RFS) of 554 days.EASIX was calculated using the formula: (LDH×creatinine)/platelet count at baseline (day –10), transplant day (day 0), and post-transplant days +7, +14, +21, +28, and +100. ΔEASIX was defined as the difference between each post-transplant value and the baseline value. OS and relapse were analyzed using Kaplan-Meier methods and Cox regression. The predictive performance of the scores was assessed using ROC analysis.Higher ΔEASIX values on days +21, +28, and particularly +100 were significantly associated with inferior OS. ΔEASIX at day +100 demonstrated the strongest predictive accuracy for OS (AUC: 0.746, p<0.001) followed by ΔEASIX at day +28 (AUC: 0.600, p= 0.044) and day +21 ΔEASIX (AUC: 0.603, p= 0.039). Kaplan-Meier analysis confirmed a significantly lower survival rate for patients with a high ΔEASIX D100 score (p<0.001). Our findings for ΔEASIX align with the dynamic prognostic approach shown by other groups using different methodologies (Armstrong et al. Blood 2022). Notably, we found that an early increase in ΔEASIX at day +7 predicted a higher risk of relapse (p=0.027). This may reflect impaired immune reconstitution and diminished graft-versus-leukemia (GvL) effect due to early post-transplant endothelial stress. Finally, dynamic changes in the EASIX score, particularly the ΔEASIX at day +100, serve as a strong independent prognostic marker for OS in acute leukemia patients undergoing allo-HCT. Unlike static pretransplant models, time-dependent ΔEASIX provides a real-time risk assessment tool, offering valuable insight for individualized post-transplant management and potentially guiding preemptive interventions such as donor lymphocyte infusion and tailored GVHD prophylaxis.