Impact of Extramedullary Disease on the Outcomes after Allogeneic Hematopoietic Cell Transplantation in Children and Young Adults with Acute Myeloid Leukemia - a CIBMTR Analysis

Background: In patients with AML, EMD can occur in the central nervous system (CNS) or other sites, in addition to bone marrow (BM) disease. Only a few retrospective studies have evaluated the impact of EMD on outcomes after alloHCT for children with AML. We hypothesized that the presence of EMD at AML diagnosis would be associated with inferior outcomes after alloHCT. Methods: The Center for International Blood and Marrow Transplant Research (CIBMTR) database was used to compare the effect of EMD in patients with AML. Based on EMD at diagnosis, patients were grouped as BM involvement (Group I), BM+CNS involvement (Group II), and BM+other EMD±CNS (Group III). The primary outcome was 1year (y) progression-free survival (PFS) with 1y overall survival (OS), non-relapse mortality (NRM), cumulative incidence (CI) of acute and chronic graft versus host disease (GVHD), and CI of relapse as secondary endpoints. Kaplan Meier estimates were used to assess survival outcomes and groups compared using log-rank test; Cox regression was used to identify covariates associated with survival.

Results: Our cohort includes 938 children (<21y; 52% male) who received first alloHCT for AML between 2008-2019. All patients were in a morphologic complete remission (CR) and confirmed to have no evidence of EMD at the time of alloHCT. Our study cohort included 630 Group I, 212 Group II and 96 Group III patients. Group III patients were younger at HCT (median age 2.9y vs 10.8y Group I, 10y Group II;), had poor-risk cytogenetics (50% vs 37% Group I, 39% Group II) and a higher pediatric disease risk index (52.1% vs 21.9% Group I, 25% Group II) and poor risk cytogenetics (50% vs 39% Group I, 37% Group II). The majority (63%) of patients had low comorbidity burden by expanded youth malignant HCT comorbidity index (score of 0-1, no difference between groups). While 67.5% of patients received chemotherapy-only conditioning, 32.1% received total body irradiation (TBI) based, and the remaining chemotherapy with low-dose TBI (4.5%) conditioning regimens. Radiation to other sites, including CNS, was reported in 6.7% of the study cohort. Donors were mostly unrelated cords (42.4%), matched unrelated (20.5%), and matched siblings (18%), with the remainder being haploidentical (6.4%) and mismatched unrelated (6.9%)Graft sources included bone marrow (39.8%), umbilical cord blood (46.4%), and peripheral blood (14.8%). GVHD prophylaxis was commonly calcineurin inhibitor-based (89%), with the remainder receiving post-transplant cyclophosphamide-based prophylaxis (6.1%) or ex vivo T cell-depleted grafts (3.2%). At a median follow of 69.8 months (range 3.4-155 months), the 1y PFS of the entire group was 67.4%. The day +100 acute GVHD, 1y NRM, relapse incidence, OS, and chronic GVHD of the entire group was 31.7%, 10.3%, 22.6%, 77%, and 29.6% respectively. Multivariable analysis of factors associated with relapse, after adjusting for disease status (includes MRD), age, and TBI use, EMD was associated with improved survival: Group II: hazard ratio (HR) of 0.697 (95% CI 0.515-0.944; p=0.012) and Group III HR of 0.695 (95% CI 0.459-1.053; p=0.087) compared to Group I cohort. There was no difference in PFS (p=0.21), NRM (p= 0.27), and OS (p=0.90) among the groups.

Conclusion: Our results suggest that the presence of EMD with BM at diagnosis did not impact clinical outcomes post-alloHCT. Although this cohort describes a subset of patients with EMD - those who were able to clear EMD prior to HCT and proceed to HCT – we observed that EMD at diagnosis among those who are otherwise able to proceed to HCT does not portend worse HCT outcomes or more relapse. It is possible that, in this setting, the ability to clear EMD prior to HCT was suggestive of chemotherapy-sensitive disease.