Analysis of Outcomes after Second Allogeneic Stem Cell Transplant in Relapsed Acute Myeloid Leukaemia in Children: A Study from the EBMT Pediatric Diseases Working Party

Background:

Relapse after a first allogeneic hematopoietic stem cell transplant (HCT1) remains the leading cause of treatment failure in children with acute myeloid leukaemia (AML). A second allogeneic stem cell transplant (HCT2) is offered to a proportion of patients failing the first allograft. It is potentially curative but carries a high risk of relapse and toxicity. The latest extensive retrospective studies, published around ten years ago, noted a prolonged overall survival rate of around 30%. Recent developments in AML management and transplantation procedures may have impacted the results of HCT2.

Methods:

Through the EBMT registry, we retrospectively evaluated transplant outcomes of patients under 18 at the time of HCT2 with relapsed AML after HCT1 who then underwent HCT2 between 2000 and 2022. The primary outcome was Leukaemia-Free Survival (LFS). To assess changes in outcomes over time, we categorised patients, according to the median year of HCT2, into two periods: 2000-2013 and 2014-2022.

Results:

A total of 345 patients were identified, with 168 receiving HCT2 from 2000 to 2013 and 177 since 2014. The median age at HCT2 was 10.4 years (Q1: 5.2; Q3: 14.5). The median time between HCT1 and relapse was 10.8 months (Q1: 6.2; Q3: 17.8). Concerning disease status at HCT2, 265 patients (76.8%) were in complete remission. Donor types included 77 matched related donors (22.4%), 86 mismatched related donors (25%) and 181 unrelated donors (52.6%). At HCT2, 73.7% of patients had a different donor than their HCT1. The stem cell sources were bone marrow (n=112, 32.6%), cord blood (n=57, 16.6%), and peripheral blood (n=175, 50.9%). The conditioning regimen was myeloablative (MAC) in 75.2% of HCT2, total body irradiation was used in 34.4%, Busulfan in 23.1%, and Treosulfan in 18.8% of transplants.

The median follow-up after HCT2 was 4.9 years (CI95%: 4.4-6). At 3 years, leukaemia-free survival (LFS), overall survival (OS), non-relapse mortality (NRM), cumulative relapse incidence (RI) and GVHD-free, relapse-free survival (GRFS) were 30.2% (CI95%: 24.9-35.5), 37.5% (CI95%: 31.8-43.2), 19.1 (CI95%: 14.9-23.7), 50.7% (CI95%: 44.9-56.2) and 20.7% (CI95%: 16.1-25.6) respectively. The incidence of grade II-IV acute GVHD (aGVHD) was 34.8% (CI95%: 29.5-40.1) and of grade III-IV 13.2% (CI95%: 9.7-17.2). Three-year extensive chronic GVHD was 11.9% (CI95%: 8.4-16.2).

The recent period (2014-2022), compared to the earlier period (2000-2013) for the time of HCT2, was significantly associated with improved LFS (3-year probability: 34.3% vs 26.3%; HR 0.65, p=0.007), OS (3-year probability: 42.9% vs 32.8%; HR 0.60, p=0.003), RI (3-year incidence: 46% vs 54.7%; HR 0.66, p=0.029) and GRFS (3-year probability: 25.2% vs. 16.1%; HR 0.65, p=0.007), NRM (3-year incidence: 19.7% vs 19%; HR 0.59, p=0.095) and GVHD related outcomes did not differ significantly between the two periods. Time from HCT1 to relapse (> 6 months) was a favourable prognostic factor for LFS (HR 0.61, p=0.004), OS (HR 0.66, p=0.024), RI (HR 0.60, p=0.01) and GRFS (HR 0.61, p=0.004). Active disease at HCT2 was a significant risk factor for LFS (HR 1.71, p=0.003), OS (HR 1.63, p=0.01), RI (HR 1.87, p=0.002) and GRFS (HR 1.71, p=0.003). Additionally, poor-risk cytogenetics was a significant risk factor for OS (HR 1.50, p=0.045).

Conclusion:

For pediatric AML patients with relapse after HCT1, HCT2 is a suitable option able to rescue a proportion of patients, especially if the time elapsing between HCT1 and relapse is longer than 6 months and if a new complete remission was achieved. Still, relapse remains a significant obstacle in this setting. Interestingly, overall survival improved over time, linked to a substantial reduction in relapse incidence.