Sequential Intensified Conditioning Regimen Allogeneic Transplantation in Younger Adult Patients with Adverse-Risk Acute Myeloid Leukemia: A Multi-Center Prospective Study from the Acute Leukemia French Intergroup on Behalf of the Filo, ALFA, and SFG

Introduction: The French Backbone Intergroup (BIG-1) trial (NCT02416388) is a large open label, multicenter phase II/III study, with several therapeutic interventions in patients (pts) aged 18-60 years (yrs) with AML. HSCT is the only curative strategy for pts with adverse risk acute myeloid leukemia (adv AML) but the prognosis remains dismal (long-term survival ranging from 10% to 25%). Sequential-based conditioning regimens (Seq) with prophylactic donor lymphocyte infusions (pDLIs) have been successful in AML pts with refractory disease (Schmid C et al, JCO 2005). We hypothesized this strategy might improve the outcome in adv AML pts in first complete remission (CR). Study design and methods: Pts were classified as adv AML based on the cytogenetics and molecular biology at diagnosis. In case of CR, the HSCT protocol strategy recommended a Seq FLAMSA (Fludarabine 30 mg/m2, high-dose cytarabine 2 g/m2, and amsacrine 100 mg/m2) followed by a RIC regimen (Busulfan 6.4 mg/kg and cyclophosphamide 100 mg/kg) from matched sibling donor (MSD), 10/10 (MUD) or 9/10 (MMUD) unrelated donors. GvHD prophylaxis consisted in rabbit ATG 5 mg/kg with CsA plus MMF. CsA was quickly tapered from day 60 and discontinued at day 120 with pDLIs administered if no GvHD. Other graft sources and conditioning regimens were nevertheless authorized. The primary endpoint was overall survival (OS) based on a unique Cox model with predefined covariates: age (median), delay between diagnosis to HSCT (+/- 6 months), Seq strategy versus other conditioning regimen, donor type (MSD or MUD versus others), cytogenetic (monosomal karyotype [MK] versus others) and complete remission (CR) at time of HSCT. A selection of significant interactions between these covariates was planned based on likelihood-ratio tests between nested models. Non-relapse mortality (NRM) and relapse-free survival (RFS) were studied (secondary endpoints). Results: Between 01/2015 and 07/2021, 2023 AML pts were included (469 adv AML). Overall, 357 pts were transplanted (58.8% male) with 200 pts receiving the protocol Seq strategy (161 pts in CR; 69 MSD, 98 MUD, 29 MMUD and 4 haploidentical HSCT). The others 157 pts received mainly a RIC regimen (n=81; 52%) (113 pts in CR; 25 MSD, 40 MUD, 17 MMUD and 75 haploidentical HSCT). Median age was 50 yrs. Median follow-up was 5 yrs. In the 112 pts who were not transplanted, OS was estimated at 42.8% (95% CI, 31.3%-53.8%) and 0% (95% CI not evaluable) at 1 and 5 yrs, respectively. In the 357 transplanted pts, OS estimates were 66.9% (95% CI, 59.1%-72.5%) and 41.2% (95% CI, 35.3%-47.1%) at 1 and 5 yrs, respectively. OS for transplanted pts with Seq in CR at time of HSCT (n=161) was 52.3% (95% CI, 44.3%-59.7%) at 5 yrs versus 45.5% (95% CI, 34.9%-55.5%) in the 115 non Seq pts. For pts not in CR, OS was 18.8% (95% CI, 7.4%-34.2%) at 5 yrs for Seq pts (n=31) versus 23.6% (95% CI, 12.1%-37.2%) for non Seq pts (n=42). While no benefit of the protocol Seq strategy was found (HR: 0.81 [0.59-1.28], p=0.214), CR at time of HSCT was associated with better OS (HR: 0.44 [0.32-0.62], p=0.0001). Two interactions were identified impacting OS: the first between age (less than 50 versus 50 yrs or more) and MK (p=0.018) and the second between the delay from diagnosis to HSCT and donor type (MSD or MUD) (p=0.084). The impact of MK was significantly worse in older pts (HRs of 1.47 [0.93-2.33], p=0.099 and 3.05 [2.05-4.54], p=0.0001 for MK in pts under 50 yrs and those aged 50 yrs or more, respectively). A non-MSD, non-MUD donor was associated with a worse OS for pts transplanted after 6 months (HR: 2.33 [1.06 - 5.10], p=0.034) but not for those receiving HSCT within less than 6 months (HR: 1.08 [0.75-1.55], p=0.68). At 5 yrs, the cumulative incidence of non-relapse mortality was 19.1% (95% CI, 15.1%-23.4%) with CR at time of HSCT as the only factor associated with less NRM (adjusted HR, 0.53 [0.32-0.90]; p=0.017). The cumulative incidence of relapse was 25.8% (95% CI, 21.0%-30.8%) at 5 yrs. The 5 yrs RFS of 39.2% (95% CI, 33.8%-44.5%) with worse RFS for pts aged 50 yrs or more with MK (HR of 2.63 [1.80-3.85]), p=0.0001, while CR at HSCT was protective (HR of 0.53 [0.38-0.73], p=0.0001). Conclusion: The 5-yrs OS of 41.2% compares favorably with published results. While a Seq regimen did not improve HSCT outcomes, best results were obtained in younger pts with adv risk AML in CR at time of HSCT irrespective of graft source for those transplanted less than 6 months after AML diagnosis.