Paper Details

Allogeneic hematopoietic cell transplantation (alloHCT) is considered the treatment with the highest curative potential for patients with European LeukemiaNet (ELN) adverse risk acute myeloid leukemia (AML). Induction chemotherapy is used to achieve morphologic remission prior to transplantation; however, the optimal number of pre-transplant chemotherapy cycles remains undefined, and to date, no dedicated study has been conducted to address this question specifically in adverse risk AML. As these leukemias are characterized by intrinsic chemoresistance, deepening remission with a second induction cycle (C2) may offer little additional benefit for patients who already achieve morphologic complete remission (CR) after the first cycle (C1), while increasing toxicity and potentially delaying or even precluding transplantation. To investigate this, we compared outcomes of patients with adverse risk AML who achieved CR1 after a single induction cycle and proceeded directly to alloHCT versus those who received C2 prior to planned alloHCT.

We performed a multicenter, retrospective study including adult patients with newly diagnosed ELN2017 and/or 2022 adverse risk AML who achieved composite CR(c) after C1 (7+3 or equivalent). We compared patients, treated according to physicians’ discretion, who proceeded directly to alloHCT after C1 (C1Allo) to patients who first received C2 before a planned alloHCT (C2Allo). Patient-level data were obtained from real-world registries across multiple European centers and the HOVON102 and HOVON132 trials. We performed 1:1 nearest-neighbor propensity score matching (PSM) on variables associated with survival, including age, TP53 mutation, complex karyotype, and estimated HCT-CI at diagnosis or pre-HCT (caliper 0.2). In subgroup analyses, we explored the relevance of measurable residual disease (MRD) assessed by flow cytometry after C1 (positivity ≥0.1% of white blood cells). Overall survival (OS), cumulative incidence of relapse (CIR), and non-relapse mortality (NRM) were estimated from a landmark defined as the start of the second treatment phase (for C1Allo: alloHCT, for C2Allo: C2) using Kaplan-Meier method, cumulative incidence function, Cox regression, and Fine-Gray method. We used Chi-squared test for between-group comparisons.

Among 608 eligible patients (238 C1Allo, 370 C2Allo), 231 matched pairs remained after PSM. Overall and individual covariable balance was good (standardized mean difference <0.1). The median age was 61 years (interquartile range: 52-65, range: 18-75), 15% had a complex karyotype, 12% had mutated TP53, and 18% had an HCT-CI score >3. Conditioning intensity between C1Allo and C2Allo was different (myeloablative: 15% vs 36%; non-myeloablative: 75% vs 55%; various sequential conditioning regimens: 9% vs 9%, p<0.01). Importantly, 10% of patients intended for C2Allo did not undergo alloHCT because of toxicity, relapse, or early death following C2.

We observed no significant difference in OS between C1Allo versus C2Allo (hazard ratio (HR): 0.81 95% confidence interval [0.60-1.10], p=0.18; 3-years: 59% vs 56%), neither for CIR (subdistribution (s)HR: 0.79 [0.54-1.14], p=0.2; 3-years: 25% vs 29%), or NRM (sHR: 1.04 [0.69-1.56], p=0.85; 3-years: 24% vs 20%). Results remained similar in a multivariable analysis restricted to all patients who underwent alloHCT, adjusted for conditioning intensity and donor type.

MRD data after C1 were available in 152 of the matched patients. MRD-positivity (n=64, MRDpos) was associated with inferior OS (HR: 1.70 [1.03-2.8], p=0.04), but not CIR (sHR: 1.31 [0.70-2.4], p=0.4), compared to MRD-negativity (n=88, MRDneg). Within both the MRDpos and MRDneg subgroups, we observed no significant difference in OS and CIR between C1Allo and C2Allo. Also after separate 1:1 PSM for both MRDpos (total matched n=40) and MRDneg (total matched n=28), results remained similar, although sample size was limited.

In this largest multicenter analysis to date, specifically evaluating ELN adverse risk AML patients achieving CR after C1, we found no benefit of a second induction cycle prior to alloHCT, irrespective of MRD status after C1. These results seem to support proceeding directly to alloHCT in these patients when a suitable donor is available, although limitations including the retrospective design and potential residual confounding of our study should be considered. Prospective validation is warranted.