Clearance of Leukemia Associated Mutations (LAMs) after Induction Therapy Is Associated with Favorable Outcomes in Intermediate Risk Acute Myeloid Leukemia: Interim Results of a Multicenter, Prospective Phase II Trial

Introduction: The optimal post-remission consolidation strategy for medically fit, intermediate-risk patients with AML in morphologic first complete remission (CR1) is controversial and can include high-dose cytarabine (HiDAC) or allogeneic hematopoietic cell transplant (alloHCT). We previously showed that patients in CR1 who cleared all Leukemia Associated Mutations (LAMs) identified by whole-exome sequencing (WES) to < 2.5% variant allele frequency (VAF) (equivalent to 5% of cells) after recovery from induction had better survival than patients with LAMs persisting above 2.5% VAF (Klco, et al. JAMA 2015). We hypothesized that intermediate-risk patients in CR1 post-induction who had LAM clearance by WES may have better outcomes than similarly treated historical controls (for whom mutation clearance did not influence treatment), and thus would do well with HiDAC consolidation; those with persistent LAMs may be more likely to benefit from alloHCT. Here, we report the interim results of a multicenter, prospective clinical trial that used molecular analysis by WES at CR1 to determine consolidation therapy (NCT02756962).

Methods: Eligible patients were ≤ 60 years with de novo AML in morphologic CR/CR with incomplete count recovery after standard 7+3 induction, and transplant eligible. To identify LAMs, we performed “real-time” WES on bone marrow and matched normal tissue at presentation. WES was repeated with post-induction recovery bone marrow (approximately 30 days after induction) to determine the VAFs of each LAM identified at presentation and reported within ~4 weeks. Patients who cleared all LAMs below 2.5% VAF were assigned to “Cohort A” and received only HiDAC for consolidation. AlloHCT was used only as salvage in the case of relapse. Patients who had persistent LAMs, regardless of mutation identity, were assigned to “Cohort B” and received HiDAC +/- alloHCT at the discretion of the treating physician. The primary endpoint compared relapse-free survival (RFS) of intermediate-risk patients (ELN 2010 criteria) in Cohort A to an historical intermediate-risk group who received chemotherapy, but not alloHCT, in CR1 (Mrozek et al, JCO 2012). Other key endpoints included overall survival (OS) of Cohort A, and RFS and OS of Cohort B patients who received alloHCT in CR1 vs. those who did not. Patients with mutant NPM1 without FLT3-ITD (favorable risk by ELN 2010 criteria) were also evaluated as an exploratory cohort.

Results: The study enrolled 107 patients between July 2016 and March 2024 from 3 centers with 100 patients evaluable for response, meeting accrual goals. With a data cut-off of 6/4/2024, the median follow-up was 32 months (range 1 – 68). We identified a median of 30 LAMs (range 3-93) per patient at presentation (median of 4 recurrently mutated AML genes and 26 clonal passenger mutations). 51 patients cleared all LAMs to <2.5% VAF and were assigned to Cohort A. The median RFS for intermediate-risk patients in Cohort A (33.1 months, 95% CI: 11.7-not reached [NR], n=33) was superior to the historical cohort (11.7 months; p<0.035). The median OS for intermediate-risk patients in Cohort A (not reached) was greater than that of the historical cohort (24.7 months; p<0.01). Of the intermediate-risk patients in Cohort A who relapsed, 8/13 (62%) received a salvage alloHCT. 49 patients had persistent LAMs (≥2.5% VAF) and were therefore assigned to Cohort B. Of the intermediate-risk patients, 22/34 (65%) received an alloHCT in CR1.The median RFS and OS for the transplanted patients has not yet been reached, but for patients who did not receive a transplant, the median RFS was 5.8 months (95% CI: 3.3-NR) and median OS was 18 months (95% CI: 9.2-NR).

Conclusions: Our interim analysis suggests that patients who clear all clonal LAMs to a VAF of <2.5% post-induction have better outcomes with HiDAC than similarly treated historical controls, for whom mutation clearance did not dictate therapy. Our results support the use of standard HiDAC as a post-remission consolidation strategy for these patients. The favorable outcomes of transplanted patients with persistent LAMs with >2.5% VAF strongly supports using alloHCT in CR1. These data suggest that a genomically-guided approach for the detection of persistent LAMs after recovery from induction chemotherapy may improve risk-stratification--and outcomes--for patients with intermediate-risk AML.