Improving Risk Assessment of AML with a Precision Genomic Strategy to Assess Mutation Clearance

The persistence of leukemic mutation(s) in AML patients who have achieved a morphologic complete remission (CR) after intensive induction chemotherapy is a strong predictor of early relapse and reduced overall survival (OS) (Klco JAMA, 2015; Morita, J Clin Oncol 2018; Jongen-Lavrencic, NEJM, 2018). There is no clinical consensus as to the optimal consolidation therapy for the ~50% of patients with intermediate-risk AML. The median relapse-free survival (RFS) for patients ≤60 years with ELN intermediate-risk disease is 0.8 years to 1.2 years, with a median OS of 1.2-2.1 years (Mrozek, J Clin Oncol, 2012). We have shown that intermediate-risk patients who clear all leukemia-associated mutations (LAMs) to a variant allele fraction (VAF) of <2.5% in first morphologic CR have a median event-free survival of 25.6 months, vs 8.8 months if they do not (HR 3.32). Median overall survival is 46.8 months if all LAMs are cleared, vs 19.3 months if they are not (HR 2.88). We hypothesized that improved post-remission risk stratification using LAM clearance can further refine risk assessment and optimize alloHCT decisions by identifying patients at lower risk of relapse, who might be expected to do well with standard chemotherapy.

Here, we report the development of a pipeline to prospectively determine the persistence of LAMs after remission-induction, and return results in a clinically actionable time-frame. We perform enhanced exome sequencing (EES) of paired skin or buccal swab (normal tissue) and bone marrow DNA to comprehensively identify all LAMs at diagnosis (Day 0) and to assess their clearance post-induction (~Day 30). EES data are generated using a CLIA-compliant assay in the CLIA-licensed environment (CLE) lab at the McDonnell Genome Institute, and results are returned to the treating physician. Intermediate risk patients ≤60 years with clearance of all LAMs (VAFs <2.5%) are assigned to receive consolidation with high-dose cytarabine (HiDAC) (Cohort A). Patients with persistence of any mutation at a VAF ≥ 2.5% are assigned to the investigator's choice arm, and are treated with HiDAC and/or alloHCT at the discretion of the treating physician (Cohort B). This stratification is part of an ongoing clinical protocol (NCT02756962) whose primary objective is to determine whether the RFS of patients who have cleared all LAM(s) post-induction (VAFs <2.5%) and are treated with HiDAC alone (Cohort A) is significantly higher than expected from a historical intermediate risk group. Measurable residual disease testing by "difference from normal" flow cytometry (lower level of detection of 0.02%, Hematologics, Seattle WA) post-induction will be correlated with clearance or persistence of mutations and clinical outcomes.

For the 23 patients sequenced to date, the mean turnaround time to issue sequencing results to the treating physician was 24 days from the time of the remission biopsy. All 23 patients had detectable LAMs at presentation (mean 28 per patient, range, 6 to 43) that could be used to track persistent disease in the day 30 remission sample. Eleven patients (48%) cleared all LAMs and received HiDAC only (Cohort A). There was no flow cytometric evidence of residual AML in Cohort A. Twelve patients (52%) had persistent LAMs (Cohort B, investigator's choice). The number of persistent leukemia-associated variants present in Cohort B ranged between 1 and 14. Surprisingly, 9 of the 12 patients with persistent LAMs by sequencing had no flow cytometric evidence of residual leukemia. Seven of 12 patients on the investigator's choice arm have received an alloHCT, and none have relapsed to date. The median follow-up for all subjects is 378 days (range, 59-683). Neither the median RFS (Fig. 1A) nor the median OS (Fig. 1B) has been reached for either cohort. While preliminary, these results suggest that patients who clear all LAMs to a VAF of <2.5% may have durable responses with HiDAC alone. The encouraging RFS seen in the investigator's choice arm (Cohort B) may reflect the decision to recommend transplant "upfront" in CR1 for patients who have molecular persistent disease.

In summary, identifying persistent LAMs after induction chemotherapy is feasible in an actionable time-frame. Early data suggest that using LAM clearance post-induction may improve current risk-stratification for intermediate-risk AML. Accrual of patients and continued follow-up are ongoing.

View largeDownload slide

View largeDownload slide

Close modal