Investigators Identify Distinct Molecular Subgroups to Guide Treatment Decisions in AML

Analyzing genetic data from more than 3,000 patients with acute myeloid leukemia (AML), researchers from Memorial Sloan Kettering Cancer Center identified 14 distinct molecular subgroups of the disease, each associated with distinct clinical presentation and outcomes. The investigators, who presented their findings at EHA2021 Virtual, used this information to develop a unified framework for disease classification and risk stratification that could be adopted in the clinic.

“Despite a detailed understanding of the genomics underlying AML pathogenesis, clinical recommendations for AML classification and risk stratification remain heavily reliant on cytogenetic findings and only consider a few well-characterized gene mutations,” the researchers, led by Yanis Tazi from Memorial Sloan Kettering Cancer Center, explained. “In the absence of informative markers, a significant proportion of AML patients are categorized as intermediate risk and display significant heterogeneity in clinical presentation and outcomes.”

Mr. Tazi and colleagues aimed to analyze genetic data from a large cohort of patients with AML to understand the role for acquired gene mutations in clinical decision support algorithms that guide disease classification, risk stratification, and treatment decisions.

In this study, the investigators mapped recurrent gene mutations across 128 genes that were implicated in myeloid pathogenesis in 2,113 patients with AML. Next, gene mutations were integrated with cytogenetic findings (fusion genes and chromosomal aneuploidies). They then performed unsupervised clustering to characterize novel molecular subgroups and their biological and clinical relevance.

The researchers looked specifically at the associations between mutations and measurable residual disease (MRD), overall survival, relapse risk, and response to hematopoietic cell transplantation (HCT). Findings were then validated in a cohort of 1,540 patients enrolled in AMLSG trials.

Of the 2,113 patients included in the analysis, 92% were assigned into one of 14 molecular classes that define distinct clinical and prognostic subgroups. Each of these subgroups, the authors noted, was associated with a distinct likelihood of response to induction chemotherapy, risk of relapse, and death over time.

These subgroups retained their independent prognostic relevance regardless of MRD status, the authors added, and were able to identify patients who would derive benefit from HCT at first remission.

The largest of these groups, representing 24% of study patients, was “secondary AML-2,” which the authors described as “a class enriched in secondary like mutations” that has a poor prognosis but high likelihood of benefiting from HCT.

Ultimately, the authors incorporated cytogenetic data and the 32 identified genes into an open-access online calculator that considers molecular, clinical, and demographic parameters to guide individualized patient management.

Study authors report no relevant conflicts of interest.


Tazi Y, Arango JE, Zhou Y, et al. A unified classification and risk stratification algorithm to support clinical decisions in acute myeloid leukemia. Abstract #S133. Presented at the EHA2021 Virtual Congress, June 9-17, 2021.