Researchers used comprehensive molecular profiling data to develop a new classification and risk-stratification system for acute myeloid leukemia (AML) that would encompass 100% of patients, according to an article in Nature Communications.
The framework includes 16 molecular subgroups of AML, classifies patients into three risk strata, “stratifies one in four [patients with AML], and achieves significant improvement in prognostic accuracy,” according to Yanis Tazi, a graduate student in the Tri-Institutional PhD Program at Cornell University, and colleagues.
Existing classification and risk-stratification is primarily reliant on cytogenetic findings, which are present in less than half of patients with AML. To more effectively classify patients, Mr. Tazi and colleagues analyzed data from 2,113 patients with AML enrolled in three of the United Kingdom’s National Cancer Research Institute trials. Using information on genetic alterations, clinical presentation, treatment response, and outcome, they developed a new framework for classification and risk stratification and validated it in an independent cohort.
Using a panel of 32 genes, the new classification assigned 92% of patients into one of 14 AML subgroups. The remaining patients were split into two classes: molecularly not otherwise specified (mNOS) or no events. In addition, the new framework refined established classes.
Among the broad findings of the updated framework is the identification of new clusters of prognostic relevance. For example, “patients with ≤2 aneuploidies (n=233, 11%), enriched for ‘MDS-related’ cytogenetic abnormalities clustered with secondary AML type mutations (sAML) such as SRSF2, SF3B1, U2AF1, ZRSR2, ASXL1, EZH2, BCOR, or STAG2, as well as novelly described here, RUNX1, SETBP1, and MLLPTD mutations.”
Adverse outcomes were specific to patients with two or more mutations as compared with those with one. Therefore, the researchers broke this group into sAML Like-1 (sAML1) for patients with single mutations and sAML Like-2 (sAML2) for patients with two or more class-defining genes.
sAML2 was identified as the second largest class, comprising 24% of patients. This class of patients was associated with high-risk disease, had poor prognosis irrespective of flow measurable residual disease negativity, and derived significant benefit from transplantation.
Mr. Tazi and colleagues also demonstrated the importance of negative molecular findings. Patients classified in the no events subgroup had favorable outcomes and were distinct from patients in the intermediate risk mNOS subgroup.
Using class membership, Mr. Tazi and colleagues created three risk strata: favorable, intermediate, and adverse. Patients with sAML1, trisomies, WT1, DNMT3A/IDH, and t(6;9) were considered intermediate risk, and sAML2 was considered adverse risk.
There were also broad implications of FLT3ITD positivity, which was associated with worse outcomes for all intermediate-risk patients and was used to upgrade risk to adverse-risk. In all, this system would re-stratify 25.5% of patients as compared with standard of care.
“Class membership provides resolution in the heterogeneity observed in clinical presentation and delivers a rationalized schema for correlative studies as compared to single biomarkers or clinical cutoffs (e.g., % blast counts), which may dichotomize or group together heterogeneous and biologically distinct nosological entities,” the researchers wrote.
The researchers acknowledged the complexity of this system and developed an open-access patient-tailored clinical decision support tool (aml-risk-model.com/calculator). Using mutations in 32 genes and cytogenetics as input variables, clinicians can assign patients to an AML class and risk group.
Any conflicts of interest declared by the authors can be found in the original article.
Reference
Tazi Y, Arango-Ossa JE, Zhou Y, et al. Unified classification and risk-stratification in acute myeloid leukemia [published online, 2022 Aug 8]. Nature Communications. doi: 10.1038/s41467-022-32103-8.
