Prognostic information is now available for patients 60 years or older with intensively treated AML and high-risk myelodysplastic syndromes (HRMDS) by the new AML60+ classification. The classification is easy to implement into routine clinical practice and offers critical information about survival following allogeneic hematopoietic cell transplantation (alloHCT) via a free online tool. The intensive chemotherapy and alloHCT benefits can be identified better by AML60+ than AML or phenotypic subsets, reported Jurjen Versluis, MD, PhD, and colleagues of Erasmus University Medical Center Cancer Institute in the Netherlands. The study results were published in the Journal of Clinical Oncology.
AML is most commonly diagnosed by hematologists in patients older than 60 years, and alloHCT is one of these patients’ most important curative consolidations. However, predicting which older AML patients will respond positively, or at all, to this potentially toxic treatment remains a challenge. Hematologists must carefully identify those who would benefit most. While devised using data from younger patients, the 2017 European LeukemiaNet (ELN2017) and ELN2022 classifications nevertheless segregate older patients with AML into Intermediate and adverse risk groups.
Dr. Versluis said, “This new model works even better than previous classification systems typographical errors in quotation marks for younger patients AML, and we hope this one will become the standard going forward for older patients AML.”
This particular study included 1,910 patients from two groups: NCRI-AML18 and HOVON-SAKK. The patients had a median age of 67 years (range: 60–84) and were treated over 12 years. The principal finding of the investigators was that low-risk features were common in the patients, including 57.3% ELN2022 adverse risk and 14.4% TP53 mutated. Two-thirds (65.3%) had MDS-related cytogenetic changes, mutations in secondary AML genes, or a combination of both, while in 9.1% of cases, secondary AML was clinician-diagnosed. The findings are plausible given its genetic heterogeneity in older patients.
To determine the importance of patient-specific genetic and cytogenetic variables for risk stratification of patients for overall survival (OS), the researchers used a machine learning random survival forest algorithm. Of the 9 variables identified, the presence of a TP53 mutation (hazard ratio (HR) = 2.42; 95% CI 1.83-3.21)), monosomal karyotype (HR = 2.06; 95% CI 1.56–2.7), and age over 65 years (HR = 1.50; 95% CI 1.31–1.72) were the three most important variables at diagnosis in predicting survival.
Finally, these 9 variables received points according to rounded HRs. TP53 was assigned three points if mutated, and one point if not; a monosomal karyotype two points; age 65 years three points; difficulty of cytogenetic reporting one point; white blood cell count >20 × 109/L, male sex, FLT3 internal tandem duplication and DNMT3A, ASXL1, and RUNX1 mutations one point each. The researchers and team used the predicted variables sequentially in a Cox regression examination for OS to define the optimal model, after they discovered a total of 9 variables and used them to construct four groups with highly distinct four-year OS: intermediate, favorable, poor, and very poor. The favorable–risk group was comprised of less adverse–risk Abstract Risk Group, consisting primarily of patients younger than 65 years old, with more favorable cytogenetics and/or molecular features, or both.
Having developed AML60+, the researchers next validated the classification in two cohorts: a validation AML cohort, and an HR-MDS test cohort. The AML development cohort was slightly younger than the AML validation cohort (66 vs. 68 years), they noted. Univariate analysis demonstrated strong prognostic separation of OS in the AML development, AML validation, and HR-MDS test cohorts by the new risk model. The investigators reported that AML60+ intermediate and very poor-risk patients had significantly better OS following alloHCT as consolidation treatment. However, the OS improvement for the poor-risk subgroup was only marginal. Based on these results, the authors concluded that AML60+ needs to be validated in other cohorts and refined as optimal treatment changes.
However, the authors point to the limitations of the study, which include its retrospective nature, across multiple patient cohorts treated over 12 years with varying intensive chemotherapy regimens. Future studies with MRD measurement are needed because their analysis didn’t account for measurable residual disease (MRD) after intensive or pre or post-alloHCT.
Reference: Versluis J, Metzner M, Wang A, et al. Risk Stratification in Older Intensively Treated Patients With AML. J Clin Oncol. 2024;42(34):4084-4094. doi:10.1200/JCO.23.02631
