Acute myeloid leukemia (AML) can occur in both children and adults. However, the genetic drivers of disease can differ depending on the patient’s age and knowing this information may inform on personalized therapeutic plans for patients. Dr. Soheil Meshinchi, a Professor in the Translational Science and Therapeutics Division at Fred Hutchinson Cancer Center, his lab, and colleagues wanted to investigate age-dependent differences in genetic drivers of AML disease to continue the advancement of AML patient care. Their findings were published recently in HemaSphere.
The researchers collected data from more than 3,000 AML patients from several sources (German Study Alliance Leukemia, Fred Hutchinson Cancer Research Center, National Cancer Institute TARGET data, and another publicly-available dataset). To uncover age-dependent genetic alterations, the Meshinchi lab stratified patients into six age groups: infants, children, adolescents and young adults, adults, seniors, and elderly. To detect mutations common to AML, cells from the bone marrow or peripheral blood of AML patients were analyzed for genetic alterations in 54 different gene targets. To help grapple with this large dataset, the researchers next employed machine learning approaches to deconvolute the multiple comparisons for each genetic alteration per person and identify trends of those genetic alterations that are based on age. The complexity of this dataset would be challenging to decipher without machine learning modeling due to the heterogeneity of the genetic variance across the dataset of more than 3,000 AML patients. Rhonda Ries, the Genomics Research Manager in the Meshinchi lab, shared her insights on their findings: “This work underscores how much the heterogeneity of AML, in terms of each individual's molecular alterations, impacts outcome. We and others have shown this previously, but this study combines data across all age groups from multiple institutions and uses machine learning approaches to predict and correlate remission status and overall survival with molecular mutations and alterations.”
Their model of the data “categorizes the molecular alterations by gene function (like signaling pathways, transcription factors, etc) amongst the 6 age groups,” explained Ries. The researchers uncovered more frequent alterations in signaling pathways and transcription factors in younger patients (infants, children, adolescents and young adults) than the older patient populations. On the opposing end, elderly patients were more likely to have genetic mutations in epigenetic regulators or splicing factors than the younger patients. This level of detail has not been seen in previous datasets—especially those that only separated patients into two groups, over or under 65 years of age.
“Since we [the Meshinchi lab] are focused on studying childhood AML, the largest proportion of alterations fall in signaling and transcription factors genes, and this continues to be an area of interest for identifying targeted therapies to inhibit these aberrant processes,” shared Ries. “Combining an individual's molecular abnormalities for risk stratification and treatment decision-making remains an active area of interest for us. We've built the largest childhood AML database of transcriptome and molecular data for patients worldwide and we continue to expand the database with more patients. Our hope is that by expanding the database of this molecularly diverse patient population, we will gain new insights into finding therapeutic targets for more personalized care.”
The highlighted work was funded by a Mildred Scheel Foundation fellowship.
Fred Hutch/University of Washington/Seattle Children's Cancer Consortium member Dr. Soheil Meshinchi contributed to this work.
Eckardt JN, Hahn W, Ries RE, Chrost SD, Winter S, Stasik S, Röllig C, Platzbecker U, Müller-Tidow C, Serve H, Baldus CD, Schliemann C, Schäfer-Eckart K, Hanoun M, Kaufmann M, Burchert A, Schetelig J, Bornhäuser M, Wolfien M, Meshinchi S, Thiede C, Middeke JM. 2025. Age-stratified machine learning identifies divergent prognostic significance of molecular alterations in AML. HemaSphere. 9(5):e70132.
Science spotlight writer Annabel Olson is a postdoctoral research fellow in the Nabet lab at Fred Hutchinson Cancer Center. Her research focuses on studying the mechanisms that drive cancer development for both genetic and virus-associated cancers. A key tool in her research is the use of targeted protein degradation to dissect dysregulated signaling pathways in cancer and to double as a relevant pre-clinical therapeutic platform.
For the Media