The accumulation of somatic mutations in several genes precedes the onset of acute myeloid leukaemia (AML). Two studies now indicate that such alterations can be detected years before diagnosis — a latency period during which monitoring of patients at risk could be improved.

Blood samples collected at the time of enrolment in large population cohorts with long follow-up durations were used in both studies. In a study published in Nature, Abelson and collaborators analysed samples from EPIC trial participants diagnosed with AML at an average of 6.3 years after enrolment (n = 95) or not diagnosed at an average of 11.6 years after enrolment (control group, n = 414). In a study published in Nature Medicine, Pinkal Desai et al. analysed samples from Women’s Health Initiative participants with or without a diagnosis of AML (n = 212 in each group; median time to diagnosis 9.6 years after enrolment).

In both studies, peripheral blood cells from individuals diagnosed with AML were more likely to harbour mutations than those from individuals in the control group; DNMT3A, TET2, and TP53 were the most frequently mutated genes in both cohorts. Alterations in several other genes were found to contribute to the risk of developing AML.

“There is a clear early genetic signal, but further refinement is needed to reliably predict a rare disease such as AML in the general population,” explains Moritz Gerstung, involved in the study published in Nature. Using a machine-learning approach, these investigators analysed non-genetic information in electronic health records from 4,500 individuals in the Clalit database. They developed a prediction model with which they were able to predict AML 6–12 months before diagnosis with 25.7% sensitivity and 98.2% specificity.

“We have to improve the accuracy of these prediction markers using additional genomic techniques (such as methylation panels) before testing clinical interventions,” clarifies Paul Brennan, involved in the Nature study. For both Gerstung and Desai, subsequent studies would involve prospective trials of potential early therapeutic interventions (such as targeted therapies) in large cohorts of individuals with a high risk of AML.

“It’s exciting that the mutation patterns leading to AML were largely similar between studies and we both identified a window of opportunity preceding an otherwise acute disease,” summarizes Duane Hassane, principal investigator of the Nature Medicine study. “This accelerates the momentum for further detection, monitoring, and prevention studies in a way that a single study could not,” he concludes.

we both identified a window of opportunity preceding an otherwise acute disease