According to results from a deep-sequencing analysis published in Nature Medicine, women with mutations of the TP53 and IDH genes in the absence of a known myeloid malignancy have a high risk for developing acute myeloid leukemia (AML) within their lifetimes. Women with spliceosome gene mutations and co-mutations of the RUNX1 and PHF6 genes, in particular, have a rapid time to AML diagnosis, compared with individuals without these mutations. Together, these mutational profiles may represent a premalignant state of AML that exists nearly a decade prior to AML diagnosis, the authors observed.
“We believe that these data will help identify patients [who] need to be followed and monitored closely based on the particular mutation they might have,” lead author Pinkal Desai, MD, of Weill Cornell Medicine in New York, told ASH Clinical News. “Clonal hematopoiesis (CH) is present in 10 to 25 percent of typical, healthy populations and many patients with CH are being identified now, but with few data about how this affects their risk of AML.”
To define “the premalignant mutational landscape” of AML, the investigators performed deep sequencing of peripheral blood samples serially collected from women participating in the Women’s Health Initiative (WHI), a large, prospective clinical trial and observational study sponsored by the National Institutes of Health.
The WHI began enrollment in 1991 and included more than 160,000 women. The present analysis included 212 women who were healthy at study baseline but developed pathologically confirmed AML during follow-up. Results from the deep-sequencing analysis among patients with AML were then compared with those from samples of 212 age-matched controls without an AML diagnosis. Participants were excluded from the analysis if they had a history of any hematologic disorder at baseline.
In the AML cohort, samples were collected from women a median of 9.6 years (range not provided) before diagnosis of AML.
The median number of mutated genes in each cohort was one (range = 0-8 mutations) and zero (range = 0-2), respectively.
The gene mutations that appeared most frequently in AML cases, compared with controls, included:
- DNMT3A (36.7% vs. 18.8%, respectively)
- TET2 (25.0% vs. 5.5%)
- TP53 (11.2% vs. 0%)
- SRSF2 (6.9% vs. 0%)
- IDH2 (6.4% vs. 0%)
- SF3B1 (5.9% vs. 1.1%)
- JAK2 (5.3% vs. 0.6%)
- ASXL1 (3.2% vs. 3.3%)
No p values were reported for these comparisons.
The authors reported that patients with AML were nearly five times as likely to present with mutations at baseline, compared with women in the age-matched cohort controls (129 [68.6%] vs. 56 [30.9%], respectively; odds ratio [OR] = 4.86; 95% CI 3.07- 7.77; p<0.001).This association was independent of age: Among women <65 years, those with AML had an OR for harboring mutations of 4.39 (95% CI 2.08- 9.61); among women ≥65 years, the OR was 6.19 (95% CI 3.25- 12.14; no p values reported).
The risk of developing AML also increased with every incremental increase in number of mutations (OR=3.27; 95% CI 2.47-4.45; p value not reported).
“[Patients] with AML demonstrated greater clonal complexity than controls,” the investigators noted, with 46.8 percent of patients in the AML cohort demonstrating co-mutations, compared with 5.5 percent of controls (OR=9.01; 95% CI 4.1-21.4; p<0.001). The authors also identified common co-occurrence patterns in patients with AML, including mutations in the following genes:
- DNMT3A with TET2
- DNMT3A with SRSF2
- TET2 with SRSF2
- IDH2 with SRSF2
Among the recurrently mutated genes, certain mutations demonstrated “increased specificity and penetrance for the development of AML,” the researchers wrote. For example, all participants with TP53 and IDH1/IDH2 mutations, as well as those with RUNX1 with PHF6 co-mutations, eventually developed AML. Results from a multivariable analysis that controlled for potential confounders like co-mutations and age also showed that the following mutations were strongly associated with the development of AML (though p values were not reported):
- TP53 (OR=47.2; 95% CI 2.5-879.1)
- IDH1/IDH2 (OR=28.5; 95% CI 1.4-562.8)
- SF3B1/SRSF2/U2AF1 spliceosome genes (OR=7.4; 95% CI 1.7-32.2)
- TET2 (OR=5.8; 95% CI 2.6- 12.9)
- DNMT3A (OR=2.6; 95% CI 1.5-4.5)
The presence of baseline mutations appeared to predict for a shorter time to AML diagnosis, shifting the median time to diagnosis from 11.9 years in patients with no baseline mutations to 8.2 years in patients with a baseline mutation (p<0.001; ranges not reported). TP53 mutation and co-mutations of DNMT3A and spliceosome genes appeared to be associated with a more rapid onset of AML: The ORs for developing AML within five years of baseline were 5.2 (95% CI 1.9-14.5; p=0.001) and 14.8 (95% CI 1.6-136; p=0.02).
According to Dr. Desai, being able to identify mutations likely to lead to AML up to 10 years prior to the actual diagnosis “provides hope that prevention strategies for AML can be envisioned. More research and prospective trials of prevention, as well as intervention strategies, is the next step.”
Because the study included only female patients, its findings may not be generalizable to the male patient population. Additional limitations of the study included the lack of sequencing assessments performed in blood and/or bone marrow at diagnosis of AML. Dr. Desai noted that the authors “were unable to correlate the mutations discovered at baseline with the mutations that may [have been] present at the time AML was diagnosed,” which is a future direction of research.
“Obviously, a mass screening of the healthy population is not recommended at this time,” Dr. Desai advised. “We are working on developing an economically feasible and psychologically sensitive way to identify and intervene in the right population.”
The corresponding authors report financial relationships with the Sandra and Edward Meyer Cancer Center, which provided financial support for the study.
Desai P, Mencia-Trinchant N, Savenkov O, et al. Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med. 2018;24:1015-23.