ORLANDO – A genome-wide study of blood and bone marrow samples from more than 1,300 adults with myeloid disorders has both confirmed the role of known or suspected driver mutations and uncovered new associations that could inform clinical care for patients with acute myeloid leukemia and myelodysplastic syndrome.
“Integration of mutational and expression data is important to refine subytpes and constellations of mutations with prognostic significance,”, of St. Jude Children’s Research Hospital in Memphis said during a late-breaking abstract session at the annual meeting of the American Society of Hematology.
Her team conducted an analysis combining full genomic sequencing and gene-expression profiles in blood and bone marrow samples from 598 adults with acute myeloid leukemia (AML) and 706 with myelodysplastic syndrome (MDS).
The goals of the study were to provide “unbiased analysis of AML and MDS by integrated genomic and transcriptome data and clinico-pathologic features and clinical outcome” and to identify and define myeloid leukemia subtypes with diagnostic, prognostic, and therapeutic significance, she said.
The median age of the MDS cohort was 73.2 years (range 23.3-93.1). According to 2016 World Health Organization criteria, 37% had a diagnosis of MDS with excess blasts, 26.3% had MDS with ring sideroblasts, 20.9% had MDS with multilineage dysplasia, 14.6% had MDS with deletion 5q, and 1.1% had unclassifiable MDS.
The median age of the AML cohort was 68 years. Of this group, 31.7% had a diagnosis of AML not otherwise specified, 29.9% had known cytogenetic alterations, 27.3% had NPM1-mutated AML, and 9.7% had RUNX1-mutated disease.
Samples from all patients underwent tumor whole-genome sequencing and whole-transcriptome sequencing.
The combined sequencing confirmed a diagnosis of AML with recurrent genetic abnormalities in 11% of cases. These patients had disease with distinct gene-expression profiles and favorable prognosis. The sequencing identified combinations of mutations in genes linked with specific AML subtypes.
For example, combinations of mutations in KIT, ZBTB7A, ASXL2, RAD21, CSF3R, and DNM2 were associated with RUNX1-RUNXT1 leukemia, whereas mutations in FLT3, DDX54, WT1, and CALR in promyelocytic leukemia/retinoic acid receptor alpha were associated with promyelocytic leukemia, and KIT and BCORL1 mutations were associated with CBFB-rearranged leukemia.
In addition to rounding up the usual genomic suspects, the investigators also identified combinations that are associated with prognosis. Notably, NPM1 mutations were found in 27.4% of AML and 1% of MDS cases, and these mutations were characterized by four gene-expression signatures that were associated with different combinations of cooperating mutations in cohesin and signaling genes, and with outcome.
They found that patients with co-occurring NPM1 and FLT3 mutations had worse prognosis than those with mutations only in NPM1, whereas patients with NPM1 mutations co-occurring with cohesin gene mutations had better outcomes.
At a briefing prior to her presentation of the data, Dr. Iacobucci explained how her group’s findings might inform treatment, including the possibility of preventing development of AML in patients with MDS.
“What we are doing, in addition to the genomic part, is also establishing a repository of patient-derived xenografts, so in this way we can have the genome information, and we can have the biological material in vivo to test different therapies,” she said.
In an interview,, from the Alfred Hospital in Melbourne, who was not involved in the genomic study, commented on the role of sequencing in treatment of patients with myeloid malignancies.
“I think the future is that as the leukemia evolves, our therapy will evolve along with it. Furthermore, we now have the potential to measure many of these mutations with much higher sensitivity than just whole-genome sequencing, so we can imagine a future whereby we can track and measure these mutations as they rise in the patient’s bone marrow or blood before the patients becomes sick with florid leukemia, and it gives us the potential to predictably alter our management before they become sick,” he said.
The study was supported by St. Jude Children’s Research Hospital and the Leukemia and Lymphoma Society. Dr. Iacobucci and Dr. Wei reported having no relevant disclosures.
SOURCE: Iacobucci I et al. ASH 2019, Abstract LBA-4.