More than a fusion gene: the RUNX1-RUNX1T1 AML

In this issue of Blood, Christen et al investigated the largest cohort to date of 331 patients with acute myeloid leukemia (AML) and t(8;21).¹ 

These patients have AML with specific morphologic features such as dysplasia in granulopoiesis (90% of patients) and eosinophilia and are mostly classified as AML with maturation (90%; formally called French-American-British [FAB] M2) or AML without maturation (10%; formally called FAB M1).²  This subtype of AML is also diagnosed by immunophenotyping that shows the coexpression of CD19 or PAX5 and CD56. The cytogenetics show a typical pattern of loss of the sex chromosome and del9q. These characteristics resulted in RUNX1-RUNX1T1–mutated AML being designated as a separate World Health Organization (WHO) entity within the category of AML with recurrent genetic abnormalities. The diagnosis is made irrespective of bone marrow blast cell counts.³ RUNX1-RUNX1T1–mutated AML also demonstrates secondary cooperating mutations in KIT, KRAS or NRAS, and ASXL1 as well as in ASXL2.^4,5 RUNX1-RUNX1T1 was one of the first fusion genes to be used for minimal residual disease (MRD) monitoring.⁶  Based on these diagnostic definitions, the best clinical practice to follow after standard chemotherapy needed to be determined, including the meaningfulness of allogeneic transplantation in first complete molecular remission (CMR).^7,8 

Today, large sequencing studies including exome sequencing or whole-genome sequencing (WES) are possible. In their article, Christen et al provide a comprehensive characterization of this specific WHO entity in 331 patients based on a screening that included 66 recurrently mutated genes. They found that 95% of patients had at least 1 additional mutation, with a mean of 2.2 driver mutations per patient. Recurrently mutated genes affecting the RAS/RTE signaling pathway were present in nearly two-thirds of patients and other epigenetic regulators in nearly half the patients. Several previously unexpected genes were found to be mutated. Data using deep sequencing (45 000×) in 62 samples from patients in complete remission demonstrated persistent mutations in 12 samples, including 5 patients who were quantitative polymerase chain reaction–negative for RUNX1-RUNX1T1 at the time of the analysis. In multivariate analysis, JAK2, FLT3-ITD^high, and KIT^high were identified as significant negative prognostic factors. Furthermore, it was demonstrated that one-third of patients studied by WES both at diagnosis and at relapse were genetically unstable and did not fully reproduce the genetic landscape of the diagnostic sample at relapse.

Therefore, this comprehensive study clearly demonstrates that patients with AML and t(8;21) at diagnosis should, according to WHO gold standards, be studied by morphology, immunophenotyping, and cytogenetics. In addition, a molecular genetic screening at diagnosis not only for RUNX1-RUNX1T1 but also for a gene panel seems to be warranted. Furthermore, findings at relapse may have implications for prognosis and especially any targeted treatment. This may be of particular importance for patients in CMR for RUNX1-RUNX1T1 but with secondary mutations still detectable in low levels (see figure).

Because the capacity for panel sequencing and WES will increase rapidly worldwide over the next few years,⁹  AML with RUNX1-RUNX1T1 should be comprehensively investigated at diagnosis, during follow-up for MRD monitoring, and at relapse to individualize treatments, including targeted approaches toward driver genes. Genetics at relapse can hold additional important information. This study definitively sets the stage.