RUNX1 Mutation Does Not Significantly Impact the Outcome of Newly Diagnosed adult AML: A Retrospective Study of Chinese AML Patients

Background

RUNX1 is one of the recurrent mutated genes in newly diagnosed acute myeloid leukemia (AML) with a frequency of 6%~15%. Although historically recognized as a provisional distinct entity, the AML subtype with RUNX1 mutations (AML-RUNX1^mut ) was eliminated from the 2022 WHO classification system. With the update of AML diagnosis and treatment protocol, different outcomes of AML-RUNX1^mut have been reported. However, an in-depth study about the prognostic value of RUNX1 mutations, epecifically the classification change related to AML-RUNX1^mut in a large well-annotated cohort is lacking.

Methods

The clinical data of 1065 AML (excluding acute promyelocytic leukemia and RUNX1 germline mutations) patients according to WHO 2016 and treated from January 2017 to December 2021 at Jiangsu Institute of Hematology were retrospectively collected and analyzed after informed consent. All AML diagnosis samples were analyzed by targeted Next Generation Sequencing (NGS) panel covering 172 frequently mutated genes in hematological malignancies. Demographic and clinical data, disease characteristics at diagnosis, first-line treatment and clinical outcome data were available for all patients. The survival time of allogeneic hematopoietic stem cell transplantation patients was censored at the time of stem cell transfusion. Cytogenetic and molecular characteristics were used to classify patients into ELN2017 and ELN2022 risk groups (excluding the effect of RUNX1 mutation).

Results

RUNX1 mutations were detected in 112 (10.5%) patients, namely, 66 (58.9%) male and 46 (41.1%) female. A total of 132 RUNX1 alterations were detected with a median variant allelic frequency of 40% (3%~93%). RUNX1 mutations were associated with older age (48.5 vs. 46.0 years old, P=0.087) and antecedent myeloid disorders (6.3% vs. 1.0%, P<0.001). One or more comutations were detected in 106 (94.7%) AML-RUNX1^mut patients. The most frequent co-gene abnormalities were FLT3-ITD and DNMT3A mutations which represented the 23.2% and 18.7%. Mutations of CEBPA bZIP, NPM1 and translocations of t(8;21), t(16;16)/inv(16) were uncommon in AML-RUNX1^mut with a frequency of 6.3%, 1.8%, 2.7%, 0%, respectively (Table 1).

Compared with AML with wild-type RUNX1 (AML-RUNX1^wt), AML-RUNX1^mut patients had a lower complete remission rate (19.6% vs. 29.7%, P=0.034), lower overall response rate (ORR) (73.5% vs. 84.3%, P=0.006) and higher no response rate (26.5% vs. 15.5%, P=0.021) in the first induction chemotherapy. When comparing groups based on the ELN2017 risk stratification criteria, no significant difference in ORR was observed between RUNX1^mut and RUNX1^wt patients in the favorable (91.2% vs. 94.0%, P=0.572), intermediate (75.0% vs. 78.4%, P=0.570) and adverse (63.3% vs. 71.4%, P=0.370 ) subgroups (Table 2).

The median duration of follow-up of the entire cohort was 8.7 months (RUNX1^wt: 8.4 months; RUNX1^mut: 10.0 months). Comparing the two groups, patients with RUNX1 mutations had a lower 5-year overall survival (OS) rate (35.4% vs. 42.3%; HR=1.563; 95% CI: 1.091~2.237, P=0.026), but no significant difference was observed in the 5-year event-free survival (EFS) rate (40.8% vs. 35.4%; HR=1.084; 95% CI: 0.750~1.570, P=0.098) (Table 2). However, when patients were stratified according to the ELN2017 guidelines (excluding the effect of RUNX1 mutation), outcomes did not differ between patients with or without RUNX1 mutation in the favorable (OS: P=0.629, EFS: P=0.137), intermediate (OS: P=0.629, EFS: P=0.114) and adverse (OS: P=0.544; EFS: P=0.714) subgroups. Similar results were obtained according to the ELN2022 guidelines. In multivariate analysis, RUNX1 was not an independent prognostic factor for OS (HR=1.353, 95% CI: 0.922~1.985, P=0.122) and EFS (HR=1.422, 95% CI: 0.975~2.075, P=0.068), while significantly worse OS and EFS were observed in patients with mutated FLT3-ITD, TP53, and DNMT3A (P-values were all less than 0.05).

Conclusion

RUNX1 mutations were rarely isolated genetic abnormities that conferred inferior outcomes. However, on the basis of ELN prognosis risk stratification, RUNX1 mutation had no impact on the clinical outcomes of AML patients in stratified subgroups. RUNX1 was not an independent prognostic factor for survival. Overall, our findings agree with the updated WHO classification system for AML that AML-RUNX1^mut should not be recognized as a distinct AML entity.