Frequency, Characteristics and Prognostic Significance of RUNX1 mutations in Patients with Acute Myeloid Leukemia, Not Otherwise Specified

Background: Somatic mutations in RUNX1 gene have been identified in a substantial proportion of patients with de novo acute myeloid leukemia (AML). It is suggested as a new candidate molecular marker and, therefore, is suggested to be routinely performed at the diagnostic stage of AML. Despite its clinical importance, however, previous cohorts have been heterogeneous in terms of cytogenetic and molecular subtypes of AML. Here, the aim of this study was to evaluate the frequency, biologic characteristics, and prognostic significance of RUNX1 mutations focusing on patients with AML, not otherwise specified (NOS).

Methods: Diagnostic samples from 202 patients with AML were analyzed for RUNX1 mutations. We excluded AML with recurrent genetic abnormalities, AML with myelodysplasia-related changes, and therapy-related AML because these entities have prognostic relevances of their own. RUNX1 mutations were detected using standard PCR techniques and direct sequencing.

Results: RUNX1 mutations were found in 27 (13.4%) patients. The mutations were clustered in Runt homology domain (13, 48.1%) and transactivation domain (9, 33.3%). Frameshift mutations were most common (52.9%), followed by missense mutations (35.3%) and nonsense mutations (11.8%). As shown in Table 1, patients with RUNX1 mutations had a lower platelet count (P = 0.03), a higher rate of trisomy 8 (P = 0.02) and trisomy 13 (P = 0.039), and a trend toward older age (P = 0.063) than patients without mutations. Presence of RUNX1 mutations and NPM1 or CEBPA mutations were mutually exclusive. At the median follow-up of 12.1 months, RUNX1 mutations predicted for shorter overall survival (OS; P = 0.007) and relapse-free survival (RFS; P = 0.003). In the multivariate analysis, RUNX1 mutation was a significant marker for inferior OS (hazard ratio, 3.037; P = 0.014) and RFS (hazard ratio, 5.699; P = 0.001).

Conclusion: The findings of our study further strengthen the previous data about RUNX1 mutations in AML. Furthermore, AML NOS with RUNX1 mutations is characterized by distinct biology and is associated with adverse clinical outcome. Our study supports the notion that RUNX1 mutational status would be integrated into diagnostic workup of AML, particularly for AML, NOS subgroup.

Table 1. Clinical and biologic features of the cohort by RUNX1 mutations

	RUNX1 mutations		P-value
	Mutated, n (%)	Wild type, n (%)
Number	27 (13.4)	175 (86.6)
Male sex	17 (63.0)	94 (53.7)	0.489
Median age, years (range)	63 (14 - 80)	55 (1 - 83)	0.063
WBC count, •109/L (median, range)	7.9 (1.1 – 133.3)	14.0 (0.8 – 231.3)	0.636
Hemoglobin, g/dL (median, range)	8.6 (5.0 – 10.6)	8.8 (4.1 – 17.3)	0.376
Platelet count, •109/L (median, range)	35 (14 – 230)	59 (9 – 900)	0.03
Blood blasts, % (median, range)	29 (0 – 94)	38.5 (0 – 93)	0.312
FAB subtypes
M0	3 (11.1)	11 (6.3)	0.609
M1, M2	21 (77.8)	129 (73.7)	0.831
M4, M5	3 (11.1)	27 (15.4)	0.767
M6, M7	0	8 (4.6)	0.546
Cytogenetic abnormalities
Normal karyotype (%)	11 (40.7)	104 (59.4)	0.106
Trisomy 8 (%)	5 (18.5)	8 (4.6)	0.02
Trisomy 11 (%)	1 (3.7)	4 (2.3)	0.823
Trisomy 13 (%)	3 (11.1)	3 (1.7)	0.039
Trisomy 21 (%)	1 (3.7)	2 (1.1)	0.866
Distribution of other mutations
FLT3-ITD	5 (18.5)	50 (28.6)	0.39
FLT3-TKD	1 (3.7)	4 (2.3)	0.823
NPM1	0	55 (31.4)	0.002
CEBPA	0	17 (9.7)	0.187
MLL-PTD	1 (3.7)	14 (8.0)	0.691