Comparison of mutational profiles and clinical outcomes in patients with acute myeloid leukemia with mutated RUNX1 versus acute myeloid leukemia with myelodysplasia-related changes with mutated RUNX1

Abstract

Studies comparing the prognostic role of RUNX1 mutations (RUNX1^mut) in acute myeloid leukemia (AML) and acute myeloid leukemia-with myelodysplasia-related changes (AML-MRC) are limited. Our study examines the genetic profile of 118 RUNX1^mut AML patients including 57 AML with RUNX1^mut and 61 AML-MRC with RUNX1^mut and 100 AML, NOS patients with wild type RUNX1 (RUNX1^wt). Results revealed that AML-MRC patients with RUNX1^mut had shorter median overall survival (OS) (11 ± 3.3 months) when compared to AML with RUNX1^mut (19 ± 7.1 months) and AML, NOS with RUNX1^wt (not reached) (p = .001). The most common concurrent mutations observed in AML-MRC with RUNX1^mut patients were DNMT3A, SRSF2, ASXL1, and IDH2 while in AML with RUNX1^mut patients were ASXL1, SRSF2, TET2, IDH2, and DNMT3A. ASXL1 and TET2 mutations appeared to adversely affect OS in AML-MRC, but not in AML with RUNX1^mut. Concurrent RUNX1/DNMT3A mutations, in contrast had negative impact on OS in AML with RUNX1^mut, but not in AML-MRC with RUNX1^mut.

Introduction

Acute myeloid leukemia (AML) is a clonal proliferation of immature myeloid precursors with a variety of clinical and immunophenotypic presentations depending on the inherited and acquired genetic aberrations [1,2]. AML is an aggressive disease with a 5-year overall survival (OS) rate of 27.4% according to the National Cancer Institute, Surveillance, Epidemiology, and End Results Program (SEER) [3]. Emerging evidence shows that the clinical prognosis of AML largely depends on several key factors including the patient’s age, coexisting comorbidities, and the precise disease sub-classification based on morphology, cytogenetics, and molecular signatures as identified by the World Health Organization (WHO) [1,4,5]. In addition to recurrent chromosomal aberrations, somatic mutations in RUNX1, NPM1, FLT3, TP53, KMT2A/MLL, and CEBPA genes are critical in risk stratification of AML patients [2,6].

As one of the key transcriptional factors, RUNX1 plays an important role in the regulation of normal hematopoietic stem cells differentiation [7]. Therefore, translocation or intragenic mutation of RUNX1 leads to impaired hematopoiesis and ultimately leukemia [8–11]. Two large cohort studies have shown that somatic RUNX1 mutations (RUNX1^mut) are frequently identified in myelodysplastic syndromes (MDS) and therapy-related AML (t-AML) [12,13]. Additionally, approximately 10% of patients with de novo AML, not otherwise specified (AML-NOS), were shown to have RUNX1^mut, particularly patients with intermediate-risk and/or non-complex cytogenetics [11,14,15]. Furthermore, AML patients with RUNX1^mut are shown to be more resistant to induction chemotherapies and associated with inferior survival outcomes when compared to AML, NOS with RUNX1^wt [1,11,14].

AML with RUNX1^mut is a new provisional entity in the 2017 WHO classification of Tumors of Hematopoietic and Lymphoid Tissue [5,16], distinguishing it from AML with recurrent genetic abnormalities, AML with myelodysplasia-related changes (AML-MRC) and t-AML [4,16]. However, studies comparing the clinical impact of RUNX1^mut and its associated genetic landscape in AML-MRC and AML with RUNX1^mut are very limited. This is clinically important for accurate risk stratification and personalized treatment approach for AML patients. Therefore, our study aimed to investigate the clinical outcomes in AML with RUNX1^mut and AML-MRC with RUNX1^mut and identify potential cooperative mutations and cytogenetic abnormalities that may have prognostic significance in these specific subgroups of AML.

Methods and material

Patient population

This study was approved by the Institutional Review Board (IRB) and Scientific Review Committee of H. Lee Moffitt Cancer Center and Research Institute (Moffitt). Between 2012 and 2017, a total of 118 AML patients harboring RUNX1^mut were detected by next-generation sequencing (NGS) using the Illumina Trusight myeloid sequencing panel (54 genes). The patients with RUNX1^mut were further categorized as AML with RUNX1^mut (n = 57) or AML-MRC with RUNX1^mut (n = 61) (Table 1). A total of 100 AML, NOS patients with RUNX1^wt were included as a control group after matching for age, sex, bone marrow (BM) biopsy, and NGS test period. Clinical parameters including diagnosis, complete blood count (CBC), molecular data, and treatment regimen were collected. t-AML and AML with recurrent genetic aberrations beyond RUNX1 (PML-RARA, RUNX1/RUNX1T1, CBFB-MYH11, BCR-ABL1 etc.) were excluded. Patients with known congenital hematological malignancies were also excluded. Germline studies for RUNX1 mutation were not performed in this study.

Table 1.

Patient demographics and frequency of genetic aberrations in patients with AML with RUNX1^mut, AML-MRC with RUNX1^mut, and AML, NOS with RUNX1^wt.

Characteristics	AML with RUNX1mut (n = 57)	AML-MRC with RUNX1mut (n = 61)	AML, NOS with RUNX1wt (n = 100)	p-Value
Median age (range, years)	67 (21–86)	72 (31–92)	60.9 (19–85)	>.05
Gender (Male: Female)	1.85	2.21	1.16	>.05
CBC data (median)
WBC, × 109/L (range)	4.5 (0.6–103.2)	3.8 (0.7–134.1)	7.1 (0.9–248)	N/A
ANC, × 109/L (range)	0.9 (0–20.7)	2.1 (0–23.8)	1.0 (0–24.9)
Hemoglobin, g/dL (range)	8.5 (6.6–12.2)	8.7 (7.4–12.1)	9.1 (4.5–13.2)
Platelets, × 109/L (range)	41 (0.5–136)	43 (2–597)	56 (8–351)
Prior MDS, MDS/MPN, MPN	20.8 (18–95.5)	22 (18–59)	50.5 (13.2–94)
Prior t-MDS	0 (0%)	42 (68.8%)	0 (0%)
Cytogenetics	(n = 51)	(n = 60)	(n = 98)
Normal karyotype	28 (54.9%)	25 (41.7%)	60 (61.2%)	.05639
<3 abnormalities	15 (29.4%)	27 (45%)	25(25.5%)	.03496*
≥3 abnormalities	NA	11(18.3%)	NA	NA
−5/del(5q)	0 (0%)	4 (6.7%)	0 (0%)	N/A
−7/del(7q)	0 (0%)	8 (13.3%)	5 (5.1%)	.0072*
+8	6 (11.8%)	7 (11.7%)	3 (3.1%)	.063618
+13	3 (5.9%)	2 (3.3%)	1 (1.0%)	.676687
−17/del(17p)	0 (0%)	1 (1.7%)	1 (1.0%)	N/A
−20/del(20q)	0 (0%)	5 (8.3%)	4 (4.1%)	N/A
Somatic mutations	(n = 57)	(n = 61)	(n = 100)
ASXL1	18 (31.6%)	23 (37.7%)	18 (18%)	.016166*
DNMT3A	19 (33.3%)	11 (18.0%)	29 (29%)	.146216
IDH2	10 (17.5%)	11 (18.0%)	12 (12%)	.49133
SRSF2	16 (28.1%)	18 (29.5%)	4 (4%)	<.00001*
TET2	5 (8.8%)	16 (26.2%)	20 (20%)	.048484*
BCOR	5 (8.8%)	2 (3.3%)	5 (5%)	.406604
BCORL	4 (7.0%)	3 (4.9%)	3 (3%)	.506597
CBL	0 (0%)	4 (6.6%)	2 (2%)	N/A
IDH1	7 (12.3%)	6 (9.8%)	9 (9%)	.803816
ETV6	6 (10.5%)	4 (6.6%)	0 (0%)	N/A
EZH2	3 (5.3%)	3 (4.9%)	0 (0%)	N/A
FLT3	7 (12.3%)	0 (%)	1 (1%)	N/A
GATA2	3 (5.3%)	1 (1.6%)	1 (1%)	.21159
KMT2A	2 (3.5%)	1 (1.6%)	2 (2%)	.692851
NRAS	7 (12.3%)	4 (6.6%)	8 (8%)	.555879
PHF6	3 (5.3%)	4 (6.6%)	3 (3%)	.555413
SF3B1	0 (0%)	4 (6.6%)	0 (%)	N/A
SETBP	2 (3.5%)	0 (0%)	2 (2%)	N/A
STAG2	2 (3.5%)	2 (2.3%)	0 (0%)	N/A
TP53	3 (5.35)	4 (6.6%)	1 (1%)	.144823
U2AF1	6 (10.5%)	6 (9.83%)	9 (9%)	.950672
WT1	3 (5.3%)	0 (0%)	2 (2%)	N/A
ZRSR2	1 (1.76%)	3 (4.9%)	1 (1%)	.474143
Othersa	1 (1.8%)	1 (1.6%) or less	1 (1%) or less	.90689
Number of mutations
<3 mutations	17 (29.8%)	12 (19.6%)	50 (59.5%)	.000267*
≥3 mutations	40 (70.2%)	49 (80.3%)	34 (40.5%)	<.00001 *
No mutations	–	–	16	N/A
Treatment
Induction chemotherapy	25 (43.9%)	10 (16.3%)	44 (44%)	.000733*
Hypomethylation agentsb	7 (12.2%)	23 (37.7%)	11 (11%)	.000049*
Allo-SCT	15 (26.3%)	13 (21.3%)	29 (29%)	.559621
Other	10 (17.5%)	15 (24.5%)	16 (16%)	.384509

Allo-SCT: allogeneic hematopoietic stem cell transplant; ANC: absolute neutrophil count; CBC: complete blood count; MDS: myelodysplastic syndromes; MPN: myeloproliferative neoplasm; t-MDS: therapy-related MDS; WBC: white blood cells. AML: acute myeloid leukemia; AML-MRC: acute myeloid leukemia with myelodysplasia-related changes; NOS: not otherwise specified.

^aOther somatic mutations include CSF3R, CEBPA, JAK2, KIT, ZRSR2, KDM6A, PDGFR, NOTCH1, ABL1, TRX, FBXW7, and RAD21.

^bHypomethylation agents include azacitidine and decitabine.

^*Indicates statistical significance.

Conventional karyotyping/fluorescence in situ hybridization (FISH)

Routine cytogenetic analysis using standard trypsin-Giemsa banding technique by Laboratory Corporation of America (Burlington, NC, USA) was performed on all patients with sufficient BM aspirate specimen and reported in accordance with an International System for Human Cytogenetic Nomenclature, 2016 (ISCH 2016, Karger). FISH probe sets designed for MDS [del(5q)/−5, del(7q)/−7, 8, del(17p)/−17, and del(20q)/−20] and AML [t(8;21)/þRUNX1-RUNX1T1, t(15;17)/PML-RARA, and inv(16)/CBFB-MYH11] were conducted based on the manufacturer’s instructions (Vysis, Downers Grove, IL, USA) at Moffitt Cancer Center.

NGS and mutation analysis

NGS was performed as previously described [17]. The panel used in our study encompasses a total of 54 genes including ABL1, ASXL1, ATRX, BCOR, BCORL1, BRAF, CALR, CBL, CBLB, CBLC, CDKN2A, CEBPA, CSF3R, CUX1, DNMT3A, ETV6/TEL, EZH2, FBXW7, FLT3, GATA1, GATA2, GNAS, HRAS, IDH1, IDH2, IKZF1, JAK2, JAK3, KDM6A, KIT, KRAS, MLL, MPL, MYD88, NOTCH1, NPM1, NRAS, PDGRA, PHF6, PTEN, PTPN11, RAD21, RUNX1, SETBP1, SF3B1, SMC1A, SMC3, SRSF2, STAG2, TET2, TP53, U2AF1, WT1, and ZRSR2. The filter setting for variant detection was set to identify any somatic, non-synonymous DNA mutations of single nucleotide variants (SNVs), insertions, and deletions. Fifty nanograms of DNA were extracted using the Qiagen AllPrep DNA extraction kit (Qiagen, Germantown, MD, USA) and quantitated by NanoDrop spectrophotometry (Agilent, Santa Clara, CA, USA). The samples were sequenced using an Illumina NextSeq 500 sequencer (Illumina, San Diego, CA, USA), with variant detection by the PierianDx bioinformatics pipeline. The variant filters were set at 5% variant allele frequency (VAF) for SNV and at 10% for insertions/deletions (indels). Any variants that met the threshold of the filters were automatically placed into a high confidence group. A second low-confidence group was established for specimens that did not meet the desired filter thresholds. In these cases, individual variants were manually interrogated by an integrated genomic viewer (IGV). If a mutation was clinically significant, it was then manually added to the final molecular pathology report. All RUNX1 mutations described were categorized as levels 1–4 variants excluding known germline polymorphisms or level 5 variants. Given that all the patients developed adult-onset AML, no additional tissue was harvested to test for RUNX1 germline mutation.

Statistical analysis

The Chi-square or Fisher exact test was used for continuous variables and Chi-squared test for categorical variables. Both tests encompassed baseline demographic data, laboratory, and BM biopsy findings and disease features. Kaplan–Meier method was used for estimated overall survival (OS) analysis and log-rank test was used for comparison of median OS. The OS was calculated from the time of AML diagnosis.

Results

Clinical characteristics of AML with RUNX1^mut and AML-MRC with RUNX1^mut

Of the 118 patients harboring RUNX1^mut, 57 patients (48.3%) were classified as AML with RUNX1^mut and the remaining 61 patients (51.7%) were categorized as AML-MRC with RUNX1^mut according to 2017 WHO classification. Patients with AML with RUNX1^mut had a median age of 67 years (range, 21–86) and were predominantly male (M:F ratio 1.85) while patients with AML-MRC with RUNX1^mut had a median age of 72 years (range, 31–92) also showing male predominance (M:F ratio 2.21). Patients with AML, NOS with RUNX1^wt had a median age of 60.9 years (range, 19–85) with a M:F ratio of 1.16. Table 1 summarizes the clinicopathologic features of these three cohorts. Among 57 patients with AML with RUNX1^mut, missense mutations constituted 38.6% RUNX1^mut followed by frameshift (36.8%) and nonsense (24.6%) mutations. Among 61 AML-MRC patients with RUNX1^mut, 45.9% had missense mutations followed by frameshift (41.0%) and nonsense (13.1%) mutations. These numbers were not statistically different between these two groups (p>.05). In this study, patients with AML with RUNX1^mut or AML, NOS with RUNX1^wt were more likely to receive induction chemotherapy than their AML-MRC with RUNX1^mut counterparts. Approximately 20–30% of AML patients from each cohort subsequently received allogeneic stem cell transplantation (Allo-SCT) (Table 1).

Concurrent mutations in AML with RUNX1^mut and AML-MRC with RUNX1^mut

The most frequent mutations in AML with RUNX1^mut were DNMT3A (33.3%), ASXL1 (31.6%), SRSF2 (28.1%), and IDH2 (17.5%); while in AML-MRC with RUNX1^mut the most common mutations were ASXL1 (37.7%), SRSF2 (29.5%), TET2 (26.2%), IDH2 (18%), and DNMT3A (18%). Other mutations were also identified in these cohorts with VAF ranging from less than 1 to 10% (Table 1). Of note, FLT3-ITD were more commonly seen in the patients with AML with RUNX1^mut as compared to AML, NOS with RUNX1^wt (7/62 versus 1/100, p = .003).

Clinical outcomes in AML with RUNX1^mut and AML-MRC with RUNX1^mut

AML-MRC patients with RUNX1^mut demonstrated shorter median OS (18.9 ± 1.9 months) when compared to patients with AML with RUNX1^mut (25.1 ± 2.8 months) or RUNX1^wt AML (39.5 ± 4.0 months) (p = .001) (Figure 1). When RUNX1^mut patients were stratified by concurrent gene mutations, TET2 appeared to adversely affect OS in AML-MRC with RUNX1^mut (RUNX1, 21.9 ± 3.0 months versus RUNX1/TET2, 9.7 ± 2.2 months, p = .009) (Figure 2(A)), but not in AML with RUNX1 mut (RUNX1, 26.3 ± 3.1 months versus RUNX1/TET2, 12.4 ± 3.2 months, p = .123) (Figure 2(B)). Concurrent DNMT3A mutation also had a negative impact on OS in AML with RUNX1^mut (RUNX1, 29.8 ± 3.5 months versus RUNX1/DNMT3A, 14.7 ± 3.5 months, p = .002), but not in AML-MRC with RUNX1^mut (RUNX1, 20.7 ± 2.9 months versus RUNX1/DNMT3A, 12.3 ± 2.1 months, p = .099) (Figure 3(A,B)). Concurrent IDH1 mutation was associated with worse clinical outcome in AML-MRC with RUNX1^mut patients (RUNX1, 19.5 ± 2.5 months versus RUNX1/IDH1, 2.6 ± 1.4 months, p = .008), but it did not affect OS in patients with AML with RUNX1^mut (RUNX1, 26.4 ± 3.1 months versus RUNX1/IDH1, 15.6 ± 7.8 months, p = .168); however, there were only seven patients with IDH1 in AML-MRC and seven patients in AML with RUNX1^mut cohorts. The combination of RUNX1/ASXL1 mutations did show a slightly worse statistically significant clinical outcome in AML patients with RUNX1^mut (RUNX1, 27.8 ± 3.6 months versus RUNX1/ASXL1, 19.4 ± 4.2 months, p = .203) (Figure 4(A)). There was a difference in clinical outcome in AML-MRC patients with RUNX1^mut (RUNX1, 20.3 ± 3.3 months versus RUNX1/ASXL1, 16.8 ± 3.5 months, p = .51) (Figure 4(B)), but this did not reach statistical significance. Other concurrent gene mutations (ETV6, EZH2, FLT3, IDH2, NRAS, SRSF2, TP53, and U2AF1, etc.) showed no effect on OS in AML with RUNX1^mut or AML-MRC with RUNX1^mut cohorts. Additionally, the presence of three or more gene mutations (including RUNX1) significantly decreased OS in patients with AML with RUNX1^mut (19.5 ± 2.9 months, versus 36.3 ± 4.2 months p = .004) (Figure 5(A)); while having no significant impact on OS in patients with AML-MRC with RUNX1^mut (17.8 ± 2.4 months versus 24.9 ± 7.5 months, p = .508) (Figure 5(B)).

Figure 1.

Overall survival curves for cases of AML with RUNX1^mut, AML-MRC with RUNX1^mut and AML, NOS with RUN1X^wt.

Figure 2.

(A) Overall survival curves for cases of AML-MRC with concurrent RUNX1/TET2 mutation compared to AML-MRC cases with RUNX1 mutation alone. (B) Overall survival curves for cases of AML with RUNX1^mut with or without concurrent TET2 mutation.

Figure 3.

(A) Overall survival curves for cases of AML with RUNX1^mut with or without concurrent DNMT3A mutation. (B) Overall survival curves for cases of AML-MRC with concurrent RUNX1/DNMT3A mutation compared to cases with RUNX1 mutation alone.

Figure 4.

(A) Overall survival curves for cases of AML with RUNX1^mut with or without concurrent ASXL1 mutation. (B) Overall survival curves for cases of AML-MRC with concurrent RUNX1/ASXL1 mutation compared to cases with RUNX1 mutation alone.

Figure 5.

(A) Overall survival curves for cases of AML with RUNX1^mut with three or more mutations versus cases with fewer than three mutations. (B) Overall survival curves for cases of AML-MRC with three or more mutations versus fewer than three mutations (including RUNX1).

Clinical impact of cytogenetic alterations in AML with RUNX1^mut and AML-MRC with RUNX1^mut

Conventional karyotyping data was retrieved in 51 of 57 patients with AML with RUNX1^mut, 60 of 61 patients with AML-MRC with RUNX1^mut, and 98 of 100 AML, NOS with RUNX1^wt patients. Patients with normal karyotype were more commonly observed in the AML regardless of RUNX1 mutation status (AML with RUNX1^mut and AML, NOS with RUNX1^wt) when compared to AML-MRC with RUNX1^mut (p < .01) (Table 1). Monosomy 7 or del(7q) was more commonly identified in AML-MRC with RUNX1^mut (13%) compared to the other two groups (0.0% in AML with RUNX1^mut and 5.1% in AML, NOS with RUNX1^wt, p = .0072). The prevalence of trisomy 8 was similar in AML with RUNX1^mut (11.8%) and AML-MRC with RUNX1^mut (11.7%), while trisomy 13 appeared in approximately 6% of patients with AML with RUNX1^mut and in 3.3% of AML-MRC with RUNX1^mut patients.

Discussion

Studies have shown that RUNX1^mut are associated with poor prognosis in AML [5,18,19]. However, whether clinical outcome of AML-MRC with RUNX1^mut is worse than AML with RUNX1^mut is yet unknown and there has been debate regarding prognostic impact of multilineage dysplasia in AML patients with RUNX1^mut. Similar to a previous study reported by Haferlach’s group [20], mild morphologic dysplasia in hematopoietic precursor did not adversely affect the clinical outcome in our patient with AML with RUNX1^mut or AML-MRC with RUNX1^mut (data not shown). However, we observed that AML-MRC with RUNX1^mut is associated with inferior OS compared to AML with RUNX1^mut. This observation is clinically important since poor treatment response and inferior survival outcome associated with RUNX1^mut could be further compromised in AML-MRC patients [21,22]. Although CPX-351 was recently approved as a front-line therapy in AML-MRC patients and hypomethylating agents are commonly used in AML patients who are not eligible for the intensive chemotherapy [23–27], there is still lack of data regarding the optimal treatment in this unique group of AML-MRC with RUNX1^mut patients. Future studies are warranted to further improve survival outcomes in this poor prognostic group.

Additionally, recent studies have shown that ASXL1, RAS, FLT3-ITD, and IDH1/2 mutations frequently concur in patients with AML with RUNX1^mut [21]. Also, the incidence of ASXL1 mutation was higher in RUNX1^mut versus AML, NOS with RUNX1^wt patients with normal karyotype (36% versus 8%) [22] and Gaidzik et al. reported that in addition to aforementioned genes, EZH2, SRSF2, SF3B1, STAG2, PHF6, and BCOR mutations frequently concur in AML with RUNX1^mut patients [14]. Similar to these studies, we observed a relatively high incidence of mutations in SRSF2, TET2, DNMT3A, ASXL1, IDH1/2, EZH2, BCOR, PHF6, STAG, and FLT3-ITD genes in RUNX1 mutant patients, either AML with RUNX1^mut or AML-MRC with RUNX1^mut group. In our study, some of these co-mutations appear prognostically important; RUNX1/TET2 co-mutations were associated with worse OS compared to RUNX1^mut alone in AML-MRC while concurrent RUNX1/DNMT3A mutations were associated with worse OS in patients with AML with RUNX1^mut than AML-MRC with RUNX1^mut. Concurrent RUNX1/ASXL1 mutations also showed negative clinical impact on OS in patients with AML with RUNX1^mut and RUNX1/IDH2 co-mutations were associated with worse clinical outcome than RUNX1^mut alone in AML-MRC patients, but not in AML with RUNX1^mut. However, these observations require cautious interpretation because of the limited number of patients included from a single institution, requiring further validation with independent patient cohorts.

Recent studies have demonstrated that AML with RUNX1^mut has a unique cytogenetic signature [20,28]. Trisomy 13 is reportedly strongly associated with RUNX1 mutation in AML (9%, 13/152) [20,29]. However, the presence of trisomy 13 was extremely low in our study appearing only in 5.9% and 3.3% of patients with AML with RUNX1^mut and AML-MRC with RUNX1^mut, respectively. Following normal karyotype, trisomy 8 was the most common aberration making up 11.8% of AML with RUNX1^mut and 11.7% of AML-MRC with RUNX1^mut.

Although RUNX1 mutations have been identified in 10–33% of MDS and AML patients [18,19,30,31], the proportion of germ-line versus somatic mutations is largely unknown. Patients with germ-line RUNX1 mutations develop early onset MDS/AML with a median age of 33 years at diagnosis whereas sporadic MDS/AML patients typically present later time in life with median age of 65–68 years [32,33]. A recent study showed that the carriers of RUNX1 germ-line mutations are typically asymptomatic, but have a high cumulative risk of developing a clonal myeloid neoplasm by age 50, suggesting that secondary mutations are necessary to trigger MDS/AML [34,35]. The median age of patients in the groups in our study was 60–72 years old and only six patients were diagnosed before 35 years. This suggests that the majority of patients in our cohort likely have RUNX1 somatic mutations, but this is inconclusive given the lack of germ-line mutation analysis.

Despite the need for studies to further elucidate the association between AML-MRC and RUNX1^mut to predict poor clinical outcome, our study provides some insight to potential points of clinical interest. We have demonstrated that patients with AML-MRC with RUXN1^mut are a unique poor prognostic group and certain co-mutations and/or cytogenetics could further affect their clinical outcomes. This warrants routine investigation by a comprehensive myeloid mutation panel to have better risk stratification and identify those that may benefit from future targeted therapies.