Clinical Characteristics and Molecular Profiling of SF3B1-Mutated Myelodysplastic Syndrome (MDS) in a Real-World Practice

Abstract

SF3B1-mutated myelodysplastic syndrome (MDS) is a distinct entity associated with a favorable prognosis. Recent data suggest that certain SF3B1 variants portend a worse prognosis. Our study aims to (1) describe SF3B1-MDS patients from a single tertiary center in Singapore and (2) determine if variant type holds prognostic value. We identified MDS patients with SF3B1 variants via next-generation sequencing (NGS) performed from 1 November 2021 to 31 October 2025 at Singapore General Hospital. Extracted genomic material from marrow or blood samples was amplified. Libraries were prepared, sequenced, and analyzed, and the hematological parameters, mutation profiles, and outcomes were evaluated. Twenty-five patients had SF3B1-MDS. Ten SF3B1 variants were found, and the three most prevalent were K700E (42%), K666N (19%), and R625C (7.7%). The median variant allele frequency (VAF) was 30% (IQR: 11–36%). Twelve patients (48%) had ≥1 co-mutations. Variant type and VAF had no impact on disease progression; only the presence of ≥1 co-mutations increased the progression chances. In our study, the analysis of SF3B1 variant type was inconclusive and showed no demonstrable statistical association with disease progression. However, the number of co-mutations affected the prognosis of patients. As SF3B1-MDS is heterogenous, further studies are needed to capture its diversity and identify features required to improve risk stratification and personalized treatment.

1. Introduction

Myelodysplastic syndrome (MDS) is a heterogenous group of clonal myeloid neoplasms characterized by ineffective hematopoiesis, leading to peripheral blood cytopenia, and a variable risk of transformation to acute myeloid leukemia (AML) [1,2]. For nearly two decades, risk stratifications were heavily reliant on the Revised International Prognostic Scoring System (IPSS-R), which used the severity and degree of cytopenias, bone marrow blasts, and cytogenetic groups to prognosticate survival and evolution to AML [3]. The landscape of MDS classification and prognostication has evolved over recent years to incorporate molecular features that better capture clinical–pathological entities and predict clinical outcomes. The International Prognostic Scoring System—Molecular (IPSS-M) incorporates individual somatic gene mutations along with hematological and cytogenetic data [4]. Treatment of MDS is often risk-adapted, varying from supportive care to higher-intensity therapies, including hypomethylating agents (HMAs) and allogeneic hematopoietic stem cell transplant (HSCT).

Splicing factor 3b subunit 1 gene, SF3B1, encodes for U2 small nuclear ribonucleoprotein and is one of the core splicing factors; it is crucial for proper spliceosome complex formation required for appropriate RNA splicing. Mutations in SF3B1 have been associated with a diversity of cancers, both non-hematological and hematological [5,6]. In particular, it has been associated with MDS and is recognized as one of the most frequently mutated genes in MDS, detected in up to a third of all MDS patients [5,7]. SF3B1 mutation has been linked to the presence of ringed sideroblasts (RSs) and erythroid dysfunction in the bone marrow, with reported mutation rates reaching up to 65% in those with RSs [8]. The latest World Health Organization (WHO) classification recognizes SF3B1-mutated MDS as a distinct MDS entity [4].

However, the presence of mutated SF3B1 is generally associated with a favorable outcome and an indolent disease course for MDS [5]. However, emerging data may suggest that not all SF3B1 mutations are the same, with certain SF3B1 variants, such as E592K and K666N, portending a worse prognosis [1,2,9,10]. Furthermore, certain co-mutations like ASXL1 are associated with a worse prognosis than having either mutation in isolation, while other co-mutations like TET2 may conversely have better prognoses than having either alone [11,12]. Lastly, de novo AML patients with SF3B1, as well as other mutations like SRSF2, USAF1, ZRSR2, ASXL1, EZH2, BCOR, and STAG2, comprise a cohort of patients known as secondary AML patients and fair clinically worse [13].

In our study, we look at the incidence of SF3B1 variants in MDS patients in a single tertiary institution and report their clinical characteristics and disease outcomes.

2. Results

A total of 25 MDS patients had SF3B1 mutations identified by NGS. Baseline demographics, hematological and bone marrow parameters, risk strata, and SF3B1 indices are listed in Table 1. Patients were stratified by the presence or absence of K700E and K666* (K666* includes the various K666 SF3B1 substitutions). The SF3B1-mutated MDS patients included 14 men and 11 women with a median age of 72 years (IQR: 68, 78). Hemoglobin levels and absolute neutrophil counts (ANCs) were similar when stratified between K666* and K700E status. Only platelet counts showed a difference in both strata; K666* SF3B1 MDS patients tend to have a lower platelet count on diagnosis with 343 (IQR: 144, 378) versus 100 (IQR: 22, 193), p = 0.007. In contrast, patients with K700E variant MDS presented a higher platelet count of 368 (IQR: 144,443), while non-K700E patients had a median platelet count of 167 (IQR: 89, 286), p = 0.03. Thirteen of the twenty-five patients had <5% blasts on the initial bone marrow aspirate, and eleven out of twenty-five had ≥5% or more RSs. For the entire population, the median percentage of RSs in the marrow was 1% (IQR: 0%, 30%). When stratified according to K666* and K700E status, the median amounts of RSs were not statistically significant. For non-K666* versus K666*, the median amounts of RSs were 28% (IQR: 0%, 48%) versus 0% (IQR: 0%, 2%), with p = 0.07. Similarly, the median amount of RSs for K700E was 28% (IQR: 0%, 54%), and for non-K700E it was 1% (IQR: 0%, 18%), with p = 0.4. Cytogenetic subgroups included normal cytogenetics, complex karyotype, monosomy 7, and trisomy 8 (+8); these cytogenetic features were incorporated into the IPSS-M risk stratification. No instances of deletion 5q (del 5q) were observed in this group. Patients with K700E tend to have “Low” or “Very Low”-risk IPSS-R scores compared to non-K700E patients. The distribution of IPSS-R scores was similar when stratified by K666* status. Patients with a higher number of co-mutations tend to have less favorable IPSS-M scores. On further analysis, 7 out of the 25 SF3B1-mutated cases (28%) fell into the IPSS-M high-/very high-risk categories.

Table 1.

Baseline characteristics and bone marrow parameters for SF3B1-mutated MDS stratified by K666* and K700E status.

	Overall	K666*			K700E
Characteristic	N = 25 1	Non-K666* N = 17 1	K666* N = 8 1	p-Value 2,3	Non-K700E N = 14 1	K700E N = 11 1	p-Value 2,3
Gender				>0.9			0.12
Male	14 (56%)	9 (53%)	5 (63%)		10 (71%)	4 (36%)
Female	11 (44%)	8 (47%)	3 (38%)		4 (29%)	7 (64%)
Race				0.8			>0.9
Chinese	18 (72%)	13 (76%)	5 (63%)		10 (72%)	8 (73%)
Indian	4 (16%)	2 (12%)	2 (24%)		2 (14%)	2 (18%)
Malay	3 (12%)	2 (12%)	1 (13%)		2 (14%)	1 (9%)
Age	72 (68, 78)	72 (70, 78)	72 (61, 77)	0.6	73 (63, 80)	72 (70, 78)	>0.9
Hemoglobin (g/dL)	8.7 (7.6, 9.5)	8.4 (7.5, 9.5)	8.8 (8.1, 9.9)	0.5	8.7 (6.9, 9.5)	8.4 (8.0, 9.5)	0.6
ANC (×109/L)	2.3 (1.4, 3.6)	2.3 (1.4, 3.6)	2.4 (1.3, 3.2)	>0.9	1.5 (1.1, 2.7)	2.8 (2.2, 3.7)	0.13
Platelets (×109/L)	255 (99, 368)	343 (144, 378)	100 (22, 193)	0.007	167 (89, 286)	368 (144, 443)	0.033
Marrow Blast (%)	1.0 (1.0, 2.0)	1.0 (1.0, 2.0)	1.0 (0.5, 9.0)	>0.9	1.0 (0.0, 6.0)	1.0 (1.0, 2.0)	>0.9
Ringed Sideroblasts (%)	1 (0, 30)	28 (0, 48)	0 (0, 2)	0.073	1 (0, 18)	28 (0, 54)	0.4
Karyotype				0.7			0.12
Normal	15 (58%)	9 (50%)	6 (75%)		9 (69.2%)	6 (46%)
-Y	3 (11.6%)	3 (16.4%)	0 (0%)		0 (0%)	3 (23.2%)
Monosomy 7	1 (3.8%)	1 (5.6%)	0 (0%)		0 (0%)	1 (7.7%)
del(11)	1 (3.8%)	1 (5.6%)	0 (0%)		0 (0%)	1 (7.7%)
i(14)	1 (3.8%)	1 (5.6%)	0 (0%)		0 (0%)	1 (7.7%)
inv(3)	1 (3.8%)	1 (5.6%)	0 (0%)		0 (0%)	1 (7.7%)
inv(12)	1 (3.8%)	1 (5.6%)	0 (0%)		1 (7.7%)	0 (0%)
trp(1)	1 (3.8%)	1 (5.6%)	0 (0%)		1 (7.7%)	0 (0%)
Trisomy 8	1 (3.8%)	0 (0%)	1 (12.5%)		1 (7.7%)	0 (0%)
der(13; 14)	1 (3.8%)	0 (0%)	1 (12.5%)		1 (7.7%)	0 (0%)
IPSS-R				0.6			0.043
Very Low	2 (8%)	2 (12%)	0 (0%)		0 (0%)	2 (18%)
Low	15 (60%)	10 (59%)	5 (63%)		8 (57%)	7 (64%)
Intermediate	2 (8%)	2 (12%)	0 (0%)		1 (7.1%)	1 (9%)
High	5 (20%)	2 (12%)	3 (37%)		5 (36%)	0 (0%)
Very High	1 (4%)	1 (6%)	0 (0%)		0 (0%)	1 (9%)
IPSS-M				0.3			0.5
Very Low	4 (16%)	4 (23%)	0 (0%)		1 (7%)	3 (27%)
Low	13 (52%)	8 (47%)	5 (63%)		8 (57%)	5 (46%)
Mod Low	1 (4%)	1 (6%)	0 (0%)		0 (0%)	1 (9%)
High	5 (20%)	2 (12%)	3 (37%)		4 (29%)	1 (9%)
Very High	2 (8%)	2 (12%)	0 (0%)		1 (7%)	1 (9%)
SF3B1 (VAF%)	30 (11, 36)	33 (21, 36)	14 (5, 36)	0.2	31 (11, 39)	30 (11, 34)	0.6
No. of Co-Mutations				0.7			>0.9
0	13 (52%)	9 (53%)	4 (50%)		7 (50%)	6 (55%)
1	4 (16%)	2 (12%)	2 (25%)		2 (14%)	2 (18%)
2	6 (24%)	5 (29%)	1 (13%)		4 (29%)	2 (18%)
≥3	2 (8.0%)	1 (5.9%)	1 (13%)		1 (7.1%)	1 (9.1%)

¹ n (%); median (Q1, Q3); ² Fisher’s exact test; ³ Wilcoxon rank sum test.

2.1. SF3B1 Variant, Clonality, and Co-Mutations

In Table 1, the median SF3B1 VAF was 30% (IQR: 11%, 36%), with no difference between the K666*/non-K666* and K700E/non-K700E strata. The breakdown of co-mutations is shown in Figure 1. The co-mutations found were BCOR, CEBPA, EZH2, FLT3, GATA2, IDH1, NRAS, TP53, ZRSR2, PHF6, DNMT3A, CSF3R, TET2, and RUNX1, with RUNX1 being the most prevalent.

Figure 1.

Co-mutations found in SF3B1-MDS patients. (A) Bar chart of all the different co-mutations found in the entire cohort; (B) co-mutations stratified by K700E and non-K700E status; (C) co-mutations stratified by K600* and non-K600* status.

Ten variants of SF3B1 were found (Table 2), with the top three most prevalent being K700E (42%), K666N (19%), and R625C (7.7%). One patient had two different SF3B1 variants: K700E and T663I. The median VAF was 30% (IQR: 11–36%). All the mutations were missense mutations. Twelve out of twenty-five (44%) patients had one or more co-mutations.

Table 2.

SF3B1 variants and type/number of co-mutations.

Variant	Nucleotide Change	Protein Change	N = 26 1 (%)
K700E	c.2098A>G	p.Lys700Glu	11 (42%)
K666*			8 (30.4%)
K666N	c.1998G>T, c.1998G>C	p.Lys666Asn	5 (19%)
K666Q	c.1996A>C	p.Lys666GIn	1 (3.8%)
K666R	c.1997A>G	p.Lys666Arg	1 (3.8%)
K666T	c.1997A>C	p.Lys666Thr	1 (3.8%)
R625C	c.1873C>T	p.Arg625Cys	2 (7.7%)
D781G	c.2342A>G	p.Asp781GIy	1 (3.8%)
E622V	c.1865A>T	p.Glu622Val	1 (3.8%)
H662Q	c.1986C>A	p.His662GIn	1 (3.8%)
T663I	c.1988C>T	p.Thr663Ile	1 (3.8%)

¹ One patient had two concomitant SF3B1 variants—K700E and T663I.

2.2. Therapy and Outcomes

Eleven out of the twenty-five patients received treatment for MDS; among them, seven received HMA, two were recruited into the MBG453 trial (TIM3 inhibitor), two received intensive chemotherapy, and three went on to receive allogeneic HSCT. At a median follow-up of 20.3 months, out of the twenty-five patients, ten patients progressed, four due to worsening cytopenias, four due to increasing bone marrow blasts > 10%, and two due to disease progression to AML. Only one patient had died at the time of analysis.

There were 11 patients with K700E. Five patients had K666N, and two patients had R625C. In a multivariate analysis (Table 3), none of the SF3B1 variant types were associated with different outcomes. When grouped together as high-risk SF3B1 variants, K666N and R625C did not have an inferior progression compared to those without the high-risk SF3B1 variant. Similarly, when comparing the K666* to the non-K666* variants, there was no difference in progression, with an HR of 1.66 (95% CI: 0.47–4.92; p = 0.4). Lastly, having K700E did not demonstrate a difference in disease progression compared to non-K700E variants, with an HR of 0.29 (95% CI: 0.06–1.38; p = 0.12). Male gender led to an HR of 5.30 (95% CI: 1.04–26.9; p = 0.044). In a similar manner, the number of co-mutations was associated with an increased chance of progression, with each additional co-mutation resulting in increasing chances of progression (HR of 2.78, with 95% CI: 1.5–5.16; p = 0.001).

Table 3.

Multivariate analysis for progression.

Characteristic	HR	95% CI	p-Value
Gender
Female	-	-
Male	5.30	1.04, 26.9	0.044
Age	0.94	0.89, 1.00	0.044
VAF (%)	0.97	0.92, 1.02	0.3
No. of Co-Mutation	2.78	1.50, 5.16	0.001
K666*	1.66	0.47, 5.92	0.4
K700E	0.29	0.06, 1.38	0.12
HR-Variant §	1.23	0.26, 5.89	0.8

Abbreviations: CI = Confidence Interval; HR = Hazard Ratio; ^§ HR-Variant defined as presence of K666N or R625 variants.

3. Discussion

3.1. Co-Mutations (Number and Type)

In this study of SF3B1-mutated MDS patients, with each increasing co-mutation, there was a 2.8-fold higher chance of disease progression requiring treatment. Disease progression was defined as a composite of worsening cytopenias, an increase in bone marrow blasts to >10%, progression to AML, or death. Only two patients progressed to AML, and given the small cohort size (n = 25), meaningful statistical analysis was not feasible for leukemic transformation alone. The co-mutations in our cohort were BCOR, CEBPA, EZH2, FLT3, GATA2, IDH1, NRAS, TP53, ZRSR2, PHF6, DNMT3A, CSF3R, TET2, and RUNX1. In the literature, SF3B1 is associated with improved prognosis; however, studies suggest that certain co-mutations can potentially change the prognosis for SF3B1-mutated MDS. Though there were multiple co-mutations detected, they were not comprehensive, and the number within the cohort was far too small to determine if specific mutations combined with SF3B1 fared worse or better than others. In contrast, the analysis by Malcovati et al. showed that in 482 SF3B1-mutated MDS cases with co-mutations, the number of co-mutations did not affect overall survival compared to cases without co-mutations [10]. And yet they did show that having two specific co-mutations of RUNX1 and EZH2 worsened the outcome [10]. Similarly, Huber et al. and Yun et al. demonstrated that RUNX1 mutations combined with SF3B1 mutations resulted in a 3.5-fold decreased survival [9,14]. Song et al. demonstrated that SF3B1 and ASXL1 had a negative prognostic factor for survival [11]. In contrast, Song et al. investigated MDS patients with SF3B1 co-mutated with DNMT3A and demonstrated similar survival to isolated SF3B1-mutated MDS but better survival than isolated DNMT3A-mutated MDS [15]. Lastly, co-mutations with TET2 did not demonstrate any significant benefit compared to having an SF3B1 mutation alone [12]. The wide range of co-mutation combinations remains complex, and further studies are needed to elucidate their significance. Thus, it would be ideal for a larger study to ascertain if SF3B1 combined with certain co-mutations, compared with having these specific mutations in isolation, would result in better or worse outcomes.

3.2. Variant Type/Hotspot

In our study, when comparing SF3B1 variants with each other, none of them had a statistically different progression—neither the high-risk K666N and R625 variants, nor the favorable K700E variant (p = 0.12). It is possible that with the increase in our cohort numbers, the positive prognosis associated with the presence of K700E may become more evident. Likewise, due to the low numbers of SF3B1-MDS patients (N = 25), with only seven having K666N (n = 5) or R625C (n = 2) variants, these variants’ effects on prognosis may not have been sufficiently forthcoming. There are an increasing number of studies that suggest that not all SF3B1-mutated MDS should be assessed as one single group, as the prognostic impact varies between the different variants. In many studies, the most frequently mutated SF3B1 variants at K700E account for 48–60% of SF3B1-mutated MDS, which is associated with the typical RS-MDS and overall improved prognosis [5,16,17,18,19]. Mutations at the K700E site led to gain-of-function mutants which confer the typical positive prognosis [18]. Conversely, though non-K700E variants are less common, these SF3B1 variants not only lack the favorable prognosis but also tend to be associated with higher-risk MDS at presentation, higher rates of leukemic transformation, and worse survival [5]. Kanagal-Shamanna et al. analyzed 94 mutated SF3B1 and 415 wild-type SF3B1-MDS genes and found that non-K700E variants had a lower ANC on presentation, a higher IPSS risk stratum, and associated RUNX1, BCOR, IDH2, and SRSF2 mutations [19]. Likewise, Liu et al. assessed K666 and R625 hotspots, which demonstrated poorer prognoses than their non-K666 and non-R625 counterparts [17]. Dalton et al. and Choi et al. demonstrated that the K666N and E592K variants have worse prognoses compared to other SF3B1 variants [1,2]. Furthermore, one study by Sakuma et al. demonstrated that there may be a gender disposition whereby K666* variants are typically associated with males, with K700E being more so with females [20]. Therefore, advancement in prognostication may require integrating both SF3B1 hotspots, as well as identifying the presence of co-mutations.

3.3. VAF and Clonality

Lastly, in our study, the clone size measured by the variable allelic frequency of SF3B1 had no overall effect on prognosis in terms of progression to AML or mortality. Jiang et al. demonstrated that SF3B1 patients with a VAF ≥ 15% had improved survival compared to those with a VAF < 15% (HR: 0.29, CI: 0.02–0.92, p = 0.044) [21]. Lachowiez et al. also found that, for SF3B1 VAF < 10%, there was a marginally decreased OS compared to VAF ≥ 10% (5 vs. 6.2 years, p = 0.04), while there was no change in leukemia-free survival [22].

3.4. Response to Treatment

There are studies that suggest that SF3B1 mutations conferred improved response to certain MDS-directed therapeutics. Idossa et al. found that SF3B1-mutated MDS was more likely to respond to lenalidomide (56 vs. 27%, p = 0.04) but did not demonstrate any difference between response to HMA and erythropoietin-stimulating agonists [23]. Luspatercept has been used in patients with SF3B1 low-risk MDS with efficacy [24,25]. Yet there are limited studies comparing SF3B1-mutated versus unmutated patients. Interestingly, in a study by Consagra et al., there was no statistical significance in hematological improvement between SF3B1-mutated versus unmutated patients (53.8 vs. 40.1%, respectively, p = NS) [26]. Furthermore, neither variant type (57.7 vs. 42.3%, p = 0.31, K700 vs. non-K700) nor VAF (52.1 vs. 47.9%, p = 0.11, VAF ≥ 38% vs. <38%) demonstrated any impact on hematological improvement. Consagra went on to show that certain SF3B1 co-mutations were associated with hematological improvement. Response rates were 0% with 5q co-mutation; 53.8% with any BCOR, BCORL1, NRAS, RUNX1, SRSF2, or STAG2 co-mutation; and 55.8% with any other co-mutations (p = 0.046) [26]. Unfortunately, our study lacked the numbers to sufficiently look at response rates by type of co-mutations.

Even when SF3B1 mutation status is known, real-world management of MDS remains individualized, due to the disease’s biological and clinical heterogeneity. Patients tend to be older, and therapeutics are frequently constrained by age-related factors, including frailty and comorbid conditions. Furthermore, treatment-related complications like iron overload from chronic transfusions can negatively affect survival. Thus, an integrative approach incorporating SF3B1 and clinical status may allow for more precise patient-centered care.

3.5. Limitations

This is a retrospective study with a low sample size, which could have led to low statistical power and increased risk of type II errors. True differences in progression may have gone undetected. Furthermore, the limited sample size may have predisposed to selection and sampling bias, as our study may not have sufficiently captured the diversity of SF3B1 mutations. Finally, stratification and multivariate analyses are limited.

4. Materials and Methods

4.1. Patients and Samples

We retrospectively identified MDS patients with SF3B1 mutations via next-generation sequencing (NGS) performed on bone marrow aspirate samples from 1 November 2021 to 31 October 2025 in the Singapore General Hospital registry. Extracted genomic DNA and RNA were amplified, and library preparation was performed with the Oncomine Myeloid Research assay (Themo Fisher Scientific, Waltham, MA, USA) to interrogate 40 DNA genes and 29 fusion driver genes. NGS libraries were sequenced on the Ion GeneStudio S5 System (Thermo Fisher Scientific, Waltham, MA, USA) and analyzed using in-house bioinformatics pipelines and the Ion Reporter^TM software, Version 5.20. The patient demographics, hematological parameters, and disease outcomes were analyzed. The SF3B1 variants, variant allele frequency (VAF), and number and type of co-mutations in each patient were analyzed. Risk stratification was performed using both IPSS-R and IPSS-M.

4.2. Statistical Analysis

Statistical analyses were conducted using R (version 2025.05.01). Patient characteristics, hematological and bone marrow indices, diagnostic risk strata, and SRFB1 parameters were analyzed. Results were reported as total numbers with percentages and medians with interquartile ranges (IQRs). Categorical variables between groups were analyzed with Fisher’s exact test, and nonparametric variables had their medians calculated and compared via the Wilcoxon ranked sum test. Disease progression was defined from the time of diagnosis until the time that the patient required treatment with a hypomethylating agent, intensive chemotherapy, and/or HSCT due to (1) worsening bone marrow blasts or (2) worsening cytopenias, or (3) until the patient transformed to AML or (4) died. Univariate Cox proportional hazards regression analysis was used to identify any association between each of the variables and progression, and this was followed by multivariate analysis. p-values < 0.05 were considered statistically significant, though inference was limited due to low event frequency.

5. Conclusions

In summary, SF3B1-mutant MDS is biologically distinct but has a considerable amount of heterogeneity depending on hotspot variations, clonality, and the number/type of co-mutations. Although SF3B1-MDS is commonly associated with an indolent course and relatively favorable outcome, variables add nuance and complexity to the prognostication of MDS. The above prompts a more integrative approach for this disease group. Reliance on the presence or absence of SF3B1 alone may lack granularity and be an oversimplification.

Nevertheless, this study is primarily descriptive and hypothesis-generating. Given the small cohort size, the prognostic significance of individual SF3B1 variants may not have been detected. Although the number of co-mutations showed a statistically significant association with disease outcome, larger studies are required to validate its prognostic relevance and to assess co-mutational burden as an adjunct prognostic factor. Additionally, interpretation of co-mutational effects is complex, as specific co-mutations influence prognosis, as reflected in the IPSS-M risk scores. Further studies examining the interactions between SF3B1 and other co-mutations are therefore needed to better define its prognostic impact. Finally, clinical variables in MDS (e.g., co-morbidities, transfusion dependence, iron overload, and recurrent infections) also carry prognostic value, and future larger studies should explore the interactions among clinical, cytogenetic, and molecular factors on disease prognosis.

Abbreviations

The following abbreviations are used in this manuscript:

AML	Acute Myeloid Leukemia
CI	Confidence Interval
HMAs	Hypomethylating Agents
HR	Hazard Ratio
HSCT	Hematopoietic Stem Cell Transplant
IPSS-R	Revised International Prognostic Scoring System
IPSS-M	International Prognostic Scoring System—Molecular
IQR	Interquartile Range
MDS	Myelodysplastic Syndrome
NGS	Next-Generation Sequencing
RSs	Ringed Sideroblasts
SF3B1	Splicing Factor 3b Subunit 1
VAF	Variant Allelic Frequency
WHO	World Health Organization

Author Contributions

Conceptualization, H.T. and C.T.; methodology, R.R.W.; software, G.F.H., S.J.K. and H.T.; formal analysis, H.T., S.J.K. and G.F.H.; investigation, S.J.K. and G.F.H.; resources, H.T., R.R.W. and T.T.T.; data curation, R.R.W. and T.T.T.; writing—original draft preparation, R.R.W.; writing—review and editing, R.R.W. and T.T.T.; supervision, H.T. and T.T.T. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the local institutional review board of Singapore General Hospital (IRB# 2018-2520, approved on 22 October 2018).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this study are available on request from the corresponding author due to data privacy reasons.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research received no external funding.