PDF
Citation
Print
INTRODUCTION
Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal hematological neoplasms characterized by ineffective hematopoiesis, manifesting as morphologic dysplasia of hematopoietic cells and peripheral cytopenia [1-3].
Several benign or malignant conditions show morphologic features and characteristics similar to MDS and MDS overlap syndromes, often making it difficult to diagnose patients correctly [4]. These MDS overlap syndromes include myelodysplastic/myeloproliferative neoplasms (MDS/MPNs) with both MDS and MPN features and hypoplastic MDS (hMDS) with aplastic anemia (AA) characteristics. In addition, MDS overlap syndromes may include secondary acute myeloid leukemia (AML) evolving from MDS, and some potential pre-MDS conditions showing cytopenia or somatic mutations but not meeting the diagnostic criteria of the MDS. However, secondary AML and potential pre-MDS states are regarded as a “continuation” or “transformation” from or into the MDS and, thus, have not been discussed in this review. This needs to be discussed as a separate topic in terms of a continuous spectrum.
The present article will provide insights into correct differential diagnoses through comparative reviews for the diagnostic criteria and genetic findings of these MDS and overlap syndromes.
Myelodysplastic syndromes
The revised 4th edition of the World Health Organization (WHO) Classification of Myeloid Neoplasms (2016) introduced refined diagnostic information for MDS subtypes (Table 1) [1, 3, 5, 6].
The terminology for adult MDS was changed in the 2016 WHO Classification and terms, such as “refractory anemia” and “refractory cytopenia,” were replaced with “myelodysplastic syndrome” followed by the appropriate modifiers: single vs multilineage dysplasia, ring sideroblasts (RS), excess blasts, or the del(5q) cytogenetic abnormality [1]. There were no changes in the classification of childhood MDS; refractory cytopenia of childhood remains a provisional entity within this category [1].
The category of ‘Refractory cytopenias with multilineage dysplasia and ring sideroblasts (RCMD-RS)’ was eliminated in the 2008 WHO Classification and was merged with RCMD because it was shown to be prognostically similar to RCMD, but lacking RS [7-9]. The discovery of mutations in the spliceosome gene SF3B1 that are associated with RS provided a link between morphology and genetics in MDS [9-12]. This combination of shared morphology (RS) and a shared underlying driver mutation (SF3B1) favors the separation of MDS with RS as distinct entities, which may have single or multilineage dysplasia [9]. Therefore, in the 2016 WHO Classification, the subtype ‘RCMD-RS’ was restored as ‘MDS with ring sideroblasts and multilineage dysplasia (MDS-RS-MLD)’ [13]. It has been shown that in cases of MDS with RS, the actual percentage of RS is not prognostically relevant [1, 11]. Therefore, according to the 2016 WHO classification, if an SF3B1 mutation is identified, a diagnosis of MDS-RS may be established if RS comprise as few as 5% of the nucleated erythroid cells, whereas at least 15% RS are still required in cases lacking a demonstrable SF3B1 mutation [1].
In cases with refractory, unexplained cytopenia but no morphological evidence of dysplasia or increased blasts, cytogenetic abnormalities listed in Table 2 (the same MDS-defining cytogenetic abnormalities listed in the 2008 WHO classification) are considered as presumptive evidence of MDS, and such cases are included in the category of MDS, unclassifiable [5, 14]. In such cases, the abnormality must be demonstrated by conventional karyotyping, not by fluorescence in situ hybridization (FISH) or sequencing technologies [1].
A very minor change in MDS with isolated del(5q) is the inclusion of one additional chromosomal abnormality (e.g., +8), but excluding monosomy 7 [13, 15, 16]. Clinical studies have not demonstrated any adverse impact of an extra chromosome on response to lenalidomide therapy [13].
Recurrent cytogenetic abnormalities identified by conventional karyotyping are found in approximately 50% of MDS cases (Table 2) [5, 17-19]. In contrast to the case with AML, in which balanced abnormalities, such as t(8;21)(q22;q22), are predominant [17, 19], the majority of the abnormalities in MDS are unbalanced changes, which result in copy-number abnormalities (CNAs), such as gains or losses of chromosomal materials [17, 20]. Complex karyotypes (3 or more chromosomal abnormalities) are frequently accompanied by TP53 mutation (approximately 40–50% cases), generally predicting adverse clinical outcomes, particularly when -5/del(5q) and del(17p) are involved [17, 21-23].
Single nucleotide polymorphism (SNP) array–based platforms can detect CNAs and copy-neutral loss of heterozygosity (LOH) or uniparental disomy (UPD), which are commonly seen in chromosomes 1p, 4q, 7q, 17p, and 21q in MDS, and in 7q and 11q in MDS/MPN [17, 24]. Combined with CNAs and chromosomal abnormalities, more than 78% to 90% of MDS patients have one or more known recurrent genetic abnormalities [17, 25, 26]. CNAs and mutations significantly correlate with the disease phenotype and clinical outcome [17].
Recurrent somatic mutations in more than 50 genes have been identified in 80–90% of MDS cases, and more than 30 driver genes involved in the pathogenesis of MDS have been identified [5, 17, 25, 26]. The genes found to be mutated in at least 5% of MDS cases are listed in Table 3 [5, 27, 28]. While a typical MDS patient harbors a median of 2 or 3 driver mutations, those with high-risk MDS (MDS with excess blasts and MDS with multilineage dysplasia) and chronic myelomonocytic leukemia (CMML) tend to show higher numbers of driver mutations than those with lower-risk MDS [MDS with single lineage dysplasia (SLD), MDS-RS-SLD, and MDS with isolated del(5q)] [17]. Mutations in several genes, such as DDX41, RUNX1, GATA2, and TP53, may be present in the germline and responsible for a predisposition to AML and MDS [17].
Some combinations of driver mutations co-occur more frequently than expected, while others are observed in a mutually exclusive manner, suggesting the presence of functional interactions between these mutations that are involved in positive and negative selections [17]. To date, del(5q) remains the only cytogenetic or molecular genetic abnormality that defines a specific MDS subtype and isolated del(5q) is associated with anemia (with or without other cytopenias) and normal-to-increased platelet counts with micromegakaryocytes [1, 17, 29]. Notably, del(5q) may exist as an isolated cytogenetic lesion; however, it is more commonly seen as a part of complex karyotypes among TP53-mutated cases [17, 23, 30].
As another example, SF3B1 mutations are strongly associated with increased bone marrow (BM) RS [1] and generally predict a favorable prognosis [17, 31, 32]. Of note, the SF3B1 mutation, which generally predicts a favorable prognosis, is found in the majority of cases with MECOM (also called EVI1, at 3q26.2)–involved rearrangements, which are uniformly associated with very poor clinical outcomes [5, 17, 33, 34].
Recently, the International Working Group for the Prognosis of MDS proposed SF3B1-mutant MDS as a distinct entity, mainly characterized by ineffective erythropoiesis, relatively good prognosis, and potential response of anemia response to luspatercept treatment (Table 4). In the proposal, RS are not required for the diagnosis of SF3B1-mutant MDS and some cytogenetic abnormalities are excluded. These abnormalities include del(5q), monosomy 7, inv(3) or abnormal 3q26 (including MECOM), and complex karyotype (≥3) [6].
SF3B1 mutations have been reported in approximately 20% of patients with the MDS with isolated del(5q), associated with a variable proportion of RS [6, 12, 25, 26, 32]. These cases are classified within the category of MDS with isolated del(5q) according to the 2016 WHO criteria (Table 1) [5, 6]. Myeloid neoplasms with concomitant JAK2 mutation and del(5q) may show overlapping myelodysplastic and myeloproliferative features [5, 35-39]; however, these mutations do not appear to alter the disease phenotype [5]. Despite insufficient data, such cases could benefit from a combined therapeutic approach with lenalidomide and a JAK2 inhibitor [35, 39, 40]. In contrast to generally favorable prognoses in SF3B1-mutated cases, patients with other splicing factor mutations, including mutations in U2AF1 and SRSF2 genes, display poor overall survival [17, 41-44].
TP53 mutations account for approximately one-fourth of the high-risk MDS cases and also comprise approximately 10% of primary AML and 50% of therapy-related MDS/AML cases [17, 45, 46]. Regardless of their clinical or pathological diagnosis, patients with TP53-mutated myeloid neoplasms uniformly show a poor prognosis [23, 47].
Despite the large number of driver mutations that have been newly detected with advanced genomics-based studies, the functional roles of these driver mutations in MDS pathogenesis remain to be elucidated [17].
Myelodysplastic/myeloproliferative neoplasms
Myelodysplastic/myeloproliferative neoplasms are a heterogeneous group of clonal hematopoietic neoplasms with features common to both MDS and MPN [1, 2]. These neoplasms include CMML, atypical chronic myeloid leukemia (aCML), juvenile myelomonocytic leukemia (JMML), MDS/MPN with RS and thrombocytosis (MDS/MPN-RS-T), and MDS/MPN, unclassifiable (MDS/MPN-U) [1].
The diagnosis and classification of these “hybrid myeloid neoplasms” can be challenging despite applying the relevant diagnostic criteria. Some patients with MDS may initially show thrombocytosis, whereas some MPN cases may have significant cytopenia(s). In addition, dysplastic features and monocytosis can be seen in MPNs, further complicating the differential diagnosis, particularly when the initial diagnostic material is not available for review [48].
Chronic myelomonocytic leukemia: CMML is characterized by a pathologic myeloproliferation of granulocytic and monocytic cells with cytopenia and myelodysplasia [39, 49]. CMML is the most common type of MDS/MPN overlap syndrome [49, 50]. In the French-American-British classification, CMML was classified as one of the 5 subtypes of MDS. It was placed in a newly created disease group, the MDS/MPN overlap category in the 2001 WHO classification system of myeloid neoplasms [49, 51]. The diagnostic criteria for CMML are summarized in Table 5 [5].
CMML has been classified into CMML-1 and CMML-2 based on the circulating and BM blast count, a prognostic factor that had been consistently shown to be associated with poor survival [49, 52, 53]. The 2016 WHO classification system further expanded these groups into CMML-0 [peripheral blood (PB) <2%, BM <5%], CMML-1 (PB 2–4%; BM 5–9%) and CMML-2 (PB 5–19%, BM 10–19%, or presence of Auer rods). This categorization identifies distinct prognostic groups with a median survival of 31, 19, and 13 months, respectively (P<0.001) [1, 49, 53]. However, the 3-tiered blast-based subgroups are somewhat controversial and it was noted that the newly proposed CMML-0 and CMML-1 categories were similar in terms of the CMML-specific Prognostic Scoring System (CPSS) [54] risk profile, cytogenetic risk groups, and mutation profiles [48]. In addition, the low concordance rate for blast enumeration was reported in CMML [48, 55, 56] and it was suggested that adding the CMML-0 subcategory does not have clear prognostic benefits [48, 57].
Moreover, CMML can be subclassified into a “myelodysplastic” variant (“MD-CMML”), defined by white blood cell (WBC) count of <13×109/L, and a “myeloproliferative” variant (“MP-CMML”), with WBC count of ≥13×109/L [39, 49, 53, 54, 58, 59]. The prognostic significance of such a classification has not been consistently reported; however, molecular data has demonstrated significant differences among MP-CMML (RAS/MAPK enriched) and MD-CMML [1, 49]. Notably, inconsistencies in categorizing patients into MP and MD-CMML prognostic groups based on a single WBC count at diagnosis can occur because of fluctuations in the WBC count in the early clinical course of CMML [49, 60].
“Oligomonocytic” CMML, which often represents a prodromal phase of the disease, is defined by relative monocytosis (≥10%), with an absolute monocyte count of 0.5–0.9×109/L, in the presence of BM findings consistent with CMML and molecular results supportive of it [39, 61-63]. Patients with oligomonocytic CMML often display a similar molecular profile to that of classic CMML, with many cases progressing to frank disease [27, 63].
BCR-ABL1-positive CML can present with monocytosis, especially with the presence of the p190 BCR-ABL1 fusion transcript [5, 39]. Monocytosis in aCML is usually absent or minimal, and when present, the monocytes should account for <10% of PB leukocytes [39]. Rarely, MPN associated with monocytosis may mimic CMML; additionally, the presence of MPN clonal markers, such as JAK2, MPL, or CALR, and BM evidence of MPN favors a diagnosis of MPN over CMML [1, 49].
Conventional metaphase karyotyping shows chromosomal abnormalities in approximately 30% of the patients with CMML, and common karyotypic abnormalities include +8 (23%) and del 7/-7q (14%), which are non-specific and not diagnostic of CMML [49, 64]. With the use of high-resolution SNP array–based karyotyping, up to 60% of the patients have been shown to have chromosomal aberrations [49, 65].
At least one mutation is detected in 86–98% of CMML cases [49]. While no mutation defines CMML, it could aid in diagnosis [27, 49]. The most frequent point mutations seen in CMML include TET2 (50–60%), SRSF2 (45–50%), and ASXL1 (40–45%), which are the hallmark molecular findings of CMML, and the combination of SRSF2 and TET2 mutations is particularly suggestive of CMML (Table 3) [27, 39, 49]. TP53 mutations are quite rare in CMML (only 1% of cases) [27, 66, 67].
TET2 mutation has been shown to be an early event in CMML pathogenesis, with the early clonal dominance of TET2 mutated cells being responsible for monocytic skewing, resulting in the CMML phenotype [27, 68]. Up to 40% of the patients have mutations in cell signaling pathways, particularly the RAS pathway genes (NRAS, KRAS, CBL) and JAK2, which are more common in the MP-CMML subtype [58, 69]. RUNX1 mutations are noted in 10–20% of patients, with additional mutations noted in other splicing complex genes (SF3B1, ZRSR2, U2AF1); moreover, mutations in genes associated with epigenetic regulation (DNMT3A, EZH2) have been observed at frequencies of <10% [27, 70, 71].
Atypical chronic myeloid leukemia, BCR-ABL1 negative: Atypical chronic myeloid leukemia, BCR-ABL1 negative is a leukemic disorder that predominantly affects the neutrophilic lineage; it is associated with neutrophilic leukocytosis and circulating immature granulocytic precursors constituting ≥10% of all leukocytes [5, 39]. The diagnostic criteria for aCML are summarized in Table 6 [5].
Prominent dysgranulopoiesis and other dysplastic features seen on BM aspirate are important features of aCML that distinguish it from CML, in addition to the lack of BCR-ABL1 translocation [1, 49, 51]. The mature neutrophils often display hyposegmented and/or bizarrely segmented nuclei, although in some cases, the dysplastic features are relatively subtle. The so-called “syndrome of abnormal chromatin clumping” is considered to be a variant of aCML. In these cases, circulating neutrophils exhibit a characteristic exaggerated clumping of nuclear chromatin, producing darker nuclei [39, 72, 73]. On metaphase cytogenetics of aCML patients, +8 and +21 are frequently found, with +13, del(20q) and i(17q) being less common [49, 74, 75].
In the 2016 WHO classification, there is an emphasis on identifying the recurrent molecular markers seen in aCML, i.e., SETBP1 (25–40%) and ETNK1 (approximately 8%) mutations (Table 3) [27, 39, 49, 76-80]. In addition, mutations commonly observed in other MDS/MPNs, including ASXL1, EZH2, TET2, SRSF2, and N/KRAS, are observed in ≥20% of aCML cases, while others are seen at lower frequencies [77, 81-83]. SETBP1 mutations were shown to correlate with monosomy 7, isochromosome i(17)(q10), ASXL1, and CBL mutations, and to be mutually exclusive of JAK2 and TET2 mutations [27, 82]. Studies on CSF3R mutations has demonstrated its high prevalence (80–90%) in chronic neutrophilic leukemia and a relatively low frequency (<1 to 10%) in aCML, with a rare incidence in other myeloid disorders [27, 49, 78, 84, 85]. Less frequently, MPN-like mutations in JAK2 (approximately 5%) and complete lack of MPL or CALR mutations have been shown in aCML [27, 49, 86].
Juvenile myelomonocytic leukemia: JMML is a clonal hematopoietic neoplasm of childhood, with a median age at presentation of 2 years [5]. As the childhood counterpart of CMML, JMML shares many clinical and molecular features with CMML, as patients present with leukocytosis and monocytosis, the latter being generally ≥1×109/L. The diagnostic criteria for JMML are summarized in Table 7 [5, 87]. JMML is an aggressive disease with progressive BM failure and short survival [39]. Cytogenetic abnormalities are seen in 30–35% of patients and sole monosomy 7 is seen in 25% of patients [88, 89].
Although its morphologic features are similar to CMML, JMML differs due to its specific and unique pathogenic background as up to 90% of patients harbor either somatic or germline mutations in one of the RAS-MAPK pathway genes (Table 3) [27, 39, 90, 91]. The most common mutations in patients with JMML occur in PTPN11, N/KRAS, NF1, and CBL genes [27, 90-97]. Activating somatic RRAS mutations have been shown in few of the cases that are negative for all 5 common mutations [90, 95, 98]. Congenital predisposition is observed and includes patients with a germline NF1 mutation (characteristic of neurofibromatosis type 1), PTPN11 mutation (observed in half of the patients with Noonan syndrome), and a germline syndrome associated with mutations in CBL [27, 92, 99]. The mutations observed in other MDS/MPNs, such as TET2, SRSF2, and JAK2, are not seen, and ASXL1 mutations are rarely reported [27, 94-96, 98, 100]. However, secondary mutations in SETBP1 and JAK3 are observed in up to 15% of patients with JMML [27, 93, 96]. MP-CMML largely occurs due to RAS pathway mutations that occur in the context of age-related clonal hematopoiesis (TET2, SRSF2, ASXL1), while JMML is a bona fide RASopathy, with additional somatic mutations, including in the epigenetic regulator genes, resulting in disease progression [88].
The differential diagnosis of the JMML phenotype includes rare myeloproliferative malignancies with receptor tyrosine kinase translocations [90]. In RAS pathway mutation-negative cases, disorders with a clinical and hematological picture mimicking that of JMML, such as infection, Wiskott–Aldrich syndrome, and malignant infantile osteopetrosis, must be excluded [5, 101-103].
MDS/MPN with ring sideroblasts and thrombocytosis: “Refractory anemia with ringed sideroblasts associated with marked thrombocytosis (RARS-T)” was included as a provisional entity in 2001 WHO classification of myeloid neoplasm, with the updated 2016 classification accepting it as a distinct entity; it is now referred to as “MDS/MPN-RS-T” [1, 49]. The concomitant presence of mutations in SF3B1, an MDS-type mutation and one of the MPN mutations, i.e., JAK2, MPL, and CALR, reinforces its categorization into this disease category [1, 49]. The diagnostic criteria for MDS/MPN-RS-T are summarized in Table 8 [5].
Macrocytic or normocytic anemia and thrombocytosis (≥450×109/L) are present and BM shows dyserythropoiesis (“MDS-like” feature) and large atypical megakaryocytes (“MPN-like” feature) similar to those observed in BCR-ABL1-negative MPN [1, 39, 49]. RS represent ≥15% of the nucleated erythroid cells. This finding is also observed in MDS-RS, which is a subtype of MDS and lacks thrombocytosis and megakaryocytic hyperplasia [39]. Large, hyperlobulated megakaryocytes and BM fibrosis are findings similar to MPN, such as essential thrombocythemia (ET) and primary myelofibrosis (PMF). However, ET generally shows no dysplasia and normocellular marrow for age [39]. MDS/MPN-RS-T shows a poorer prognosis than both MDS-RS and ET, and differential diagnoses are clinically important [39, 76].
The majority of the patients with MDS/MPN-RS-T have a normal karyotype (82.6%), while molecular aberrations are seen in >95% of the patients [49]. MDS/MPN-RS-T shows a molecular profile that combines the genetic features of MDS and MPN (Table 3) [27]. The mutations in SF3B1 have been shown to be a hallmark of MDS-RS, occurring in >80% of cases, and the mutations have been demonstrated in a similar proportion of patients with MDS/MPN-RS-T [12, 27, 32]. As the second most common molecular finding and accounting for the proliferative features of the disease, JAK2 mutations are observed in over 50% of patients, and somatic mutations in ASXL1, TET2, DNMT3A, and SETBP1 genes occur in about 10–30% of patients, while MPL and CALR mutations are observed in <5% of patients [27, 32, 104-108]. SF3B1 is concomitantly mutated with JAK2 in 50% of the cases, while cases with wild type SF3B1 tend to harbor ASXL1 and JAK2 mutations, and are negatively associated with SRSF2 and U2AF1 mutations [49, 105]. Although studies for these mutations are not required for the diagnosis of MDS/MPN-RS-T, the detection of these mutations supports the diagnosis and has prognostic implications as JAK2 mutated cases appear to have a better prognosis than the triple negative cases [5, 32, 39, 109, 110].
Spliceosome mutations are thought to represent the founder mutation in the MDS/MPN-RS-T, with nearly all SF3B1 wild-type cases harboring a mutation in an alternative splicing gene [27, 105]. The allelic frequency of SF3B1 is typically the highest among all identified mutations and the acquisition of an MPN-associated mutation (JAK2, CALR, or MPL) has been suggested to lead to the evolution of MDS-RS to MDS/MPN-RS-T [32, 105, 110].
MDS/MPN, unclassifiable: As the least characterized disease of the group, MDS/MPN-U includes cases of myeloid neoplasms showing both myeloproliferative and myelodysplastic features at the time of the initial presentation, but which do not entirely fulfill the diagnostic criteria for any specific subtype of MDS/MPN, MDS, or MPN [5, 27, 39]. Cases of MDS/MPN-U are more heterogeneous, not forming a unique subset of diseases; however, similar to other entities, these cases must present with cytosis and cytopenia that is not attributable to a known antecedent history of MDS or MPN [48]. In the 2016 WHO classification of myeloid neoplasms, MDS/MPN-U continues to be a provisional entity [1, 49]. The diagnostic criteria for MDS/MPN-U are summarized in Table 9 [5].
The differential diagnosis between MDS/MPN-U and aCML can be challenging, as MDS/MPN-U can show overlapping features with aCML. This differential diagnosis has relevant prognostic implications, since aCML shows poorer overall survival and shorter acute leukemia free survival than MDS/MPN-U [39, 86]. On PB, aCML generally shows higher WBC counts than those observed in MDS/MPN-U. The platelet count can be increased in MDS/MPN-U, while thrombocytopenia is more common in aCML.
There is no recurrent cytogenetic feature of MDS/MPN-U, with the majority (49%) of the patients having diploid karyotype, and +8 is seen in 15% of patients [49, 50, 111, 112].
There are no molecular genetic findings specific for MDS/MPN-U (Table 3) [5, 27]. The ASXL1 mutation has been identified as the most frequent molecular abnormality, occurring in 29–56% of patients with MDS/MPN-U [27, 113-115]. Mutations in TET2, SRSF2, and JAK2 are seen in approximately 25% of the patients, with mutations in other splicing genes—SETBP1, EZH2, and N/KRAS—each seen in approximately 10–15% of the patients [27, 80, 86, 111, 113-115]. TP53 mutations have been reported in 8–9% of cases, higher than what is reported in other MDS/MPNs, while a variety of other genes recurrently mutated across myeloid malignancies are seen at low frequencies [27, 114, 115]. The common driver mutations of MPN (i.e., JAK2, CALR, MPL) tend to exclude the diagnosis since they may be seen in MPN cases or in the disease progression. However, in cases without a previous history of MPN they might be useful to confirm the diagnosis [39, 108, 109, 116].
Hypoplastic MDS and aplastic anemia
Hypoplastic MDS represents approximately 10–15% of patients, and defined by a BM cellularity <25% on trephine biopsy or by an inappropriately reduced cellularity for their age in younger patients [117, 118]; however, their diagnosis is still an object of debate and has not been clearly established as a specific subtype in the 2016 WHO classification [117].
Hypocellularity in MDS may lead to difficulties in the differential diagnosis with AA [5]. Dysmegakaryopoiesis, dysgranulopoiesis, and the identification of sideroblasts (on BM aspirate) or clusters of blasts (identified by CD34 immunostaining of BM biopsy sections) are helpful in this distinction [5, 117, 119]. Mild, isolated dyserythropoiesis is very common in AA but cannot be used as a distinctive feature [117]. In younger patients, inherited BM failure syndromes should be considered and it is important to investigate the family history, extramedullary manifestations, and specific genes associated with BM failure with germline predisposition according to the 2016 WHO classification [1, 117].
Hypoplastic MDS is characterized by BM hypoplasia, a low rate of progression to acute leukemia, and a poor response to conventional MDS therapies [117, 120]. The clinical and laboratory markers situate hMDS in the middle of the clinical spectrum between normo/hypercellular MDS and AA (Table 10) [117].
The genomic landscape of hMDS resulted in-between AA and non-hMDS in terms of the number of somatic mutations (AA<hMDS<non-hMDS), variant allele frequency (AA< hMDS<non-hMDS), and involved genes (Table 10) [117]. The largest study illustrated that 38% of patients with h-MDS harbored at least one somatic mutation, albeit with a lower number of mutations per patient when comparing h-MDS with non-hMDS [118, 121]. In the study, SF3B1, TET2, DNMT3A, and SRSF2 mutations were the most prevalent in h-MDS [118, 121]. In addition, a significantly higher prevalence of PNH clones was present among the patients with h-MDS compared with non-hMDS [118, 121]. It is reported that 20% to 35% of AA patients have somatic mutations associated with hematologic malignancies, most characteristically in the ASXL1, BCOR, and BCORL1 genes [122].
Hypoplastic MDS may have both pathogenetic mechanisms similar to that of MDS (selective growth advantage of somatically mutated clonal hematopoietic progenitor cells) and AA (immune-mediated destruction of marrow precursors) [117]. The integration of laboratory and genetic features enabled the segregation of hMDS patients into two distinct groups: one highly consistent with the profile of myeloid neoplasms and the other more closely resembling AA, with no evidence of clonal disease [117, 118]. In summary, hMDS more likely represents a mixture of entities along a spectrum rather than a homogeneous in-between category [117].
CONCLUSION
MDS/MPNs have the characteristics of both MDS and MPN; however, their molecular characteristics usually do not show an “in-between” form, and it has been reported that they show various spectra within the group. Notably, MDS/MPN-RS-T is regarded as a “true hybrid” neoplasm in terms of both clinicopathological and molecular characteristics. Similarly, it is possible that hMDS may not be an AA-like hypocellular form of MDS but a heterogeneous group.
The advances in next-generation sequencing technology have led to the discovery of frequent mutations in MDS and overlap syndromes; however, most of the mutations are not specific for the diagnosis of these diseases. The established diagnostic criteria for these diseases based on clinical, morphologic, and laboratory features are still useful combined with the genomic data.
It is expected that a further revised classification of MDS and the overlap syndromes will focus on the roles of the mutations as therapeutic targets and prognostic indicators.
Unknown 
< 10% 
10-20% 
20-30% 
30-50% 
>50%
XML Download