Myeloproliferative neoplasms (MPNs) are a group of related hematopoietic disorders characterized by a chronic disease state in which effective hematopoiesis is accompanied by enhanced expression of erythroid, myeloid, and/or megakaryocytic cell lines. These malignancies can be driven by a number of genetic mutations, including JAK2, MPL, and CALR.
Although the use of next-generation sequencing (NGS) techniques has greatly enhanced the knowledge surrounding a number of different malignancies, the roles and clinical implications for many mutations remain poorly defined in cases of MPNs.
To address these knowledge gaps, a sequencing study was undertaken at Massachusetts General Hospital (MGH) and Brigham and Women's Hospital (BWH) in 143 BCR-ABL1-negative MPN patients at those institutions from 2014 to 2017. The results from this study, which was headed by Valentina Nardi, MD, a staff pathologist at Massachusetts General Hospital, were recently published (Hum Pathol 2019;86:1-11).
Regarding their findings, Nardi observed, “SF3B1 mutations are prevalent within myeloid malignancies in myelodysplastic syndromes (MDS), acute myeloid leukemia (AML), and MDS/MPN with ring sideroblasts and thrombocytosis (MDS/MPN-RS-T). However, among MPNs, these mutations seem to be more prevalent in cases of primary myelofibrosis.”
MPNs are often characterized by overproduction of red blood cells, white blood cells, and/or platelets. MPNs have been classified by the World Health Organization (WHO) into a number of subcategories: chronic myeloid leukemia, BCR-ABL1-positive (CML), chronic neutrophilic leukemia, polycythemia vera (PV), primary myelofibrosis (PMF), essential thrombocythemia (ET), chronic eosinophilic leukemia, and MPN, unclassifiable (MPN-U). Recently, a review was published regarding the roles of known mutations (e.g., JAK2, MPL, and CALR, etc.) in MPNs (Semin Hematol 2019;56:7-14).
PMF typically presents as an excess of abnormal megakaryocytes with progressive marrow fibrosis. This disease is divided into two different stages: advanced/fibrotic (f-PMF) and early/prefibrotic (p-PMF). Similar to ET, canonical mutations in JAK2, CALR, or MPL are commonly observed. The presence of a small number of patients with so-called “triple negative” (TN) cases of PMF, which do not display canonical mutations to the three aforementioned genes, are also documented. Generally, the presence of mutated CALR in PMF is associated with a better survival relative to patients carrying JAK2 or MPL mutations, while TN cases display the worst outcomes.
In the peripheral blood, PV typically presents with an excess of red blood cells, while the marrow is characterized by hypercellularity. More than 98 percent of all PV cases show mutations to JAK2; of these, roughly 95 percent were the JAK2 V671F mutation in exon 14. Less frequently, in roughly 3 percent of cases, JAK2 mutations, including point mutations and small insertions or deletions, are found in exon 12. The JAK2 V671F mutation does not appear to be disease-specific, as it may also be present in PMF and ET (~55% each), and infrequently in MPN/MDS. The presence of mutations SRSF2, ASXL1, and IDH2 in PV have associations with worse clinical outcomes.
The presence of excessive platelets in the peripheral blood characterizes ET. The presence of JAK2 V617F (50-60%), CALR exon 9 (25-35%), and MPL W515 (~3%) mutations are often found in ET. Roughly 12 percent of ET cases lack these canonical mutations and are thus termed TN, although it is worth noting that the TN term may be somewhat of a misnomer in some instances, as there are documented ET cases in which noncanonical somatic or germline mutations to MPL and JAK2 have been identified (International Agency for Research on Cancer, 2017).
This subcategory of MPNs includes diseases that are not classifiable as one of the other six subcategories of MPNs. The frequency of MPN-U can vary widely in different studies, differing by more than 20 percent. However, most studies will typically show an incidence of 10-15 percent or even less.
Splicing factor 3B subunit 1 (SF3B1), which is a component of the U2 small nuclear ribonucleoprotein complex, is mutated in 80 percent or more of cases of myelodysplastic syndrome with ring sideroblasts and in MDS/MPN with RS and thrombocytosis (MDS/MPN-RS-T). However, mutations to this gene are much less frequent in MPNs, being reported in approximately 10 percent of cases.
Both MPNs and MDS/MPN-RS-T share a number of clinical and morphologic features, as some MPN patients may also display anemia and thrombocytosis. The ability to distinguish between SF3B1-mutant MPNs and MDS/MPN-RS-T is important when one considers disease prognosis and clinical management. MDS/MPN-RS-T, unlike “pure” MPNs, results in significant dysplasia in erythroid-derived cells. While mutation of SF3B1 appears to confer a favorable prognosis in MDS, the relevance of this mutation in the outcomes of those patients with MPNs is not known.
“In this study, we describe and characterize the prevalence and the clinicopathologic features of SF3B1-mutated MPNs and compare them with SF3B1-WT MPNs,” Nardi stated.
MPN cases from between 2014 and 2017 were identified at MGH and BWH via a search of their pathology databases. From these, 143 cases (MGH—36 patients and BWH—107 patients) of BCR-ABL1-negative MPNs in chronic phase were identified via retrospective review. Importantly, these cases excluded chronic eosinophilia leukemia/idiopathic hypereosinophilic syndrome, as well as those patients in accelerated or blast phase (i.e., greater than 10% bone marrow or peripheral blood blasts).
The study series included both newly and previously diagnosed MPN patients receiving a follow-up bone marrow biopsy. All cases had bone marrow biopsy, aspirate, and concurrent NGS-based panel testing results available. Classification of all cases, which was performed after review of the clinical records and the pathologic findings, was done using 2017 WHO criteria (International Agency for Research on Cancer 2017; 29-59, 93-94, 109-11).
All cases had diagnoses confirmed via review of peripheral blood smears, Wright-Giemsa–stained bone marrow aspirates, and hematoxylin and eosin– and Giemsa–stained slides of marrow core biopsies.
Targeted DNA sequencing was performed using either the Rapid Heme Panel or the Heme SNaPshot NGS v1 assay. The panel, which is a 95-gene panel based on the TruSeq Custom Amplicon kit, was utilized for 133 patients. The Heme SNaPshot NGS v1 assay, which is based on the Trusight Myeloid Sequencing Panel, targets 54 genes and was utilized for 10 patients. Both assays use the same target enrichment methods for and share 48 gene targets. Between these two assays, 101 distinct genes were interrogated; internal validation studies showed that there was 97 percent concordance for the mutations identified in samples analyzed using both assays.
The clinical features of 143 chronic phase BCR-ABL1-negative MPN patients for whom targeted DNA sequencing studies had been performed were analyzed. For these patients, the disease breakdown was as follows: PMF—51 cases (including f-PMF—35 cases and p-PMF—16 cases); ET—39 cases (including post-ET myelofibrosis—10 cases); PV—32 cases (including post-PV myelofibrosis—11 cases); and MPN-U—21 cases. A majority of patients, 87 (61%), received a new diagnosis, while 57 (39%) had been previously diagnosed with MPN.
Mutations to JAK2 were most frequently observed, being present in 100 cases, while mutations to CALR and MPL were present in 27 and seven cases, respectively. A total of 10 patients were termed TN, meaning that their MPNs had unmutated JAK2, CALR, and MPL.
The frequencies of the driver mutations among the different MPN states were generally in agreement with previously reported literature data. A total of 77 of the 143 cases showed mutations in 25 additional genes, while only four cases (four of the 10 TN cases) displayed no mutation. The mutational load did not differ significantly between the different types of MPNs or between fibrotic (MF grade 2-3) and nonfibrotic (MF grade 0-1) cases.
“However, f-PMF cases did show a marginally higher average number of mutations compared to ET cases,” Nardi observed.
It is of interest to note that mutations to a specific gene, ASXL1, did show a statistically significant association (p=.01) with f-PMF, as compared to p-PMF. No significant difference in mutation frequency among the MPN subtypes (p=.2) was noted in the 15 cases having mutated SF3B1.
There were a total of 15 patients having SF3B1-mutated MPNs; the largest number of these were nine cases of PMF, of which six were f-PMF and three were p-PMF. In other cases, the following MPNs were present: post-ET myelofibrosis—two cases; MPN-unclassified—two cases (with one case each of MF2 and MF3 fibrosis); post-PV myelofibrosis—one case; and PV—one case. The vast majority (13 of 15) had been previously diagnosed as MPN, with a duration between testing and diagnosis ranging from 2 to 34 years. Only two of the SF3B1-mutated MPNs were from newly diagnosed patients, both of which were p-PMF cases.
The presence of SF3B1 mutations was noted in approximately 10 percent of MPN diagnosed at two large academic medical centers. These mutations were more frequently observed in patients with advanced fibrosis, both primary (PMF and MPN-U with fibrosis) and secondary (post-ET myelofibrosis and post-PV myelofibrosis).
Although the authors noted that the possibility of referral bias could not be excluded, the study's findings would seem to suggest that SF3B1 mutation is not a rare finding in MPN, especially in PMF. Interestingly, MPNs having mutated SF3B1, with the exception of a subset that have RS present, are nearly identical in morphological, clinical, and molecular presentation to MPNs having wild-type (i.e., non-mutated) SF3B1.
“It is important,” Nardi noted, “[to] be aware that a subset of otherwise classic MPN cases may exhibit SF3B1 mutations and RS either at presentation or during disease evolution.”
Although a subset of cases with mutant SF3B1 presented with RS, no association with myelodysplastic progression was observed. She then added, “Patients having SF3B1-mutated MPNs frequently exhibit marrow fibrosis.”
The authors state it is crucial that the bone marrow and peripheral blood be carefully examined for morphologic dysplasia in order to obtain a proper diagnosis and avoid misdiagnosing patients with MDS/MPN-RS-T.
Regarding the study limitations, the authors cite the lack of a long follow-up period that did not allow conclusions to be drawn about patient outcomes, including overall survival. Since NGS-based mutational analyses are a fairly recent addition to the routine workup for MPN patients, the ability to reliably assess the impact of SF3B1 mutations upon prognosis is currently somewhat limited, given the long disease courses associated with MPNs.
Those having MPNs in the accelerated and blastic phases were excluded, as it was thought that their inclusion would introduce a selection bias for patients that had not experienced progression during the varying follow-up periods between first diagnosis and the time point included in the study.
In addition, since most participants with SF3B1 mutations had prior diagnoses, some even receiving therapies, comparison with patients having wild-type SF3B1 could be confounded, as the latter group were predominantly newly diagnosed. “However, the results of our study suggest that acquisition of SF3B1 mutations in MPN is not associated with a myelodysplastic evolution,” Nardi noted.
When asked about the larger implications of their observations about the presence of SF3B1 mutations in MPNs, she replied, “We wanted to stress that they do occur in patients with MPN. These patients, besides the likelihood of having a fibrotic marrow, have no other distinguishing features we could detect from SF3B1 wild-type patients, and also we wanted to make sure one would not jump to the conclusion the patient may have MDS or MDS/MPN-RS-T, which occasionally can be hard to distinguish from an advanced/fibrotic MPN.”
Regarding the potential of targeted therapies, Nardi explained, “There are spliceosome inhibitors now in clinical trials for patients with MDS and AML, so yes, perhaps this may be a valid target to go after.”
Richard Simoneaux is a contributing writer.