Clinical and molecular features of patients with prefibrotic primary myelofibrosis previously diagnosed as having essential thrombocythemia in Japan

Objective: Prefibrotic/early primary myelofibrosis (pre‐PMF) and essential thrombocythemia (ET) exhibited different features of bone marrow; however, this is not always easy to judge objectively, making pathologists' distinction often suboptimal. In the WHO 2008 criteria, pre‐PMF was not defined as a subgroup of PMF; therefore, affected patients were at a higher risk of misdiagnosis with ET. In this study, we examined the prevalence of pre‐PMF patients among those previously diagnosed with ET in Japan. Method: We reviewed bone marrow specimens and clinical and molecular parameters of patients who were previously diagnosed with ET by the WHO 2008 criteria. Results: Among 107 ET patients, 13 patients were redefined as having pre‐PMF. Pre‐PMF patients exhibited a higher frequency of MPL mutation and increased platelet counts compared to true ET patients. Molecular analysis revealed the frequencies of high‐risk molecular mutations, such as ASXL1, EZH2, and SRSF2, were significantly increased in pre‐PMF patients than those in true ET patients. Conclusion: These results demonstrated the value of reexamining clinical records for patients diagnosed with ET by the WHO 2008 criteria and emphasized that adequate examinations of patients' bone marrow are crucial for an accurate diagnosis of pre‐PMF and ET.

Keywords: essential thrombocythemia; JAK2 V617F; MPL; myeloproliferative neoplasms; prefibrotic primary myelofibrosis

In the revised 4th edition of the World Health Organization (WHO) classification system for tumors of hematopoietic and lymphoid tissues (hereafter referred to as the WHO 2017 criteria), primary myelofibrosis (PMF) consists of the following two subgroups: prefibrotic/early PMF (pre‐PMF) and overt PMF.[1] Although both overt PMF and pre‐PMF are characterized by proliferation of predominantly abnormal megakaryocytes and granulocytes, they are different in terms of the degree of fibrosis in bone marrow. Overt PMF is distinguished by marked reticulin and/or collagen fibrosis, and pre‐PMF is characterized by no or minimal reticulin fibrosis. Because of the prominent megakaryocytic proliferation and minimal reticulin fibrosis in the bone marrow of patients with pre‐PMF, bone marrow specimens from patients with pre‐PMF and essential thrombocythemia (ET) are similar. Therefore, identifying the presence of aberrant megakaryocytes, a unique feature of PMF, is crucial for differentiating between bone marrow specimens from patients with these two diseases. Such differentiation is very important clinically because the life expectancy is significantly lower for pre‐PMF patients than that for patients with ET due to the rapid progression of fibrosis and a higher incidence of leukemia.[[2]] In the 4th edition of the WHO classification (WHO 2008 criteria), the possibility of a pre‐PMF (MF grading = 0) as a subgroup of PMF was described.[4] However, due to the difficulty of distinguishing megakaryocyte morphologies and the omission of pre‐PMF as a self‐standing diagnostic entity in WHO 2008 criteria, patients defined as having ET by the WHO 2008 criteria include those who would be diagnosed with pre‐PMF.[5] To investigate the frequency and clinical and molecular features of pre‐PMF among patients previously diagnosed with ET in Japan, we retrospectively reviewed the pathological results of bone marrow biopsies from patients who were previously diagnosed with ET.

Bone marrow biopsy specimens from patients who were previously diagnosed with ET by the WHO 2008 criteria between January 2010 and December 2017 at Juntendo University Hospital were examined by a central reviewer (M. I., a hematopathologist and an author of this manuscript). Bone marrow biopsy specimens were taken at initial diagnosis except for a subset of patients (n = 9) who was previously diagnosed in other institutions or whose (n = 2) bone marrow specimens at initial diagnosis were not available. In the course of the review process, only age, sex, peripheral blood cell counts, and bone marrow cell counts were supplied to the central reviewer. This study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of Juntendo clinical research and trial center (IRB: 17‐114).

Driver gene mutations such as JAK2 V617F, CALR exon 9, and MPL W515K/L were screened as previously reported.[[6]] For triple‐negative cases, whole‐exome sequencing for JAK2 and MPL was performed by next‐generation sequencing (NGS). Mutations on the ASXL1, EZH2, SRSF2, IDH1, and IDH2 genes were examined by NGS (Morishita S et al, manuscript in preparation).

A total of 107 patients who were previously diagnosed with ET according to the WHO 2008 criteria were enrolled in the study. The mutational statuses were as follows: JAK2 V617F (n = 58, 54.2%), CALR exon 9 (n = 25, 23.4%), MPL exon 10 (n = 8, 7.5%), and triple‐negative (n = 16, 15.0%). Through a central review, the pathologist classified the specimens as follows: 87 ET, 13 pre‐PMF, three indistinguishable ET and polycythemia vera (PV; ET/PV), one PV, one atypical chromic myeloid leukemia (aCML), and two non‐myeloproliferative neoplasm (MPN) specimens (Figure). After further evaluation of the clinical records and molecular data, we defined patients based on the WHO 2017 criteria as follows: 87 true ET, 13 pre‐PMF, 4 PV, 1 MPN unclassified (MPN‐U), and two reactive thrombocytosis patients (Figure A). Four patients whose bone marrow demonstrated PV or ET/PV were diagnosed with PV because their hemoglobin values met the WHO 2017 criteria and they harbored the JAK2 V617F mutation. The degree of inconsistency between diagnoses based on the WHO 2008 criteria and the WHO 2017 criteria was primarily caused by the different hemoglobin values given for the definition of PV in the WHO 2008 criteria (18.5 g/dL for males, 16.5 g/dL for females) and the WHO 2017 criteria (16.5 g/dL for males, 16.0 g/dL for females). The patient whose bone marrow specimen was diagnosed as aCML was positive for the JAK2 V617F mutation. Because patients harboring MPN‐associated mutations, such as those in JAK2, CALR, and MPL, are essentially excluded from receiving a diagnosis of aCML,[1] this patient was diagnosed with MPN‐U.

To compare clinical features between pre‐PMF patients and patients with true ET at the initial diagnosis, 12 pre‐PMF and 85 true ET patients whose precise blood count data at the first diagnosis were available were examined. Despite a previous study in which patients diagnosed with ET exhibited significantly lower leukocyte counts, platelet counts, and serum lactate dehydrogenase levels and higher hemoglobin levels than pre‐PMF patients,[5] only a significantly lower platelet count was observed in the true ET patients compared to that in the pre‐PMF patients in our cohort (P = 0.035; Table). Bone marrow fibrosis was observed in 69.2% (n = 9) of the patients with pre‐PMF but in only 27.6% (n = 24) of the patients with ET (P = 0.003; Table), which is consistent with previous observations.[5]

Comparison of the clinical parameters between true ET and pre‐PMF patients

1 Abbreviations: ET, essential thrombocythemia; Hb, hemoglobin; LD, lactate dehydrogenase; pre‐PMF, prefibrotic primary myelofibrosis; WBC, white blood cell count.

• 2 Median values and ranges are indicated. Eighty‐five ET and 12 pre‐PMF patients were analyzed; however, for age, all patients were analyzed.
• 3 Average values and ranges are indicated.
• 4 The chi‐square test was used for the male to female ratio and bone marrow fibrosis. The remaining variables were analyzed by the Mann‐Whitney test.

JAK2 V617F, CALR exon 9, and MPL exon 10 mutations were found in 53.8%, 15.4%, and 15.4% of the pre‐PMF patients, respectively, and in 52.9%, 26.4%, and 6.9% of the true ET patients, respectively (Figure B). Because we previously observed a higher JAK2 V617F allele burden in patients with PMF than that in patients with ET,[6] we examined whether such a difference existed between pre‐PMF and ET patients. Unlike a previous study in which patients in the prefibrotic stage of PMF exhibited a significantly higher JAK2 V617F allele burden compared to patients with ET,[9] the average JAK2 V617F allele burden among the pre‐PMF patients (42.6%) was only slightly higher than that among the true ET patients (31.4%; P = 0.091; Table). Patients harboring MPL mutation showed a tendency to be rediagnosed as pre‐PMF compared to those with other mutations (Figure C). In addition, unlike patients harboring the driver mutation, a subset of triple‐negative patients were rediagnosed as non‐MPN (Figure C).

To further investigate the molecular features of pre‐PMF and true ET, genetic risk factors for a poor prognosis in PMF patients, such as mutations in ASXL1, EZH2, SRSF2, IDH1, and IDH2,[10] were examined in 21 randomly selected true ET patients and 10 pre‐PMF patients. We found that the number of patients who harbored genetic risk factors was significantly higher among the pre‐PMF patients than that among the ET patients (P = 0.045 by the Mann‐Whitney test; Figure D). Although the worse outcome of pre‐PMF compared to ET [5] is probably not solely due to, and thus cannot be fully explained by, the presence of these gene mutations, these data implied the importance of distinguishing between these sets of patients.

In the present study, we reexamined the bone marrow specimens and clinical and molecular parameters of patients who were previously diagnosed with ET based on the WHO 2008 criteria. As a result, the prevalence of pre‐PMF defined according to the WHO 2017 criteria in Japanese patients who were previously diagnosed with ET was 12.1%, which is comparable to reports in the previous studies.[[5], [11]] Higher leukocyte counts, platelet counts, and serum lactate dehydrogenase levels and lower hemoglobin levels were associated with pre‐PMF rather than ET [[5], [12]] and these hematological parameters were proposed to predict pre‐PMF among ET patients.[14] Although we observed a similar trend in the above parameters in our cohort, only platelet counts showed significance (Table). The variation between cohorts may reflect the different ethnic groups that were studied.

The frequencies of JAK2 V617F mutations between pre‐PMF and true ET patients were similar as previously described,[5] while the frequency of MPL exon 10 mutations was greater in pre‐PMF vs true ET patients (Figure B), which is consistent with a recent report.[15] Several studies have reported that ET patients harboring MPL mutations exhibited a higher likelihood of progression to secondary myelofibrosis than those who were negative for MPL mutations.[[16]] Considering the higher frequencies of MPL mutations among patients redefined as having pre‐PMF rather than ET, ET patients harboring MPL mutations may have included pre‐PMF patients in the previous studies.

In conclusion, we found that the prevalence of pre‐PMF among Japanese patients who were previously diagnosed with ET according to the WHO 2008 criteria was comparable to that reported in a Caucasian cohort. Molecular analysis revealed that the frequencies of mutations on genes associated with a poor prognosis were significantly increased in pre‐PMF patients compared to those in true ET patients, demonstrating the value of reexamining clinical records for patients previously diagnosed with ET according to the WHO 2008 criteria. This study emphasized that despite the use of driver mutations in diagnosing MPN, adequate examinations of patients' bone marrow are crucial for an accurate diagnosis.

The authors would like to thank the members of the Department of Hematology at Juntendo University Graduate School of Medicine for their encouragement of this study. This work was supported in part by the Grant for Cross‐disciplinary Collaboration, Juntendo University (30‐26) (to YE), and JSPS KAKENHI Grants #15K15368 (to NK, MA, and SM), #16K19203 (to SM), and #17K09943 (to NK, MA, and MI). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The authors have no conflicts of interest to declare.