Clinical features of children with polycythemia vera, essential thrombocythemia, and primary myelofibrosis in Japan: A retrospective nationwide survey

Abstract Background Philadelphia‐negative (Ph‐negative) myeloproliferative neoplasms (MPNs), including polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF), are exceptionally rare during childhood. Thus, clinical features of pediatric Ph‐negative MPNs remain largely unknown. This study was therefore performed to address this. Methods We performed a retrospective study to collect clinical information of children diagnosed with Ph‐negative MPNs from 2000 to 2016 using questionnaires in qualified institutions in Japan. The results obtained from the questionnaire survey were then combined with those from the national registry data. Results Among 50 children identified, five had PV, 44 had ET, and one had PMF. Median age at diagnosis was 14.0, 9.0, and 0 years, respectively. Male to female ratio was 4:1, 21:23, and 1:0, respectively. Detection rates of the JAK2 V617F variant were 0/5 in PV and 9/39 in ET. Frequencies of complications, such as thrombosis and subsequent leukemia, were lower than complication frequencies in adults. We identified two children who developed subsequent leukemia, which has not been reported previously, and one of them died. Conclusion This is the first nationally representative survey of pediatric Ph‐negative MPNs. Given its rarity, an international collaboration with comprehensive genetic analyses might be needed to fully elucidate the clinical and genetic features.

Although hematologic findings are similar between cases of adult and pediatric MPNs in terms of clinical findings and genetic events, considerable differences have been indicated by several studies [3][4][5][6].
The diagnostic criteria for MPNs are defined by clinical features, laboratory findings, histopathology, and genetic analysis, which are basically utilized in the evaluation of adults [7][8][9]. Recently, three driver mutations in JAK2, CALR, and MPL were incorporated into the diagnostic criteria, but the prevalence of these mutations in pediatric patients is reported to be much lower than that in adults [6,10]. Consequently, the diagnosis of childhood MPN often relies on clinical and laboratory findings in combination with the exclusion of secondary thrombosis, rather than using molecular markers. However, given the low incidence of the disease, studies on clinical and laboratory features of children with MPNs on a national scale are scarce.

Prognostic factors also differ between adults and children with
MPNs. For instance, older age and past thrombotic events are often regarded as risk factors for thrombosis [11,12], but children being younger rarely have past histories of thrombotic events. There is currently no consensus on risk factors associated with complications or poor prognoses for children with MPNs, nor on the optimal clinical management of this group of patients [1,2].
In this article, we present the results of a retrospective survey of pediatric MPNs in Japan that was performed on a national scale. Our objective was to clarify the clinical characteristics, incidence, complications, and types of driver mutations among children with MPNs.

JSPHO registry survey
In addition to the questionnaires, we also referred to the JSPHO national registry [14], and the data of patients diagnosed with Phnegative MPNs during the 2000-2016 period were extracted. Patient information from the registry, such as date of birth and patient initials, was used to identify duplications with the data from the questionnaire surveys. For nonduplicated patients, their clinical information was collected by sending the same questionnaire to the physicians in charge of these patients. Patients who did not consent to this questionnaire survey were excluded from further analyses. The national registry data included diagnosed patients who were <20 years of age.

Overview
Responses to the questionnaires were received from 136 institutions, including all 106 board-certified educational institutes of JSPHO.
A total of 73 patients with Ph-negative MPNs from 44 institutions were reported. From the data available in the JSPHO national registry, 83 patients with Ph-negative MPNs were identified. The data from these two cohorts were combined and the duplicate cases or those with inadequate information were excluded. One patient who did not consent to this study was excluded from further analyses.
Ultimately, 57 cases were subjected to further analysis ( Figure 1).
One patient who experienced polycythemia with a family history and germline EPOR mutation was also excluded. Among the patients who were reported to be ET patients, platelet numbers decreased to normal values (<450 × 10 9 /L) during the follow-up period in six cases. Therefore, these patients were thought to have reactive thrombocytosis.
Of the patients ultimately included in the study, five were diagnosed with PV, 44 were diagnosed with ET, and one was diagnosed with PMF. One patient with PV and one patient with ET of the 50 total  (23)

Clinical characteristics
There was a predominance of males among the patients with PV and a slight predominance of females among patients with ET. Median age at diagnosis was 14.0 years old for patients with PV and 9 years old for patients with ET.

Biological characteristics
Results of blood analyses at presentation are presented in Table 1.
For patients with PV, the median leukocyte count was 8.0 × 10 9 /L, hemoglobin concentration was 193 g/L, and platelet count was 178 × 10 9 /L. For patients with ET, the median leukocyte count was 11.5 × 10 9 /L, hemoglobin concentration was 125 g/L, and platelet count was 1432 × 10 9 /L.

Complications
No patient with PV experienced a hemorrhagic or thrombotic event.
Of the five patients with PV, three received phlebotomy and three received no therapy during the follow-up period. One patient who developed subsequent leukemia had received hydroxyurea prior to leukemia development (Table 2).
Four patients with ET experienced hemorrhage and three patients with ET experienced thrombotic events. Several patients with ET required additional therapies, such as anagrelide or hydroxyurea (Table 2). Two patients who developed myelofibrosis and subsequent acute myeloid leukemia were identified, one patient with PV and one with ET.
The clinical courses of these two patients are summarized in Figure 2. Patient 2 was a 14-year-old boy diagnosed with ET according to an elevated platelet count (10,100 × 10 9 /L), bone marrow findings, and the detection of JAK2 V617F variant. He had no karyotype abnormality.
Initially, he was treated with anagrelide. He experienced myelofibrosis 6 months after the initial diagnosis and was then prescribed ruxolitinib.
Acute myeloid leukemia developed 5 months after the development of myelofibrosis. He received an human leukocyte antigen-haploidentical peripheral blood stem cell transplantation and was alive and in complete remission at last follow-up.

Primary myelofibrosis
Only one patient was diagnosed with PMF. The patient was male and at diagnosis was 11 months of age. He presented with no hemorrhage or thrombosis. This patient received immunosuppressive therapy followed by a stem cell transplantation. He was still alive at his most recent follow-up, 12 years after the initial diagnosis.

Comparison between ET and reactive thrombocytosis
Among cases who were reported to be ET patients, platelet numbers decreased to normal values (<450 × 10 9 /L) during the follow-up period in six cases. These patients were thought to have reactive thrombocytosis, and thus were excluded from the MPN group. Here, we compared F I G U R E 2 Clinical course of two patients with pediatric MPNs who developed subsequent leukemia. Each number over the arrow indicates the time after the diagnosis of their MPNs Abbreviations: ET, essential thrombocythemia; FAB, French-American-British; HU, hydroxyurea; IFN, interferon; PV, polycythemia vera. the clinical characteristics between ET patients and these patients with reactive thrombocytosis. The median age at diagnosis of patients with reactive thrombocytosis (7.5 years old) was less than that of patients with ET (9 years old). For blood analyses, similar results were obtained for patients with reactive thrombocytosis and patients with ET. One patient with reactive thrombocytosis showed 46, XY, inv(9)(p11q13), which was regarded as a normal variant [15]. The JAK2 V617F variant was present in none of the seven patients with reactive thrombocytosis. No patient was tested for CALR or MPL mutations. No patients with reactive thrombocytosis experienced either hemorrhage or thrombosis, and the majority of these patients were controlled with aspirin or observation only (Table 2).

DISCUSSION
To the best of our knowledge, this is the first nationally representative survey aimed at clarifying the clinical aspects of pediatric Ph-negative MPNs. We initially investigated the characteristics of children with Phnegative MPNs in Japan using a questionnaire survey. Responses to the questionnaires were received from all 106 board-certified educational institutes of JSPHO. Next, in an attempt to capture additional cases, we supplemented our approach by accessing data from the JSPHO national registry. The results from the two methods were combined, allowing us to perform the comprehensive survey reported.
Males predominated the patients with PV, whereas there was a slight female predominance in patients with ET, which was consistent with previous reports of large studies [3,4]. Median ages of these patients were also similar to the results of a recent systematic review [3]. Therefore, the biological characteristics of the patients included in our current study appear to be compatible with those of patients included in previous reports.
Blood analysis results at the time of diagnosis for the patients with PV in our current study differed from those of the previous systematic review; however, they were similar to those of a large Italian study [3,4]. According to a systematic review [3], the prevalence of JAK2 V617F mutation was only 24% among children with PV, which is much lower than that of adult patients [3,6,7,10]. Considering this number, we think that it is not surprising that our five PV patients did not have JAK2 V617F mutation. However, definitive conclusions cannot be drawn due to limitation in patient sample size.
In the current study, platelet numbers decreased to normal values during the follow-up period in six cases of thrombocythemia. These patients were thought to have reactive thrombocytosis, and thus were excluded from the MPN group. When we compared these patients with patients with ET, the blood analysis results at diagnosis appeared similar for the patients with ET, those with reactive thrombocytosis, and the patients included in the systematic review [3]. In addition, for patients with ET, frequencies of three driver mutations in our study were similar to the results reported in the systematic review [3]: specifically, JAK2 V617F 28% versus 31%, CALR 8% versus 10%, and MPL 0% versus 2%, respectively. However, the prevalence of all three major driver muta-needing further evaluation. Second, genetic testing of patients with MPNs in Japan is not centralized. As a result, there were several patients included in our study that did not receive sufficient genetic analysis. Consequently, we were unable to determine the precise genetic profile of children with MPNs. Comprehensive and centralized genetic analyses of children with MPNs should be performed in the future.
In summary, our current data revealed the clinical characteristics of children with MPNs and suggested that complications for children with MPNs are less frequent than those for adult patients. Therefore, the prognosis of children with MPNs is relatively good. However, it should be kept in mind that some children with MPNs may develop myelofibrosis and subsequent leukemia. Finally, given the rarity of pediatric Ph-negative MPNs, international collaborations including comprehensive genetic analyses might be needed to fully elucidate the clinical and genetic features of these hematopoietic disorders.

AUTHOR CONTRIBUTIONS
HI and AS designed the study, performed the evaluation, and wrote the paper. YM, NH, and SH provided patient information and important clinical data. All authors revised the manuscript for important intellectual content and approved the final version of the manuscript.