Polycythemia Vera: An Appraisal of the Biology and Management 10 Years After the Discovery of JAK2 V617F

Abstract

Polycythemia vera (PV) is a chronic myeloproliferative neoplasm that is associated with a substantial symptom burden, thrombohemorrhagic complications, and impaired survival. A decade after the seminal discovery of an activating mutation in the tyrosine kinase JAK2 in nearly all patients with PV, new treatment options are finally beginning to emerge, necessitating a critical reappraisal of the underlying pathogenesis and therapeutic modalities available for PV. Herein, we comprehensively review clinical aspects of PV including diagnostic considerations, natural history, and risk factors for thrombosis. We summarize recent studies delineating the genetic basis of PV, including their implications for evolution to myelofibrosis and secondary acute myeloid leukemia. We assess the quality of evidence to support the use of currently available therapies, including aspirin, phlebotomy, hydroxyurea, and interferon. We analyze recent studies evaluating the safety and efficacy of JAK inhibitors, such as ruxolitinib, and evaluate their role in the context of other available therapies for PV. This review provides a framework for practicing hematologists and oncologists to make rational treatment decisions for patients with PV.

INTRODUCTION

Polycythemia vera (PV) is a clonal hematologic malignancy characterized by a pronounced symptom burden, including fatigue, pruritus, and symptomatic splenomegaly, along with an increased risk of thrombosis and the potential for evolution to myelofibrosis (MF) and secondary acute myeloid leukemia (sAML). The history of PV was previously divided into three eras.¹ The first era was marked by its initial descriptions by Vaquez in 1892 and Osler in 1903, and the introduction of therapeutic phlebotomy (TP). The second era, beginning in 1939, was characterized by diagnostic advances, allowing for distinction of relative and absolute polycythemia through measurement of red cell mass. Importantly, this era witnessed the first application of radioactive phosphorous as a treatment for PV, and the subsequent report of acute myeloid leukemia (AML) in patients treated with [³²]P. The third era marked the efforts of the Polycythemia Vera Study Group (PVSG), spanning from 1967 to 1997, which led to the development of formal diagnostic criteria and raised awareness about the effect of TP, long-term consequences of certain cytoreductive therapies, and consequently, the introduction of hydroxyurea (HU).

In 1997, PV entered the molecular biology era,¹ and the discovery of the JAK2 V617F mutation in 2005² marked the start of the fourth era of PV. The discovery of JAK2 V617F improved understanding of PV pathogenesis, facilitated diagnostic capability, and led to the reclassification of PV as a myeloid neoplasm. This landmark finding not only reinvigorated research efforts into understanding the genetic basis of PV but also ushered in the development of novel therapeutics (Fig 1). To date, the defining features of each era are relevant in the current molecular era of PV. TP remains a cornerstone of management, with the target hematocrit level being validated only recently.³ Distinction of PV from secondary causes of erythrocytosis remains paramount, especially in patients lacking a JAK2 mutation. Associations between cytoreductive treatments and leukemia continue to polarize the field. Because 10 years have passed since the seminal discovery of JAK2 V617F, it is important to critically reappraise what is known about the pathogenesis and management of PV.

Fig 1.

Polycythemia vera (PV): the JAK2 discovery era, 2005 to present. ET, essential thrombocythemia; IFN, interferon; MF, myelofibrosis; MPN, myeloproliferative neoplasm; peg-IFN, pegylated IFN.

EPIDEMIOLOGY

The prevalence of PV is estimated to be 44 to 57 cases per 100,000 persons,⁴ and approximately 148,000 persons are living with PV in the United States. This may represent an underestimation, given that JAK2 V617F can be latent and PV asymptomatic, as evidenced by its detection in large, unselected population surveys.⁵ These data indicate that PV is not rare, but rather, a relatively common hematologic malignancy. Although the median age at diagnosis is 61 years, PV has been diagnosed in all age groups.^6–8 PV was historically thought to occur more commonly in males, but females may account for 75% of patients younger than 40 years in whom PV is diagnosed.^6,8 Younger patients with PV appear to be at risk for unique complications, including abdominal venous thrombosis, and their disease appears to transform to MF and/or sAML as frequently as it does in older patients.⁸

DISEASE PATHOGENESIS: GENETIC UNDERPINNINGS OF PV

The genetic basis of PV was largely unknown until the discovery of an activating mutation, V617F, in the JAK2 gene encoding tyrosine kinase in 2005.^2,9–11 The V617F mutation, located within the autoinhibitory pseudokinase domain of protein JAK2, leads to JAK2 kinase hyperactivity and aberrant cytokine-independent signaling through the JAK-STAT axis. It is now established that approximately 96% of patients with PV are positive for JAK2 V617F. In addition, about 2% to 3% harbor mutations in exon 12 of JAK2,^12,13 whereas rare patients with erythrocytosis and low erythropoietin (Epo) levels may carry mutations in LNK (SH2B3), a negative regulator of JAK-STAT signaling.¹⁴ Clearly, activation of JAK2 and consequent downstream signaling is a hallmark of PV.

JAK2 V617F also occurs in approximately 50% of patients with essential thrombocythemia (ET) and PMF.¹⁵ These findings indicate that JAK2 V617F is not specific to PV and raise the question of how a single mutation can result in such phenotypic diversity. JAK2 V617F mutant allele burden is typically highest in PV and lowest in ET, and JAK2 V617F homozygosity resulting from an acquired uniparental disomy is more common in PV than ET.^16–18 Patients with JAK2 V617F–positive ET tend to have higher hemoglobin (Hgb) levels and lower platelet counts than patients with JAK2 V617F–negative ET.^19,20 Therefore, higher levels of JAK2 V617F may create a more PV-like phenotype.

Although differences in JAK2 V617F allele burden appear to play an important role in differentiating myeloproliferative neoplasm (MPN) disease phenotypes, additional genetic factors likely contribute as well. Recent studies have revealed additional mutations in many patients with PV, although none are nearly as prevalent as JAK2, nor do any appear to be entirely specific to PV.^21,22 Mutations in TET2 are present in a subset of MPNs, including PV, and they can occur either before (ie, pre-JAK2) or after the acquisition of JAK2 V617F.^23,24 Recent studies demonstrated that the order of acquisition of TET2 versus JAK2 mutations in MPNs may modulate disease phenotype.^25,26 Of note, patients in whom the JAK2 mutation was acquired first were more likely to present with a diagnosis of PV than with ET.

Large-scale genomic studies have clarified the spectrum of genetic changes in PV.^21,22 Nangalia et al²¹ performed exome sequencing on 151 patients with MPN including 48 with PV. The median number of mutations in PV was 6.5, which was approximately half that found in 39 patients with MF. These findings suggest that a relatively small number of pathogenic mutations are needed to initiate and establish PV. Although CALR mutations were identified in most patients with ET and MF lacking JAK2 V617F, CALR mutations have rarely been observed in PV.²⁷ TET2 mutations were found in our (8.3%) of 48 patients in the PV cohort, whereas SF3B1 and DNMT3A mutations were each identified in two patients. In a separate study of 31 patients with PV, exome sequencing revealed mutations in several additional genes, including ASXL1.²² In MF, ASXL1 mutations have been shown to be associated with worse overall survival and increased likelihood of transformation to sAML.²⁸ However, the implications of such mutations in PV are unclear. Several of these recurrently mutated genes play roles in epigenetic regulation (eg, TET2, ASXL1, DNMT3A) or RNA splicing (eg, SF3B1), and they are also mutated in other MPNs as well as myelodysplastic syndromes (MDS). Overall, these recently reported studies have highlighted the genetic complexity of PV, but the specific influence of individual mutations on PV disease phenotype remains incompletely understood.

DIAGNOSIS

The diagnosis of PV is currently based on the World Health Organization 2008 criteria, which include the presence of a JAK2 mutation and increased red cell mass, in which sex-based Hgb values are used as a surrogate: greater than 16.5 g/dL in women and greater than 18.5 g/dL in men. Other criteria include consistent bone marrow histology, low Epo levels, and the formation of endogenous erythroid colonies (EEC).²⁹ Unfortunately, EEC testing is not routinely available and has been difficult to standardize. In addition, Epo levels are normal in 20% of cases and can also be suppressed in ET.³⁰ A proposed alternative system opposed using a Hgb threshold and instead suggested testing of nuclear red cell mass;³⁰ this was based on studies that upgraded ET to PV when this diagnostic test was used.^31–33 Reports describe decreased survival in patients with a masked PV presentation, that is, those with consistent features but Hgb levels below the World Health Organization threshold, as a result of a higher rate of transformation to MF and/or sAML compared with those with overt PV.³⁴ It is unclear if masked PV is an early form or distinct PV phenotype with a different prognosis.

New diagnostic criteria have been proposed that incorporate JAK2 mutational testing, allowing for a lower Hgb threshold of 16 g/dL and 16.5 g/dL in women and men, respectively. The guidelines omit EEC testing, include characteristic bone-marrow histopathological abnormalities as a major criterion, and retain suppressed Epo levels.³⁵ Red cell mass and plasma volume studies, which are sensitive and specific determinants of blood volume not confounded by plasma volume expansion, are not available to most practitioners.

NATURAL HISTORY

Major complications of PV include its thrombotic tendency and its potential for later evolution to MF and/or sAML. More recently, a pronounced symptom burden, which often includes fatigue, pruritus, and symptoms from splenomegaly has been appreciated (Fig 2). This symptom burden negatively affects quality of life and is prominent, even in classically low-risk patients.³⁶

Fig 2.

Patient-reported symptoms in polycythemia vera. Incidence of each symptom is estimated based on data from Scherber et al,^35a Emmanuel et al,^35b Johansson et al,^35c and Abelsson et al.^35d

Historically, thrombosis has been the most common consequence, leading to serious morbidity and implicated in as many as 37% of deaths.^37,38 The Gruppo Italiano Studio Policitemia followed 1,213 patients for as long as 33 years (mean, 6 years) and found thromboses in 41% of patients, including 20% with a thrombotic event as the presenting symptom. They found another 6% in the 2 years before diagnosis and 19% during follow-up.³⁷ Historically reported yearly thrombosis rates have been 2.5% to 4%. However, more recent studies show lower rates of thrombosis, perhaps because of more aggressive therapy.^3,7 Predominant is arterial thrombosis, particularly cerebrovascular accidents, whereas acute coronary syndromes have been the most common cause of death. Venous thrombosis is less common than arterial thrombosis, but it has been shown to adversely survival.⁷ Hepatic vein thrombosis, or Budd-Chiari syndrome, and thromboses of the portal and/or mesenteric vein are also strongly associated with PV and are often a presenting feature. Microvascular symptoms, including headache, visual change, dizziness, and erythromelalgia, are also prominent and affect quality of life (Fig 2).

Traditional risk factors include advanced age and previous history of thrombosis (Table 1). Although the magnitude of risk is greater in older patients, younger patients are not protected. In a retrospective study of 120 patients with PV younger than 45 years at diagnosis, thrombosis rates were similar to those of patients with PV older than 65 years at diagnosis (27% v 31% [younger v older]).⁸ However, the involved vascular bed differed, because venous events were more common in younger patients, especially women presenting with abdominal venous thrombosis (AVT). This observation suggests that young age and sex should be considered in the assessment of thrombosis risk.

Table 1.

Risk Factors Associated With Increased Morbidity and Mortality in Patients With Polycythemia Vera

Risk Factors
For thrombosis
Age > 60 years
Previous history of thrombosis
Leukocytosis39–42*
Increased JAK2 V617F allele burden6,43–45*
High-risk gene expression profile46*
For transformation to myelofibrosis or secondary acute myeloid leukemia
Older age47
Longer disease duration48
Leukocytosis41
Exposure to phosphorus-32, pipobroman, or chlorambucil7,48
Risk factor associated with decreased survival
Older age7
Leukocytosis7
History of venous thrombosis7
Abnormal karyotype7

^*Emerging or controversial risk factor.

Leukocytosis is an emerging but inconsistent risk factor for thrombosis (Table 1).^39–42 The degree of thrombocytosis has not been correlated with thrombotic risk.⁴⁹ Associations with JAK2 V617F allele burden have also been inconsistent. Some investigators reporting that increased thrombosis rates were associated with higher allele burden, whereas others found no association.^6,43–45 A high-risk gene expression profile has been associated with an increased thrombotic risk and other adverse outcomes despite similar leukocyte counts and JAK2 V617F allele burdens.⁴⁶

Transformation to MF and/or sAML is a significant cause of mortality in PV and is likely largely related to disease duration (Table 1). In large studies, progression to MF was reported in 9% to 21% of patients with PV, while evolution to sAML has been documented in 3% to 10% of patients with PV.^7,48,50 Risk factors for transformation include older age and leukocytosis.^41,47 An increased risk of sAML has been associated with certain cytoreductive therapies.^7,48 The effect of the JAK2 V617F allele burden on disease evolution is not well established.⁵¹ A potential role for mutations in other genes (eg, ASXL1) in modulating transformation risk is under investigation. Most prognostic systems developed for primary MF have excluded patients with post-PV MF. In an Italian study of 647 patients with PV, the median survival of 68 patients (10%) who developed MF was 5.7 years.⁴⁷ Multivariable analysis revealed three risk factors that adversely affected survival: WBC count greater than 30 × 10⁹/L, Hgb concentration less than 10.0 g/dL, and platelet count less than 100 × 10⁹/L. The prognosis for sAML after MPN, including PV, is dismal, with median survival of approximately 5 months.⁵² Overall life expectancy for patients with PV is variable but significantly shorter than that of age-matched control subjects. A recent retrospective study demonstrated that survival varied from 10.9 to 27.8 years in high- and low-risk patients, respectively, and that it was most highly influenced by advanced age, leukocytosis, and history of venous thrombosis.⁷

THERAPEUTIC DECISION MAKING

Goals of Therapy

The primary goals of treatment are to reduce the risk for thrombosis, ameliorate the PV symptom burden, and prevent MF and/or sAML transformation. Unfortunately, current therapies are not effective for this latter purpose, and controlling blood counts may have less of an effect than anticipated. The European Leukemia Net (ELN) expert panel defined a complete response as a hematocrit less than 45% without phlebotomy, platelet count less than 400 × 10⁹/L, leukocyte count less than 10 × 10⁹/L, normal spleen size, and no disease-related symptoms.⁵³ However, in a study of 261 patients with PV who were given HU and followed up for a median of 4.4 years,⁵⁴ no association was observed between achieving an ELN or a hematocrit response and better survival or fewer vascular complications. No particular platelet or leukocyte count could be shown to be protective against thrombosis.⁵⁵ The ELN response criteria were updated in 2013, but they have not been validated to predict survival or progression to MF or AML.⁵⁶ In the subsequent sections, PV treatments are carefully reviewed.

ANTIPLATELET THERAPY

The routine use of aspirin for thromboprophylaxis is largely based on the European Collaboration on Low-dose Aspirin in Polycythemia Vera, or ECLAP, study in which 253 patients were randomly assigned to receive aspirin 100 mg per day and 265 patients, to receive placebo (Table 2). Total follow-up was 1,478 person-years.⁵⁷ Although the study did not meet its primary endpoint, which was cumulative rate of nonfatal myocardial infarction, nonfatal stroke, or death from cardiovascular causes, the observed trend was strongly in favor of aspirin therapy (relative risk [RR], 0.41; P = .09). The second primary endpoint was the cumulative rate of nonfatal myocardial infarction, nonfatal stroke, pulmonary embolism, major venous thrombosis, or death from cardiovascular causes; it did achieve statistical significance (RR, 0.4; P = .03). A strong trend toward reducing the risk of major cerebrovascular events (RR, 0.32; P = .08) and major venous thrombotic events (RR, 0.49; 95% CI 0.24 to 0.74; P = .003) was observed. Evaluation of safety end points revealed no statistically significant difference in the incidence of major hemorrhages (RR, 1.62; P = .6).⁵⁷ To our knowledge, no published data describes the use of other antiplatelet agents, including clopidogrel.

Table 2.

Summary of Selected Clinical Trials Involving Patients With Polycythemia Vera

Study	No. of Subjects	Treatment	Results and Notes
PVSG 01	431	TP v phosphorus-32 plus TP v chlorambucil plus TP	Increased risk of thrombosis with TP alone but decreased risk of transformation to myelofibrosis or secondary acute myeloid leukemia and better overall survival; note initial HCT goal was < 50% and subsequently adjusted to < 45%
CYTO-PV	365	Low HCT, goal < 45% v high HCT, goal 45% to 50%	Lower HCT target associated with significantly lower rate of cardiovascular death and major thrombosis
ECLAP	518	Aspirin 100 mg per day v placebo	Aspirin significantly lowered the combined risk of nonfatal myocardial infarction, nonfatal stroke, pulmonary embolism, major venous thrombosis, or death from cardiovascular causes; no significant difference in incidence of major hemorrhage
French PEG-IFN-α2a	37	PEG-IFN-α2a	95% CR at 12-month median follow-up JAK2 V617F allele burden decreased from median of 45% to 3% after 36 mo; molecular CR in 30% after 4 years
MDACC PEG-IFN-α2a	43	PEG-IFN-α2a	76% CR and 18% CMR at 42-month median follow-up; failure to achieve CMR was associated with mutations other than JAK2 V617F
RESPONSE	222	Ruxolitinib v BAT	Primary endpoint, a composite of phlebotomy independence and spleen volume reduction, was achieved in 21% of the ruxolitinib group v 1% of the BAT group; symptom improvement of > 50% occurred in 49% of the ruxolitinib group v 5% of the BAT group

Abbreviations: BAT, best available therapy; CR, complete response; CMR, complete molecular response; CYTO-PV, Cytoreductive Therapy in Polycythemia Vera; ECLAP, European Collaboration on Low-Dose Aspirin in Polycythemia Vera; HCT, hematocrit; MDACC, MD Anderson Cancer Center; PEG-IFN, pegylated interferon; PSVG, Polycythemia Vera Study Group; RESPONSE, randomized, open label, multicenter phase III study of efficacy and safety in polycythemia vera subjects who are resistant to or intolerant of hydroxyurea: JAK inhibitor INC424 tablets versus best available care; TP, therapeutic phlebotomy.

PHLEBOTOMY

The notion that increased blood viscosity is a consequence of elevated hematocrit as the basis for hypercoagulability serves as the rationale for phlebotomy, which remains an essential component of modern PV therapy.⁵⁸ Recently, after decades of debate, a hematocrit target of 45% or less was associated with a lower risk of cardiovascular morbidity and mortality than that associated with a goal hematocrit of 45% to 50% (Table 2).³ Of note, patients in the lower hematocrit arm may have had more aggressive HU use and a lower leukocyte count, factors that may modify the effect of TP.

However, historical studies raised concerns about TP. In the PVSG 01 study, 431 patients with PV were randomly assigned to receive TP alone, [³²]P plus TP, or chlorambucil plus TP (Table 2).⁵⁹ Thrombosis-free cumulative survival was worse in the TP-alone arm. The need for more than four therapeutic phlebotomies a year in the TP-alone arm was associated with increased thrombotic risk. This led to the belief that frequent TP requirements were detrimental and necessitated cytoreductive therapy. However, overall survival and risk of evolution to MF and sAML were lower in the TP alone arm. Certain concerns about TP may be unfounded. Iron deficiency is a consequence of repeated TP; but, in fact, it should be viewed as a therapeutic goal to further limit erythropoiesis. Although iron deficiency can be associated with a number of clinical signs and symptoms, including glossitis, dysphagia, cheilosis, koilonychia, fatigue, global weakness, cognitive deficits, neuromuscular disturbances, and pica syndrome, these rarely prompt treatment discontinuation.⁶⁰

CYTOREDUCTIVE THERAPY: BALANCING RISKS AND BENEFITS

Cytoreduction is typically prescribed for high-risk patients, that is, those with advanced age and/or thrombosis history. Uncontrolled disease-related symptoms, including those due to splenomegaly; intolerance of phlebotomy; extreme thrombocytosis; and progressive leukocytosis may also prompt therapy.⁶¹ HU is considered a first-line cytoreductive therapy in PV, but evidence to support its efficacy in preventing thrombosis is typically extrapolated from studies of ET.^62,63 Its use in PV is based on a perceived balance of safety and efficacy rather than data from randomized studies showing thrombosis reduction. Rates of thrombosis appeared to be lower than those of historical control subjects treated with TP alone (9.8% v 32.8%), and lower rates of sAML (5.9%) were noted for HU compared with chlorambucil (17%) and [³²]P (10%) from the PVSG 01 and PVSG 08 protocols.⁶⁴ The rate of sAML in historical control subjects treated with TP was 1.5%, but this was not statistically different from the rate in patients treated with HU.

Data from large studies have not implicated single-agent HU as a cause of AML.^38,65 In a recent retrospective, multicenter study of 1,545 patients with PV, HU was implicated only when it was used in combination with pipobroman.⁷ In a population-based study, researchers utilized the Swedish MPN registry of 11,039 patients to compare 162 patients with MPN (110 with PV) and AML and/or MDS transformation to 242 matched control subjects (176 with PV). They found no association between any cumulative dose of HU and transformation.⁶⁶ In a randomized study, the French Polycythemia Study Group compared HU and pipobroman. Higher rates of AML/MDS and shortened survival were observed in patients treated with pipobroman during long-term follow-up over 16 years.⁶⁷ Rates of AML and/or MDS in the HU arm at 10, 15, and 20 years were higher than expected at 6.6%, 16.5%, and 24.2%, respectively. However, this may reflect the natural history of PV with long follow-up.

The controversial association between HU and AML will be difficult to resolve with certainty, given inherent challenges to the conduct of a definitive clinical trial. Intrinsic as well as therapy-related contributions likely exist.⁶⁸ The risk of leukemia with cytoreductive therapies cannot be evaluated by absolute terms but by the magnitude of risk,⁶⁸ which appears to be lower with HU than with other agents such as chlorambucil, busulfan, [³²]P, and pipobroman.^{7,65,66,69,70} These latter treatments may still have a role in the treatment of elderly patients with PV or those with medical comorbidities that are likely to substantially shorten survival in their own right.^71,72 Careful patient selection, improved sophistication in assessing vascular risk, and establishment of evidence-based or consensus-driven guidelines for rational use of cytoreductive agents, including HU, are needed.

EMERGING TREATMENTS

Interferon Therapy

Renewed enthusiasm for use of pegylated interferon (PEG-IFN), an agent without known leukemogenic effects, originated from the French phase II study of PEG-IFN-α-2a (PEG-IFN-α-2a) in patients with newly diagnosed PV (Table 2).⁷³ PEG-IFN-α-2a was started at 90 μg weekly for 2 weeks, with a dosage escalation in the absence of hematologic response. At 12 months, 95% of patients achieved a complete hematologic response (CHR). Only three patients (8%) had stopped treatment. Sequential monitoring of JAK2 V617F allele burden in 29 patients showed a decrease in 26 (90%), with a median decrease from 45% to 22%, 5%, and 3% after 12, 24, and 36 months, respectively. Complete molecular response (CMR), that is, undetectable JAK2 V617F, was achieved in seven patients. CMR lasted from at least 6 to at least 18 months and persisted after discontinuation in five patients. It should be noted that the clinical relevance of allele burden reduction is not yet clear. No vascular events were recorded.⁷³ In a recent update, the cumulative incidence of CMR was 14% and 30% of patients at 2 and 4 years, respectively.⁷⁴ Moreover, 27% of patients treated maintained CHR after PEG-IFN-α-2a was discontinued.

Another group evaluated PEG-IFN at the dosage of 90 μg weekly in 43 patients with newly diagnosed and previously treated disease (Table 2).⁷⁵ After a median follow-up of 42 months, CHR was achieved in 76% of patients with a median time of 40 days; CMR was reported in 18% of patients. No patient who achieved CMR had a relapse during the course of treatment with PEG-IFN-α-2a. No grade 4 toxicities occurred, and grade 3 adverse effects were infrequent. To date, 32% of patients have been removed from the study. In 18% of patients, treatment discontinuation was directly related to PEG-IFN-α-2a. Two patients remained in CHR and CMR 4 and 18 months after treatment discontinuation. Patients who did not achieve CMR had a higher frequency of mutations outside the JAK-STAT pathway and were more likely than others to acquire new mutations during therapy. Patients with both JAK2 V617F and TET2 mutations at the onset of therapy had a higher JAK2 V617F allele burden and a less significant reduction in JAK2 V617F allele burden than those of patients with a JAK2 mutation and wild-type TET2.⁷⁵

Although this is considered first line,⁶¹ high-quality evidence supporting this recommendation is not yet available. Rather, randomized trials of PEG-IFN-α-2a versus HU in patients with newly diagnosed PV as well as those with HU intolerance/resistance or previous AVT are ongoing (ClinicalTrials.gov identifiers: NCT01259856 and NCT01259817). The completion of such trials is necessary to enable an objective evaluation of the long-term tolerability, safety, and therapeutic value of PEG-IFN therapy for patients with PV, including the potential to reduce thrombotic risk or evolution to MF and/or sAML.

JAK1/2-INHIBITION THERAPY

In a phase II clinical trial, of 34 patients with PV treated with the JAK1/2 inhibitor ruxolitinib, 97% achieved a clinicohematologic response, with a complete response achieved by 59% according to modified 2009 ELN criteria.⁷⁶ Durable hematocrit control, reduction in leukocyte and platelet counts, and a reduction in palpable splenomegaly of 50% or greater by week 24 were observed, along with improvement in the PV symptom burden. Within 4 weeks of treatment, most experienced reduction of 50% or greater in the severity of pruritus, night sweats, and bone pain. After 3 years of ruxolitinib, nearly one quarter of patients achieved a 50% or greater reduction in the JAK2 V617F allele burden.⁷⁶

A randomized, open-label, phase III clinical trial of ruxolitinib versus best available therapy (BAT) was subsequently conducted in patients with PV with splenomegaly, HU resistance or intolerance, and an ongoing requirement for phlebotomy (Table 2).⁷⁷ The primary endpoint was the composite of independence from phlebotomy and reduction in splenic volume. Overall, 222 patients were randomly assigned to receive ruxolitinib or BAT, which most commonly included HU despite the fact that HU intolerance or resistance was an entry criterion. The combined primary endpoint was achieved by 21% versus 1% of patients (P < .001) randomly assigned to receive ruxolitinib versus BAT, respectively. At week 48, 91% of patients maintained their response. Treatment with ruxolitinib was associated with greater benefit for both hematocrit control without phlebotomy (ruxolitinib, 60%; BAT, 20%), and splenic volume reduction (ruxolitinib, 38%; BAT, 1%). Symptom burden, as assessed by using the MPN symptom-assessment form, improved by greater than 50% in 49% versus 5% of patients randomly selected to be given ruxolitinib versus BAT. Among patients receiving ruxolitinib, grade 3 or greater adverse events included anemia in 1.8% and thrombocytopenia in 5.5%. These results led to the approval of ruxolitinib in December 2014 by the US Food and Drug Administration for use in patients with PV and HU resistance or intolerance.

Because of the study design, the investigators could not address the effect of ruxolitinib therapy on the incidence of thromboembolic events or disease progression.⁷⁷ Important issues that may influence use beyond the treatment of patients with HU resistance or intolerance are a lack of information about long-term safety, with a potentially increased risk of herpes simplex virus noted; the high cost; and whether the natural history can be altered by JAK 1/2-inhibition in PV. Therefore, this agent is currently considered a second-line therapy.

CONCLUSION

Despite important breakthroughs in our understanding of PV, several important concepts require further insight. Although the discovery of JAK2 V617F represented a major breakthrough, our understanding of the genetic basis for PV remains incomplete; likely contributing factors include additional host modifiers and additional somatic mutations. PV diagnostic criteria continue to evolve, with consideration of the appropriate Hgb threshold. Thrombosis risk assessment remains generic rather than personalized, and the ongoing challenge is to identify surrogate markers that can enable individualized assessment of thrombosis risk. The same challenge exists in identifying early on those with a potential for aggressive disease, yet additional prognostic indicators, apart from disease duration, JAK2 allele burden, and leukocytosis are needed. With regard to therapeutic goals, meeting response criteria has less of an impact than anticipated; further, the natural history of PV creates inherent limitations to trial design that make it difficult to discern both long-term toxicity and efficacy. Although HU is considered a front-line cytoreductive therapy, its use is based more on historical, nonrandomized data, ease of administration, and opinion.⁷⁸ Similarly, PEG-IFN is intriguing, but has also been prematurely adopted as a front-line agent before the scrutiny of a randomized, phase III clinical trial.⁷⁸

With regard to second-line therapy, patients with HU resistance or intolerance now have JAK inhibition as an approved option. Lack of information about long-term safety, high cost, and whether the natural history can be altered by JAK 1/2-inhibition in PV are important questions that may affect use beyond the patient population that is HU resistant or intolerant. Despite expanding treatment options, a need remains for new therapeutic interventions in PV, and many novel strategies can be seen as a potential complement to JAK inhibition, including HSP90 inhibition, targeting of other activated pathways (eg, PI3K, ERK, AKT/mTOR) or targeting epigenetic modifiers such as histone deactyleases.⁷⁹ Finally, preclinical data has demonstrated that an MDM2 inhibitor combined with PEG-IFN-α-2a can synergistically increase apoptosis and reduce JAK2 V617F allele burden in xenograft models.⁸⁰ Such a clinical trial is now underway (No. NCT02407080). Hopefully, the therapeutic potential of emerging treatments will be realized, and the next era will be characterized by significant alteration in the natural history of PV.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Disclosures provided by the authors are available with this article at www.jco.org.

AUTHOR CONTRIBUTIONS

Conception and design: Brady L. Stein, Stephen T. Oh, Alison R. Moliterno, John Mascarenhas, Ronald Hoffman

Collection and assembly of data: Brady L. Stein, Stephen T. Oh, Dmitriy Berenzon, Gabriela S. Hobbs, Marina Kremyanskaya, Raajit K. Rampal, Camille N. Abboud, Mark Heaney, Elias J. Jabbour, Rami Komrokji, Alison R. Moliterno, Ellen K. Ritchie, Lawrence Rice, John Mascarenhas, Ronald Hoffman

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Polycythemia Vera: An Appraisal of the Biology and Management 10 Years After the Discovery of JAK2 V617F

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or jco.ascopubs.org/site/ifc.

Brady L. Stein

Consulting or Advisory Role: Incyte, Sanofi Oncology

Speakers' Bureau: Incyte

Stephen T. Oh

Consulting or Advisory Role: Incyte, Gilead, CTI Biopharma, Pfizer

Speakers' Bureau: Incyte

Research Funding: Incyte (Inst), Gilead (Inst), CTI Biopharma (Inst)

Dmitriy Berenzon

Consulting or Advisory Role: Incyte

Gabriela Hobbs

No relationship to disclose

Marina Kremyanskaya

Research Funding: Incyte

Raajit Rampal

Consulting or Advisory Role: Incyte

Camille N. Abboud

Honoraria: Teva, ARIAD

Consulting or Advisory Role: Gerson Lehrman Group

Speakers' Bureau: Teva

Research Funding: Pfizer (Inst), Merck (Inst), Novartis (Inst), Sanofi/Genzyme (Inst), CTI Biopharma (Inst), Cellerant Therapeutics (Inst)

Kenneth Adler

No relationship to disclose

Mark Heaney

Consulting or Advisory Role: Blueprint, Incyte, Novartis

Elias Jabbour

No relationship to disclose

Rami Komrokji

Consulting or Advisory Role: Incyte, Celgene

Speakers' Bureau: Incyte, Celgene

Research Funding: Incyte (Inst), Celgene (Inst)

Travel, Accommodations, Expenses: Incyte, Celgene

Alison R. Moliterno

Honoraria: Incyte

Consulting or Advisory Role: Incyte

Ellen K. Ritchie

Consulting or Advisory Role: Incyte

Speakers' Bureau: Celgene, Incyte

Travel, Accommodations, Expenses: Celegene

Lawrence Rice

Consulting or Advisory Role: Amgen, Achillion, Apellis, Alexion, Synergy, Sideromics

Speakers' Bureau: Novartis, Amgen

John Mascarenhas

No relationship to disclose

Ronald Hoffman

Research Funding: Geron, Incyte, Novartis, Roche