Optimal clinical management of patients with primary myelofibrosis and post-essential thrombocythemia/polycythemia vera myelofibrosis is a challenge, given the typically advanced age of presentation and variability of the disease course and prognosis. Current medical therapeutic options have not demonstrated an impact on the disease course, which exceeds the palliation of disease-related extramedullary hematopoiesis and alleviation of cytopenias. In contrast, allogeneic stem cell transplantation (SCT) can lead to ‘cure’ but is limited due to patient's age or comorbidities. Currently, in patients, who are reasonable candidates, SCT (frequently with a reduced intensity conditioning regimen) is employed for intermediate- to high-risk disease. Current pharmaco-medical therapy is used as a bridge to transplant, or instead of transplant in poor transplant candidates. Pathogenetic insights, especially the discovery of the Janus kinase (JAK)2V617F mutation, have ushered in a host of new potential therapeutic agents that may augment the role of medical therapy. Similarly, the boundaries of transplantation continue to alter with strategies that decrease conditioning-related toxicity, improved antimicrobial prophylaxis and decreased graft-versus-host disease. The potential for continued improvements in both medical and transplant therapy suggests that for the immediate future the optimal choices for an individual patient will remain potentially volatile and present complex decisions.
Disease presentation and natural history
The BCR-ABL-negative myeloproliferative disorder of primary myelofibrosis (PMF)1 and the therapeutically indistinguishable advanced forms of essential thrombocythemia and polycythemia vera (that is, post-ET/PV myelofibrosis) is an extremely challenging disorder for both patients and clinicians. Indeed myelofibrosis is a very heterogeneous disorder with respect to age of onset, presenting features, phenotypic manifestations and prognosis. Therapeutic options for this debilitating and frequently fatal disorder range from observation and supportive care to medical therapy or even hematopoietic stem cell transplantation (SCT). Optimal management for an individual patient is a very complex decision-making process involving a thorough estimate of the individual's prognosis, the ability to tolerate a therapy and the risks and expectations of pharmaco-medical therapy, splenectomy or transplant-based therapy. Clinically meaningful advances in both medical and transplantation therapy continue to alter appropriate treatment choices in this disorder.
Patients presenting with myelofibrosis range from being asymptomatic (discovered on investigation for occult causes of leukocytosis or splenomegaly) to being severely debilitated. The median age at diagnosis is typically in the seventh decade of life (median 67 years),2 although patients may clearly present in the third to fifth decades of life.3, 4 Symptoms related to PMF can be broken down into one of three main categories: myeloproliferative, cytopenias and constitutional.
The natural history of myelofibrosis is quite variable as patients may experience morbidity or mortality directly from the clinical consequences of myelofibrosis or from transformation to acute myeloid leukemia (AML) (that is, PMF or post-ET/PV blast phase). Death, however, may well arise from the complications of myelofibrosis (that is, cytopenias) exacerbating underlying comorbidities such as coronary artery disease, or given the median age of the population afflicted, death from an unrelated comorbidity is always possible. Curative therapy in PMF is currently possible only with allogeneic hematopoietic SCT (AHSCT). AHSCT is, however, associated with a risk of mortality as well as morbidity that influences its application. Essential to the decision-making process in determining the appropriateness of low- or high-intensity therapies for PMF patients is the issue of prognosis. The prognosis for PMF patients is quite variable, with various prognostic features helpful to stratify the suspected outcomes in PMF patients. The various prognostic systems currently in use are discussed below (see Table 6 for comparison of PMF prognostic systems).
Current strategies for stem cell therapy
Stem cell transplantation can be performed with either autologous or allogeneic (syngeneic) stem cells. However, allogeneic SCT is the only curative treatment approach.
Only a few studies have investigated autologous SCT after high-dose chemotherapy in patients with myelofibrosis.5, 6, 7 In a pilot study, 21 patients received peripheral blood SCT after myeloablative conditioning with busulfan (16 mg per kg body weight). The median time of leukocyte and platelet engraftment was 21 days, however, graft failure was seen in 14% of the cases, and some patients had delayed engraftment up to 96 days for leukocyte recovery, and more than 200 days for platelet recovery. Three patients (14%) died from nonrelapse causes. Erythroid response without transfusion for more than 8 weeks was seen in 10 of 17 patients, and symptomatic splenomegaly improved in 7 of 10 patients. Four of eight patients showed some improvement of thrombocytopenia, and a reduction of bone marrow fibrosis was seen in 29% of the cases. In a small series of three patients who received autologous SCT after conditioning with treosulfan (42 g m−2 body surface), a prolonged leukocyte reconstitution of 28–38 days was seen, and a significant reduction of spleen size was noted.5 Overall, autologous SCT is a potential treatment approach that can relieve disease-related symptoms such as splenomegaly, but the curative potential is very unlikely.
Allogeneic SCT after standard myeloablative conditioning
Despite the increased use of allogeneic SCT in treatment of hematological malignancies, major concerns about performing this treatment approach in patients with PMF early arose due to the nature of bone marrow histopathology, which is distorted by fibrosis and it was hypothesized that this might lead to a higher risk of engraftment failure. The first small reports and case reports in the early 1990s, however, demonstrated that engraftment was feasible and regression of bone marrow fibrosis was achievable after myeloablative conditioning.8, 9 Furthermore, in relapsed patients after allografting, a graft-versus-myelofibrosis effect could be demonstrated by donor-lymphocyte infusions.10, 11 Larger retrospective studies including more than 50 patients with myelofibrosis were reported by Guardiola et al.12 in a combined analysis of European and American centers, and from Deeg et al.13 reporting the results of the Fred Hutchinson Cancer Center in Seattle. The latter study was recently updated and included 95 patients.14 In the retrospective European-American study, Guardiola et al.12 reported on 55 patients with myelofibrosis who underwent conventional allogeneic SCT. The median time from diagnosis to transplantation was 21 months (range, 2–266 months). Most of the patients received conditioning regimens, including total-body irradiation (TBI). Matched related donors were used in the majority of the patients (n=49). According to the Lille risk score, 76% had intermediate- or high-risk disease. Splenectomy prior to transplantation was done in 27 patients. Graft failure occurred in 9% of the patients and nonrelapse mortality at 1 year was 27%. The 5-year overall and disease-free survival were 47 and 39%, respectively. Patients with low risk according to the Lille score had better overall survival than patients with intermediate- or high-risk disease (85 versus 45–30%). In a multivariate analysis, hemoglobin <10 g per 100 ml, abnormal karyotype, high-risk disease according to the Lille score and presence of osteosclerosis had an adverse impact on survival.
The Seattle group recently updated its results of allogeneic SCT in patients with Philadelphia chromosome-negative myeloproliferative diseases.14 The study included 104 patients, and standard myeloablative conditioning was used in 95 patients. Most patients had myelofibrosis (n=62), or advanced ET (n=18), or PV (n=12). The median age was 49 years (range, 18–70 years), and stem cell donors were related (n=59), syngeneic (n=3) or unrelated donors (n=45). The conditioning regimens mainly consisted of TBI plus cyclophosphamide, or busulfan plus cyclophosphamide. Nonrelapse mortality at 5 years was 34%. The overall survival at 7 years was 61%. Improved survival in a multivariate analysis was seen for patients conditioned with a targeted busulfan/cyclophosphamide regimen, younger patients, patients with a low comorbidity score and patients with high platelet count at transplantation. The outcome of 320 patients with myelofibrosis after myeloablative conditioning reported to the International Bone Marrow Transplant Registry and to the National Marrow Donor Program showed a day 100 mortality of 22% after sibling transplantation, of 42% after matched unrelated donor transplantation and of 27% after alternative family donor transplantation, respectively. The overall survival at 5 years was 39, 31 and 31%, respectively. Patients with favorable status (Karnofsky index >90%, absence of blasts in peripheral blood) had survival of 81% at 5 years.15 The results of studies with standard myeloablative conditioning followed by allografting are listed in Table 1.12, 13, 14, 16, 17, 18, 19, 20, 21
Taking together the results of the published studies on conventional allogeneic SCT, the median age of the patients was between 38 and 54 years and significantly lower than the median age of about 67 years for patients with myelofibrosis at the time of diagnosis. The 5-year survival in the larger studies was 47 and 61%, respectively,12, 14 indicating the curative potential of standard myeloablative allograft. However, the nonrelapse mortality at 1 year ranged from 20 to 48% (Table 1), and one of the most significant prognostic factors for impaired survival was increasing age of the patient. In the European-American study, patients <45 years of age experienced 62% survival, whereas patients older than 45 years had a survival of only 14% mainly due to a higher treatment-related mortality22 In one study, SCT from an unrelated donor was associated with a worse outcome, while other studies did not show differences between related and unrelated SCT.13, 14, 22 Other factors for improved survival in multivariate analysis were conditioning with a targeting busulfan regimen,13, 14 high platelet count and low comorbidity index,14 low risk according to the Dupriez (Lille) score,12, 13 normal karyotype,12, 13 hemoglobin >10 g per 100 ml12 and non-osteosclerosis 12 (Table 2).
Allogeneic SCT after dose-reduced conditioning
Allogeneic SCT after standard myeloablative conditioning chemotherapy has been shown to be a curative treatment approach in patients with myelofibrosis, but the major limitation of this approach is that it can most reliably be performed only in younger patients with good performance status. The introduction of so-called ‘non-myeloablative’ or ‘dose-reduced’ or ‘toxicity-reduced’ conditioning regimen is based on the concept of shifting eradicating tumor cells from high-dose chemotherapy to the immunologically mediated graft-versus-tumor effect. The potential advantages are less treatment-related morbidity and mortality, and a broader application in elderly patients. Evidence for an immunologically mediated graft-versus-myelofibrosis effect comes from reports in relapsed patients after allogeneic SCT that show a remarkable reduction of bone marrow fibrosis after donor-lymphocyte infusion.10, 11 The feasibility of dose-reduced conditioning in patients with myelofibrosis was first reported in a small series of case reports.23, 24, 25, 26 In the two largest studies published so far, 27, 28 patients up to their seventh decade of age were included. In the German study,28 21 patients with a median age of 53 years (range, 32–63 years) were included. The conditioning regimen consisted of busulfan (10 mg per kg), fludarabine (180 mg m−2) and anti-thymocyte globulin (ATG; Fresenius, Graefelfing, Germany) (30 mg per kg for related and 60 mg per kg for unrelated donors), followed by SCT from related (n=8) or unrelated (n=13) donors. No primary graft failure was observed, and leukocyte and platelet engraftment were seen after a median of 16 days and 23 days, respectively. Complete donor chimerism was seen in 95% of patients at day +100. Acute graft-versus-host disease (GvHD) grade II–IV and grade III/IV was observed in 48 and 19% of the patients, respectively. Chronic GvHD occurred in 55% of the patients. Nonrelapse mortality was 16% at 1 year. After a median follow-up of 22 months (range, 4–59 months), the 3-year estimated overall and disease-free survival was 84%. The second study, from the MPD-Research Consortium, also included 21 patients, with a median age of 54 years (range, 27–68 years). Different conditioning regimens were used, including melphalan plus fludarabine, cyclophosphamide plus fludarabine, thiotepa plus fludarabine and TBI (2 Gy) in combination with fludarabine. All patients had intermediate or high risk according to the Lille score. One graft failure was observed. More than 95% donor chimerism was seen in 18 patients. Nonrelapse mortality was 10%, and overall 2-year survival was 87%. Table 3 shows other published results including three or more patients. The most commonly used regimens were busulfan/fludarabine based, and melphalan/fludarabine based. In comparison to the reported myeloablative transplantations (Table 1), the median age of the patients was more than 10 years older with 51–58 years. The nonrelapse mortality was lower than 20%, and overall survival after a relatively short follow-up was between 84 and 100%. Only in one small study,30 including nine patients, the nonrelapse mortality exceeded 40%, and the overall survival was only 56% at 1 year. A retrospective comparison between conventional and reduced-intensity conditioning regimen was performed in 26 patients from the Swedish Group for Myeloproliferative Disorders.29 Despite the fact that in the reduced-intensity group (n=10), the median age was 14 years older than in the myeloablative group (n=17), the nonrelapse mortality was lower in the reduced-intensity conditioning regimen than in the myeloablative group (10 versus 30%). In a prospective trial of the Chronic Leukemia Working Party of the European Group for Blood and Marrow Transplantation, 104 patients with PMF and a median age of 55 years (range, 32–68) were included. The risk profile distribution was low risk with constitutional symptoms (18%), intermediate risk (n=58%) and high risk (n=19%). All but three patients received peripheral blood stem cells as graft source either from related (n=31) or unrelated donor (n=69). All but one (1%) patient showed leukocyte and platelet engraftment after a median of 18 and 21 days, respectively. The median duration of leukocyte aplasia was 9 days (range, 3–21). Acute GvHD grade II–IV occurred in 19%, and severe aGvHD grade III/IV in 7%, while chronic GvHD was seen in 32% of the patients. Nonrelapse mortality at 1 year was 19% (95% CI, 11–27%) and significantly lower for patients younger than 50 years of age (0 versus 27%; P=0.004) and for patients with low risk versus intermediate/high-risk disease (0 versus 27%; P=0.02). The cumulative incidence of relapse at 3 year was 29% (95% CI, 15–43%). After a median follow-up of 16 months, the 3-year overall and event-free survival was 70% (95% CI, 60–80%) and 55% (95% CI, 42–68%). Significant factors for improved 3-year overall and event-free survival were age <50 years (92 versus 62%, P=0.003 and 79% versus 46%, P=0.004), and low versus intermediate/high risk (100 versus 62%, P=0.01 and 72 versus 48%, P=0.02), while no impact on survival was seen for cytogenetic abnormalities, Janus kinase (JAK)2 mutation status and donor (related versus unrelated).31 In another single center study including 39 patients who received a dose-reduced conditioning consisting of thiotepa and cyclophosphamide favorable factors for improved survival were human leukocyte antigen-identical donor, Karnofsky-Index of 100%, time from diagnosis to transplantation of <1 year and previous splenectomy.32
Even if the follow-up of these studies is rather short, they demonstrated that reduced conditioning is effective and feasible with acceptable toxicity even in older patients.
Currently available pharmaco-medical therapies for myelofibrosis
Currently available medical options are all considered in the armamentarium for myelofibrosis with the following caveats: (1) No medical therapy has been clearly proven to prolong survival, (2) no medical therapy has been proven to alter the natural history of the disorder, and given these first two caveats and (3) all current medical options are inherently palliative in intent. Given this latter fact, the benefit of utilizing currently available pharmaco-medical options must be weighted against their known toxicities. Since these medical options are explicitly chosen to palliate aspects of myelofibrosis, we will consider them in the context of the symptom they aim to palliate. For further details see Table 4).
In a meta-analysis of published data on the use of erythropoietin (Epo) in PMF, Rodriguez et al.70 reported a response rate of 33% with drug doses of up to 600 units per kg per week. Patients with endogenous serum-Epo levels of <125 mU ml−1 had the highest likelihood of response. In a more recent study, Cervantes et al.59 reported an overall response rate of 45% in 20 PMF patients treated with 30 000 units per week. A serum-Epo level of <125 U l−1 was again associated with a significantly higher likelihood of response in a multivariate analysis. Thus, Epo-treatment should be restricted to patients with anemia and inadequate Epo levels (that is, below the above-mentioned threshold. Darbopoietin has also been tried in a pilot study of 20 myelofibrosis patients with a 40% response rate (initial dose 150 μg per week increased to 300 μg per week), again with responses only in those patients with inadequate baseline serum Epo values.60
Androgens have been beneficial for the palliation of myelofibrosis-associated anemia.71 The most comprehensive analysis has been performed using 600 mg per day of danazol in which a 37% response rate was observed, with a median time to peak response of 5 months (range, 1–9 months). Patients who were anemic, yet not transfusion requiring were the most likely to respond.72 Currently, a trial of androgens is reasonable in palliating patients with myelofibrosis-associated anemia (after being satisfied that there is no occult prostatic cancer). There are no data comparing the palliative benefits of Epo to androgen therapy or using those agents in combination in PMF patients with anemia.
Tumor necrosis factor (TNF)-α is implicated in the pathogenesis of PMF-associated cytopenia as well as that of constitutional symptoms.73 Etanercept (Enbrel; Immunex, Seattle, WA, USA) is a TNF-α antagonist that has been shown to be helpful in abrogating TNF-α-associated disease features in rheumatoid arthritis. In a pilot study involving 22 patients with PMF, etanercept treatment (25 mg twice weekly, subcutaneous injections) for up to 24 weeks was well tolerated and produced 60% response in constitutional symptoms.74 However, only modest benefit was observed in either improving anemia or reducing splenomegaly (20%).
The diagnosis of extramedullary hematopoiesis (EMH) is based upon history, clinical examination and supportive radiographic studies (ultrasound, radionuclide, computed tomography and MRI). In the asymptomatic patient, there does not appear to be any advantage to uncovering occult EMH.
Hydroxyurea is an oral, well-tolerated, nonspecific myelosuppressive agent that can reliably control the leukocytosis as well as thrombocytosis associated with PMF.65 Treatment of leukocytosis is clinically useful only if it is extreme and symptomatic. We advise control of thrombocytosis in the presence of risk factors for thrombosis. Splenomegaly is not as responsive to hydroxyurea and might require a higher dose (2–3 g per day). Hydroxyurea therapy may potentially exacerbate anemia or thrombocytopenia (if present), but hydroxyurea-associated anemia might be ameliorated by the concomitant use of Epo therapy.
Alkylating agents have activity in PMF by causing a direct, nonspecific myelosuppression, and therefore may potentially palliate symptoms associated with myeloproliferation. A clinical trial was recently reported in PMF utilizing low doses of melphalan.67 Over a 7-year period, 104 patients with PMF were treated with 2.5 mg of oral melphalan three times a week with 66% of patients achieving a response after a median of 7 months of therapy. However, it was concerning that blastic transformation occurred in 26% of the study cohort.
Therapeutic splenectomy in PMF may result in extreme thrombocytosis, leukocytosis and accelerated hepatomegaly. Palliative benefit from 2-chlorodeoxyadenosine, a purine nucleoside analogue, has been reported to provide beneficial cytoreduction in such instances.75 The drug has been successfully used in PMF in either of two schedules commonly employed for this agent: 4 to 6 monthly cycles of treatment with either 0.1 mg per kg per day intravenously by continuous infusion for 7 days, or 5 mg m−2 intravenously over 2 hours for 5 consecutive days. In a recent update of our experience, responses were observed in 55, 50, 55 and 40% of patients for hepatomegaly, thrombocytosis, leukocytosis and anemia, respectively.67 Responses were frequently durable and lasted for a median of 6 months after discontinuation of treatment.
Therapy with interferon (IFN)-α has been utilized in patients with PMF based on its cytoreductive properties. IFN-α suppresses the proliferation of megakaryocytes and fibroblasts and has shown significant inhibition of collagen production in experimental models of myelofibrosis.76 Moreover, IFN-α inhibits in vitro and in vivo fibrogenic cytokines such as transforming growth factor β (TGF-β) and platelet-derived growth factor (PDGF), and in vitro IFN-α has shown an antiproliferative effect on hematopoietic progenitor stem cells.77 Aside from control of leukocytosis and/or thrombocytosis in patients exhibiting hyperproliferative features and sporadic instances of bone marrow fibrosis regression,78 clinical trials of IFN-α in PMF have been generally disappointing.79 Results have been equally underwhelming with IFN-γ,80 and pegylated IFN-α (PEG-IFN-α) 81 due to lack of tolerance and negligible response rates. In one study, subcutaneous PEG-IFN-α-2b (PEG-Intron) was administered weekly to 36 patients with BCR-ABL-negative MPDs, including 11 patients with PMF.80 None of the patients with PMF responded.
Medical therapy of PMF (post-ET/PV blast phase)
Blast phase of myelofibrosis occurs in approximately 10–15% of patients. In a recent series of 91 consecutive myelofibrosis patients who experienced blood pressure (BP), transformation was clinically heralded by organomegaly, worsening constitutional symptoms, anemia, thrombocytopenia and leukocytosis.82 Therapeutically these patients represent a very serious challenge. Leukemic Transformation (LT) was found to be fatal in 98% of patients after a median of 2.6 months (range, 0–24.2). Survival was equally poor regardless of whether patients received strictly supportive care, low intensity therapy (that is, low-dose cytarabine) or induction. Among 24 patients who received AML-like induction therapy (that is, cytarabine/idarubicin; mitoxantrone/etoposide and high-dose cytarabine), 42% had a brief return to chronic phase and endured a 33% treatment-related mortality. No complete remission was observed, and no activity was observed with the agent gemtuzumab. No patient was successfully brought to allogeneic transplant, suggesting that transplant should be considered well in advance of BP if desired. (further details are as per authors63, 66, 68, 69 given in Table 4).
Evolving pharmaco-medical therapy: what is next?
Improved understanding? Improved outcomes with targeted therapy?
Advances in the medical therapy of primary and post-ET/PV myelofibrosis will need to arise from the use of therapies specifically targeting the pathogenetic mechanisms of disease. Recent and ongoing trials have focused upon the cytokine milieu, the increased angiogenesis or the fibrosis inherent in this disorder. Further targeted strategies have focused upon tyrosine-kinase inhibition, aberrant farnesylation or DNA hypermethylation. Future strategies also hope to yield clinical benefit through the inhibition of the constitutively active JAK2-tyrosine-kinase mutation in the majority of MPD patients.
Targeting the aberrant cytokine milieu of myelofibrosis
The group of immunomodulatory cytokine inhibitors and antiangiogenic agents are collectively known as IMIDs. Initial pilot studies with thalidomide in PMF were dose escalating in nature beginning at doses of 100 mg per day.61 Subsequent low-dose (50 mg per day) thalidomide with a prednisone taper 62 resulted in significant responses observed for anemia (67%), thrombocytopenia (75%) and splenomegaly (33%), however without any definitive improvement in marrow or karyotypic abnormalities. Subsequently, lenalidomide (a second-generation, more potent cytokine inhibitor IMID) has been evaluated in 68 patients with symptomatic PMF with overall response rates of 22% for anemia, 33% for splenomegaly and 50% for thrombocytopenia,64 however with improved marrow histology in responders. Mirroring the activity of lenalidomide in del(5q) myelodysplastic syndrome, PMF patients with an abnormality of chromosome 5 seem to respond best to this agent.83 Given the promising results obtained with lenalidomide, a randomized, placebo-controlled, international clinical study to determine the activity of pomalidomide (20 000-fold more potent than thalidomide) in inhibiting TNF-α (with or without a prednisone taper) in PMF and post-ET/PV MF is currently under way with results eagerly anticipated (Table 5).
Additional angiogenesis inhibitors evaluated with minimal clinical activity and no impact on intramedullary manifestations of disease in myelofibrosis include both vatalanib (PTK787/ZK 222584: an oral inhibitor of the VEGF receptor-1 (VEGFR-1) and VEGFR-2 tyrosine kinases)92 and SU5416 (synthetic inhibitor of VEGFR-2 kit, and flt-3).93 The ability of the nonspecific ‘angiogenesis inhibitors’ such as thalidomide and lenalidomide to provide more profound responses (in the absence of improvements in angiogenesis) suggests that the clinical activity is independent of the inhibition of angiogenesis (further details as per authors 84, 85, 89, 90 in Table 5).
Inhibition of fibrogenesis
TGF-β1 plays a central role in the prominent bone marrow stromal reaction observed in patients with PMF. Pirfenidone (inhibitor of PDGF, TNF-α and TGF-β) was tested in a prospective study, but failed to show any significant clinical activity in PMF, in which 16 of 28 (57%) patients were withdrawn because of disease progression or drug intolerance, and only one patient had an improvement in anemia and splenomegaly.94 Similarly, clinical trials have been initiated based on preliminary data in a murine model of myelofibrosis suggesting that the proteasome-inhibitor bortezomib inhibits activation of the nuclear factor (NF)-κB pathway and decreases plasma concentration of TGF-β1, thus inhibiting the development of myelofibrosis.86 Several compounds that inhibit TGF-β1-mediated signaling are being explored preclinically in myelofibrosis, including GC-1008,88 AP-11014,95 LY580276 (IC50 175 nM),96 pyrazole 2 (IC50 18 nM),97 SB 505124 (IC50 47 nM)98 and SD-208.99
Inhibition of tyrosine kinases
The use of the tyrosine-kinase inhibitor imatinib mesylate for patients with PMF was based on its inhibitory activity against PDGF-mediated signaling and the reduction of bone marrow fibrosis and microvessel density observed in patients with chronic myeloid leukemia (CML), for which imatinib is a standard therapy.100 Results from all the phase II trials of imatinib in patients with PMF reported to date are modest to toxic.101, 102 Dasatinib (BMS-354825) is a dual Src- and Abl-kinase inhibitor, and a 300-fold more potent Abl-kinase inhibitor than imatinib,87 and effectively inhibits PDGF receptor-β (IC50 28 nM).102 Based on these preclinical data and on the results obtained with imatinib, clinical trials are underway.
Farnesyl transferase inhibitors
Tipifarnib (R115777), a non-peptidomimetic farnesyl-transferase inhibitor, was administered at 600 mg orally twice daily for 4 weeks of every 6-week cycle to eight patients with PMF.103 Two of them had a significant decrease in splenomegaly, one had normalization of white blood cell count and differential and one became transfusion independent.87 A second larger trial 104 was performed among 34 patients (with a lower 600 mg per day dosing). Tipifarnib resulted in little improvement in anemia, but achieved a clinically relevant decrease in organomegaly in 11 patients (33%)—splenomegaly in three, hepatomegaly in five and both in an additional three-, many of whom had previously failed hydroxyurea. Responses observed did not significantly correlate with reductions in bone marrow fibrosis, osteosclerosis, neo-angiogenesis or resolution of baseline karyotypic abnormalities. Whether this latter benefit is independent of tipifarnib-induced myelosuppression is unclear.
Targeting DNA hypermethylation
Aberrant cytosine phosphate guanine island hypermethylation in regulatory areas of tumor suppressor genes leading to inactivation is commonly encountered in human cancer.105 Methylation of p15INK4B, p16INK4A and the retinoic acid receptor β has been observed in advanced-stage PMF.106 Azacitidine and decitabine are DNA methyltransferase inhibitors that induce reactivation of methylated genes.91 Both agents are approved by the US Food and Drug Administration for the treatment of patients with MDS, and are currently being investigated in phase II studies in PMF.
Inhibition of JAK2
Although the currently identified molecular defects in MPD patients, such as the JAK2V617F,36, 37, 38, 40 do not yet fully explain many issues of MPD pathogenesis, they provide an exciting and hopefully more fruitful therapeutic target. There have already been multiple reports of agents in development, which have demonstrated the ability to inhibit the aberrant JAK2V617F, along with the wild type JAK2, such as TG101209,107 Go6976,108 erlotinib,109 MK0457110 and CEP-701.111 Intriguingly, in primary cells from wild-type JAK2 patient, who have the c-MPLW515L/K, growth inhibition can similarly be accomplished by JAK2 inhibitors, such as TG101209.107 These latter observations suggest the possibility that even in JAK2 wild-type MPD patients, a growth dependence on the JAK–signal transducers and activators of transcription (STAT) pathway may exist and agents targeting this pathway may be active regardless of the JAK2 mutation status. This hypothesis is further supported by the continual discovery of aberrations in this pathway, as in exon 12 of JAK2 gene in JAK2V617F-negative PV patients.112 Several clinical trials with these intriguing agents are scheduled to begin in 2007 and early 2008.
Who, when or how to transplant?
Since allogeneic SCT is increasingly used as curative treatment option even in older patients, the ever-present morbidity and mortality must be carefully balanced with patient life expectancy in order to offer optimal management. Several risk scores for myelofibrosis have been developed. The most widely used risk-assessment model is the Lille (or ‘Dupriez’) score,113 which distinguishes low, intermediate and high risk according to the hemoglobin (<10 g per 100 ml) and white blood cell count (<4 × 109 per liter or >30 × 109 per liter) with overall survival of 93 months, 26 months and 13 months, respectively. Another scoring system (Cervantes score)4 includes hemoglobin (<10 g per 100 ml), circulating blasts and constitutional symptoms, and it distinguishes low risk (none to one adverse factor: median survival, 176 months), and high risk (two to three adverse factors: median survival, 33 months). More recently, another risk-assessment score for patients with PMF was reported,114 which includes hemoglobin (<10 g per 100 ml), white blood cell count (<4 × 109 per liter or >30 × 109 per liter), platelet count (<100 × 109 per liter) and monocytes (1 × 109 per liter). Patients with no risk factor had a median survival of 173 months, while patients with one risk factor had a median survival of 61 months, and those patients with two or more risk factors had a median survival of 26 months. A model from the Japanese Research Committee for Idiopathic Hematopoietic Disorders proposed four risk factors for myelofibrosis: male gender, circulating blasts 1%, constitutional symptoms and hemoglobin <10 g per 100 ml. Patients with no or only one risk factor had a median survival of 292 months, while patients with two to four risk factors had a median survival of 66 months only.115 The only risk factor which includes age as risk factor is the Cologne score.116, 117
In Table 6 the different risk models are shown. This risk-assessment model might help to select patients for allogeneic SCT. However, other prognostic factors for decision making such as comorbidities, or cytogenetic abnormalities118 are not included in any of the models. Furthermore, risk features of one model cannot be translated into another model, for example, a patient with high risk according to the Cervantes score can have low risk according to the Lille score and vice versa. Therefore, given the low number of published results about allogeneic SCT (standard conditioning as well as reduced-intensity conditioning), all recommendations are based more on assumptions than on facts.
Age should not be a limiting factor for allogeneic SCT. More important are comorbidities such as cardiac or pulmonary diseases or other diseases, which can be summarized in a comorbidity score, which influences treatment outcome more significantly than the biological age.119 For patients with high risk according to the Cervantes score or intermediate/high risk according to the Lille score, allogeneic SCT should be evaluated. Since the life expectancy in industrialized countries is about 80 years for women and 75 years for men, patients up to the age of 70 years and older with low comorbidity score are suitable candidates for allogeneic SCT. Since patients with low-risk disease according to the Cervantes- or to the Lille score have a very low treatment-related mortality, at least in younger patients early transplantation should be considered. Unfortunately, all risk models are static models and do not take into account the dynamics of the disease. Therefore, our recommendation is to carefully monitor young patients with low-risk disease regarding the dynamics and to perform allogeneic SCT if clear signs of progress occur (decrease in hemoglobin, increase in lactate dehydrogenase, constitutional symptoms, rapid fibrosis in bone marrow, and so on), even if per definition the patients remained low-risk according to the above-mentioned risk scores.
The time point for performing allogeneic SCT is crucial. The so far limited literature suggests that transplantation in an early phase of the disease (for example, Lille risk score—low) resulted in a lower transplant-related morbidity and mortality and less risk of relapse. However, the risk of treatment-related mortality even in the era of reduced intensity conditioning for low-risk patients is still about 10% and should be balanced on an individual basis with survival time without early transplant. In contrast, transplantation in a late phase of the disease (Lille risk score—high, Cervantes high or transformation in acute leukemia) is associated with an increase risk of treatment-related mortality as well as risk of relapse. If the patient has almost developed AML, the chance of curing the disease by allogeneic SCT is considerably decreased. More work is necessary to develop dynamic risk models to determine the kinetics and dynamics of the progression in order to determine the optimal time point for allogeneic SCT with low treatment-related mortality and low risk of relapse.
It has been thought reasonable to perform allogeneic SCT with standard myeloablative conditioning in younger patients (<45 years) and dose-reduced conditioning in elderly patients (50–70 years). However, given the low mortality thus far reported after dose-reduced conditioning in comparison with standard conditioning, it seems questionable whether an age cut-off is wise. Studies in mice and later in humans provided evidence that curing hematological diseases by allogeneic SCT is not mainly related to high-dose conditioning regimen but rather to the immunologically mediated graft-versus-tumor effect of donor T cells. Since the toxicity of the conditioning regimen prior to allogeneic SCT contributes to treatment-related morbidity and mortality, an efficacious strategy to reduce treatment-related mortality might be to minimize toxicity. Such regimens were called non-myeloablative, but several dose-reduced regimens are still myeloablative. This approach can provided ensure allogeneic engraftment and complete donor chimerism by shifting tumor reduction from the preparative regimen to the action of immunocompetent donor cells to induce a graft-versus-myelofibrosis effect. Using modern methods such as highly sensitive JAK2 mutation screening after allografting, adoptive immunotherapy with donor lymphocyte infusion can be performed to eradicate minimal residual disease. Thus, for the majority of patients with intermediate risk, a dose-reduced conditioning is the regimen of choice, irrespectively of the age. Other factors which have to be taken into account to determine the intensity of the optimal conditioning regimen are (1) comorbidity index, which indicates how intensive a conditioning regimen is acceptable with regard to toxicity and risk of mortality for the patients and (2) the risk of relapse, which—if it is very high—requires a more intensified conditioning.
After many years of therapeutic fatalism, the introduction of low-intensity allogeneic SCT and new drugs as well as the discovery of JAK2V617F mutation offers a realistic curative chance and new hopes for molecular target therapy for patients with myelofibrosis. The choice of ‘optimal’ therapy in a patient with myelofibrosis is unlikely ever to be taken without thoughtful consideration and some consternation. Medical decision making will remain deciding at what point does the impact of medical therapy (with presumably less toxicity than SCT) exceed the net benefit of transplant. The intersection of these two forces will continue to change with, hopefully, decreased toxicity of transplant, and, hopefully, the ability of medical therapy to alter the natural history of the disease not just palliate symptomatology. Would we be able to reach a point where transplant is clearly the front-line therapy, or only medical therapy is considered? Perhaps, a scenario more likely will evolve that will mirror the therapy of CML: in essence, front-line therapy with medical therapy (perhaps a JAK2 inhibitor, although yet to be proven) with allogeneic SCT as second-line therapy for those with an inadequate response. Similar to CML, targeted therapies will probably play a larger role for those with a ‘chronic phase’ of the disease (that is, having arisen from the initiating genetic lesion likely). Those with more advanced phase (that is, equivalent to an accelerated or PMF-BP) will require transplantation initially or after a brief trial of a theoretically efficacious therapy. Similar to CML blast phase, we have clearly demonstrated PMF BP to be characterized by complex, and high-risk, karyotypic abnormalities, which are unlikely to be responsive to inhibition of primary targets. The ultimate goal for all patients should be cure, which can be achieved currently only by allogeneic SCT. Despite the marked improvement in performing allogeneic SCT, there is still a considerable risk of treatment-related morbidity and mortality. If palliation is the treatment aim, several drugs can relieve treatment-related symptoms, or cytopenia. In the near future, the treatment of myelofibrosis will be further improved by defining the optimal time point for allogeneic SCT by including remission-inducing drugs to make nontransplant candidates fit to become a transplant candidate, and reduce further treatment-related morbidity by developing effective post-transplant strategies, and by defining the role of the new JAK2-tyrosine-kinase inhibitors. We conclude, a more individualized, risk-adapted treatment approach including the available drug medication, low intensity allografting, as well as new JAK2 inhibitors will be the subject of the future of clinical research to cure more myelofibrosis patients with less risk from treatment: the present and future hold promise for patients and clinicians responding to the challenges of myelofibrosis.
We thank the staff of the clinics for providing excellent care of our patients and the medical technicians for their excellent work in the laboratories and Richard Brummel for proof editing.