Allogeneic Stem Cell Transplantation for Myelofibrosis: Reversing the chronic phase in the JAK inhibitors’ era?

Abstract

At present, allogeneic hematopoietic stem cell transplantation (allo-SCT) is the only curative treatment for patients with myelofibrosis (MF). Unfortunately a significant proportion of candidate patients are considered transplant ineligible due to their poor general condition and advanced age at time of diagnosis. The approval of the first JAK inhibitor, Ruxolitinib, for patients with advanced MF in 2011 has had a qualified impact on the treatment algorithm. The drug affords substantial improvement in MF-associated symptoms and splenomegaly but no major effect on the natural history. There has, therefore, been considerable support to assess the drug’s candidacy in the peri-transplant period. The drug’s precise impact on clinical outcome following allo-SCT is currently not known; nor is the drug’s long term efficacy and safety known. Considering the rarity of MF and the small proportion of patients who are undergoing allo-SCT, well designed collaborative efforts are required. In order to address some of the principal challenges, an expert panel of laboratory and clinical experts in this field was established, and an independent workshop held during the 54^th American Society of Hematology’s meeting in New Orleans, USA, on 6^th December 2013 and the European Hematology Association meeting in Milan, Italy on 13^th June 2014. This document summarizes the results of these efforts.

Introduction

Although allogeneic hematopoietic stem cell transplant (allo-SCT) is the only treatment which accords long term remission and potential cure for patients diagnosed with myelofibrosis (MF), it is not available for a significant majority of patients due to the advanced age at diagnosis (1). The term MF comprises a rather motley collection of several clonal but clinically distinct, BCR-ABL1-negative myeloproliferative neoplasms (MPNs) subtypes (2). The World Health Organization (WHO) classifies MF as primary myelofibrosis (PMF), post-polycythemia vera myelofibrosis (PPV-MF) or post-essential thrombocythemia myelofibrosis (PET-MF) (3). Following the unravelling of somatic mutations in JAK2, MPL and CALR genes over the past decade, we have an improved understanding of the complex genomic landscape of MPNs, though precise details on driver mutations remain incomplete (4–7). These seminal observations confirmed the principal role of an hyperactive JAK-signal transducers and activators of transcription (STAT) intracellular signaling in the pathogenesis of these disorders (8). Thereafter, many efforts led to the introduction of JAK inhibitors into the clinics in 2007 (9).

There was considerable enthusiasm surrounding this, largely based on the very impressive results achieved in patients with the BCR-ABL1-positive MPN, chronic myeloid leukemia (CML), following treatment with tyrosine kinase inhibitors (TKIs) (10). Current clinical experience, however, has not yielded as impressive results with JAK2 inhibitors, but rather a qualified progress leading to the licensing of ruxolitinib, a selective JAK1 and JAK2 inhibitor, for patients with advanced MF (11–13). The drug affords substantial symptomatic benefits, reduction in splenomegaly and improved survival, but no significant impact on the malignant clone. Furthermore approximately 50% of patients discontinue ruxolitinib therapy at 3 years due to loss of clinical benefit or adverse side effects (11).

Allo-SCT remains the only treatment which accords long term remission and potential cure to MF patients. The historical results of conventional allo-SCT have, however, for the most part, been confounded by substantial non-relapse transplant-related mortality (NRM) (14–16). This is probably due, at least in part, to the older age of patients with MF and the increasing presence of significant co-morbid conditions. Efforts to improve these results have led to general improvements in allo-SCT technology and an increased use of reduced-intensity conditioning (RIC) preparative regimens. RIC is associated with a lower NRM, but a higher risk of relapse, compared with standard myeloablative conditioning (MAC), and the current results suggest a qualified success (Table 1) (17–21). The notion of using ruxolitinib in order to reduce constitutional symptoms and splenomegaly prior to allo-SCT appears attractive. What is not known is the drug’s impact on short and long-term outcomes following allo-SCT; nor is the precise timing of allo-SCT or the role of allo-SCT specific prognostic and predictive scores in the JAK inhibitor era known. In order to address some of these challenges, an expert panel comprising of laboratory and clinical experts in MPNs was established, and an independent workshop held during the 54^th American Society of Hematology’s annual meeting in New Orleans, USA, on 6^th December 2013. Several discussion groups were held thereafter with the penultimate at the European Haematology Association’s annual meeting in Milan, Italy, on 13^th June 2014; additionally a search of published literature (popular search engines, including PubMed, Google scholar and the NCI database) since 1990 was performed. Herein, we summarize our findings and assess the suitability of JAK inhibitors for patients with MF who are candidates for allo-SCT.

Table 1:

Comparative analysis of current retrospective studies of reduced intensity conditioning (RIC) and myeloablative conditioning (MAC) allo-SCT for myelofibrosis (MF) patients

Reference	N	Years	Preparative regimen	Median age (range)	OS	GVHD	NRM	Relapse
Guardiola et al.14	55	1979–1997	MAC (TBI, 63%)	42 (4–53)	5 years, 47%	aGVHD, 60% cGVHD, 60%	27%	18%
Kerbauy et al.15	104	1990–2005	MAC, 95; RIC, 9	49 (18–70)	2 years, 70%; 5 years, 61%	aGVHD, 64% cGVHD, 84%	32%	10.5%
Abelsson et al.16	92	1982–2009	MAC, 40; RIC, 52	46 MAC 55 RIC	2 years, 55%/75%; 5 years, 49%/59%	aGVHD, 75% MAC 28% RIC	32% 24%	10%
Nivison-Smith et al.17	57	1993–2005	MAC, 40; RIC, 17	47 (16–71)	2 years, 59%/66%; 5 years, 58%	aGVHD, 37%	25%	12%
Kroger et al.19	103	2002–2007	RIC, all (BU/Flu)	55 (32–68)	5 years, 67%	aGVHD, 27% cGVHD, 49%	19%	29%
Gupta et al.20	233	1997–2010	RIC, all	55 (19–79)	5 years, 47%	aGVHD, 37% (day 100) cGVHD-42%	18%	43%

Clinical impact of bone marrow histology

While bone marrow (BM) histology clearly plays a critical role in the diagnosis and classification of MF (22,23), there is some debate as to the methodology used in reporting BM fibrosis and also the integration of serial BM samples in the optimal treatment algorithm of MF patients. Importantly, although MF is an acquired clonal hematopoietic stem cell disorder, BM fibrosis represents a reactive phenomenon characterized by the presence of increased collagen fibers in a disorderly manner (24). Histologically this is best assessed by reticulin and collagen (trichrome) stains of the bone marrow trephine biopsy, and graded by the four-grade semi-quantitative European consensus system (25,26). A recent International and European LeukemiaNet (ELN) effort has proposed a set of BM features that characterize therapy response by BM morphology (27). In analogy to the WHO grading concept for fibrosis, this proposal defines reproducible scoring systems for the grading of collagen deposition and osteosclerosis. Other abnormalities include increased microvessel density and new bone formation, with the eventual development of thickened bone trabeculae (osteosclerosis) Following a successful allo-SCT for MF, the clonal hematopoeisis is promptly replaced by normal donor stem cells that colonize the marrow and produce normal hematopoiesis (Figure 1A and 1B); in contrast, changes in the bone marrow stroma, which remains of the recipient origin, take considerably longer (28). Several studies, both prospective and retrospective, have examined these changes and their time course after allo-SCT, and these data are summarized in Table 2 (29,30).

Figure 1:

A,B: Bone marrow biopsy from 50-year old female with post-PV MF prior to allo-SCT. The marrow is markedly hypercellular for age (100% cellularity) and contains large clusters of abnormal megakaryocytes with hyperchromatic and “cloud-like” nuclei, as well as smaller forms with hyperlobulated nuclei. Reticulin stain shows grade 2 of 3 fibrosis (A, hematoxylin & eosin; B, reticulin stain). C,D: Bone marrow biopsy taken 12 months after allo-SCT shows cellularity with normal range for the patient’s age (60% celluarity), with a normal myeloid:erythroid ratio and morphologically normal megakaryocytes that do not cluster. Reticulin stain shows resolution of reticulin fibrosis (grade 0). Chimerism studies on bone marrow showed >98% donor origin (C, hematoxylin & eosin; D, reticulin stain).

Table 2:

Pathology of primary myelofibrosis (PMF) and changes in pathologic features observed after allo-SCT

Feature	Abnormality in PMF	Resolution after allo-SCT	Time when resolution occurs
Cellularity	Increased	Yes (up to 100%)	Upon engraftment
Megakaryocyte morphology	Hyperchromatic, ‘cloud-like’ nuclei, tight clustering	Variable; proportion with resolution unknown	Upon engraftment
Myeloid:erythroid ratio	Increased	Variable; proportion with resolution unknown	Upon engraftment
Reticulin fibrosis	Increased	Yes (40–93%)	6–23 months
Microvasculature	Increased density	Yes (50–100%)	3–12 months
Bone trabeculae	Thickened with new bone formation	Variable (0–50% at 12 months)	12 months or longer

Following a successful allo-SCT, the initial change in BM stroma is a reduction in the reticulin fibrosis grade, which may exhibit heterogeneity and has posed challenges in reporting (Figure 2A and 2B) (31). This heterogeneity may impede determination of any reduction in fibrosis grade in sequential post-SCT biopsies in comparison to the pre-SCT biopsy. Regardless, complete resolution of reticulin fibrosis has been confirmed in about 90% of MF patients by 12 months following both MAC and RIC transplants (14,17,32–34). Resolution of osteosclerosis is more controversial, and has been reported to occur in some studies, but not others (33,35).

Figure 2:

A,B: Bone marrow biopsy taken 12 months after SCT shows cellularity with normal range for the patient’s age (60% cellularity), with a normal myeloid:erythroid ratio and normal megakaryocytes that do not cluster. Reticulin stain shows resolution of reticulin fibrosis (grade 0). Chimerism studies on bone marrow showed >98% donor origin (A, hematoxylin & eosin; B, reticulin stain).

While successful allo-SCT in patients with MF thus leads to an eventual reversal of the abnormal BM stromal changes and fibrosis in most patients, conversely, relapse of the MPN clone following allo-SCT is associated with return of bone marrow stromal abnormalities stimulated by the abnormal clone. Although some studies use complete resolution of reticulin fibrosis to define a complete ‘histopathologic’ response to allo-SCT, in the presence of full donor chimerism, the significance of persistent fibrosis >12 months after allo-SCT is uncertain and may not necessarily be indicative of persistent or relapsed disease (14,36). More recently, the importance of the rate of decline of the BM fibrosis has been demonstrated, with patients having a rapid regression have the best outcomes (Table 3) (15,37). In the study conducted by Krӧger and colleagues, patients who achieved a complete or near-complete resolution of fibrosis on day +100 post allo-SCT had the best overall survival and this was independent of the IPSS score at the time of transplant.

Table 3:

Dynamics of bone marrow (BM) fibrosis regression (adapted, with permission, from Krӧger et al, Biol Blood Marrow Transplant, 2014;20:812–815)

Time (days from allo-SCT)	Day 0	Day +30	Day +100
BM fibrosis—grade 0	0%	6%	25%
BM fibrosis—grade 1	0%	15%	29%
BM fibrosis—grade 2	28%	35%	27%
BM fibrosis—grade 3	72%	44%	18%

Impact of the timing of transplant

Prior to the introduction of JAK inhibitors into the clinics, it was generally accepted that allo-SCT should be offered to MF patients, perhaps as early as possible following the diagnosis, in order to abrogate the dire consequences of transformation to acute myeloid leukemia (AML). Following the qualified responses accorded by ruxolitinib, there has been considerable uncertainty about the precise timing of allo-SCT. Additionally there is much uncertainty about discontinuing ruxolitinib prior to allo-SCT. Interruption or discontinuation of ruxolitinib, for any reason, often leads to the return of constitutional symptoms to baseline in less than a week and regrowth of spleen shortly thereafter. There are also reports of serious withdrawal symptoms and thus it would appear desirable to continue the drug for as long as possible prior to transplantation (38). These challenges are now being addressed in a prospective multi-center phase II study, MPD-RC114 [NCT01790295], in which ruxolitinb is administered to eligible MF patients for 60 days prior to definitive allo-SCT employing RIC with fludarabine (Flu) and busulfan (Bu) RIC for matched donors and Flu-Bu and antithymocyte globulin (ATG) for mismatched related and unrelated donors. The study mandates ruxolitinib taper to commence 9 days prior to stem cell infusion and continues slowly over a period of 4 days, with the last dose of ruxolitinib to be taken 1 day prior to start of conditioning therapy.

Results following an allo-SCT in general appear to be the best in younger patients who are transplanted at an earlier disease stage and better risk-scores, with the notable exception of the recent European Group for Blood and Marrow Transplantation (EBMT) study (discussed below) (15,39,40). In patients in whom transplant is delayed, there is an increasing risk of progressive splenomegaly, portal hypertension, and iron overload as a consequence of frequent transfusions and the risk of transformation to poor-risk AML.

Impact of pre-transplant conventional therapy

The standard initial treatment for most, if not all, MF patients who present with symptomatic disease is oral hydroxyurea (hydroxycarbamide) at a dose of 500mg to 2000mg per day or subcutaneous alpha interferon (IFN), particularly pegylated-IFN (41–43). Hydroxyurea has not been considered to impact on the clinical outcome of allo-SCT, nor does it appear to have any substantial effect on the natural history of MF, although there remains some debate as to its contribution towards leukemic transformation. In contrast, pegylated-IFN has shown to improve BM fibrosis, reduce splenomegaly and lead to a molecular response in several small studies (44,45). Neither drug appears to affect the time to engraftment, nor affect the incidence or severity of acute or chronic graft-versus-host disease (GvHD).

Impact of targeting the JAK/STAT pathway

Currently there is a paucity of clinical data for patients with MF who have received initial therapy with JAK inhibitors and are subsequently treated with allo-SCT. Several on-going efforts, such as the MPD-RC 114 study (discussed above) are now assessing the potential impact of the use of ruxolitinib during the peritransplantation period for MF patients. The notion of using ruxolitinib as initial therapy leading to a definitive allo-SCT is supported by the substantial and durable improvement in constitutional symptoms and reduction of splenomegaly in two randomized phase III trials (11,12); these results were confirmed on longer follow-up and some intriguing data on the drug’s effects on BM fibrosis noted (Figure 2A and 2B) (46). The adverse events attributed to ruxolitinib appear to be relatively mild and generally manageable. Anemia is the most common side effect, sometimes with thrombocytopenia, and may require dose adjustments. It is therefore possible, though not confirmed, that initial treatment with this drug for a pre-determined period, might improve the general condition, splenomegaly and performance status of candidate patients, factors which may also improve allo-SCT results.

An interesting compromise would be to offer a transplant only to patients who have ruxolitinib resistance or intolerance or patients who start developing significant cytopenias or transfusion dependency on JAK inhibitor therapy. This may allow patients responding to ruxolitinib to continue until resistance/intolerance, or indeed, progression of disease, are noted. The disadvantages with such an approach are the potential risks for leukemic transformation and the prolonged use of a therapy which has yet to demonstrate an effect on the malignant clone itself. There are additional financial implications and the longer term effects of the drug are not known. A recent report documented four patients who developed AML and one who developed chronic myelomonocytic leukemia on ruxolitinib for 3 to 18 months. The time form diagnosis to leukemic transformation to AML was 2, 3, 4, 10 and 20 years. Four of these 5 patients had been treated with hydroxyurea for varying duration, ranging from few months (2 patients) to several years (2 patients). Four of the 5 patients manifested extra-medullary disease without bone marrow involvement (3 patients with leukemia cutis and one with granulocytic sarcoma) (47). With these few patients, and variable duration of prior hydroxyurea exposure, it is not possible to establish a definitive causal relationship between prior exposure to cytotoxic therapy and the emeregence of the myeloid sarcoma. This raises the possibility of evolution of a more aggressive clone in some patients whilst on ruxolitinib therapy and also raises the concern that in patients who become resistant to ruxolitinib, the phenotypic features of disease may be more aggressive and conceivably, make a previously transplant-eligible patient ineligible. Clearly additional work is required to establish the long-term efficacy and safety of the drug in the clinics.

There are additional dilemmas with regards to the potential use of pre-transplant JAK inhibitors. Ruxolitinib has remarkable anti-inflammatory and immunomodulating activity, but the drug also affects dendritic cell differentiation and impairs T-cell and NK cell activation (48,49). The dendritic cell impairments might result in increased infection rates, while the NK cell dysfunction may impair graft-versus-leukemia (GvL) effect (50). Conversely, JAK inhibitors appear to reduce acute GvHD in murine models by increasing T-regulatory cells (Treg), which has led to the successful treatment of six patients with ruxolitinib for steroid-refractory acute GvHD (51). Another JAK inhibitor, tofactinib, a selective JAK3 inhibitor, has been found to reverse CD8+ T-cell mediated mucocutaneous GvHD-like disease in a murine model (52).

Impact of targeting the epigenome

Studies on the remarkable complexity of genetic architecture in MPNs have discovered an unexpected degree of differences in individual cases at the epigenetic level. However, mutations in genes encoding epigenetic regulators are uncommon in MPN and no clear patterns have emerged so far (53). Aberrations in ASXL1, EZH2, SRSF2 and IDH1/2 appear to have a predictive impact on the overall and leukemia-free survival, suggesting that the MPN epigenome is clinically relevant; the greatest impact appears to be with the ASXL1 mutation (13,53). In addition, promoter specific hypermethylation of candidate genes such as the chemokine receptor CXCR4 has been linked to the constitutive migration of CD34⁺ cells in PMF (54). Global methylation profiling in MF revealed a distinct methylation signature, and in patients who transform to AML, it has been noted that the number of differentially methylated regions increase significantly and the aberrant genes are involved in the ‘interferon pathway’ (55, 56). Collectively, these and related observations suggest the importance of the epigenome in MF patients and there are now several research efforts assessing the potential role of DNA methyl transferase (DNMT) inhibitor and histone deacetylase (HDAC) inhibitors in MF.

Preclinical studies utilizing a DNMT inhibitor and an HDAC inhibitor sequentially in primary MF cells in vitro and in NOD/SCID mouse models in vivo reveal that these agents can inhibit the malignant clone, upregulate the transcription of CXCR4 in primary MF CD34+ cells and reverse the abnormal stem cell trafficking in MF, resulting in homing of stem cells to the marrow, rather than the spleen (57,58). These findings suggest a possible role of hypomethylating therapy prior to allo-SCT in an effort to improve outcomes. Several pilot clinical studies suggest clinical activity in both early and late phases of the disease (59–63). These agents also have the potential to harness an immunomodulatory effect that could synergize with the GvL effect, making them of additional interest (64,65). The notion of administering these agents at a lower dose/intensity to harness their immune modulatory effects which include the up-regulation of cancer testes antigens and augmentation of Tregs, following allo-SCT is additionally attractive (66).

At present there are several on-going clinical trials, assessing a combination of JAK inhibitors and epigenetic modulators in both transplant and non-transplant settings, based on pre-clinical evidence of synergy (67). Recent small series of MF patients suggest the efficacy and safety of a combination of ruxolitinib and DNMT inhibitors (68). In addition, work is underway to identify specific inhibitors of mutated components of the epigenome, and such inhibitors are now in early phase clinical trials (69,70). The recent observation of the induction of cellular differentiation by an IDH1 mutant inhibitor in a refractory AML xenograft model and the candidacy of TET2 mutations as a predictive biomarker for response to hypomethylating agents in MDS patients are also of some interest (71,72).

Impact of patient-related factors

At present the very considerable advances in the understanding of the genomic landscape in MF appear not to have been validated sufficiently for adaptation in treatment algorithms to assess candidacy for allo-SCT compared to conventional therapy (73). The Lille score, first introduced in 1996 for patients with PMF, remains the best studied risk-score tool for transplant candidates with BCR-ABL1 negative MPNs (74). More recently, the International Prognostic Scoring System (IPSS) was introduced to assess PMF patients at time of diagnosis (75). The Dynamic-IPSS (DIPSS) and DIPSS-plus (DIPSS plus cytogenetic information) were subsequently introduced to help refine prognosis for these patients at any time during disease evolution (Table 4) (76,77). These risk-scores are applied at diagnosis and are based on hematological parameters, such as anemia, leukocytosis and peripheral blood blasts, as well as clinical features, such as age and presence of constitutional symptoms. Comparative studies of the usefulness of the various scoring systems in allo-SCT suggest a qualified application at best and most experts advocate the use of transplant-specific risk-models instead. The Seattle group and the Essen group have demonstrated the usefulness of the DIPSS score in terms of predicting NRM, relapse and survival (78,79). Krӧger and colleagues proposed a relatively simple model which added age, JAK2 status and constitutional symptoms to the Lille model to predict for survival following a RIC-allo-SCT (80). The Genoa group developed a RIC-allo-SCT risk-model based on donor type, transfusion requirement and splenomegaly, but this has not attracted wider use (81). Finally, the Gratwohl risk-score is a useful tool for risk-assessment for patients with most hematological malignancies being considered for an allo-SCT, but unfortunately no patients with MF were included in its development (82)

Table 4:

Risk Stratification for Myelofibrosis (MF); (courtesy of Prof A Rambaldi)

Variable	Lille	IPSS	DIPSS	DIPSS plus
Age >65 years		√	√	√
Constitutional symptoms		√	√	√
Hb <10 g/dL	√	√	√	√
Leukocyte counta	√	√	√	√
Circulating blasts ⩾1%		√	√	√
Platelet count <100 × 109/L				√
RBC transfusion need				√
Unfavorable karyotype +8, −7/7q-, i(17q), inv(3), −5/5q-, 12p-, 11q23 rearr.		√

It is actually quite remarkable that risk stratification methods for patients with MF per se have not been studied in a validated prospective manner; rather the risk scores tested in patients with PMF, have been extrapolated into the treatment-decision making process for all MF patients, a cohort which also includes patients with PPV-MF and PET-MF. There is therefore an urgent need to develop MF-specific risk scores not only for potential allo-SCT candidates, but also for non-transplant therapies. Efforts to develop risk scores, including transplant-specific scores, based on integrated molecular-risk profiles, for all MPNs are now being developed (Figure 3).

Figure 3:

A Risk-Score based on Genomic Abnormalities in Myelofibrosis (MF); although mutations in ASXL1, SRSF2, EZH2, predicted poorer outomes (high-risk), only ASXL1 mutations were considered significant in the context of IPSS, DIPSS and DIPSS-plus; (Courtesy of Prof A Vannucchi)

The impact of disease-associated splenomegaly on allo-SCT outcomes in hematological malignancies has been debated often. In general, patients with splenomegaly appear to have delayed engraftment, which nevertheless does not appear to affect the overall survival (36,83). Interestingly, a French study noted a favorable survival for male patients who underwent pre-transplant splenectomy, while a German study demonstrated a higher relapse rate in the splenectomized patients (17,84). Additionally, there appears to be no agreement as to the actual size of the spleen which might, or might not, serve as a trigger for pre-transplant therapy. There is now an emerging consensus to avoid splenic irradiation or splenectomy immediately prior to transplant due to the intrinsic complications of radiation and splenectomy and thus the notion of reducing splenomegaly medically is indeed attractive.

Results of treatments for MF patients who are either in frank AML transformation, or at high risk of transformation, including conventional cytotoxic induction regimes or novel induction regimens, of both higher and lower intensity prior to an allo-SCT, remain suboptimal, with a median survival of less than 5 months (85–87). Current efforts are investigating diverse agents, including hypomethylating agents, JAK and MEK inhibitors, BCL-X_L and BCL-2 inhibitors and clofarabine (88–93).

Impact of the transplant-procedure and donor related factors

Several prospective multicenter studies have now confirmed the usefulness of RIC-SCT using related or unrelated donors for younger and older patients with MF. In a large retrospective EBMT study, outcomes from 250 consecutive patients (median age 56 years) with PPV-MF (n=120) or PET-MF (n=130) who received allo-SCT, either in MF ‘chronic’ phase (n=193) or MF transformed to AML (n=57), were analyzed (39). The study accords an opportunity to assess the impact of the various transplant-procedure related factors in a ‘real-world’ cohort of MF patients, including those who had transformed to AML. It is of interest that the non-transplant risk scores of 61% of the study cohort were unknown at the time of transplant. 117 patients received allografts (115 matched; 2 mismatched) and 133 unrelated allografts (124 matched; 9 mismatched); 80 patients underwent conventional conditioning and 170 received RIC. Over half of the transplanted patients had an interval from diagnosis to allograft of 10 years or more. The 3-year overall survival was 55%, with a relapse incidence of 32%, following a median post-transplant follow-up period of 13 months. In uni-variante analysis, older age (>55-years), transplant in AML, and an unrelated donor, were the factors which affected survival (Figure 5). In this study, there appear to be no statistically significant difference between the conventional and RIC regimens, nor was there a significant impact on outcomes in T-depleted versus the non-T-depleted cohorts.

The Center for International Blood and Marrow Transplant Research (CIBMTR) published a report on the outcomes of 233 patients with PMF who received RIC-allo-SCT (20). Since the study excluded patients with PPV-MF, PET-MF and those who had transformed to AML, it offers an assessment of the role of RIC allo-SCT in PMF, using a validated risk-assessment tool (DIPSS) which addresses the course of disease. The study cohort DIPSS-risk stratification was: low (12%), intermediate-1 (49%), intermediate-2 (37%), high (1%) and unknown (1%). 79 patients received fully matched related allografts and 154 received unrelated allografts: 104 fully matched, 40 partially matched and 10 mismatched. The median age was 55years (range 19–79) and the median follow-up of survivors was 50 months (range 3–134). A variety of RIC regimens were used: Flu-Total Body Irradiation (22%); Flu-Melphalan (Flu-Mel) (28%); Flu-Bu (38%); ‘others’ (12%). 123 of 233 patients (53%) received in vivo T-cell depletion: ATG (110 patients) and alemtuzumab (13 patients). The 5-year overall survival was 56% for matched related, 48% for matched unrelated and 34% for partially matched or mismatched unrelated cohorts. Multivariable analysis confirmed donor type as the sole independent factor influencing survival; importantly, age at transplant did not appear to affect survival. There was a trend for patients with intermediate-2 disease to have higher NRM, mainly from GvHD, infections and organ failure, resulting in an inferior survival compared to intermediate-1 and low risk patients. The study lends support to the use of RIC allo-SCT for PMF with the best results accorded to the low and intermediate-1 risk categories and those who have a suitably matched related donor available. The recruitment of just 1% DIPPS-high risk patient in this CIBMTR effort is of note.

Another international effort, the MPD-RC 101 study, has analyzed outcomes in a smaller cohort of 66 patients who received RIC-allografts between 2007 and 2011 (94). The study cohort comprised all MF subtypes and included about 35% intermediate-2/high risk DIPSS. Following a median follow-up of 25 months, 24 of 32 patients who had received a related allograft (30 matched; 2 mismatched) following Flu-Mel conditioning, were alive compared to 11 of 34 who had received an unrelated allograft (25 matched; 9 mismatched) following Flu-Mel-ATG conditioning. The NRM was 22% and 59%, and the secondary graft failure 3% and 12%, for related versus unrelated cohorts, respectively. Survival did not appear to correlate with the degree of donor matching, age at transplant, or JAK2 status at diagnosis. The investigators speculated that the increased risk of graft failure and death in the unrelated cohort might be due to the in vivo T-cell depletion, which was not offered to the related patients.

Efforts have also assessed the suitability of umbilical cord blood (UCB) allo-SCT for adults with MF (95). The results of 35 patients with MF, including 7 who had transformed to AML, who received UCB-allo-SCT were reported to Eurocord; 24 of 35patients received RIC with Flu-cyclophosphamide and total body irradiation, with or without ATG. The median age of this cohort was 54 years. Following a median follow-up of 24 months, the overall survival was 44%; 15 of 24 patients relapsed. It is of interest that in this small series of patients, outcomes for patients with AML was similar to those in chronic phase. The results of this Eurocord series are sufficiently encouraging to suggest further formal exploration of this donor source.

The subtype of MF, patient CMV status, use of T-cell depletion and specific RIC regimen do not appear to influence post-allo-SCT outcomes (38). Currently there appears to be no firm consensus with regards to the prognostic significance of the JAK2, CALR and MPL mutations, one of which is present in most MF patients, on the outcomes following an allo-SCT. There is one preliminary report suggestive of best outcomes in CALR positive and worst outcomes in JAK2/MPL/CALR negative (‘triple-negative’) patients following allo-SCT (96). In this effort comprising of 119 patients with MF, a multivariable analysis revealed shorter survival and higher NRM for the ‘triple-negative’ cohort compared to patients with any mutations. These observations have now been validated by several recent efforts confirming the notion of patients with JAK2 wild-type (WT), MPL WT and CALR WT, to have poor outcomes (97,98).

Impact of patient reported outcomes

Despite the integration and subsequent validation of patient reported outcomes (PROs) in the non-transplant treatment algorithms for MF, their use in the clinical decision-making process for transplant remains a challenge. There are few studies assessing quality of life, financial burden and symptoms in MF patients undergoing allo-SCT. The development of the Myelofibrosis Symptom Assessment Form (MF-SAF): an evidence-based brief inventory to measure quality of life and symptomatic response to treatment in myelofibrosis. facilitated a platform for the objective quantification of these parameters and was incorporated in ruxolitinib MF-registration studies (11,12). In an effort to assess the role of PROs in assessing MF patients” candidacy and post-transplant progress, SYMPTOMS, a prospective questionnaire-based study was recently commenced (99–102). The study evaluates the quality of life of patients undergoing allo-SCT, utilizing MPN-SAF TSS, the FACT-BMT, Global Impression of Change, and a financial questionnaire at baseline, day +30, day +100 and day +365.

Conclusion

Following the qualified results accorded by ruxolitinib for patients with MF, many efforts are appraising the current treatment algorithm for newly diagnosed patients and assess the impact of JAK inhibition and other candidate therapies on transplant outcomes (103–107). Currently allo-SCT remains the only treatment modality which appears to be associated with long term remissions and potential cure. Transplant-related mortality and morbidity, both GvHD- and non-GvHD-associated, remains a problem with both RIC and MAC regimens. The notion of having access to a drug which results in substantial symptomatic benefit, significant reduction in splenomegaly, and an improvement in survival appears attractive. However, the fact that ruxolitinib does not appear to affect the risk of AML and does not reduce the malignant clone, suggest that allo-SCT performed as early as possible affords the best chance of a long term remission and potential cure for most patients. Indeed, recent assessment of transplant activity suggests an increase in the utilization of allo-SCT, particularly with RIC regimens.

There is now an urgent need to develop suitable risk stratification methodology to optimize the decision-making process for treating patients with MF. This should encompass both transplant and non-transplant treatments, since none of the current methods have been validated for patients with PPV-MF and PET-MF, who comprise about 60% of all MF patients. Efforts are underway to develop a molecular-based risk score which should help navigate the treatment algorithms commensurate to AML transformation-risks. MF patients have distinct subclones which persist for years and emerging evidence suggest that the order in which the various mutations are acquired influences the timing of disease presentation and may also affect response to therapy. Furthermore, novel molecular insights into the evolution of the various MPN genotypes and pathologic subtypes into MF and AML pose interesting clinical questions. A principle challenge now is to decide how exactly allo-SCT efforts can be improved; topical challenges include refinement in risk stratification tools prior to transplant, role of peri-transplant JAK inhibition and indeed other candidate therapies, evaluation of alternative donor sources, and reduction in NRM and primary and secondary graft failure. In addition, several significant challenges have been identified in the methodology and the selection of the surrogate end-points of some of the recent clinical trials. A recent proposal led by the European LeukemiaNet (ELN) and the International Working Group for MPN Research and Treatment (IWG-MRT) appears to be a reasonable starting point to develop prospective randomized trials which may help to develop definitive guidance for the selection of suitable peri-transplant therapy, if any, in due course (108).