Chapter 86: Primary Myelofibrosis

Primary myelofibrosis is a chronic clonal myeloid neoplasm characterized by (1) anemia; (2) neutrophilia and thrombocytosis or, in a minority, thrombocytopenia and leukopenia; (3) splenomegaly; (4) immature granulocytes, increased cluster of differentiation (CD) 34+ cells, erythroblasts, and teardrop-shaped red cells in the blood; (5) marrow fibrosis; and (6) osteosclerosis. The disorder originally was described by Heuck¹ in 1879 under the title “Two Cases of Leukemia and Peculiar Blood and Bone Marrow Findings.” In his monograph, Silverstein traced the history of the concepts set forth during the first half of the 20th century to explain the pathogenesis of this disease, including its origin in the marrow, the appearance of extramedullary hematopoiesis, and the relationship of fibrosis to hematopoietic changes.² The disease occurs, often, de novo but can develop as a later phase of polycythemia vera (Chap. 84), essential thrombocythemia (Chap. 85), or, rarely, chronic myelogenous leukemia (Chap. 89), or an atypical myeloproliferative disease (Chap. 89). It may be preceded by a prefibrotic phase (see “Special Clinical Features: Prefibrotic Primary Myelofibrosis” below). Numerous designations for the disease have been proposed or used, and different names were preferred in different countries.³ Primary myelofibrosis has been designated the most recent “official” name of the disease by a working group on classification of the World Health Organization.³ This compromise selection is unfortunate as the fibrosis is secondary, not primary, and the choice omits focusing on the central pathologic change: a clonal myeloid disease with the singular finding of neoplastic (dysmorphic) megakaryocytopoiesis.⁴ It is a disease of cells, not fibers, and is the only cancer in the medical lexicon not so designated; rather, it is named for an epiphenomenon in the extracellular matrix.

The discovery that the mutated Janus kinase 2 (JAK2) gene mutation can play a role in the causation and behavior of several myeloproliferative diseases⁵ and that approximately 50 percent of cases of primary myelofibrosis have a mutation of the JAK2 gene⁶ or, far less commonly, thrombopoietin receptor (MPL) gene has led to better understanding of the pathogenesis of the disease and its relationship to other myeloproliferative diseases. The JAK signaling pathways also represent an important new target for therapy. The subsequent observation of mutations of the calreticulin (CALR) gene in approximately 70 percent of patients without a JAK2 or MPL mutation has provided deeper insights into the pathogenesis of myelofibrosis, the opportunity to examine interactions among somatic mutations in the cells of patents, and the opportunity for new therapeutic approaches (see “Etiology and Pathogenesis” below).⁷

INCIDENCE

Age and Sex

Primary myelofibrosis characteristically occurs after age 50 years. The median age at diagnosis is approximately 65 to 70 years,^{8,12,13,14,17} but the disease also can occur in neonates and children.^{2,12,14,18,19,20} In infants, the disorder can mimic the classic disease or show certain features but not others, such as absence of hepatosplenomegaly.¹⁹ Familial infantile myelofibrosis mimics the adult disease and in some cases is transmitted by autosomal recessive inheritance.^20,21,22,23 The occurrence of primary myelofibrosis in children usually is in the first 3 years of life.^20,24,25 In young children, girls are afflicted with the disease twice as frequently as boys.¹⁹ In young and middle-age adults, the disease is similar to that in older subjects, although the proportion of indolent cases may be higher.^18,20,26 In adults, the disease occurs with about equal frequency in men and women.^{8,12,13,14,15,16} Like virtually all clonal myeloid diseases,²⁷ primary myelofibrosis can cluster in families, suggesting transmission of an unidentified predisposition gene.^28,29,30 A large Swedish study found a significant relative risk (five- to sevenfold) for a familial occurrence of another myeloproliferative disease neoplasm, although not specifically primary myelofibrosis. The latter finding may relate to the small number of cases of primary myelofibrosis in that study.¹⁷ The incidence of the disease is approximately 0.5 cases per 100,000 population per year in northern European countries.^31,32,33,34 A survey in Olmstead County, Minnesota reported an incidence of 1.5 case per 100,000 population per year and a median age of onset of 67 years,³⁵ similar to other studies (see above in this section).

EXOGENOUS FACTORS

Exposure to high concentrations of benzene^36,37,38 or very-high-dose ionizing radiation³⁹ preceded the development of primary myelofibrosis in a very small number of cases in the past. Epidemiologic studies have not determined a relative risk of myelofibrosis after high-dose benzene exposure. Lower-level benzene exposure was not found to be associated with the risk of myeloproliferative disease in a comprehensive study.⁴⁰ Thus, the few case reports must be viewed with caution in assessing causation. Benzene in exposures greater than 40 to 200 ppm-years is associated with an increased relative risk of acute myelogenous leukemia (AML)^41,42 but not of chronic myelogenous leukemias (CMLs; Chap. 89) These levels of exposures are far above the current levels permitted by the Occupational Safety and Health Administration.

IMMUNE MECHANISMS

Reports of myelofibrosis in patients with lupus erythematosus, have suggested the possibility of immunologic-mediated hyperplasia of marrow connective tissue (see “Immune and Inflammatory Manifestations” below).² These forms of myelofibrosis are different from the monoclonal multipotential hematopoietic cell disease, which is the principal type considered in this chapter.

CLONAL MYELOID DISEASE, ANIMAL MODELS, AND ACTIVATING MUTATIONS

The disease arises from the neoplastic transformation of a single hematopoietic multipotential cell, a conclusion derived from the presence of clonal cytogenetic abnormalities in patients with an identifiable chromosomal abnormality and in studies in women with primary myelofibrosis who also were heterozygous for isotypes A and B of glucose-6-phosphate dehydrogenase (G6PD).^43,44 Although the nonhematopoietic tissues of these patients expressed both isotypes, each patient had blood cells with only one G6PD isotype. The findings strongly imply the blood cells of each patient arose from only one transformed stem cell. Furthermore, chromosome studies of colonies of hematopoietic progenitor cells in primary myelofibrosis established that the same clonal cytogenetic abnormality is present in erythroblasts, neutrophils, macrophages, basophils, and megakaryocytes.⁴⁵ These studies were confirmed by (1) examining X-linked restriction fragment length polymorphisms in women with primary myelofibrosis with heterozygosity for X chromosome-linked genes^46,47 and (2) verifying the presence of a mutation of codon 12 of the N-RAS gene in five blood cell lineages of a patient with the disease.^48,49 Lymphocyte derivation from the clone has been noted using mutation in codon 12 of the RAS gene as the marker.⁴⁸ Using fluorescence in situ hybridization (FISH) analysis, T and B lymphocytes were found to be derived from clonal expansion of a multipotential hematopoietic cell in three of four patients with primary myelofibrosis with a 13q– or 20q– clonal cytogenetic abnormality.⁵⁰ Primary myelofibrosis can be distinguished from secondary myelofibrosis in women by clonality studies.⁵¹ The discovery of the JAK2^V617F mutation has permitted this marker to be used in assessing clonality (as will mutations in CALR and MPL as their determination reaches the clinic). Mutated JAK2-containing cells were identified in all blood lineages of patients with primary myelofibrosis and in the common multipotential hematopoietic cell.⁵²

The neoplastic hematopoietic stem cells in primary myelofibrosis containing the JAK2^V617F mutation behave differently from the same cell population in polycythemia vera when studied in nonobese, diabetic, severe combined immunodeficient mice. Although studied in a nonhuman environment, the findings are consistent with the presence of JAK2^V617F mutations in three phenotypically different myeloproliferative diseases: polycythemia vera, essential thrombocythemia, and primary myelofibrosis.⁵³

Animal models allowed the derivation of the murine myeloproliferative leukemia (mpl) virus, carrying the oncogene v-mpl, which in mice produced a syndrome having features of a mixed idiopathic myelofibrotic–polycythemic disorder (Chap. 111).⁵⁴ The availability of v-mpl led to the isolation of the thrombopoietin receptor and its ligand thrombopoietin.⁵⁵ Later models of myelofibrosis and osteosclerosis, mimicking some of the important features of human primary myelofibrosis, were induced in mice by retroviral-mediated overexpression of thrombopoietin.^56,57 The concomitant high levels of fibroblastic factors (transforming growth factor [TGF]-β₁ and platelet-derived growth factor [PDGF]) resulted in intense fibrosis.⁵⁸ In this model, increased osteoprotegerin was thought to be the principal cause of osteosclerosis.⁵⁹ The disease was cured by murine hematopoietic stem cell transplantation.⁵⁶ These findings were extended and led to the discovery of the human analogue of v-mpl, c-MPL that encodes the human thrombopoietin receptor. Two gain-of-function somatic mutations, MPL^W515L and MPL^W515K, were found to be associated with primary myelofibrosis and essential thrombocythemia.^60,61 These gain-of-function mutations may also be germline and in such a case are associated with familial (hereditary) thrombocytosis.⁶²

A syndrome in mice that results from the GATA-1 (low) mutation also leads to a phenotype that closely simulates human myelofibrosis. The mice gradually develop anemia, teardrop poikilocytes, myeloid immaturity, marrow fibrosis, extramedullary hematopoiesis, and overexpression of profibrotic cytokines in marrow.⁶³ GATA-1 is a transcription factor required for normal megakaryocyte development. GATA-1 deficiency in mice leads to increased megakaryocytic proliferation, followed by myelofibrosis and osteosclerosis, as a result of exaggerated elaboration of fibroblast-inducing and osteoblast-stimulating factors.^64,65

The discovery in 2005 that an activating somatic G→T point mutation (V617F) in JAK2 was associated with the three major myeloproliferative diseases—polycythemia vera, essential thrombocythemia, and primary myelofibrosis—has rapidly led to a fuller understanding of the pathogenesis of these diseases.^48,66 A dominant, gain-of-function mutation in the gene JAK2 residing on chromosome 9p24, which encodes the JAK2 tyrosine kinase, is present in approximately 50 percent of patients with primary myelofibrosis, in approximately 95 percent of patients with polycythemia vera (Chap. 84), and in approximately 60 percent of patients with essential thrombocythemia (Chap. 85), but is absent in healthy individuals.^67,68 In confirmation, the expression of the mutated human JAK2 gene transferred into mice can induce a myeloproliferative disease with features characteristic of the human disorders.^{69,70,71,72,73,74} Homozygosity results from allelic duplication as a result of uniparental disomy of chromosome 9p, not loss of the normal allele corresponding to the mutation.⁷²

It is not yet precisely known how JAK2^V617F, the most prevalent mutation, links the three diseases and what modifiers explain the dramatically different phenotype and expected survival of the patients with polycythemia and primary myelofibrosis. At least four other modifying factors have been proposed to account for the different phenotypes and the apparent absence of the mutation in a high proportion of patients with primary myelofibrosis: (1) gene dosage, (2) germline modifiers, (3) predisposition alleles, and (4) additional somatic mutations.^66,68,73 An example of each influence follows. There is an increasing JAK2^V617F allele burden from essential thrombocythemia, to polycythemia vera, to primary myelofibrosis.⁷⁴ For example, the JAK2^V617F allele burden may be a key determinant of the degree of myeloproliferation and myeloid metaplasia reflected by significantly higher levels of white blood cell counts, CD34+ cell counts, lower platelet counts, and a higher frequency of splenomegaly in homozygous polycythemia vera patients compared to their heterozygous counterparts. These findings are consistent with JAK2^V617F-positive chronic myeloproliferative disorders as a biologic continuum with phenotypic presentation in part influenced by JAK2^V617F mutational load.⁷⁴ Myeloproliferative neoplasm predisposition alleles could provide a selective advantage for the development of mutations in the JAK2 signaling pathway. A number of pre-JAK2 alleles, which arise in cells prior to JAK2 mutation, may contribute to the phenotype displayed.⁶⁸

A mutation in the thrombopoietin receptor gene, MPL, (MPL^515L/K) has been found in some mutant JAK2-negative patients with primary myelofibrosis. The finding of an activating JAK2 (~50 percent of patients) or MPL (~4 percent of patients) mutation, which employs JAK2 for signaling, reinforces the critical role of unregulated activation of the JAK-STAT (signal transducer and activator of transcription) signaling pathway in the pathogenesis of primary myelofibrosis.^75,76,77 Using next-generation gene whole-exome sequencing, primary myelofibrosis patients who did not have mutated JAK2 or MPL genes (approximately half of all patients) were only found to have an occasional somatic mutation.⁷⁸

It came as a surprise that nearly a decade after the description of the JAK2 mutation in primary myelofibrosis and related myeloproliferative diseases, two groups reported, simultaneously, the presence of CALR mutations in approximately 70 percent of patients with primary myelofibrosis who did not carry the JAK2 or MPL mutations (35 percent of total population).^79,80 The CALR mutations were either from deletions or insertions, explaining the failure to identify them by Sanger sequencing technology. The 52-bp deletion (type 1 mutation) or 5-bp insertion (type 2 mutation) are the most frequent types, accounting for approximately 85 percent of the CALR mutations. The CALR mutations are missense for the gene’s final domain, encoding the calcium binding site.

Patients with primary myelofibrosis and JAK2, CALR, or MPL mutations may have an additional two to four somatic mutations. Some of the principal genes mutated are TET2, ASXL1, DNMT3A, EZH2, and IDH1, which are implicated in epigenetic regulation, and TP53 and CBL.⁷ As many as five somatic mutations could be identified in 3 percent of patients, whereas 12 percent of patients with primary myelofibrosis did not have an identifiable somatically mutated gene, using targeted next-generation whole-exome gene sequencing.⁷ However, the proportion of these additional somatic mutations was less that that detected by similar studies in polycythemia vera,⁸¹ suggesting that these sequencing technologies, capable of uncovering only approximately 1 percent of the genome, may leave unidentified additional somatic and germline mutations present in a much greater proportion of primary myelofibrosis patients.

Isolated genetic findings in individual patients have included (13q14) deletions, mutation or overexpression of the retinoblastoma gene,^82,83 NF1 (17q11) deletions,⁸⁴ RAS mutations in approximately one in 20 patients studied, and occasional patients with mutations in KIT.⁷¹ Uniparental disomy has been found on chromosomes 9p (site of JAK2, creating a second allele of V617F) and 1p.⁸³

HMGA2, a gene on chromosome 12, normally is not expressed in humans and is implicated in mesenchymal tumors. HMGA2 was expressed in 12 of 12 patients with idiopathic myelofibrosis studied, implying that, if confirmed, expression of this gene in myeloid cells may play a role in the disease.⁸⁷

CENTRALITY OF CD34+ CELL EGRESS AND NEOPLASTIC MEGAKARYOCYTOPOIESIS

Neoplastic megakaryopoiesis is the most prominent alteration in this clonal myeloid disease and is responsible for most of its major manifestations. Constitutive mobilization and circulation of CD34+ cells are prominent features of the clonal expansion. This phenomenon, probably, is the result of epigenetic methylation of the CXCR4 promotor, a resultant decrease in CXCR4 mRNA, decreased expression of CXCR4 on CD34+ cells, and their resultant enhanced migration into the blood in primary myelofibrosis patients.⁸⁵

Circulating CD34+ cells in patients with primary myelofibrosis generate about 24-fold the number of megakaryocytes in culture than do CD34+ cells from normal subjects, express increased levels of BCL-XL, and have delayed apoptosis.^86,87 Media conditioned with CD61+ cells (presumptive megakaryocytes) elaborated greater quantities of growth factors and proteases, including TGF-β and metalloprotease-9, than did CD61+ cells generated from normal CD34+ cells.

Circulating CD34+ cells in patients with primary myelofibrosis also had a higher expression of eight genes (CD9, GAS2, DLK1, CDH1, WT1, NFE2, HMGA2, and CXCR4) than did normal CD34+ cells. These genes or subsets of them are likely related to disease pathogenesis and were shown to be related to specific manifestations in patients (e.g., increased expression of CD9 and DLK1 with platelet count and WT1 with severity score).⁸⁸

ENHANCED ANGIOGENESIS AND SPLENIC ENDOTHELIAL CELLS

Microvessel density and marrow blood flow are increased in patients with myelofibrosis. These changes may be related to an increase in circulating endothelial cell progenitors.⁸⁹ The endothelial cells examined by laser microdissection, cell sorting, or cell culture from spleen capillaries and splenic vein contained the JAK2^V617F mutation in 12 of 18 patients studied who had the mutation in their granulocytes. A definitive explanation for this finding has not been established.⁹⁰

DYSFUNCTION OF HEMATOPOIESIS

Neoplastic myeloproliferation usually is the dominant marrow abnormality in the granulocytic and megakaryocytic lineages resulting in intensely cellular marrows and mild to moderate blood granulocytosis and thrombocytosis. Hypocellular hematopoiesis, resulting from exaggerated apoptosis of very early precursors, can be present initially or emerge later, leading to granulocytopenia and/or thrombocytopenia. Anemia is a frequent finding and results from a combination of decreased erythropoiesis, shortened red cell survival, and the effects of splenomegaly on the distribution of red cells in the circulation. Hemolysis can be a prominent factor in some cases. Neoplastic megakaryocyte expansion and intense dysmorphogenesis of megakaryocytes are constant features of the disease. Even in intensely fibrotic marrows with severe decreases in erythroid and granulocytic precursors, clusters of megakaryocytes are easily found interspersed between collagen bundles. The term megakaryocytic myelosis, one of the many synonymous terms for the disease, exemplifies the constancy of this finding. The dominance of megakaryopoiesis may relate to the average fivefold overexpression of FKBP51 in megakaryocytes in primary myelofibrosis and the marked predisposition of CD34+ cells to differentiate into megakaryocytes (see “Centrality of CD34+ Cell Egress and Neoplastic Megakaryocytopoiesis” above). FKBP51 increases resistance to apoptosis, possibly by an effect through the calcineurin pathway.⁹¹ Thus, the designation “chronic megakaryocytic leukemia” would be a more accurate designation for primary myelofibrosis, if one used an internally consistent classification.⁷ Although elevated levels of thrombopoietin (and interleukin [IL]-6 and IL-11) are found in the serum of patients with primary myelofibrosis, their etiologic role in the human disease is unresolved.⁹² A marked increase in the thrombopoietin receptor MPL is observed on the platelets and megakaryocytes of a proportion of patients with primary myelofibrosis.⁹³ Despite the animal models of thrombopoietin-induced myeloproliferation and osteomyelofibrosis and the apparent abnormality of MPL receptor sites on human megakaryocytes, autonomous megakaryocyte growth, characteristic of human primary myelofibrosis marrow in culture, has not been associated with either an autocrine effect of MPL ligand (thrombopoietin) or of a mutation in MPL.

FIBROPLASIA

Four of the five major types of collagen⁹⁴ are present in normal marrow: type I in bone, type III in blood vessels, and types IV and V in basement membranes. The fine reticulin fibers that appear after silver impregnation of marrow are principally type III collagen. They do not stain with trichrome dyes. The thicker collagen fibers are principally type I collagen and stain with trichrome dyes, but do not impregnate with silver. The amount of the very fine fibrous network barely perceptible in normal marrow that is stained by silver impregnation techniques⁹⁵ increases in the marrow of patients with primary myelofibrosis (Table 86–1 and Fig. 86-1C).⁹⁶ The fibrous network contains collagen and occasionally progresses to include thick collagen bands that are evident with trichrome stains. Collagen types I, III, IV, and V are increased in myelofibrosis, but type III collagen is increased uniformly and preferentially.^{97,98,99,100,101,102,103,104} The latter occurrence accounts for the increased plasma concentration of procollagen III aminoterminal peptide, a component of collagen type III, which is cleaved during collagen biosynthesis.^96,101,102 Serum prolyl-hydroxylase and marrow and plasma fibronectin also increase in patients with idiopathic myelofibrosis or myelofibrosis from other causes.^98,99 Several other matrix materials are increased in marrow or plasma (Tables 86–1 and 86–2).^105-119

Marrow fibrosis in primary myelofibrosis is most closely correlated with increased neoplastic and dysmorphic megakaryocytes in the marrow. Even densely fibrotic marrow with little residual granulopoiesis or erythropoiesis usually has numerous megakaryocytes scattered throughout the fibrotic areas.^96,120 The increased pathologic emperipolesis (the entry of neutrophils and other marrow cells into the canalicular system of megakaryocytes), evident in human primary myelofibrosis and in mouse models, suggests this may be an additional mechanism of α-granule injury and release of TGF-β and PDGF.¹²¹ Animal models also indicate that marrow monocytes and macrophages may play a subsidiary role in the induction of fibrosis.^121,122,123 Secretion of PDGF, basic fibroblast growth factor (bFGF), and TGF-β from monocytes that are part of the clone have the potential to act as myeloproliferative growth factors and profibrotic cytokines.¹¹⁴

The increased content of marrow collagen types I and III results from release of fibroblast growth factors, which include PDGF,^123,124 epidermal growth factor,¹²⁶ endothelial cell growth factor,¹²⁶ TGF-β,^114,127,128 and bFGF,^114,129 each of which is present in megakaryocyte α granules. Other factors, such as tumor necrosis factor α, IL-1α, IL-1β, and lysyl oxidase, which can be released from marrow cells, also can stimulate fibroblasts.^130,131,132 Platelet factor 4, also derived from megakaryocytes, inhibits collagenase and could contribute to collagen accumulation,¹²⁰ although studies showing a poor correlation between plasma platelet factor 4 concentration and marrow fibrosis have dampened enthusiasm for the role of this factor.¹³³ Substance P, a peptide that acts as a neurotransmitter and a modulator of immune and hematopoietic functions, is increased in the fibrotic marrow and colocalizes with fibronectin. It is angiogenic and is a fibroblast mitogen.¹¹⁵ Its precise role in the complex interactions among fibroblasts, cytokines, and matrix protein deposition is not clear. The high urinary excretion of platelet-derived calmodulin, a putative fibroblast growth factor, in patients with myelofibrosis has added this compound to the array of factors that may contribute to the fibroplasia.¹³⁰ The plasma level of matrix metalloprotein III is decreased and the level of tissue inhibitor of metalloproteinase is increased in patients with idiopathic myelofibrosis.¹³⁴ The expression of matrix metalloproteinase-14 in marrow increases by nearly two orders of magnitude as fibroplasia progresses during the course of the disease; and, megakaryocytes and endothelial cells are the major sources of this protein.¹⁰⁷ Neutrophil collagenase (matrix metalloproteinase-8) content is decreased early in the disease.¹⁰⁷ Bone morphogenetic proteins (BMPs) also are implicated as a contributory factor in fibroplasia. BMP1, BMP6, BMP7, and BMP-receptor 2 are increased in marrow in myelofibrosis as a result of release from megakaryocytes and stromal cells. These proteins are activators of latent TGF-β₁ and processors of collagen precursors. In addition, TGF-β₁ induces release of BMP6.¹⁰⁸

This complex combination of alterations contributes to matrix deposition. The pathogenetic role of released growth factors in fibroplasia is not completely understood. Generalizations from in vitro experiments or correlation between two variables provide only a limited perspective. For example, TGF-β can stimulate or inhibit fibroblast growth, depending on the repertoire of other growth factors in the environment.^127,128

Fibroplasia is associated with an increase in the number and size of marrow sinuses,¹¹¹ the number of endothelial cells,¹³⁵ an increase in vascular volume in the marrow,¹⁰⁶ and an increase in blood flow through the marrow.^136,137 These processes are responsible for the increase in marrow collagen types IV and V and laminin synthesized by endothelial cells in the marrow of patients.¹²⁶

The fibroblastic proliferation in marrow is not an intrinsic part of the abnormal clonal expansion of hematopoiesis.¹³⁸ In cases of primary myelofibrosis in which G6PD isoenzyme studies or chromosome karyotyping establish monoclonal growth of hematopoietic cells, marrow fibroblasts contain both G6PD isoenzymes and do not share the clonal chromosome abnormality.¹³⁹ The findings strongly imply that the fibroblasts differentiate from a primordial cell different from the neoplastic hematopoietic stem cell in primary myelofibrosis and that fibroblast proliferation and enhanced collagen synthesis are secondary results of abnormal hematopoiesis.

EXTRAMEDULLARY HEMATOPOIESIS

Extramedullary hemopoiesis is almost always present in liver and spleen, where it contributes to organ enlargement.^8,9,10 Escape of progenitor cells from marrow and their lodgment in other organs contributes to extramedullary blood cell formation. Reversion of the liver and spleen to their fetal hematopoietic functions (metaplasia) is not a major factor in extramedullary hematopoiesis, and quantitatively significant, effective hematopoiesis does not occur outside of the marrow (see also “Extramedullary (Fibrohematopoietic) Tumors” below).

PRESENTING SYMPTOMS

Some patients are asymptomatic at the time of diagnosis; in which case the disease is detected by medical examination for an unrelated reason. In symptomatic patients, fatigue, weakness, shortness of breath, pruritus, and palpitations are nonspecific but frequent complaints.^{9,10,11,12,13} Fatigue is the most frequent self-reported complaint and is disproportionate to the degree of anemia. Weight loss is common, but anorexia is less so, and fever and night sweats may occur. The term constitutional symptoms, often used in studies of the response to treatment of myelofibrosis, refers specifically to the aggregate occurrence of fever, weight loss, and night sweats.¹⁴⁰ A dragging sensation in the left upper abdomen caused by an enlarged spleen or early satiety from encroachment of the spleen on the stomach may occur. Severe left upper quadrant or left shoulder pain can occur from splenic infarction and perisplenitis. Patients may report unexpected bleeding. Occasionally, bone pain is prominent, especially in the lower extremities. Fever, weight loss, cachexia, night sweats, and bone pain are more frequent later in the course of the disease and are related to the increase in circulating inflammatory cytokines that are a key feature of the disease (see “Immune and Inflammatory Manifestations” below).

PRESENTING SIGNS

Hepatomegaly is detectable in two-thirds of patients, and splenomegaly is present on palpation or imaging studies in almost all patients at the time of diagnosis.^8,9,10,11,12 The spleen is mildly enlarged in one-fourth, moderately enlarged in half, and massively enlarged in approximately one-fourth of patients. Muscle wasting, peripheral edema, and purpura are present infrequently. Bone tenderness may be present. The latter signs may develop in a larger proportion of patients over the course of the disease.

Neutrophilic dermatosis, a syndrome that closely mimics the raised and tender plaques of Sweet syndrome, may occur.^141,142,143 It can be the presenting or a significant complicating feature, and can progress to bullae or pyoderma gangrenosum.^141,144 The dermatopathology of neutrophilic dermatosis is different from leukemia cutis and is unrelated to infection or vasculitis. The predominant histologic lesion is an intense polymorphonuclear neutrophilic infiltrate.

Skin infiltrates related to hematopoietic cells (leukemia cutis) are uncommon.¹⁴⁵ These cutaneous lesions may have myeloid cells with giant cells carrying CD61 markers characteristic of megakaryocytes.^146,147 Skin lesions representing cutaneous fibrohematopoietic tumors may occur.

SPECIAL CLINICAL FEATURES

Prefibrotic Primary Myelofibrosis

The presenting findings of the clonal myeloid diseases are changing because of more and earlier access to healthcare in industrialized countries (see Table 86–2). A subset of patients, perhaps as many as 25 percent, with primary myelofibrosis present without overt reticulin fibrosis in the marrow.^148,149 Blood hemoglobin may be normal and white cell count mildly elevated. The classic findings of frequent teardrop red cells, myelocytes, and nucleated red cells in the blood film and palpable splenomegaly often are absent. Thrombocytosis is a constant finding. Essential thrombocythemia is closely simulated, but observation eventually shows evolution to primary myelofibrosis. The most important distinction with essential thrombocythemia is the nature of the megakaryocytic expansion.¹⁵⁰ In primary myelofibrosis, bizarre changes are evident with wide variation in megakaryocyte size, from very small to giant size cells. Nuclear lobulation is abnormal, with bulky multilobulation, hypolobulation, and free megakaryocyte nuclei in the marrow spaces. In essential thrombocythemia, megakaryocytes are increased but they do not display the dysmorphia observed in myelofibrosis. The prefibrotic disease usually evolves into fully developed myelofibrosis over a period of years. Investigators, evaluating histopathology in a blind fashion, have confirmed the entity of prefibrotic myelofibrosis and this abnormality predicts for progression to an overt primary myelofibrosis and, thus, has an impact on the risk of progression to acute leukemia and prognosis in general.¹⁵¹

Extramedullary (Fibrohematopoietic) Tumors

The appearance of symptoms or signs leading to (1) identification of a mass on imaging regardless of location, (2) appearance of signs or symptoms of an effusion in the thorax or abdomen, (3) unexpected neurologic signs, or (4) another finding that appears unexpected in a patient with primary myelofibrosis should be considered a fibrohematopoietic (extramedullary) tumor(s) until proven otherwise. Foci of hematopoiesis may become clinically apparent as fibrohematopoietic tumors in the adrenal glands,^152,153 renal parenchyma,^154,155,156 and lymph nodes.^157,158,159 Tumors composed of hematopoietic tissue, sometimes with intense fibrosis, can develop in the bowel,^{160,161,162,163} breast,^164,165,166 liver,^167,168 lungs,^169,170,171 mediastinum,¹⁷² pleura and mesentery,^169,171,173 skin,^174,175 synovium,¹⁷⁶ thymus,¹⁶⁹ thyroid,¹⁷⁷ thorax,¹⁷⁸ prostate,¹⁷⁹ spleen,¹⁸⁰ or urinary tract.^{178,181,182,183,184}

Extramedullary hematopoiesis in the intracranial or intraspinal epidural space can lead to serious neurologic complications, including subdural hemorrhage,¹⁸⁵ delirium,^185,186 increased intracranial pressure,¹⁸⁷ orbital apex syndrome,¹⁸⁸ papilledema,¹⁸⁹ cerebral tumor,¹⁹⁰ coma,¹⁹¹ motor and sensory impairment,^192,193 spinal cord compression,^194,195 and limb paralysis.^195,196 Intraspinal myelography,^{193,194,195,196,197,198,199} computed axial tomography,^{185,187,191,192,193,194,195,196,197,199} positron emission tomography after ⁵²Fe infusion,¹⁸⁶ and magnetic resonance imaging (MRI)^198,199 each has been used to define the location and nature of the masses.

Hematopoietic foci on serosal surfaces can produce effusions, sometimes massive, in the thorax,^178,180 abdomen,^{172,173,201,202} and pericardial space.^{203,204,205,206} The effusion fluid often contains megakaryocytes, immature granulocytes, and, occasionally, erythroblasts.^207,208,209 Splenectomy is sometimes followed by extramedullary hematopoietic tumors in soft tissues,²¹⁰ in body cavities, or on serosal surfaces,²⁰⁹ perhaps as a result of an increase in circulating hematopoietic progenitors²¹¹ and loss of the filtration function of the spleen. In rare cases, extramedullary soft-tissue megakaryoblastic tumors simulate the myeloid sarcoma (synonyms: chloroma, granulocytic sarcoma) of other types of myelogenous leukemia.^212,213

Portal Hypertension and Varices and Pulmonary Arterial Hypertension

In patients with primary myelofibrosis, there can be a massive increase in splenoportal blood flow and a decrease in hepatic vascular compliance or the presence of hepatic vein thrombosis, either of which can result in severe portal hypertension, ascites, esophageal and gastric varices, intraluminal gastrointestinal bleeding, and hepatic encephalopathy.^214,215,216 The hepatic venous pressure gradient, normally less than 6 torr, is markedly elevated.²¹⁷

Perisinusoidal fibrosis,^218,219,220 collagen bundles in the spaces of Disse,²¹⁹ perisinusoidal fibroplasia,^{218,219,220,221} and foci of hematopoietic cells^219,222 appear to contribute to the decreased sinusoidal compliance. Portal vein thrombosis is a complication of primary myelofibrosis and occasionally precedes disease onset.²²³

Rarely, portal hypertension is accompanied by pulmonary hypertension and may result from pulmonary fibrosis¹⁷¹ or hydrodynamic factors.²²⁴ Pulmonary arterial hypertension, also, may be the principal problem.^225,226 Although as many as one-third of patients with primary myelofibrosis have an elevated systolic pulmonary artery pressures (>35 torr), the fraction that is symptomatic is very small. Elevated vascular endothelial growth factor (VEGF) levels, elevated circulating endothelial cells, and elevated marrow microvessel density in patients suggest that proangiogenic factors may contribute to the hypertension.²²⁷ Contrariwise, secondary myelofibrosis with polyclonal hematopoiesis and normal blood CD34 cell concentrations frequently occurs in patients with primary pulmonary hypertension.²²⁸

Immune and Inflammatory Manifestations

Abnormalities of humoral immune mechanisms have been observed in up to half of patients with primary myelofibrosis.^{229,230,231,232,233,234} The array of immune products and events reported includes anti–red cell antibodies,^{233,234,235,236,237} antiplatelet antibodies,^238,239 antinuclear antibodies,^229,230,234 elevated plasma-soluble IL-2 receptor,²⁴⁰ anti-Gal (galactoside determinants) antibodies,²⁴¹ anti–γ-globulins,^229,230,234 antiphospholipid antibodies,^234,242 antitissue or organ-specific antibodies,^231,233 and circulating immune complexes,^{234,243,244,245} as well as complement activation,^234,246 immune complex deposition,²³¹ interstitial immunoglobulin deposition,²³¹ increased numbers of marrow plasmacytoid lymphocytes,^231,243 and development of amyloidosis.^{244,245,246,247}

Inflammatory cytokines, including IL-1β, IL-6, IL-8, tumor necrosis factor (TNF)-α, TNF receptor II (TNFRII), and C-reactive protein also are markedly elevated and play a role in the constitutional symptoms seen in patients with progressive disease,²⁴⁸ which explains the often rapid, symptomatic improvement in patients treated with JAK2 inhibitors before splenic size is reduced.

Occasional reports of nonclonal secondary myelofibrosis associated with lupus erythematosus,^{249,250,251,252,253,254} vasculitis,²⁵⁵ polyarteritis nodosa,^234,255 ulcerative colitis,²⁵⁶ scleroderma,²⁵⁷ biliary cirrhosis,^237,258,259 Sjögren syndrome,²⁶⁰ and acute reversible myelofibrosis responsive to glucocorticoids,²⁶¹ although fundamentally different processes from primary myelofibrosis, have raised the possibility that immune mechanisms play a role in the development of marrow fibrosis in some circumstances.

Bone Changes

A large proportion of patients have osteosclerosis at diagnosis or develop osteosclerosis during the course of the disease,^{11,12,13,14,15,262,263,264,265} as reflected by increased bone density on imaging studies and histomorphometric analysis of a bone biopsy (Table 86–3).^{263,264,265,266,267,268} The proximal femur and humerus, pelvis, vertebrae, ribs, and skull may be involved. MRI can uncover evidence of new bone formation and periosteal thickening. Lumbar spine dual-energy x-ray absorption studies and quantitative computed tomography provide evidence for increased bone formation, bone thickening, and higher proportions of cancellous and of woven bone.^268,269 Osteolytic lesions are rare²⁷⁰ and may reflect a myeloid sarcoma.²⁷¹ Periostitis, although infrequent, can lead to debilitating bone pain.²⁷²

Thrombosis

The risk of arterial and venous thrombosis is elevated in patients with primary myelofibrosis, although not to the degree seen in polycythemia vera or essential thrombocythemia.²⁷³ Approximately 10 percent of patients with myelofibrosis will develop a significant thrombotic event during the first 4 years of the disease. The two principal risk factors are an elevated leukocyte count and age, but not platelet count.²⁷⁴ In a large multicenter study of 707 patients with primary myelofibrosis, thromboses occurred in 7.2 percent of patients over the period of observation, or 1.8 percent patient-years. The combination of the JAK2 mutation, leukocytosis, and age predicted the highest incidence of thrombosis.²⁷⁵ The cardiovascular risk factor, such as hypertension, hypercholesterolemia, or smoking, further increase thrombotic risk. Multiple thrombotic episodes may occur; and, the thrombotic event may occur at or just before diagnosis.

Noncirrhotic splanchnic vein thrombosis includes hepatic vein thrombosis (Budd-Chiari syndrome) and portal vein thrombosis, which may occur with minimal evidence of a clonal myeloproliferative disease. In the past marrow examination or evidence of erythropoietin-independent colony growth was used to determine if an occult or incipient myeloproliferative disease may underlie the thrombosis. Now JAK2 mutational analysis can be done and is positive in 35 percent of seemingly idiopathic hepatic vein thrombosis and 25 percent seemingly idiopathic portal vein thrombosis.²⁷⁶ Presumably, future use of CALR gene mutational analysis will increase the proportion of cases of hepatic vein thrombosis indicative of an underlying occult myeloproliferative disease.

BLOOD CELL COUNTS AND MORPHOLOGY

The range of values for blood cell counts at the time of diagnosis is very broad. Normocytic–normochromic anemia is present in most, but not all, patients (see Table 86–2).^{8,9,10,273,274,275,276,277,278,279,280} Mean hemoglobin concentration in a series of patients at diagnosis was approximately 9 to 12 g/dL (range: 4–20 g/dL).^{8,9,10,11,12,13,14,15,16,279,280} Anisocytosis and poikilocytosis are a constant finding. In all cases, teardrop-shaped red cells (dacryocytes) are present in sufficient number to be found in every oil immersion field (see Fig. 86-1). Nucleated red cells are present in the blood film of most patients and average 2 percent of nucleated cells (range: 0 to 30 percent). The percentage of reticulocytes is mildly increased but may vary widely in a given case. A decreased blood hemoglobin may be attributed in part to the expansion of plasma volume and a higher than normal proportion of the red cell volume in an enlarged spleen. Ineffective erythropoiesis can result in a decrease in red cell mass.²⁷⁷ Erythroid hypoplasia is present in many patients.^281,282 In some patients, hemolysis may be prominent, and polychromatophilia and very elevated reticulocyte counts can occur.^278,279 The antiglobulin (Coombs) test usually is negative, but red cell autoantibodies can develop and lead to immune-mediated hemolysis,^{234,235,236,283} which rarely has been a presenting finding of the disease.²³⁶ Occasional patients have had a positive acid hemolysis and sucrose hemolysis test, reflecting a concurrent clone of cells consistent with paroxysmal nocturnal hemoglobinuria.²⁸⁴ Acquired hemoglobin H disease, coincident with typical white cells and platelet changes of myelofibrosis, can occur²⁸⁵ and results in hemolysis, hypochromic–microcytic red cells, marked poikilocytosis, and hemoglobin H inclusions that stain with brilliant cresyl blue. Red cell aplasia, in association with primary myelofibrosis, has been observed.^280,286

The total white cell count usually is usually moderately elevated as a result of granulocytosis.^{8,9,10,11,12,13,14,15,16} The mean total blood white cell count was 10,000 to 14,000/μL (10 to 14 × 10⁹/L) in four large studies. Neutropenia, however, is present in approximately 20 percent of patients at the time of diagnosis.^{8,9,10,11,12,13,14,15,16} The range of white cell counts was 400 to 237,000/μL (0.4 to 237.0 × 10⁹/L) at the time of diagnosis.^{8,9,10,11,12,13,14,15,278,279} Myelocytes and promyelocytes are present in small proportions in the blood film in most patients, and a low proportion of blast cells (0.5 to 2.0 percent) may be found in the blood film. The blood blast cells range from 0 to 20 percent at the time of diagnosis. In patients with blast counts at the high end, which is unusual at presentation, the disease merges with or may progress rapidly to AML. Hypersegmentation, hyposegmentation (acquired Pelger-Huët anomaly), and abnormal granulation of neutrophils may be present.^{8,9,10,11,12,13,14,15,16} Neutrophil alkaline phosphatase scores may be elevated (25 percent of patients) or decreased (25 percent of patients).²⁸⁷ The percentage of basophils may be slightly increased.²⁷⁹ The mean platelet count in patient series ranges from 175,000 to 580,000/μL (175 to 580 × 10⁹/L) at the time of diagnosis. Individual platelet counts can range from 15,000 to 3,215,000/μL (15 to 3215 × 10⁹/L).^{8,9,10,11,12,13,14,15,16,278,279} The platelet count is elevated above the normal upper limit in approximately 40 percent of patients.²⁷⁹ Mild to moderate thrombocytopenia is present in approximately one-third of patients at the time of diagnosis, particularly if splenomegaly is massive. Giant platelets and abnormal platelet granulation in the blood film are characteristic features of the disease.

Approximately 10 percent of patients present with pancytopenia because of severe impairment of hematopoiesis affecting each cell lineage, coupled with sequestration in a massively enlarged spleen. Pancytopenia usually is associated with intense marrow fibrosis.

Increased concentrations of multipotential,^288,289 granulocytic,^290,291 monocytic,²⁹¹ erythroid,²⁹² and megakaryocytic²⁹³ progenitor cells are present in the blood of patients, as measured by clonogenic assays in semisolid cultures. The frequency of hematopoietic progenitor cells in the blood is correlated with the extent of marrow reticular fiber density.²⁹³ Megakaryocytes also are present in the systemic venous blood.²⁹⁴ An increase in blood CD34+ cells is very characteristic of primary myelofibrosis, and the concentration of these cells lends weight to the diagnosis. The height of the CD34+ cell count is correlated with the extent of disease and disease progression. Greater than 15 × 10⁶/L blood CD34+ cells is virtually diagnostic of primary myelofibrosis, and patients with greater than 300 × 10⁶/L CD34+ cells have more rapid progression of disease than patients with fewer CD34+ cells.²⁸⁹

Endothelial progenitor cells (CD34+CD133+ and VEGF receptor 2–positive cells) are significantly higher in the blood of primary myelofibrosis patients than of normal subjects.⁸⁹

Mild lymphocytopenia resulting from decreased CD3+, CD4+, CD8+, and CD3–/CD56+ T cells is the rule.²⁹⁵

FUNCTIONAL ABNORMALITIES OF BLOOD CELLS

The neutrophils of some patients have impaired phagocytosis, oxygen consumption, nitroblue tetrazolium reduction, and hydrogen peroxide generation, and decreased myeloperoxidase^296,297 and glutathione reductase activities.²⁹⁷ CD34+ cells have impaired in vitro differentiation to natural killer cells, which appears to be related to a dysregulation in control of IL-15.²⁹⁸

Bleeding time can be prolonged disproportionately to the platelet count.^299,300 Platelet abnormalities include impaired aggregation in response to epinephrine, depletion of dense granule adenosine diphosphate content,³⁰¹ decreased platelet lipoxygenase pathway activity,³⁰² and others.^303,304 The correlation of bleeding or thrombosis with platelet functional abnormalities is weak.^303,304 The lupus anticoagulant has been present, rarely.²⁴²

MARROW EXAMINATION

Morphology

In the fibrotic phase, marrow aspiration often is unsuccessful because of the fibrosis.^{8,9,10,11,12,13,14,15,16,96,97} The marrow biopsy specimen usually is cellular and shows granulocytic and megakaryocytic hyperplasia (see Fig. 86-1).^{8,9,10,11,12,13,14,15,16,288,289} Erythroid cells may be decreased, normal, or increased in number. Silver stain usually shows an increase in reticular fibers, and in half of patients a striking increase in reticular fibers is seen.²⁸⁹ Hematoxylin and eosin stains of the biopsy specimen may show mild collagen fibrosis; occasionally the fibrosis is extreme (see Fig. 86-1). Collagen fibrosis may be more evident using a Gomori trichrome stain with which collagen characteristically stains green. In intensely fibrotic marrows, cellularity may be markedly decreased but megakaryocytes usually remain evident.²⁸⁹ Giant megakaryocytes and micromegakaryocytes, abnormal nuclear lobulation, and naked megakaryocyte nuclei are present.^{8,9,10,11,12,13,14,15,16,305} Thrombopoietin receptors are decreased on megakaryocytes and platelets.⁹³ Granulocytes may show hyperlobulation and hypolobulation of the nucleus, acquired Pelger-Huët anomaly, nuclear blebs, and nuclear–cytoplasmic maturation asynchrony.³⁰⁶ Clusters of blasts and CD34+ cells are often present. Dilated marrow sinusoids are common. Intrasinusoidal, immature hematopoietic cells, and megakaryocytes are present.⁹⁸ As a reflection of the high blood flow to marrow-bearing bone and the widened sinusoidal system, microvessel density is significantly increased in approximately 70 percent of patients.^306,307 Histomorphometric analysis of marrow biopsies permit detection of osteosclerosis.^263,265,266 Grading of the degree of myelofibrosis has used the Bauermeister scale,³⁰⁸ which assesses fibrosis on a scale of 0 to 4, and the revised European grading scale of 0 to 3.³⁰⁹ Digital imaging may be used for less subjective, quantitative grading of fibrosis or osteosclerosis, if its utility is confirmed in additional studies.³¹⁰

The marrow in the prefibrotic stage usually has no or slight reticular fibrosis. The marrow is cellular and there is often an increase proportion of late neutrophil precursors (myelocytes, metamyelocytes, bands). Myeloblasts and CD34+ cells are inconspicuous. Erythropoiesis may be slightly decreased. Increased and abnormal megakaryocytopoiesis is the hallmark of this phase. Clusters of megakaryocytes are present. Megakaryocytes are large and admixed with small megakaryocytes. Nuclei are often ballooned and have scalloped margins. Bare megakaryocyte nuclei are present. Megakaryocyte involvement is facilitated by staining the marrow with a megakaryocyte marker such as CD61.

Cytogenetic Findings

Chromosome abnormalities of hematopoietic cells are evident in approximately 40 percent of patients at the time of diagnosis.^{311,312,313,314,315,316} The most frequent findings are partial trisomy 1q, interstitial deletion of a segment of the long arm of chromosome 13, del(13)(q12–22), which bears the retinoblastoma gene,^{312,313,314,317} del 20q, and trisomy 8.³¹⁸ Involvement of chromosome 5, 6, 7, 9, 13, 20, or 21 occurs with heightened frequency.³¹⁸ The 5q– abnormality is more prevalent in primary myelofibrosis than any other chronic myeloproliferative disorder. Abnormality of chromosome 12 resulting from several translocations or deletion or inversion occurs in approximately 3 percent of patients.³⁰¹ The del(13) and der(6)t(1;6)(q21–23;p21.3) are associated with myelofibrosis but are not exclusively seen in patients with primary myelofibrosis.³²⁰ Aneuploidy as a result of monosomy or trisomy is common. Pseudodiploidy, manifested by partial deletions and translocations, occurs. Patients with the clinical features of typical primary myelofibrosis very rarely have the Philadelphia (Ph) chromosome in their marrow cells.³²¹ Approximately 15 percent of patients present with unfavorable karyotypes, including three or more abnormalities, +9, –7/7q–, 5/5q–, i(17q), inv(3), 12p–, or 11q23. An unfavorable karyotype is associated with a sixfold greater risk of acute leukemic transformation than a favorable karyotype.³²² With increasing knowledge of the chromosomes commonly affected, interphase FISH of blood cells is used to look for prevalent abnormalities, compensating for the technical difficulties of harvesting cell suspensions, given the intense marrow fibrosis.³¹⁴ Clonal chromosomal abnormalities found in hematopoietic cells have not been observed in marrow fibroblasts.¹³⁹

Magnetic Resonance Imaging

Marrow fibrosis alters the hyperintensity of T1-weighted images that normally results from marrow fat. As cellularity and fibrosis progress, hypointensity of T1-weighted and T2-weighted images develops. MRI does not distinguish between primary myelofibrosis and secondary causes of fibrosis,^264,323,324 but the clinical distinctions usually are very evident from the results of prior physical, blood, and marrow examinations and the evidence for mutations in JAK2, CALR, or MPL. Patchy or diffuse osteosclerosis is a common finding, as are “sandwich vertebrae,” so called because of marked radiodensity of superior and inferior margins of the vertebral body. MRI can identify the uncommon periosteal reactions that usually occur in the distal femur, proximal tibia, or ankle. The reactions represent expansion of marrow cellularity into normally inactive regions of long bones or extramedullary space-occupying lesions of fibrohematopoietic tissue.²⁶⁴ The findings of sodium fluoride (¹⁸F) positron emission tomography can be virtually specific for osteosclerosis of primary myelofibrosis.³²⁵

PLASMA AND URINE CHEMICAL CHANGES

Serum levels of uric acid, lactic dehydrogenase, bilirubin, alkaline phosphatase, and high-density lipoprotein frequently are elevated.^{8,9,10,11,12,13,14,15,16} Serum cholesterol is often decreased and is a negative prognostic factor.^326,327,328 Serum levels of albumin and cholesterol frequently are decreased.³²⁹ Hypocalcemia³³⁰ or hypercalcemia³³¹ may occur. Plasma levels of thrombopoietin and IL-6 are elevated but do not correlate with either platelet or megakaryocyte mass.^332,333 Elevated thrombopoietin is not explained by increased marrow hematopoietic or stromal cell production.³³⁴ Serum-soluble IL-2 receptor³³⁵ and serum VEGF³³⁶ levels are increased. Urinary excretion of calmodulin is approximately three times normal.¹³⁰ The serum contains evidence of increased collagen (see Table 86–1) and bone (see Table 86–2) synthesis.

CML (Chap. 89) should be considered in the differential diagnosis of primary myelofibrosis. In CML, the white cell count is usually greater than 30,000/μL (30 × 10⁹/L) in almost all patients and greater than 100,000/μL (100 × 10⁹/L) in half of patients. In myelofibrosis, the white cell count usually is less than 30,000/μL (30 × 10⁹/L) at the time of diagnosis. In CML, red cell shape usually is normal or slightly perturbed. In myelofibrosis, teardrop poikilocytes are present in every oil immersion field and exaggerated anisocytosis and anisochromia are often prominent. The marrow in CML shows intense granulocytic hyperplasia, with almost 100 percent cellularity and usually no or very slight fibrosis.³³⁷ In myelofibrosis, the marrow has mildly increased cellularity or is hypocellular, with moderate to marked reticulin fibrosis. Occasionally, patients with CML develop intense marrow fibrosis and dysmorphic blood cell changes that make distinction between the two diseases difficult on merely morphological grounds.³¹⁹ However, the Ph chromosome or the BCR-ABL fusion gene is present in CML and is absent in primary myelofibrosis; whereas, the JAK2^V617F or the CALR, or the MPL mutation is present in approximately 90 percent of cases of primary myelofibrosis and absent in CML. Most cases are readily separable based on the aforementioned distinctions.

Patients with primary myelofibrosis may have pancytopenia or bicytopenia and in that respect mimic patients with oligoblastic myelogenous leukemia (myelodysplasia [MDS]; Chap. 87). Contrariwise, patients with oligoblastic leukemia rarely have intense fibrosis.³³⁸ Prominent splenomegaly is expected in patients with primary myelofibrosis but not in patients with oligoblastic myelogenous leukemia, which helps to distinguish the former from the latter patients. The absence of a high frequency of teardrop-shaped red cells, nucleated red cells, and striking anisopoikilocytosis in the blood film mitigates against primary myelofibrosis.

Because some patients with primary myelofibrosis have platelet counts greater than 450,000/μL (450 × 10⁹/L), the diagnosis of primary thrombocythemia may be considered. Moreover, some patients may be in transition from thrombocythemia to myelofibrosis. The anisopoikilocytosis, nucleated red cells, and myeloid immaturity in the blood film characteristic of myelofibrosis are not present in patients with thrombocythemia. Marrow fibrosis usually is insignificant in thrombocythemia, and splenic enlargement often is absent or slight. For these reasons, a clear distinction usually exists between the two disorders.^279,339 The prefibrotic phase of primary myelofibrosis may mimic essential thrombocythemia, but the more prominent splenomegaly and the more disordered and characteristic dysmorphic megakaryopoiesis in primary myelofibrosis can be used to distinguish the two entities by an experienced hematopathologist.³⁴⁰ Careful observation of disease evolution is, also, important.³⁴¹

Hairy cell leukemia (Chap. 93), when associated with shape abnormalities of red cells, pancytopenia, splenomegaly, and fibrotic marrow, can closely mimic primary myelofibrosis.^336,342 Usually, careful scrutiny of the blood and marrow by microscopy, histochemistry, and cell immunophenotype shows evidence of the abnormal mononuclear (hairy) cells characteristic of the disease.

Hepatic disease can be associated with cytopenias and splenomegaly, although the specific blood and marrow findings usually make the distinction with primary myelofibrosis obvious. In a review of 170 cases of splenomegaly in a county hospital, hepatic disease was the second most common cause of massive splenomegaly after primary myelofibrosis.³⁴³

Primary autoimmune myelofibrosis is characterized by intense marrow fibrosis and an increase in marrow polyclonal T and B lymphocytes.^344,345 Serologic or clinical evidence of lupus erythematosus or other connective tissue diseases is absent, giving primary autoimmune myelofibrosis a definitive diagnostic niche. Cytopenias that occur may be immune mediated (e.g., immune hemolytic disease), and the blood cell findings (anisopoikilocytosis, nucleated red cells, myeloid immaturity) characteristic of primary myelofibrosis usually are absent. The marrow may be cellular with increased megakaryocytes, but strikingly dysmorphic megakaryocytopoiesis is absent. Splenomegaly, a nearly constant feature of primary myelofibrosis, usually is absent. Polyclonal hyperglobulinemia may be present.

Patients with sporadic idiopathic or familial pulmonary hypertension have significant marrow fibrosis. They can be distinguished from patients with primary myelofibrosis with pulmonary hypertension by the latter’s high-circulating CD34+ cell count, the presence of clonal platelets and granulocytes, a high frequency of dacryocytes in the blood film, and a JAK2^V617F mutation.³⁴⁶

Metastatic carcinoma, especially derived from carcinoma of breast or prostate^{347,348,349,350,351,352} or disseminated mycobacterial infection,^353,354 can induce reactive marrow fibrosis and occasionally simulate primary myelofibrosis. Demonstration of metastatic carcinoma cells or mycobacteria in the marrow indicates the etiology. Other disorders reported with secondary myelofibrosis include mastocytosis,^{355,356,357,358} angioimmunoblastic lymphadenopathy,³⁵⁹ angiosarcoma,³⁶⁰ lymphoma,^361,362,363 multiple myeloma,^364,365,366 renal osteodystrophy,³⁶⁷ hypertrophic osteoarthropathy,³⁶⁸ gray platelet syndrome,³⁶⁹ systemic lupus erythematosus,^{251,252,253,254} polyarteritis nodosa,²⁵⁶ hypereosinophilic syndrome,^370,371 kala azar,³⁷² primary thrombocytopenic purpura,³⁷³ thrombotic thrombocytopenic purpura,³⁷⁴ tretinoin administration,³⁷⁵ neuroblastoma,³⁷⁶ giant lymph node hyperplasia,³⁷⁷ vitamin D-deficiency rickets,^{378,379,380,381} Langerhans cell histiocytosis,³⁸² acute promyelocytic leukemia,^383,384 and malignant histiocytosis.³⁸⁵ Correction or amelioration of the primary disorder can lead to disappearance of the marrow fibrosis.

Lymphoma,^386,387 chronic lymphocytic leukemia,^388,389 hairy cell leukemia,^342,390 systemic mastocytosis,³⁹¹ macroglobulinemia,³⁹² amyloidosis,^244,245 myeloma,^393,394 malignant teratoma,³⁹⁵ and essential monoclonal gammopathy³⁹⁶ can coincide with primary myelofibrosis.

All clonal hematopoietic diseases (AML, CML, oligoblastic myelogenous leukemia [MDS], lymphomas) may have increased marrow reticulin fibers but only infrequently have collagen fibrosis.³⁹⁷ Acute megakaryoblastic leukemia is accompanied by intense marrow fibrosis (Chap. 88). Approximately 15 percent of patients with polycythemia vera, whether treated by phlebotomy, alkylating agents, or ³²P, develop a clinical state indistinguishable from primary myelofibrosis during 20 years of observation (Chap. 84).^398,399,400 Essential thrombocythemia may evolve into a myelofibrotic stage, estimated to occur in approximately 7 percent of cases (Chap. 85). This estimate is complicated by the question of whether some cases of essential thrombocythemia actually were distinguished as early (prefibrotic) primary myelofibrosis.^340,341 Sideroblastic anemia has progressed to primary myelofibrosis.⁴⁰¹ Rarely, primary myelofibrosis reverts to polycythemia vera, with disappearance of marrow fibrosis.^402,403 Even more rarely, primary myelofibrosis, carrying the JAK2 mutation, has undergone clonal evolution to BCR-ABL–positive CML or vice versa.^404,405

DECISION TO TREAT

A proportion of asymptomatic patients remains stable for years and they do not require specific treatment. Constitutional symptoms, anemia, thrombocytopenia, and splenomegaly are the principal initial reasons for therapy. A hemoglobin less than 10 g/dL,^12,14,20 a white count less than 4000/μL (4.0 × 10⁹/L) or greater than 30,000/μL (30.0 × 10⁹/L),¹³ a platelet count under 100,000/μL (100 × 10⁹/L), and blood blasts above 1 percent of total leukocytes^12,20,406 predict more rapid progression of disease. In addition, patients may have a loss of sense of well-being, fatigue, night sweats, loss of weight, low-grade fever, and loss of functionality as a result of the accompanying elaboration of inflammatory cytokines. Staging protocols may be useful in comparing concurrent and sequential clinical trial results (see “Course and Prognosis” below). In an individual patient under the care of a clinician experienced in the disease, following the disease for evidence of progression is a very important additional factor in determining the timing of treatment, even in patients deemed at higher risk by formulaic techniques, especially before introducing allogeneic stem cell transplantation with its morbidity and potential mortality in this age group.

RED CELL TRANSFUSION

Patients with severe anemia or with moderate but symptomatic anemia may require periodic red cell transfusions (Chap. 138). Alternative mechanisms for raising the blood hemoglobin concentration include use of recombinant erythropoietin and androgen therapy.

RECOMBINANT HUMAN ERYTHROPOIETIN FOR ANEMIA

Serum erythropoietin levels usually are appropriate to the severity of anemia in patients with myelofibrosis.⁴⁰⁷ Thus, use of erythropoietin for anemia is, usually, disappointing. In some studies, patients selected by their inappropriately low serum erythropoietin levels (<125 U/L) for the degree of anemia, beneficial effects can result.^408,409

ANDROGENS AND GLUCOCORTICOIDS FOR ANEMIA

Severe anemia may improve with androgen therapy in some patients.⁴¹⁰ Testosterone, oxymetholone, and fluoxymesterone have been used but have virilizing effects. In addition, they have the potential for hepatic injury and other side effects. Danazol, 600 to 800 mg/day orally for up to 6 months, can be used. The drug is tapered to the minimum effective dose or discontinued if no significant response occurs. Improvement may be limited to a decreased frequency of red cell transfusion. Androgens often are used after splenectomy if anemia returns and requires transfusion of red cells. They are more effective in splenectomized patients or those with less splenic enlargement. Patients undergoing androgen therapy should have periodic assessment of liver size by physical examination, measurement of liver function tests, and, if appropriate, ultrasonographic imaging to detect liver injury (e.g., peliosis) or tumors.⁴¹¹ Evaluation of male patients for prostatic enlargement or cancer is prudent before starting androgen therapy. Patients with significant hemolytic anemia may benefit from glucocorticoid therapy. Prednisone 25 mg/m² per day orally can be tried. If tolerated, the dose can be continued for 1 to 2 months and then tapered gradually. In children, high-dose glucocorticoid therapy can ameliorate marrow fibrosis and improves hematopoiesis.^412,413

DRUG THERAPY FOR MYELOPROLIFERATION, SPLENOMEGALY, OR CYTOPENIAS

A variety of drugs have been used for treatment of massive splenomegaly, thrombocytosis, or constitutional symptoms.

JAK2^V617F Kinase Inhibitors

Because JAK2 mutations and the resulting effects on JAK-STAT signaling are thought to be a key factor in the clonal expansion leading to primary myelofibrosis in at least 50 percent of patients, an effort to synthesize and test inhibitors of the mutant JAK2 protein product were conducted.⁴¹⁴ Early studies of the JAK2 kinase inhibitor TG101209, an oral, bioavailable small molecule, and a potent inhibitor of JAK2 kinase, showed its ability to inhibit JAK2^V617F-dependent phosphorylation of STAT3 and STAT5, inhibit colony growth of cells harboring JAK2 and MPL mutations, and to have therapeutic effects in a nude mouse model of JAK2^V617F-induced myeloproliferative disease.⁴¹⁵ The effects of several JAK2 inhibitors have been described.^416,417,418 Although their effects are somewhat different, the most striking and consistent effect with each agent is a decrease in spleen size and reversal of constitutional symptoms. They often suppress blood cell counts, and thrombocytopenia can be dose-limiting. At least initially, the anemia worsens although this laboratory deterioration may be transient, lasting only for few months. Typically, in spite of progression of anemia, most treated patients report decreased fatigue. The reduction in large spleen size and the improvement in the quality of life of many patients have been dramatic and were similar in those with and without a JAK2 mutation. This effect may be explained by the drug’s ability to inhibit JAK1 and JAK2 isoforms, the former having a role in cytokine elaboration. These agents may decrease morbidity and mortality, prolonging survival (see “Course and Prognosis” below).^{419,420,421,422,423}

In 2011, ruxolitinib, an oral JAK2 inhibitor, was approved by the FDA for use in patients with intermediate or high-risk myelofibrosis. It decreases spleen size, fatigue, night sweats, pruritus, and red cell transfusion requirements, and can result in weight gain in a significant proportion of patients. The principal dose-limiting side effect is a decreased platelet count. Although some patients may have worsened anemia or neutropenia, the net effect often was beneficial, with improvement in fatigue and other symptoms. Headache, dizziness, and diarrhea also may occur but are usually manageable without discontinuing the drug. After 6 months of treatment approximately 40 percent of treated patients have a significant decrease in spleen size and constitutional symptoms. The initial drug trials were limited to patients with a platelet count at the initiation of ruxolitinib therapy of 100 × 10⁹/L; however, newer studies using lower starting doses of ruxolitinib and gradually incrementing those doses have indicated that patients with platelet counts of between 50 and 100 × 10⁹/L may receive similar benefits from carefully incremented drug doses. Table 86–4 shows a suggested approach to initial dosage. Of all available therapeutic modalities for primary myelofibrosis, ruxolitinib is the only therapy that has shown benefit in clinical trials that included a comparison group given a placebo.

Hydroxyurea

Hydroxyurea is a commonly used agent for exaggerated accumulation of platelets, occasional very high leukocyte counts, troublesome areas of extramedullary hematopoiesis, and symptomatic splenomegaly.^424,425,426 Hydroxyurea can, inconsistently, decrease the size of the spleen and liver, decrease or eliminate constitutional symptoms of night sweats or weight loss, and occasionally lead to an increase in hemoglobin concentration, a decrease of platelet counts, and a decrease in the degree of marrow fibrosis. Patients with myelofibrosis often do not have the marrow tolerance to chemotherapy of patients with other chronic myeloproliferative diseases. Hydroxyurea can be administered in doses of 0.5 to 1.0 g/day or 1 to 2 g orally two to three times per week, depending on the level of pretreatment blood cell counts. Patients should be evaluated for dose adjustment at least every week for 1 month and, if appropriate, eventually extended to evaluation every 3 months. Although alkylating agents, especially busulfan and other cytotoxic agents, have been used successfully, they have largely been replaced by hydroxyurea. Use of alkylating agents has resurfaced with the suggestion that melphalan or busulfan may be useful as therapy.^427,428

Thalidomide and Lenalidomide

Thalidomide is poorly tolerated at optimal doses of approximately 800 mg/day. Most patients receive about half that amount and are tapered to the lowest effective dose. One study of 14 patients found the drug was not beneficial and had high toxicity rates.⁴²⁹ Other studies found some decrease in spleen size and improvement in blood hemoglobin and platelet counts in a minority of patients receiving up to 600 mg/day.^430,431 In subsequent studies, lower doses of thalidomide (50 mg/day) coupled with prednisone were more tolerable and resulted in improvement of anemia and thrombocytopenia in about half of patients, with sustained improvement in some patients after treatment was stopped.⁴³² The thalidomide congener lenalidomide may supersede thalidomide use. Lenalidomide has provided responses in a significant minority of patients.^{433,434,435,436} The drug can result in marked improvement in hemoglobin concentration or avoidance of a requirement for transfusion (22 percent of patients treated), improvement in platelet count (50 percent of patients treated), and decrease in spleen size (33 percent of patients treated). Neutropenia and thrombocytopenia were the most troubling side effects.⁴³³ The drug has also been useful in patients with primary myelofibrosis who have a 5q– cytogenetic abnormality.^435,436 Another thalidomide congener, pomalidomide, has completed a phase 3 trial and has not been shown to decrease transfusion requirements.⁴³⁷

Cyclosporine, Etanercept, Imatinib Mesylate, and Tipifarnib

Cyclosporine has been used to achieve a serum level of 100 to 200 ng/mL in severely anemic patients with evidence of immune abnormalities (positive Coombs test, antinuclear antibodies).⁴³⁸ Three of six patients responded with an increased hemoglobin concentration. Cyclosporine has been used with apparent success in a single patient with myelofibrosis and red cell aplasia.⁴³⁹

TNF-α has been proposed as a target to inhibit its possible effects in the pathogenesis of primary myelofibrosis.⁴⁴⁰ Of 20 patients treated with soluble TNF-α receptor (etanercept), 12 had improvement in constitutional symptoms (fever, night sweats, fatigue, weight loss), and four had improved blood counts and decreased spleen size.^441,442

Imatinib mesylate for treatment of myelofibrosis has been examined on empirical grounds and has been largely ineffective in influencing the disease course.^442,443 Modest doses have not been well tolerated, and responses have been infrequent and insubstantial.

The farnesyl transferase inhibitor tipifarnib is not well tolerated.⁴⁴⁴ Although it may decrease spleen size, it has shown no advantages over hydroxyurea.

Interferons

Interferon-α and interferon-γ act synergistically to inhibit myeloproliferation.⁴⁴⁵ Interferon-α has been used extensively for treatment of CML prior to the availability of mutant tyrosine kinase (BCR-ABL) inhibitors (Chap. 89). Interferon-α has not been used extensively in primary myelofibrosis, but has been useful for treatment of splenic enlargement, bone pain, and thrombocytosis in select patients.⁴⁴⁶ Trials comparing interferon therapy with hydroxyurea or other therapy have not been reported.⁴⁴⁷ Hydroxyurea is easier to use (oral versus parenteral) and has less-frequent and less-severe side effects than interferon, especially in older patients. A polyethylene glycol conjugated interferon-α preparation may prove more practical and tolerable for use in patients with myelofibrosis. Although largely ineffective in later stages of myelofibrosis, it has shown efficacy in the early myeloproliferative stage of primary myelofibrosis with mild to moderate marrow fibrosis.^{448,449,450,451}

Serosal Implants

Cytarabine Ascites resulting from peritoneal hematopoietic implants has been treated with intraperitoneal cytarabine.⁴⁵² Intrasplenic cytarabine administered via a splenic artery catheter has resulted in significant improvement in a patient (see also “Radiotherapy” below).⁴⁵³

IMMUNE-RELATED FIBROSIS

Intravenous Immunoglobulin

Although autoimmune or systemic lupus erythematosus-related myelofibrosis has responded to glucocorticoids or intravenous immune globulin^251,254 and a variety of other fibrotic disorders occasionally respond,⁴⁵⁴ primary myelofibrosis does not have a sustained response to such therapy because the fundamental lesion is the hematopoietic multipotential cell neoplasm, neoplastic megakaryocytosis, severe megakaryocytic dysmorphia, and cytokine release with resultant fibrogenesis and, sometimes, osteogenesis.

BISPHOSPHONATES FOR BONE DISEASE

Debilitating bone pain can be a vexing problem in some patients with osteosclerosis and periostitis. Dramatic improvement in bone pain and hematopoiesis after etidronate 6 mg/kg per day on alternate months⁴⁵⁵ or clodronate 30 mg/kg per day for several months, during which marked improvement was still present 33 months later,⁴⁵⁶ highlight the potential usefulness of this family of drugs for bone symptoms.⁴⁵⁷

RADIOTHERAPY

Radiotherapy can be useful for patients with primary myelofibrosis in several situations. For example, in the presence of severe splenic pain (splenic infarctions) or massive splenic enlargement with contraindication to splenectomy (e.g., thrombocytosis), repeated doses of 0.5 to 2.0 Gy to the spleen can ameliorate the pain.⁴⁵⁸ Splenic radiation can result in further cytopenias or worsening cytopenias, especially thrombocytopenia, referred to as an abscopal effect on marrow production, perhaps because of the circulation of large numbers of CD34+ cells exposed in the spleen. Other situations in which radiation may be useful are ascites resulting from myeloid metaplasia of the peritoneum,⁴⁵⁹ focal areas of severe bone pain (periostitis or the osteolysis of a myeloid sarcoma),^272,458,460 and extramedullary fibrohematopoietic tumors,^157,458 especially of the epidural space.¹⁹¹ Low-dose radiation to the liver for symptomatic hepatomegaly and ascites provides only short-term relief.^458,461 Low-dose radiotherapy to the lung has been used successfully to palliate the effects of pulmonary hypertension thought to result from extensive extramedullary hematopoiesis in the organ. Low-dose radiotherapy has relieved signs of respiratory insufficiency, especially hypoxemia,²²⁶ but in several unreported instances known to the authors, this approach has led to deterioration of pulmonary function.

SPLENECTOMY

Splenectomy has been important in the management of primary myelofibrosis.⁴⁶² The major indications for splenectomy include (1) painful enlarged spleen (~50 percent of patients), (2) excessive transfusion requirements or refractory hemolytic anemia (~25 percent of patients), (3) portal hypertension (~15 percent of patients), and (4) severe thrombocytopenia (~10 percent of patients).

Patients who have a prolonged bleeding time or coagulation times are at serious risk for hemorrhage with surgery and should not undergo the procedure unless the abnormalities can be corrected by platelet transfusion and factor replacement therapy. Evidence of low-grade intravascular coagulation, such as elevated D-dimer levels, may require prophylactic heparin therapy and platelet transfusion should excessive bleeding occur.

Removal of the spleen in patients with primary myelofibrosis may be difficult. Usually the spleen is adherent to neighboring serosal surfaces and structures (e.g., inferior surface of left hemidiaphragm) and has numerous collateral vessels and very dilated splenoportal arteries and veins. Immediate postoperative mortality is a function of surgical experience and skill and of the rapidity of recognition of postoperative complications. In experienced hands, perioperative mortality is approximately 10 percent. Postoperative morbidity from hemorrhage, subphrenic hematoma, subphrenic abscess, injury to the tail of the pancreas, pancreatic fistulas, or portal vein stump or mesenteric vessel thrombosis occurs in approximately 30 percent of patients. Infection, especially, pneumonia occurs in approximately 10 percent of patients. Later postoperative changes include liver enlargement (sometimes massive), extramedullary hematopoietic tumors, thrombocytosis, and a decrease in teardrop-shaped red cells. Leukemic blast transformation occurs in approximately 15 percent of patients after splenectomy. Hydroxyurea or aspirin and anagrelide may be useful for exaggerated thrombocytosis (Chap. 85). The morbidity and mortality from splenectomy and the modest extension of life have led to increasing conservatism regarding its use. However, splenectomy can improve the condition for which it was performed in approximately 50 percent of patients. Median survival after splenectomy has been approximately 18 months.

PORTAL-SYSTEMIC VASCULAR SHUNT SURGERY

Circulatory dynamic studies are performed at the time of surgery in patients undergoing operation for portal hypertension and bleeding varices or refractory ascites. In patients in whom the hepatic wedge pressure elevations result from markedly increased blood flow from the spleen to the liver, the preferred treatment procedure for portal hypertension is splenectomy. In patients who have portal hypertension resulting from intrahepatic block or hepatic vein thrombosis and who have a hepatic venous pressure gradient well above the upper limits of normal (6 torr), a splenorenal shunt can be performed⁴⁶³ or, to avoid abdominal surgery, a transjugular intrahepatic portosystemic shunt can be used.^464,465 Variceal sclerotherapy or variceal ligation has been used to treat bleeding varices resulting from portal hypertension.

HEMATOPOIETIC STEM CELL TRANSPLANTATION

Marrow transplantation is the only curative approach to primary myelofibrosis. Marrow transplantation therapy has been used increasingly in younger patients with a poor prognosis (e.g., severe anemia and leukopenia or exaggerated leukocytosis) who have a histocompatible sibling.^{466,467,468,469,470,471,472,473} The median age in most studies is approximately 50 years, whereas the median age of all patients is approximately 70 years.³²² Patients engraft at a rate similar to the rate of patients with hematologic diseases without marrow fibrosis (Chap. 23). The decision to use full-conditioning allogeneic transplantation is a function of the patient’s age (<50 years), the severity of the blood cell and marrow abnormalities, and the likelihood of a protracted indolent course without transplantation. Younger patients, especially those younger than age 50 years, with a DNA-based matched sibling donor, progressive disease, and poor prognostic findings, such as hemoglobin less than 10 g/dL, blast cells greater than 1 percent of blood cells, unfavorable cytogenetics (e.g., abnormalities involving chromosome 5, 7, or 17, or cells with three or more abnormalities) are usually considered for transplantation. Although a large spleen may slightly delay the expression of donor granulopoiesis, on average the results in patients with a spleen is the same as those who had prior splenectomy.^471,474 In addition, the latter procedure incurs significant risk of morbidity or mortality.

Patients younger than age 50 years who are transplanted with stem cells from a matched sibling donor have a lower posttransplantation mortality and have better outcomes than those older than age 50 years.⁴⁶⁰ Transplant-related mortality using full-conditioning regimens is approximately 35 to 40 percent and 5-year survival is approximately 50 percent in patients younger than 50 years of age.

Donor lymphocyte infusion in patients who have lost donor dominance of hematopoiesis and a return of myelofibrosis can result in regression of fibrosis and return to normal hematopoiesis for at least 6 and 20 months at the time of reporting.^475,476

In JAK2^V617F-positive patients, real-time polymerase chain reaction analysis can permit a sensitive determination of residual JAK2^V617F-positive cells after transplantation. In a study of patients receiving low-intensity conditioning, 17 of 21 patients became JAK2^V617F-negative, and in one case donor lymphocyte infusion eliminated JAK2^V617F-positive cells.⁴⁷⁷

The option of nonmyeloablative transplantation in older patients is gaining favor.^{470,471,478,479,480,481,482} Several reports of lower posttransplantation mortality and salutary outcomes have led some to consider this approach as the preferred approach in patients older than 50 years of age, and perhaps in younger patients as well. One study showed that the outcome of nonmyeloablative transplantation was dramatically better than myeloablative transplant.⁴⁷⁰ The study compared 17 patients receiving a myeloablative conditioning regimen and 10 receiving nonmyeloablative conditioning. The median age was 50 years (age range: 5 to 63 years) at transplantation. After a median followup of 55 months, 20 patients were alive. The transplantation-related mortality was 10 percent in the nonmyeloablative group and 30 percent in the myeloablative group. There was no difference in survival for high- or low-risk patients or between sibling and unrelated donor transplantations. This study confirmed prior smaller comparisons of myeloablative and nonmyeloablative stem cell transplantation.⁴⁸³

A large cooperative study of reduced conditioning transplantation found the most favorable results in transplants using matched-related donors and in intermediate stage disease.⁴⁸²

Autologous blood stem cells mobilized with granulocyte colony-stimulating factor (G-CSF) and administered after busulfan conditioning produced clinical benefit, including improved erythropoiesis, improved platelet counts, and decreased splenic size, in a plurality of 21 patients with primary myelofibrosis whose age range was 45 to 75 years. The 2-year actuarial survival rate was 61 percent.⁴⁸⁴ 

The rate of disease progression is associated with at least 16 variables measured at the time of diagnosis, namely: (1) older age; (2) severity of anemia; (3) exaggerated leukocytosis (>25 × 10⁹/L) or leukopenia (<4.0 × 10⁹/L); (4) constitutional symptoms of fever, sweating, or weight loss at the time of diagnosis; (5) proportion of blast cells in the blood (≥1 percent); (6) male gender; (7) severity of thrombocytopenia; (8) proportion of CD34+ cells in the blood; (9) the presence of the V617F mutation in JAK2; (10) monocytosis; (11) a decreased proliferating cell nuclear antigen index and a decreased apoptotic index by in situ end labeling; (12) degree of liver enlargement; (13) extent of marrow fibrosis; (14) postsplenectomy spleen histology; (15) WT1 expression in CD34+ cells; and (16) certain clonal cytogenetic abnormalities, especially involving chromosome 5, 7, or 17, or with three or more abnormalities. Abnormalities such as 13q or 20q did not affect patient survival compared to those patients without cytogenetic alterations.

Each retrospective study has found a different subset of these factors to be significant prognostic factors. The most consistent predictive variables appear to be advanced age, severity of anemia, and higher-risk clonal cytogenetic abnormality at the time of diagnosis, each of which represents a poor prognostic indicator.^{8,12,14,16,312,314,485,486,487,488,489,490}

In a study of more than 1000 consecutive cases of myelofibrosis at seven centers, among which the median survival was 69 months, variables (1) through (5) (listed above) proved to be the most useful in dividing patients into four risk-factor categories. Shorter survival was observed in patients who were older than age 65 years, had a hemoglobin concentration less than 10 g/dL, a leukocyte count greater than 25 × 10⁹/L (25,000/μL), a blood blast cell count equal to or greater than 1 percent, and constitutional symptoms. The patients were assigned to a risk group based on the number of risk factors present. If no risk factors were present, risk was low; if one factor was present, risk was low-intermediate; if two risk factors were present, risk was high-intermediate; and if three or more risk factors were present, risk was high. The application of these variables could distinguish among patients with low risk with a survival of 135 months, low-intermediate risk with a survival of 95 months, high-intermediate risk with a survival of 48 months, and high risk with a survival of 27 months.⁴⁹¹ Overall, the 5-year survival of patients with primary myelofibrosis is approximately 40 percent of the survival expected for healthy age- and sex-matched controls.⁴⁹²

Some data suggest that the use of JAK2 inhibitors in the treatment of primary fibrosis may not only decrease symptomatology and spleen size, but may prolong survival.^493,494 An analysis of 100 patients receiving ruxolitinib compared to 350 carefully matched patients who did not receive ruxolitinib showed a median 18-month prolongation in survival in patients with poorer prognosis disease. At the upper range some patients had a 4- to 5-year prolongation of life.⁴⁹⁴ In addition, the clinical improvement in vitality and functionality can be dramatic.

The major causes of death are infection, hemorrhage, postsplenectomy mortality, and acute leukemic transformation.^{495,496,497,498,499} Acute leukemia occasionally is preceded by the development of myeloid sarcomas.^{35,460,499,500} Evolution of the disease to acute lymphocytic leukemia or lymphoma may occur.^501,502 An increased risk of progression to leukemia has been reported in splenectomized patients.⁵⁰³ Progression to acute leukemia is associated with a blast count greater than 3 percent and a platelet count less than 100,000/μL (100 × 10⁹/L) at the time of diagnosis. Treatment with erythropoietin or androgens is associated with increased risk of progression to acute leukemia, as well.⁵⁰⁴ The administration of JAK2 inhibitors may provide improved efficacy for the treatment of AML evolving in patients with primary myelofibrosis. Rare spontaneous remissions of apparent primary myelofibrosis are documented.^505,506

A variety of gene mutations have been associated with prognosis in patients with primary myelofibrosis, either overall survival or risk of conversion to acute myelogenous leukemia. These include the following gene mutations: IDH, EZH, ASXLI, and SRSF2.^507,508 The A3669G (rs6198) single nucleotide polymorphism of the glucocorticoid receptor contributes to the erythrocytosis in polycythemia and contributes to the risk of evolution to AML.⁵⁰⁹

About 5 to 10 percent of patients with primary myelofibrosis do not have a mutation in JAK2, CALR, or MPL. These so-called “triple-negative” patients with myelofibrosis appear to have a less favorable prognosis as do those whose neoplastic cells have wild type CALR and an ASXL1 mutation.^509A

Primary myelofibrosis in infants and children has a more varied pathobiology than in adults. Patients have been followed for decades without requiring significant treatment,⁵¹⁰ and spontaneous remission has been described.⁵¹¹ Because of its variable course, conservative management may be appropriate while the course of the disease is followed.