Chronic Myeloproliferative Neoplasias

Mutations and prognosis in primary myelofibrosis

Article metrics


Patient outcome in primary myelofibrosis (PMF) is significantly influenced by karyotype. We studied 879 PMF patients to determine the individual and combinatorial prognostic relevance of somatic mutations. Analysis was performed in 483 European patients and the seminal observations were validated in 396 Mayo Clinic patients. Samples from the European cohort, collected at time of diagnosis, were analyzed for mutations in ASXL1, SRSF2, EZH2, TET2, DNMT3A, CBL, IDH1, IDH2, MPL and JAK2. Of these, ASXL1, SRSF2 and EZH2 mutations inter-independently predicted shortened survival. However, only ASXL1 mutations (HR: 2.02; P<0.001) remained significant in the context of the International Prognostic Scoring System (IPSS). These observations were validated in the Mayo Clinic cohort where mutation and survival analyses were performed from time of referral. ASXL1, SRSF2 and EZH2 mutations were independently associated with poor survival, but only ASXL1 mutations held their prognostic relevance (HR: 1.4; P=0.04) independent of the Dynamic IPSS (DIPSS)-plus model, which incorporates cytogenetic risk. In the European cohort, leukemia-free survival was negatively affected by IDH1/2, SRSF2 and ASXL1 mutations and in the Mayo cohort by IDH1 and SRSF2 mutations. Mutational profiling for ASXL1, EZH2, SRSF2 and IDH identifies PMF patients who are at risk for premature death or leukemic transformation.


Primary myelofibrosis (PMF) is a myeloproliferative neoplasm1 whose cardinal features include extensive extramedullary hematopoiesis, associated with marked hepatosplenomegaly, anemia that is often transfusion-dependent and profound constitutional symptoms.2 Median survival is estimated at <6 years with causes of death, including leukemic transformation, progressive cachexia, vascular events and infections.3 At present, the only treatment modality with curative potential is allogeneic stem cell transplantation, but this particular treatment is still associated with a high rate of mortality and morbidity.4 Other treatment options are primarily palliative and include splenectomy, involved field radiotherapy,5 erythropoiesis stimulating agents, androgen preparations and thalidomide and its analogs to treat anemia, hydroxyurea and JAK inhibitors to treat splenomegaly or constitutional symptoms.6, 7, 8 However, none of these drugs have been shown to either reverse bone marrow fibrosis or induce genetic remissions and their effect on survival is uncertain.

Physicians taking care of patients with PMF are often faced with therapeutic decisions that include the indication and timing of allogeneic stem cell transplantation or investigational drug therapy. Such decisions rely upon currently available prognostic scoring systems, recently developed by the International Working Group for myeloproliferative neoplasm Research and Treatment (IWG-MRT), including the International Prognostic Scoring System (IPSS),3 which is applicable at time of diagnosis, and the dynamic IPSS (DIPSS) that can be applied at any time during the disease course.9 Both IPSS and DIPSS use five adverse factors, including age >65 years, hemoglobin <10 g/dl, leukocyte count >25 × 109/l, circulating blasts 1% and presence of constitutional symptoms, in order to distinguish among low, intermediate-1, intermediate-2 and high-risk patients with respective median survivals of 11.3, 7.9, 4.0 and 2.3 years, per IPSS, or not reached, 14.2, 4.0 and 1.5 years, per DIPSS. DIPSS was recently modified into DIPSS-plus10 by incorporating three additional DIPSS-independent risk factors: platelet count <100 × 109/l,11 red cell transfusion need12, 13 and unfavorable karyotype;14, 15 median survival for the low, intermediate-1, intermediate-2 and high-risk categories were 15.4, 6.5, 2.9 and 1.3 years. The current study seeks to improve upon these prognostic models by incorporating recently described molecular markers (somatic mutations), some of which have already been shown useful in the prognostic assessment of patients with acute myeloid leukemia (AML).16

Patients and Methods


This international collaborative study, approved by the individual center institutional review boards, included 879 patients with PMF diagnosed according to the 2008 WHO criteria;1 patients with the ‘prefibrotic’ variant were excluded. The main objective was to evaluate the prognostic relevance of recently discovered mutations in PMF in the context of currently used prognostic models. In order to examine prognostic relevance of mutations detected at the time of initial diagnosis and further validate their value when encountered at time of tertiary center referral, we used two independent patient cohorts. Initial analysis was performed on 483 European patients studied within 1 year of diagnosis; the seminal observations were subsequently validated in an independent cohort of 396 patients studied at time of referral (including both newly and previously diagnosed patients) to the Mayo Clinic and receiving treatment according to current practice.5

Mutational analysis

For the European cohort, the Sanger technique was used to detect mutations across the entire coding region of EZH2 and TET2, and regions previously described as mutational hotspots for DNMT3A, CBL, ASXL1, IDH1, IDH2, SRSF2 and MPL; JAK2V617F was detected by real-time PCR(RT-PCR). Similar techniques were used for analyzing the Mayo samples for mutations found to be prognostically relevant in the European cohort: ASXL1, SRSF2, EZH2, IDH1 and IDH2 (see Supplementary Methods and Supplementary Table S1 for details).

Gene expression profiling

Gene expression profiling was performed on total cellular RNA isolated from CD34+ cells of a subset of patients from the European cohort, randomly selected among those defined at ‘mutationally high-risk’ or ‘mutationally low-risk’ as described in Supplementary Methods.

Statistical analysis

The main end points were overall and leukemia-free survival. Actuarial survival curves were estimated by the method of Kaplan–Meier and compared by the log-rank test. For the European cohort, survival was measured from the date of diagnosis to the date of death or last follow-up and for the Mayo cohort from time of referral to date of death or last follow-up. Patients who underwent allogeneic stem cell transplantation were censored at the time of transplant. Multivariable analyses were conducted with the Cox proportional hazards regression model. The prognostic strength of mutations to predict transformation into AML was evaluated in the framework of competing risks; that is, the cumulative incidence of progression to AML was calculated by taking AML-unrelated death as competing risk.17 Multivariable analyses of factors predicting progression to AML were also performed within the framework of competing risks by the sub-hazard regression method.18 The mutual exclusivity of pairs of mutations was estimated with two-by-two contingencies tables and Fisher’s exact test. Association between mutations and clinical and biological features was assessed by the χ2 test. P-values <0.05 were considered significant.


Clinical presentation and disease course

Clinical and laboratory characteristics of all 879 study patients, stratified into European (n=483) and Mayo (n=396) cohorts, are outlined in Table 1. Median follow-up was 3.7 years (range 0–27.9) for the European and 3.4 years (range 0–26) for the Mayo cohorts. During follow-up, 157 (32.5%) deaths were recorded in the European cohort and 246 (62%) in the Mayo cohort; 75 (16%) and 46 (12%) leukemic transformations were documented in the European and Mayo cohort, respectively.

Table 1 Clinical and laboratory characteristics of 879 patients with PMF stratified into European (n=483) and Mayo Clinic (n=396) cohorts

For the European cohort, median overall survival from time of diagnosis was 9.7 years (95% CI: 7.9–12.2) (Supplementary Figure S1A); the respective values for IPSS low, intermediate-1, intermediate-2 and high-risk disease were 22.8 years (95% CI: 12.2- not reached), 10.5 years (95% CI: 7.12–23.7), 6.2 years (95% CI: 5.2–9.5) and 2.5 years (95% CI: 1.7–2.8) (P<0.0001) (Supplementary Figure S1B). Unfavorable karyotype had an IPSS-independent detrimental effect on overall survival (HR: 6.46; 95% CI: 3.4–12.5; P<0.001) (Supplementary Figures S2A and B) and, in univariate analysis, on leukemia-free survival (HR: 3.45; 95% CI: 1.03–11.6; P=0.033) (Supplementary Figure S2C); the latter retained borderline significant in multivariable analysis (HR: 3.3; 95% CI: 0.98–11.3; P=0.053). For the Mayo cohort, median survival from time of referral was 5 years and significantly different among DIPSS-plus risk groups (P<0.0001); the median survivals for low, intermediate-1, intermediate-2 and high-risk groups were 18.7, 8.3, 5.2 and 2.3 years (Supplementary Figure S3).

Frequency and distribution of mutations

Among patients in the European cohort, 382 (79%) displayed at least one somatic mutation (see Supplementary Figure S4 for details); the frequency of individual mutations and their interrelationships are shown in Figure 1a. MPL mutations were less frequent (0.7% versus 12%; P<0.001) in JAK2V617F-mutated patients; IDH1 and IDH2 mutations, which were mutually exclusive, were more frequent in DNMT3A-(27% versus 5%; P=0.001) or SRSF2-(36% versus 7%; P<0.001) mutated patients. ASXL1 mutations were more frequent in EZH2-(11% versus 3%; P<0.002), TET2-(11% versus 4%; P=0.043), DNMT3A-(11% versus 2%; P=0.022), CBL-(11% versus 2%; P<0.001) and SRSF2-(13.5% versus 7%; P=0.035) mutated patients. Thirty-one patients (6.4%) displayed >2 mutations (Supplementary Table S2 and Supplementary Figure S5). As of their prognostic relevance in the European cohort (see below), mutation analyses in the Mayo cohort were focused on ASXL1, SRSF2, EZH2, IDH1 and IDH2; the corresponding mutational frequencies were 31% (85 out of 279 studied cases), 15% (52 of 358), 6% (15 of 270), 3% (10 of 374) and 2% (7 of 374). Significant clustering was noted for SRSF2 with both IDH1 (P=0.0002) and IDH2 (P=0.0009) mutations. One patient displayed four mutations (ASXL1, SRSF2, EZH2 and IDH2) and four patients displayed three mutations (two displayed ASXL1, SRSF2 and EZH2 and two others displayed ASXL1, SRSF2 and IDH1 mutations).

Figure 1

The frequency and the pair-wise co-occurrence of mutations in the 483 PMF patients included in the European cohort are represented by a Circos diagram in (a). Co-occurring mutations are indicated in the clockwise direction. In the Circos representation, the length of the arc corresponds to the frequency of mutation in the first gene (color coded) and the width of the ribbon corresponds to the frequency of patients who also had a mutation in the second gene. The frequency of mutations in this cohort is shown on the right side. (b) shows the prevalence proportion of individual mutations in the four IPSS risk categories.

Clinical and cytogenetic correlates of mutations

In the European cohort ASXL1, SRSF2 and EZH2 mutations were significantly (P<0.05) enriched in the IPSS high-risk group with mutational frequencies of 42%, 25% and 12%, respectively (Figure 1b). Also in the Mayo cohort ASXL1 and SRSF2 mutations were more enriched in the DIPSS-plus high-risk category (45%, P=0.0002 and 22%, P=0.004, respectively), with no difference for EZH2 mutations.

In the European cohort, JAK2V617F mutation was associated with older age (P=0.005), higher hemoglobin level (P=0.0001) and increased leukocyte count (P=0.001); EZH2 mutations with leukocytosis (P=0.01) and 1% circulating blasts (P=0.02); ASXL1 mutations with leukocytosis (P<0.001), 1% circulating blasts (P=0.001), anemia (P=0.01), splenomegaly (P=0.007) and constitutional symptoms (P<0.01); SRSF2 mutations with older age (P<0.001), leukocytosis (P<0.01), 1% circulating blasts (P=0.012) and constitutional symptoms (P=0.009). There were no significant associations with cytogenetic risk groups.

In the Mayo cohort, ASXL1 mutations clustered with older age (P=0.01), constitutional symptoms (P<0.0001), leukocytosis (P=0.007) and 1% circulating blasts (P=0.02); SRSF2 mutations with older age (P=0.002) and anemia/transfusion need (P<0.05); EZH2 mutations with 1% circulating blasts (P=0.003). Also in the Mayo cohort, ASXL1 mutations were more likely to occur in the presence of normal karyotype (37% versus 19%; P=0.003) and were not different between favorable and unfavorable cytogenetic categories. EZH2, SRSF2 and IDH1/2 mutational frequencies were similar in patients with normal versus abnormal karyotype.

Overall and leukemia-free survival correlates of mutations

European cohort

Univariate analysis disclosed significant correlations between shortened survival and EZH2 (P=0.0003), ASXL1 (P<0.0001) and SRSF2 (P<0.0001) mutations (Figure 2a). A multivariable analysis that included all three mutations as covariates confirmed their inter-independent prognostic value: HR was 1.91 (95% CI: 1.1–3.36; P=0.025) for EZH2, 2.21 (95% CI: 1.57–3.11; P<0.0001) for ASXL1 and 2.6 (95% CI: 1.63–41.6; P<0.0001) for SRSF2 mutations. Risk of leukemia was significantly increased in EZH2-(P=0.03), ASXL1-(P<0.0001), SRSF2-(P=0.007) and IDH1- or IDH2-(P<0.0001) mutated patients (Figure 2b). In multivariable analysis that included the four mutations as covariates, ASXL1 (HR: 2.5, 95% CI: 1.5–4.1; P<0.0001), SRSF2 (HR: 2.73, 95% CI: 1.34–5.55; P=0.005) and IDH1 or IDH2 (HR: 2.66, 95% CI: 1.10–6.47; P=0.03), but not EZH2 (HR: 1.98, 95% CI: 0.88–4.46), mutations remained significant.

Figure 2

Kaplan–Meier estimates of overall survival in patients of the European series depending on the mutational status (wild-type or mutated) of EZH2, ASXL1, SRSF2 and IDH1 or IDH2 mutation is presented in (a). The impact of mutations in EZH2, ASXL1, SRSF2 and IDH1 or IDH2 on the cumulative risk of transformation to acute leukemia using a competitive risk analysis is shown in (b). The P-values reported in each panel refer to the comparison of mutated and wild-type subjects. A comparison of overall survival by Kaplan–Meier estimates (c) and the cumulative risk of transformation to acute leukemia (d) in patients considered at ‘mutationally high-risk’ (that is, those mutated at least in one of ASXL1, EZH2, SRSF2, IDH1 or IDH2) versus ‘mutationally low-risk’ (that is, no mutation) is presented.

As EZH2, ASXL1 and SRSF2 mutations predicted overall survival, while ASXL1, SRSF2 and IDH1 or IDH2 predicted leukemic transformation, all independent of each other, we considered the four mutations (ASXL1, EZH2, SRSF2 and IDH1/2) as being detrimental for disease outcome, in general, and accordingly classified the European patients into those who displayed at least one (‘mutationally high-risk’) or none (‘mutationally low-risk’) of the four mutations. Median survival was significantly shorter in the mutationally high-risk compared with the low-risk category (81 versus 148 months; P<0.0001) (Figure 2c). Multivariable analysis confirmed the independent prognostic value for survival of distinguishing mutationally low from high-risk groups in the context of IPSS (P<0.0001), cytogenetic risk stratification (P<0.0001) or both (P=0.04). However, when the four mutations were considered individually, only ASXL1 mutations retained their significance in the context of IPSS (HR: 2.02; P<0.001).

Leukemia-free survival was significantly shorter in the mutationally high-risk group (P<0.0001) and the cumulative risk of leukemia using competitive risk analysis was significantly increased (HR: 2.96; 95% CI: 1.85–4.76; P<0.0001) compared with mutationally low-risk (Figure 2d); this was confirmed by multivariable analysis that included IPSS (P<0.001), unfavorable karyotype (P<0.007) or both (P=0.024) (Supplementary Table S3). The significant difference in overall and leukemia-free survival between the mutationally high- and low-risk groups was apparent when lower-risk categories (low plus intermediate-1 risk patients) were considered separately from higher-risk categories (intermediate-2 plus high-risk patients) (Figures 3a and b).

Figure 3

Kaplan–Meier estimates of overall survival in patients at ‘mutationally high-risk’ compared with ‘mutationally low-risk’ in the settings of IPSS categories; a lowest and a highest IPSS category were considered, the first including IPSS low plus Intermediate-1, the latter Intermediate-2 plus high-risk categories (a). In (b), the cumulative risk of transformation to acute leukemia in the same patient populations as in (a) is shown.

Mayo Clinic cohort

Univariate analysis identified ASXL1 (HR: 2.0, 95% CI: 1.5–2.7; P<0.0001), SRSF2 (HR: 1.6, 95% CI: 1.1–2.4; P=0.01), EZH2 (HR: 1.9, 95% CI: 1.1–3.4; P=0.03) and IDH1 (HR: 2.5, 95% CI: 1.1–5.6; P=0.02) mutations as being prognostically detrimental for survival. Prognostic significance was sustained for all but IDH1 mutations in multivariable analysis that included all four mutations. However, when each one of the four mutations was individually evaluated in the context of DIPSS-plus, only ASXL1 mutations remained significant (HR: 1.4, 95% CI: 1.0–1.9; P=0.04). The same held true when all four mutations and DIPSS-plus were included in the multivariable model as covariates (HR: 1.4, 95% CI: 1.0–2.0; P=0.046). Notably, the prognostic relevance of ASXL1 mutations was mostly restricted to patients with either intermediate-2 risk (P=0.03) or intermediate-1 plus low-risk disease (P=0.07) unlike high-risk disease (P=0.78) (Figure 4a).

Figure 4

Three-tiered dynamic DIPSS-plus stratified survival data in 279 patients with PMF sub-stratified according to ASXL1 mutation status (a) and according to ASXL1/SRSF2/IDH1 mutation status (b).

Both univariate (P=0.0004 and P=0.0001, respectively) and multivariable (including karyotype and thrombocytopenia as covariates; P=0.0002 and P=0.02, respectively) analysis for leukemia-free survival identified SRSF2 and IDH1 mutations as being significant. Considering the prognostic relevance of ASXL1 mutations for survival and SRSF2 and IDH1 mutations for leukemia-free survival, the Mayo cohort was divided into mutationally high (ASXL1, SRSF2 or IDH1 mutated) and mutationally low (none of the above) risk categories, which showed significant difference in both overall (multivariable HR accounting for DIPSS-plus: 1.4, 95% CI: 1.0–1.9; P=0.04) and leukemia-free (multivariable HR accounting for karyotype and thrombocytopenia: 3.2, 95% CI: 1.6–6.4; P=0.001) survival. Significant differences in survival between the mutationally high- and low-risk disease groups were again not apparent in DIPSS-plus high-risk disease (P=0.39) but were restricted to low plus intermediate-1 risk disease categories (P=0.02; Figure 4b).

Association of ‘mutationally high risk’ status with a specific molecular signature

To explore whether these specific molecular assets associated with unique gene expression profiles, we compared CD34+ cells from 9 mutationally high and 11 mutationally low-risk patients, randomly selected in the European cohort. As shown in the heat map in Figure 5, the two mutational groups displayed a distinct molecular signature, that was characterized by 39 differentially expressed genes, 27 overexpressed and 12 downregulated (Supplementary Table 4).

Figure 5

Gene expression profile was analyzed in CD34+ cells purified from mutationally high-risk (n=9) and low-risk (n=11) patients randomly selected in the European cohort. The heat map was computed on the gene list provided in Supplementary Table 4 using the clustering algorithm included in the Partek GS package (Partek Incorporated, St Louis, MO, USA) by means of euclidean distance and average linkage. Gene coloring is based on normalized signals, as shown at the bottom of the figure; green indicates reduced expression, red increased expression. Gene symbol is indicated on the left. Low-risk and high-risk groups clustered separately in the dendrogram shown at the top of the heat map.


In this international collaborative project, we studied two independent cohorts of patients with PMF in order to clarify the prognostic relevance of newly described mutations, including TET2,19 CBL,20 IDH1 or IDH2,21 ASXL1,22 EZH2,23 DNMT3A24 and SRSF2.25 We used the first cohort of 483 European patients studied within 1 year of diagnosis, to select out which one of these mutations predicted worse outcome. Overall and leukemia-free survivals were inter-independently predicted by ASXL1, SRSF2 or EZH2 mutations and ASXL1, SRSF2 and IDH1 or 2 mutations, respectively. The observations in terms of overall survival were validated in an independent cohort of 396 Mayo Clinic patients studied at the time of their referral, whereas only SRSF2 and IDH1 mutations were associated with leukemic transformation in the Mayo cohort. When both cohorts were subjected to multivariable analysis that included recently developed prognostic models for survival in PMF (IPSS3 in the European cohort and DIPSS-plus10 in the Mayo cohort), only ASXL1 mutations retained their significance. IPSS- or DIPSS-plus-independent prognostic value was also demonstrated by classifying patients into mutationally ‘high’ or ‘low’ risk groups defined as the respective presence or absence of at least one of the aforementioned mutations that were prognostically relevant for either overall or leukemia-free survival.

We believe that there is a major practical relevance of the observations from the current study in terms of disease prognostication and therapeutic decision making. The currently used scores for survival prognostication in PMF include IPSS,3 DIPSS9 and DIPSS-plus; the latter incorporates karyotype,14, 15 consistent with the major prognostic contribution of karyotype in other myeloid malignancies including AML26 and MDS.27 Conceivably, inv,3 i(17q) or monosomal karyotype were identified as predictors of a >80% 2-year mortality in PMF.28 In line with these observations, the current study identifies additional submicroscopic genetic changes, such as ASXL1 mutations, which cluster with normal karyotype and yet portend a poor prognosis. Accordingly, whereas the outcome of patients with IPSS/DIPSS-plus high-risk disease may not be affected by the presence of ASXL1 mutations, the apparently lower-risk patients might not fare as well as expected, an observation that might affect therapeutic decisions. Similarly, the presence of SRSF2 or IDH1 mutations appears to predict leukemia independent of currently known risk factors including thrombocytopenia and unfavorable karyotype.

The potential prognostic value of IDH,29 EZH230 and SRSF231 mutations in PMF were previously suggested by single center studies. We reasoned that an integrated and comprehensive mutational analysis of the most commonly occurring mutations, performed in large well-characterized cohort of subjects with PMF, could allow to identify mutationally defined subgroups of patients resulting in improved prognosis assessment and providing a background for the interpretation of results obtained with conventional therapy, including allogeneic stem cell transplantation and particularly novel drugs. As was the case in the current study, SRSF2, EZH2 and IDH1 mutations have previously been associated with poor outcome in MDS32, 33, 34, 35 and IDH1 in distinct subtypes of AML.36, 37, 38, 39 The detrimental effect of ASXL1 mutations has also been shown in MDS,34, 35, 40 AML41, 42, 43 and CMML.44 Therefore, ASXL1 mutations represent a relatively frequent and independent prognostic biomarker in myeloid malignancies. Additional value is also suggested for EZH2, IDH1 and SRSF2 mutations. These observations suggest an untapped resource of genetic alterations for disease prognostication and warrant the inclusion of mutation screening for ASXL1, EZH2, SRSF2 and IDH1 as laboratory correlative studies in future clinical trials and prospective observational studies. Other potential prognostic biomarkers in PMF include nullizygosity for JAK2 46/1 haplotype,45 low JAK2V617F allele burden,46, 47 increased serum IL-8 and IL-2R levels and excess serum-free light chain levels.41, 48

The current study identifies mutations and mutational combinations that have clinical and prognostic correlates in PMF and could eventually prove useful for advancing the understanding of disease mechanisms, as suggested by the association of a specific gene expression signature in the CD34+ cells of an exploratory cohort of subjects with their mutational asset. These results are also consistent with previous suggestions that mutations affecting epigenetic regulation might be prognostically more relevant than those involved with JAK-STAT signaling. Regardless, mutational profiling in PMF might clarify disease heterogeneity, refine current prognostic models, inform therapeutic decisions and provide clinical trials with a tool for better patient stratification and monitoring of the effects of novel drugs.


  1. 1

    Vardiman JW, Thiele J, Arber DA, Brunning RD, Borowitz MJ, Porwit A et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood 2009; 114: 937–951.

  2. 2

    Tefferi A . Myelofibrosis with myeloid metaplasia. N Eng J Med 2000; 342: 1255–1265.

  3. 3

    Cervantes F, Dupriez B, Pereira A, Passamonti F, Reilly JT, Morra E et al. New prognostic scoring system for primary myelofibrosis based on a study of the International Working Group for Myelofibrosis Research and Treatment. Blood 2009; 113: 2895–2901.

  4. 4

    Scott BL, Gooley TA, Sorror ML, Rezvani AR, Linenberger ML, Grim J et al. The Dynamic International Prognostic Scoring System for myelofibrosis predicts outcomes after hematopoietic cell transplantation. Blood 2012; 119: 2657–2664.

  5. 5

    Tefferi A . Primary myelofibrosis: 2013 update on diagnosis, risk-stratification, and management. Am J Hematology 2013; 88: 141–150.

  6. 6

    Verstovsek S, Mesa RA, Gotlib J, Levy RS, Gupta V, DiPersio JF et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med 2012; 366: 799–807.

  7. 7

    Harrison C, Kiladjian JJ, Al-Ali HK, Gisslinger H, Waltzman R, Stalbovskaya V et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med 2012; 366: 787–798.

  8. 8

    Pardanani A, Gotlib JR, Jamieson C, Cortes JE, Talpaz M, Stone RM et al. Safety and efficacy of TG101348, a selective JAK2 inhibitor, in myelofibrosis. J Clin Oncol 2011; 29: 789–796.

  9. 9

    Passamonti F, Cervantes F, Vannucchi AM, Morra E, Rumi E, Pereira A et al. A dynamic prognostic model to predict survival in primary myelofibrosis: a study by the IWG-MRT (International Working Group for Myeloproliferative Neoplasms Research and Treatment). Blood 2010; 115: 1703–1708.

  10. 10

    Gangat N, Caramazza D, Vaidya R, George G, Begna K, Schwager S et al. DIPSS plus: a refined Dynamic International Prognostic Scoring System for primary myelofibrosis that incorporates prognostic information from karyotype, platelet count, and transfusion status. J Clin Oncol 2011; 29: 392–397.

  11. 11

    Patnaik MM, Caramazza D, Gangat N, Hanson CA, Pardanani A, Tefferi A . Age and platelet count are IPSS-independent prognostic factors in young patients with primary myelofibrosis and complement IPSS in predicting very long or very short survival. Eur J Haematol 2010; 84: 105–108.

  12. 12

    Tefferi A, Siragusa S, Hussein K, Schwager SM, Hanson CA, Pardanani A et al. Transfusion-dependency at presentation and its acquisition in the first year of diagnosis are both equally detrimental for survival in primary myelofibrosis—prognostic relevance is independent of IPSS or karyotype. Am J Hematol 2010; 85: 14–17.

  13. 13

    Elena C, Passamonti F, Rumi E, Malcovati L, Arcaini L, Boveri E et al. Red blood cell transfusion-dependency implies a poor survival in primary myelofibrosis irrespective of IPSS and DIPSS. Haematologica 2011; 96: 167–170.

  14. 14

    Hussein K, Pardanani AD, van Dyke DL, Hanson CA, Tefferi A . International Prognostic Scoring System-independent cytogenetic risk categorization in primary myelofibrosis. Blood 2010; 115: 496–499.

  15. 15

    Caramazza D, Begna KH, Gangat N, Vaidya R, Siragusa S, Van Dyke DL et al. Refined cytogenetic-risk categorization for overall and leukemia-free survival in primary myelofibrosis: a single center study of 433 patients. Leukemia 2011; 25: 82–88.

  16. 16

    Patel JP, Gonen M, Figueroa ME, Fernandez H, Sun Z, Racevskis J et al. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 2012; 366: 1079–1089.

  17. 17

    Grunkemeier GL, Jin R, Eijkemans MJ, Takkenberg JJ . Actual and actuarial probabilities of competing risks: apples and lemons. Ann Thorac Surg 2007; 83: 1586–1592.

  18. 18

    Fine JP, Gray RJ . A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc 1999; 94: 496–509.

  19. 19

    Delhommeau F, Dupont S, Della Valle V, James C, Trannoy S, Masse A et al. Mutation in TET2 in myeloid cancers. N Eng J Med 2009; 360: 2289–2301.

  20. 20

    Grand FH, Hidalgo-Curtis CE, Ernst T, Zoi K, Zoi C, McGuire C et al. Frequent CBL mutations associated with 11q acquired uniparental disomy in myeloproliferative neoplasms. Blood 2009; 113: 6182–6192.

  21. 21

    Green A, Beer P . Somatic mutations of IDH1 and IDH2 in the leukemic transformation of myeloproliferative neoplasms. N Engl J Med 2010; 362: 369–370.

  22. 22

    Carbuccia N, Murati A, Trouplin V, Brecqueville M, Adelaide J, Rey J et al. Mutations of ASXL1 gene in myeloproliferative neoplasms. Leukemia 2009; 23: 2183–2186.

  23. 23

    Ernst T, Chase AJ, Score J, Hidalgo-Curtis CE, Bryant C, Jones AV et al. Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet 2010; 42: 722–726.

  24. 24

    Abdel-Wahab O, Pardanani A, Rampal R, Lasho TL, Levine RL, Tefferi A . DNMT3A mutational analysis in primary myelofibrosis, chronic myelomonocytic leukemia and advanced phases of myeloproliferative neoplasms. Leukemia 2011; 25: 1219–1220.

  25. 25

    Yoshida K, Sanada M, Shiraishi Y, Nowak D, Nagata Y, Yamamoto R et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature 2011; 478: 64–69.

  26. 26

    Frohling S, Schlenk RF, Kayser S, Morhardt M, Benner A, Dohner K et al. Cytogenetics and age are major determinants of outcome in intensively treated acute myeloid leukemia patients older than 60 years: results from AMLSG trial AML HD98-B. Blood 2006; 108: 3280–3288.

  27. 27

    Greenberg PL, Tuechler H, Schanz J, Sanz G, Garcia-Manero G, Sole F et al. Revised international prognostic scoring system for myelodysplastic syndromes. Blood 2012; 120: 2454–2465.

  28. 28

    Tefferi A, Jimma T, Gangat N, Vaidya R, Begna KH, Hanson CA et al. Predictors of greater than 80% 2-year mortality in primary myelofibrosis: a Mayo Clinic study of 884 karyotypically annotated patients. Blood 2011; 118: 4595–4598.

  29. 29

    Tefferi A, Jimma T, Sulai NH, Lasho TL, Finke CM, Knudson RA et al. IDH mutations in primary myelofibrosis predict leukemic transformation and shortened survival: clinical evidence for leukemogenic collaboration with JAK2V617F. Leukemia 2012; 26: 475–480.

  30. 30

    Guglielmelli P, Biamonte F, Score J, Hidalgo-Curtis C, Cervantes F, Maffioli M et al. EZH2 mutational status predicts poor survival in myelofibrosis. Blood 2011; 118: 5227–5234.

  31. 31

    Lasho TL, Jimma T, Finke CM, Patnaik M, Hanson CA, Ketterling RP et al. SRSF2 mutations in primary myelofibrosis: significant clustering with IDH mutations and independent association with inferior overall and leukemia-free survival. Blood 2012; 120: 4168–4171.

  32. 32

    Patnaik MM, Hanson CA, Hodnefield JM, Lasho TL, Finke CM, Knudson RA et al. Differential prognostic effect of IDH1 versus IDH2 mutations in myelodysplastic syndromes: a Mayo Clinic study of 277 patients. Leukemia 2012; 26: 101–105.

  33. 33

    Thol F, Kade S, Schlarmann C, Loffeld P, Morgan M, Krauter J et al. Frequency and prognostic impact of mutations in SRSF2, U2AF1, and ZRSR2 in patients with myelodysplastic syndromes. Blood 2012; 119: 3578–3584.

  34. 34

    Bejar R, Stevenson KE, Caughey BA, Abdel-Wahab O, Steensma DP, Galili N et al. Validation of a prognostic model and the impact of mutations in patients with lower-risk myelodysplastic syndromes. J Clin Oncol 2012; 30: 3376–3382.

  35. 35

    Bejar R, Stevenson K, Abdel-Wahab O, Galili N, Nilsson B, Garcia-Manero G et al. Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med 2011; 364: 2496–2506.

  36. 36

    Abbas S, Lugthart S, Kavelaars FG, Schelen A, Koenders JE, Zeilemaker A et al. Acquired mutations in the genes encoding IDH1 and IDH2 both are recurrent aberrations in acute myeloid leukemia: prevalence and prognostic value. Blood 2010; 116: 2122–2126.

  37. 37

    Paschka P, Schlenk RF, Gaidzik VI, Habdank M, Kronke J, Bullinger L et al. IDH1 and IDH2 mutations are frequent genetic alterations in acute myeloid leukemia and confer adverse prognosis in cytogenetically normal acute myeloid leukemia with NPM1 mutation without FLT3 internal tandem duplication. J Clin Oncol 2010; 28: 3636–3643.

  38. 38

    Green CL, Evans CM, Hills RK, Burnett AK, Linch DC, Gale RE . The prognostic significance of IDH1 mutations in younger adult patients with acute myeloid leukemia is dependent on FLT3/ITD status. Blood 2010; 116: 2779–2782.

  39. 39

    Thol F, Damm F, Wagner K, Gohring G, Schlegelberger B, Hoelzer D et al. Prognostic impact of IDH2 mutations in cytogenetically normal acute myeloid leukemia. Blood 2010; 116: 614–616.

  40. 40

    Thol F, Friesen I, Damm F, Yun H, Weissinger EM, Krauter J et al. Prognostic significance of ASXL1 mutations in patients with myelodysplastic syndromes. J Clin Oncol 2011; 29: 2499–2506.

  41. 41

    Pardanani A, Lasho TL, Finke CM, Rajkumar SV, Singh PP, Ketterling RP et al. Polyclonal immunoglobulin free light chain levels predict survival in myeloid neoplasms. J Clin Oncol 2012; 30: 1087–1094.

  42. 42

    Schnittger S, Eder C, Jeromin S, Alpermann T, Fasan A, Grossmann V et al. ASXL1 exon 12 mutations are frequent in AML with intermediate risk karyotype and are independently associated with an adverse outcome. Leukemia 2013; 27: 82–91.

  43. 43

    Metzeler KH, Becker H, Maharry K, Radmacher MD, Kohlschmidt J, Mrozek K et al. ASXL1 mutations identify a high-risk subgroup of older patients with primary cytogenetically normal AML within the ELN Favorable genetic category. Blood 2011; 118: 6920–6929.

  44. 44

    Gelsi-Boyer V, Trouplin V, Roquain J, Adelaide J, Carbuccia N, Esterni B et al. ASXL1 mutation is associated with poor prognosis and acute transformation in chronic myelomonocytic leukaemia. Brit J Haematol 2010; 151: 365–375.

  45. 45

    Tefferi A, Lasho TL, Patnaik MM, Finke CM, Hussein K, Hogan WJ et al. JAK2 germline genetic variation affects disease susceptibility in primary myelofibrosis regardless of V617F mutational status: nullizygosity for the JAK2 46/1 haplotype is associated with inferior survival. Leukemia 2010; 24: 105–109.

  46. 46

    Tefferi A, Lasho TL, Huang J, Finke C, Mesa RA, Li CY et al. Low JAK2V617F allele burden in primary myelofibrosis, compared to either a higher allele burden or unmutated status, is associated with inferior overall and leukemia-free survival. Leukemia 2008; 22: 756–761.

  47. 47

    Guglielmelli P, Barosi G, Specchia G, Rambaldi A, Lo Coco F, Antonioli E et al. Identification of patients with poorer survival in primary myelofibrosis based on the burden of JAK2V617F mutated allele. Blood 2009; 114: 1477–1483.

  48. 48

    Tefferi A, Vaidya R, Caramazza D, Finke C, Lasho T, Pardanani A . Circulating interleukin (IL)-8, IL-2R, IL-12, and IL-15 levels are independently prognostic in primary myelofibrosis: a comprehensive cytokine profiling study. J Clin Oncol 2011; 29: 1356–1363.

Download references


This work was supported by AIRC ‘Special Program Molecular Clinical Oncology 5 × 1000’ to AGIMM group (project no. 1005; a detailed description of the AGIMM project is available at AMV and RM were also supported by AIRC (IG9034 and 12055, respectively). AMV, RM, MC were supported by a FIRB project, no. RBAP11CZLK.

Author information

Correspondence to A M Vannucchi or A Tefferi.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Preliminary presentation of results has been done at the Annual Meeting of the American Society of Hematology, Atlanta 2012.

Supplementary Information accompanies this paper on the Leukemia website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article


  • myelofibrosis
  • prognosis
  • mutations
  • ASXL1
  • prognostic score

Further reading