Finke, Christy M.

Abstract Background : Refractory anemia with ring sideroblasts and thrombocytosis (RARS-T) is a provisional entity in the MDS/MPN overlap syndromes, with diagnostic features of RARS, along with a platelet count > 450 x 10(9)/L and large atypical megakaryocytes. Mutations involving SF3B1 (~80%), JAK2 (~60%), TET2 (~25%) and ASXL1 (~15%) have been reported. The frequency of thrombotic events in RARS-T is thought to be similar to that of essential thrombocytosis (ET) (3.6 vs 3.9/100 patient years; Hematologica 2012; 1036-41). However, unlike in ET, it remains unclear as to whether thrombotic events in RARS-T impact overall (OS) or thrombosis-free (TFS) survival. Methods : 82 patients with WHO-defined RARS-T were included in the study. All patients had bone marrow (BM) biopsies, iron stains for detection of BM ring sideroblasts (RS) and cytogenetics performed at diagnosis. Mutational analysis for JAK2, CALR, MPL, SF3B1, SRSF2 and ASXL1 was carried out on BM DNA obtained at the time of diagnosis. Details of type of thrombotic event and cardiovascular (CV) risk factors including; hypertension, diabetes, smoking and dyslipidemia were obtained by careful review of the medical record. OS was calculated from time of initial diagnosis to the time of last follow-up or death. TFS was calculated from the time of diagnosis of RARS-T to development of thrombosis in uncensored patients and date of last follow up or death in patients censored for thrombosis. Conventional statistics were utilized for all analyses. Results: Among the 82 study patients, 46 (56%) were males and median age was 72 years (range, 48-93). At a median follow-up of 26.5 months; 48 (59%) deaths and 2 leukemic transformations were documented. Median OS was 44 months. Karyotype was diploid in 61 (74%) and analyzed mutational frequencies were; SF3B1 67% (28/42), JAK2 V617F 40% (25/61), ASXL1 21% (9/41), SRSF2 6% (2/31), MPL 1% (2/41) and CALR 0, respectively. Cardiovascular risk factors were present in 52 (63%). a) Thrombotic events and their impact on thrombosis free survival: Eight (10%) patients had a thrombotic event prior to or at the time of diagnosis of RARS-T (venous-8, arterial-0). Nine (11%) patients developed subsequent thrombotic events (venous-7, arterial-2), with no fatalities. In univariate analysis, lower hemoglobin level (p=0.008), lower BM RS % (p=0.04), history of thrombosis (p=0.02), and absence of SF3B1 mutations (p=0.017) were associated with an inferior TFS. Age (p=0.07), JAK2 mutation status (p=0.56) and CV risk factors (p=0.95) did not have prognostic impact. Given the association between SF3B1 mutations and BM RS, in a multivariable model that included these two as covariates, only the absence of SF3B1 mutations retained prognostic significance (p=0.02). Absence of SF3B1 mutation remained significant when history of thrombosis or lower hemoglobin level were introduced into the multivariable model. b) Risk factors for overall survival: In univariate analysis, hemoglobin <10 gm/dl (p=0.007), presence of circulating blasts (p=0.002) and presence of CV risk factors (p=0.04) were associated with an inferior OS. Neither history of thrombosis (p=0.59) nor SF3B1 (p=0.12) mutations affected OS. In multivariable analysis, hemoglobin <10 gm/dl (p=0.03) [HR 1.9; 95% CI 1.1-3.5] and circulating blasts (p=0.002) [HR 1.3; 95% CI 1.1-1.6] remained significant. Conclusions: In contrast to what is observed in ET, thrombosis history does not appear to affect either overall or thrombosis-free survival in RARS-T. The current study suggests a lower thrombosis risk in SF3B1 -mutated patients with RARS-T, compared to their SF3B1 wild-type counterparts. Disclosures No relevant conflicts of interest to declare.

Abstract Abstract 4009 Trisomy 8 is the most common among sole cytogenetic abnormalities in both acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS). In the very first paper published on isocitrate dehydrogenase (IDH) mutations in AML, 13 of the 16 IDH1 mutations detected were associated with normal karyotype, 2 with trisomy 8 and one with trisomy 13. Trisomy 8 was also recurrent in patients with IDH1/IDH2-mutated post-MDS AML. In the current study we examined the prevalence and disease distribution of IDH1 and IDH2 mutations in a large (n=157) group of patients with hematologic malignancies and isolated trisomy 8. The Mayo Clinic cytogenetic database allowed identification of 157 patients with isolated trisomy 8. Archived bone marrow cell pellets were used to extract DNA for IDH1 and IDH2 mutation analysis. Eighteen IDH mutations were identified: 15 IDH2 (14 R140Q and one R140W) and 3 IDH1 (2 R132C and one R132G). Seventeen of the 18 IDH mutations occurred in myeloid malignancies whereas one (IDH2R140W) occurred in a patient with angioimmunoblastic lymphoma who was not previously exposed to chemotherapy or radiotherapy, and in whom the IDH2 mutation disappeared after effective lymphoma chemotherapy. Among the 17 IDH-mutated myeloid malignancies, disease-specific IDH1/IDH2 mutational frequencies were as follows: 27% (3/11) for post-MDS AML, 25% (3/12) for therapy-related MDS/AML, 15% (8/54) for de novo MDS, 13% (2/15) for de novo AML and 3% (1/32) for myeloproliferative neoplasm (MPN). In contrast, IDH mutational frequencies were significantly lower among 64 additional patients with AML or MDS without isolated trisomy 8: 7% in de novo AML (n=28), 0% in de novo MDS (n=21), 0% in post-MDS AML (n=11) and 0% in therapy-related MDS/AML (n=4). In the 54 patients with trisomy 8-associated de novo MDS, prognosis was similar between IDH mutated (n=8; median survival 14 months) and unmutated (n=46; median survival 16 months) cases (p=0.7). The majority of IDH-mutated cases with de novo MDS belonged to high risk MDS disease category. However, 3 of the 8 IDH-mutated patients with de novo MDS and 2 of the 3 with therapy-related MDS did not display excess bone marrow blasts. The current study suggests a possible association between IDH mutations and trisomy 8 in AML and MDS but not in MPN or MDS/MPN. The fact that the mere presence of trisomy 8 did not result in a more than expected incidence of IDH mutations in MPN or MDS/MPN makes it unlikely that such an association would be secondary to trisomy 8-associated genetic or biologic changes. Instead, it is possible that IDH mutations, which have been shown to cluster with high-risk disease in both MPN and MDS associated with 5q-, promote a selective advantage for the survival of the clone that harbors trisomy 8. Consistent with this contention, we were able to demonstrate in one of our patients with relapsed AML, the presence of IDHR132C both at initial AML diagnosis and time of relapse, whereas the trisomy 8 abnormality was seen only at the time of relapse. Regardless, the presence of molecular heterogeneity among patients with trisomy 8-associated AML or MDS might explain the controversial prognostic influence of the specific cytogenetic abnormality. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 4105 Background: JAK2 mutations are found in approximately 99% of patients with polycythemia vera (PV) and 60% of those with essential thrombocythemia (ET) or primary myelofibrosis (PMF). It is currently assumed that other mutations relevant to JAK signaling contribute to the pathogenesis of JAK2 mutation-negative myeloproliferative neoplasms (MPN). The same might hold true for some cases of “idiopathic erythrocytosis” associated with subnormal serum erythropoietin level (sEpo). LNK is a plasma membrane-bound adaptor protein whose function includes inhibition of wild-type and mutant JAK2 phosphorylation. LNK exon 2 mutations were recently described in two patients with JAK2V617F-negative ET or PMF. Both mutations involved the LNK pleckstrin homology (PH) domain; one was a 5 base-pair deletion and missense mutation leading to a premature stop codon (603_607delGCGCT; 613C>G) and the other a missense mutation (622G>C; glutamic acid to glutamine substitution; E208Q). Objectives: i) To estimate the prevalence of LNK mutations in chronic and blast phase MPN ii) To determine if LNK mutations are mutually exclusive of other MPN-associated mutations iii) To test the hypothesis that LNK mutations might contribute to the pathogenesis of JAK2 mutation-negative PV or otherwise unexplained erythrocytosis Methods: LNK mutation analysis was performed on bone marrow or blood cells using modified primers for amplifying the SH2 and PH domains, using previously published methods (Oh et al. Blood First Edition Paper, prepublished online April 19, 2010; DOI 10.1182/blood-2010-02-270108). Results: LNK mutation studies were performed in 172 patients; paired chronic-blast phase samples were analyzed in 26 cases. Diagnoses in the 172 study patients were as follows: 78 chronic-phase MPN (39 JAK2V617F-positive PV, 25 PMF and 14 ET) enriched for TET2, IDH, JAK2V617F, or MPL mutations, 61 blast-phase MPN (41 blast-phase PMF, 11 blast-phase PV and 9 blast-phase ET), 25 JAK2V617F-negative PV and 8 “unexplained erythrocytosis” associated with subnormal sEpo and negative for Epo receptor (EPOR) mutations. Ten novel heterozygous LNK mutations, all but one affecting exon 2 in the PH domain, were identified: 6 missense mutations involving codons 215, 220, 223, 229 and 234, one nonsense mutation involving codon 208 (622G>T leading to glutamic acid to a stop codon substitution; E208X), one synonymous mutation involving codon 208, and 2 deletion mutations involving exon 2 (685-691_delGGCCCCG) or exon 5 (955_delA). LNK mutations were most frequent in blast-phase MPN, occurring in 6 (9.8%) of the 61 informative patients; chronic-phase sample analysis in 4 of these revealed the same mutation in only one case. Mutant LNK was detected in chronic-phase samples only in 2 additional patients with blast-phase MPN. Among the 8 LNK-mutated blast-phase MPN cases, 7 had blast-phase PMF. The clinical phenotype in the remaining 2 LNK-mutated patients was that of an isolated erythrocytosis without overt features of PV. JAK2V617F was documented in 3 and IDHR140Q in 1 LNK-mutated patients. Conclusions: LNK mutations i) target an exon 2 ‘hot spot' in the PH domain spanning residues E208-D234, ii) might be most prevalent in blast-phase PMF iii) are not mutually exclusive of JAK2 or IDH mutations and iv) might be part of the “missing link” in the pathogenesis of JAK2 mutation-negative “idiopathic” erythrocytosis. Disclosures: No relevant conflicts of interest to declare.

Abstract Background : Anemia is one of the most prominent symptoms in primary myelofibrosis (PMF) and is often associated with inferior quality of life and survival. Current drugs, including JAK inhibitors, are suboptimal in the treatment of PMF-associated anemia and better information on its pathogenesis is critical for the development of more effective drugs. In the current study of JAK2/CALR/MPL annotated patients with PMF, we examined its correlation with both "driver" and "non-driver mutations", as well as cytogenetic abnormalities, in order to gain better insight into its pathogenesis. Methods: Study patients were selected based on availability of "driver" mutation information. PMF diagnosis was according to World Health Organization criteria (Blood. 2009;114:937). Previously published methods were used for CALR, JAK2 and MPL mutation analyses and determination of CALR variants (Blood. 2014;124:2465). Considering their relatively high mutational frequency in PMF, subsets of patients were also screened for ASXL1, spliceosome component (SF3B1, U2AF1, SRSF2, ZRSR2) and TET2 mutations. Cytogenetic analysis and reporting was done according to the International System for Human Cytogenetic Nomenclature. Statistical analyses considered clinical and laboratory parameters obtained at time of first referral at the Mayo Clinic. Results : Analysis was conducted on 722 patients (median age 64 years; 64% males). DIPSS-plus risk distribution was 14% low, 17% intermediate-1, 37% intermediate-2 and 33% high. All patients were annotated for JAK2/CALR/MPL mutations as well as CALR variants; 477 harbored JAK2, 139 CALR and 41 MPL mutations; 65 patients were triple-negative. The 139 CALR -mutated patients included type 1/type 1-like (n =115) and type 2/type 2-like (n =24). Non-driver mutations screened included ASXL1 (n =480; mutated 38%), SRSF2 (n =474; mutated 14%), U2AF1 (n =457; mutated 16%), SF3B1 (n =328; mutated 8%), ZRSR2 (n =180; mutated 11%) and TET2 (n =180; mutated 18%). Karyotype was normal in 60%, favorable in 28% and unfavorable in 12%. Anemia was defined as being absent (normal sex-adjusted hemoglobin level; n =110; 15%), mild (hemoglobin level of ≥10 g/dL but below sex-adjusted normal value; n =263; 36%), moderate (hemoglobin level below 10 g/dL but not transfusion-dependent; n =108; 15%) and severe (transfusion-dependent anemia; n =241; 33%). Presence of at least mild anemia was associated with abnormal karyotype (p=0.006) with no difference between favorable and unfavorable abnormalities, U2AF1 (p=0.002), TET2 (p=0.02) and ASXL1 (p=0.04) mutations; other significant associations included male sex and older age. Presence of moderate to severe anemia was associated with U2AF1 (p<0.0001), SRSF2 (p=0.007) and driver mutations other than CALR type 1/type 1-like (p<0.0001). Presence of severe anemia was associated with U2AF1 (p<0.0001), SRSF2 (p=0.003) and non-CALR driver mutations (17% incidence in both types of CALR variants vs 51% in triple negative, 35% JAK2, 39% MPL mutated cases; p<0.0001). An association with older age but not gender was also noted for both moderate to severe and severe anemia (p<0.0001). During multivariable analysis, independent associations with moderate to severe anemia were confirmed for U2AF1 (p<0.0001), SRSF2 (p=0.007) and age (p=0.0001), but not driver mutation profile (p=0.30). A similar analysis for severe anemia also identified U2AF1, SRSF2 and age as being significantly relevant. Conclusions : The current study identifies older age and the spliceosomal mutations U2AF1 and SRSF2 as having strong and independent association with moderate to severe anemia in PMF. Targeting the spliceosome machinery or its mutant components offers a potential approach in the treatment of PMF-associated anemia. Disclosures Pardanani: Stemline: Research Funding.

Abstract Background: In a previously published phase-1/2 study, momelotinib (a JAK 1/2 inhibitor) induced 53% anemia, 39% spleen and >50% constitutional symptoms response rates in patients with myelofibrosis (MF) (Pardanani A, et al. Leukemia. 2013;27:1322). The particular study included 100 patients from the Mayo Clinic who experienced 43% spleen and 44% anemia response rates (Pardanani et al. Leukemia. 2015;29:741); CALR and ASXL1 mutations independently predicted spleen but not anemia response. In the current study, we examined the impact of momelotinib therapy on DIPSS-plus- and molecular risk-adjusted survival. Methods: The current study represents sponsor-independent analysis of long-term survival among 100 consecutive Mayo Clinic patients who received momelotinib therapy as part of a larger (n =166) phase-1/2 study (NCT00935987). Safety and efficacy data from NCT00935987, as well as protocol design and drug dosing and schedule, have already been formally reported and not reiterated here (Pardanani A, et al. Leukemia. 2013;27:1322; Pardanani A, et al. Leukemia. 2015;29:741; Abdelrahman RA, et al. BJH. 2015;169:77). All statistical analyses considered clinical and laboratory parameters obtained at the time of entry to study. Survival analysis was considered from the date of study entry to date of death or last contact. Cox proportional hazard regression model was used for multivariable analysis of survival. Samples from time of study entry were used for mutation screening. Driver mutations were classified into two prognostically-relevant groups: CALR type 1/type 1-like (favorable) vs all others including CALR type 2/type 2-like, JAK2, MPL and triple-negative (unfavorable). Results: 100 patients with MF (median age 66 years; 58% males) were treated at the Mayo Clinic between 11/20/09 and 11/10/10; 64 patients had primary MF, 22 post-polycythemia vera MF and 14 post-essential thrombocythemia MF. Baseline data: DIPSS-plus risk distribution was 63% high, 36% intermediate-2 and 1% intermediate-1 (JCO 2011;29:392); for the purposes of the current analysis, the single patient with intermediate-1 risk disease was included in the intermediate-2 category. 54 (54%) patients displayed abnormal karyotype. 73 (73%) patients harbored JAK2, 16 (16%) CALR, 7 MPL and 4 "triple-negative". Among the 16 CALR -mutated cases, 13 were classified as "type 1/type 1-like". 94 patients were screened for ASXL1 mutations; 41 (44%) were mutated. Follow-up data was updated in July 2015, providing a minimum follow-up of over 4.5 years since the last patient registration on the protocol. At a median follow-up of 3.2 years, 67 (67%) deaths and 13 (13%) leukemic transformations were documented. Median survival from the time of study entry was 3.2 years and was affected by driver mutation prognostic groups (median survival "not reached" for "favorable" vs 3 years for "unfavorable"; p=0.01), ASXL1 mutational status (median survival 3.9 years in unmutated vs 2.5 years in mutated; p=0.02) and DIPSS-plus risk stratification (median survival 2.6 years in high-risk vs 4 years in intermediate-2 risk; p=0.07). In order to assess the impact of momelotinib on long-term outcome, we compared survival of the study cohort (n =100) with that of 442 consecutive Mayo Clinic patients with high or intermediate-2 risk primary MF (median age 67 years; 66% males). There was no significant difference between this control group of PMF patients versus the momelotinib patient cohort in distribution of age (p=0.56), sex (p=0.12), JAK2/CALR/MPL mutational status (p=0.41) or ASXL1 mutational status (p=0.9); however, the momelotinib cohort had a higher fraction of patients with DIPSS-plus high-risk disease (63% vs 47%; p=0.004). Overall survival was similar between the momelotinib cohort and the control group of PMF patients who were not treated with momelotinib (Figure 1; p=0.99); the difference in survival remained not significant during multivariable analysis that included risk stratification by DIPSS-plus (p=0.32), "favorable" vs "unfavorable" driver mutational status (p=0.78) or ASXL1 mutational status (Figure 2; p=0.61). Conclusions: In the current long-term study, we could not demonstrate a survival impact from momelotinib therapy in myelofibrosis. Although adjusted for clinical and molecular risk status, our study is retrospective and cannot be relied upon to discount marginal survival effect. Disclosures Al-Kali: Novartis: Research Funding. Pardanani:Stemline: Research Funding.

Abstract Background : In essential thrombocythemia (ET) , ̴ 85% of patients harbor one of three "driver" mutations, with mutational frequencies of approximately 58%, 23% and 4%, for JAK2, CALR and MPL, respectively; ̴ 15% are wild type for all three mutations and are operationally referred to as "triple negative" (Blood. 2014;124:2507). In one of the original descriptions on CALR mutations, CALR -mutated patients with ET, compared to their JAK2-mutated counterparts, were reported to have better survival (NEJM. 2013;369:2379). However, this observation was not supported by subsequent studies while other reports suggested differential prognostic effect from distinct CALR variants in myelofibrosis (Blood. 2014;124:2465). In this study, we sought to clarify the impact of all three mutations, and CALR variants, on overall (OS), myelofibrosis-free (MFS) and leukemia-free (LFS) survival. Methods: Patientswere selected from our institutional database of myeloproliferative neoplasms, based on availability of mutational status inforomation. ET diagnosis was according to WHO criteria (Blood. 2009;114:937). Published methods were used for CALR, JAK2 and MPL mutation analyses and determination of CALR variants (Blood. 2014;124:2465). Kaplan-Meier survival analysis was considered from the date of diagnosis to date of death or last contact. MFS and LFS calculations considered fibrotic or leukemic transformation events as uncensored variables, respectively. Cox proportional hazard regression model was used for multivariable analysis. Results : A total of 502 patients (median age 59 year; 61% females) met study eligibility criteria. Median levels of hemoglobin, platelet count and leukocyte counts were 13.7 g/dL, 893 x 10 (9)/L and 8.8 x 10(9)/L, respectively. All patients were annotated for JAK2/CALR/MPL mutations as well as CALR variants; 324 harbored JAK2, 111 CALR and 13 MPL mutations; 54 patients were triple-negative. The 111 CALR-mutated patients included type 1 (n=55), type 2 (n=41) or other (n=15) CALR variants. At a median follow-up time of 9.9 years, 172 (34.3%) deaths, 42 (8.4%) fibrotic progressions, 15 (3%) blast transformations and 12 (2.4%) polycythemic conversions were documented. In univariate analysis, survival data appeared significantly better in "triple negative" patients (median not reached) and inferior in MPL-mutated cases (median 8.5 years) whereas median survival times were similar for JAK2 (18.5 years) and CALR (22.1 years) mutated cases (Figure 1; p=0.0006). However, the difference in survival was no longer apparent (p=0.60) during multivariable analysis that included age and sex, which are known to differentially cluster with specific driver mutations; in the current study, median age/sex distributions for "triple-negative", CALR, JAK2 and MPL mutated cases were 44 years/72% females, 48 years/46% females, 60 years/65% females, 70 years/46% females, respectively (p=<0.0001/0.0007). Of note, both age and sex were independently predictive of shortened survival. OS data remained unchanged when CALR-mutated patients were further stratified into type 1 vs type 2 vs other CALR variants, with similar survival data between the three CALR mutation groups (p=0.98). In univariate analysis, MPL-mutated patients were significantly more prone to fibrotic progression (Figure 2; p=0.0083). The prognostic relevance of MPL mutations to MFS remained significant when age and sex were included in multivariable analysis (p=0.008). In the current cohort, univariate analysis identified lower hemoglobin and lower platelet count as the only other risk factors for fibrotic progression. Multivariable analysis confirmed the independent prognostic relevance of MPL mutations (p=0.003), lower hemoglobin level (p=0.0009) and lower platelet count (p=0.0094) for MFS. There was no significant difference in LFS among the four driver mutational categories (p=0.9): 9 events in JAK2, 6 in CALR, none in triple negative and none in MPL mutated cases. Among the 6 leukemic transformations in CALR-mutated cases, three were type 1, two type 2, and one other CALR variants. Conclusions : Age- and sex-adjusted survival is similar among ET patients with JAK2 vs CALR vs MPL vs "triple-negative" mutational status. Survival is also similar between patients with distinct CALR variants. MPL -mutated patients with ET might be at a higher risk of fibrotic progression. Figure 1. Figure 1. Figure 2. Figure 2. Disclosures Pardanani: Stemline: Research Funding.

Abstract Background : In primary myelofibrosis (PMF), ̴ 88% of patients harbor JAK2, CALR or MPL, with respective mutational frequencies of approximately 60%, 22% and 6%; ̴ 12% are wild type for all three driver mutations and are labelled as being "triple negative" (Blood. 2014;124:2507). It is now well-established that CALR-mutated patients with PMF display significantly better survival, compared to all the other mutational categories; in addition, preliminary data have suggested inferior survival for "triple-negative" cases (Leukemia. 2014;28:1472; Blood. 2014;124:1062) and the possibility that the favorable impact of CALR mutations might be restricted to CALR type 1 or type 1-like variants (Blood. 2014;124:2465). In the current study, we sought to clarify the prognostic impact of driver mutations, and CALR variants, on overall (OS) and leukemia-free (LFS) survival, in a two-center study involving 1118 patients. Methods: A total of 1118 patients with PMF from the Mayo Clinic (n =722) and University of Florence, Italy (n =396) were included in the current study. Study patients were selected based on availability of mutation information. PMF diagnosis was according to World Health Organization criteria (Blood. 2009;114:937). Previously published methods were used for CALR, JAK2 and MPL mutation analyses and determination of CALR variants (Blood. 2014;124:2465). Kaplan-Meier survival analysis was considered from the date of first referral for the Mayo cohort and date of diagnosis for the Florence cohort. Leukemic transformation events replaced death as the uncensored variable during LFS analysis and Cox proportional hazard regression model was used for multivariable analysis. Results : Analysis was first conducted on the Mayo cohort of 722 patients (median age 64 years; 64% males). DIPSS-plus risk distributions were 14% low, 17% intermediate-1, 37% intermediate-2 and 33% high. All patients were annotated for JAK2/CALR/MPL mutations as well as CALR variants; 477 harbored JAK2, 139 CALR and 41 MPL mutations; 65 patients were triple-negative. The 139 CALR -mutated patients included type 1/type 1-like (n =115) and type 2/type 2-like (n =24). At a median follow-up time of 3.1 years, 439 (61%) deaths and 63 (8.7%) blast transformations were documented. In univariate analysis, survival in patients with CALR type 1/type 1-like mutations was significantly longer than every other mutational category: HR (95% CI) were for triple-negative 2.7 (1.8-4.0), JAK2 2.7 (2.0-3.7), CALR type 2/type 2-like 2.3 (1.3-4.2) and MPL 1.9 (1.1-3.1) (Figure 1); these differences were sustained during multivariable analysis that included ASXL1 mutations (n =480) and DIPSS-plus, and were also apparent in the Florence cohort of 396 cases that included 251 JAK2, 53 CALR type 1/type 1-like, 21 CALR type 2/type 2-like, 50 triple-negative and 21 MPL mutated cases (Figure 2). There was no difference in survival among the Mayo cohort across mutational categories not including CALR type 1/type 1-like variants (p=0.33). Because specific driver mutations in PMF differentially cluster with age and sex, multivariable analysis including these two parameters was also performed and confirmed the significant survival advantage of CALR type 1/type 1-like over triple-negative, JAK2 and CALR type 2/type 2-like mutated cases. Based on the above information, we divided driver mutational categories into "type 1/type 1-like" (median survival 10.3 years) and "all other mutational categories" (median survival 3.9 years; p<0.01). The difference in survival between these two groups remained significant (HR 2.1, 95% CI 1.5-2.9) during multivariable analysis that included both DIPSS-plus (P<0.01) and ASXL1 mutation (HR 1.8, 95% CI 1.4-2.2). LFS was similar between these two groups (p=0.4) but it was inferior in triple-negative cases compared to CALR type 1/type 1-like variants (HR 2.8, 95% CI 1.3-6.2), MPL (HR 4.6; 95% CI 1.04-20.4) and JAK2 (HR 2.5, 95% CI 1.3-4.8) but not CALR type 2/type 2-like variants (p=0.56). There was no difference in LFS among mutational categories not including triple-negative cases (p=0.33). Conclusions : In PMF, driver mutations might be classified into two prognostically distinct categories: "favorable" (type 1/type 1-like) and "unfavorable" (all other mutational categories). In addition, triple-negative cases might be at a higher risk of leukemic transformation. Disclosures Pardanani: Stemline: Research Funding. Vannucchi:Novartis: Other: Research Funding paid to institution (University of Florence), Research Funding; Shire: Speakers Bureau; Baxalta: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau.

Abstract Background: Addition of sex combs-like 1 (ASXL1) is frequently mutated (mutational frequency 14-29%) and is prognostically relevant in myelodysplastic syndromes (MDS) (Bejar, NEJM 2011, Thol, JCO 2011). The prevalence and prognostic impact of ASXL1 mutation in MDS with ring sideroblasts (MDS-RS) is not known. MDS-RS is defined by the presence of ≥15% RS in bone marrow (BM), with refractory anemia with ring sideroblasts (RARS) being the prototype but may also be seen in other categories. Methods : Our institutional database was reviewed to identify patients with WHO-defined primary MDS with ≥1% BM RS. Pathology slides, including iron stains, were reviewed to accurately quantify BM RS percentage and confirm WHO morphologic categories. All patients were annotated for their mutational status including ASXL1, JAK2, MPL and IDH with a subset for SF3B1 by PCR sequencing performed on BM specimens obtained at diagnosis. Results : i) Patient characteristics: A total of 76 MDS patients displayed ≥1% BM RS (median age 72 years; 76% males); 51 (67%) patients had ≥15% BM RS. IPSS-R risk categories for the entire 76 study patients were 30% very low, 37% low, 14% intermediate, 11% high and 8% very high; IPSS-R karyotype was normal in 63%, very good/good risk 17%, intermediate risk 12%, and poor/very poor in 8%; 3% had monosomal karyotype. ii) Prevalence of ASXL1 mutations: Twenty-one (28%) of the 76 study patients were ASXL1 mutated; ASXL1 mutational frequencies were 25% (16/63 patients) in the absence and 38% (5/13 patients) in the presence of excess blasts (P =0.34). When considering only those patients with ≥15% BM RS (n =51), ASXL1 mutations were detected in 12 (24%) patients with mutational frequencies of 24% (11/46 patients) in the absence and 20% (1/5 patients) in the presence of excess blasts (P =0.84); ASXL1 mutational frequencies were 13% (3/23 patients) in RARS and 37% (7/19 patients) in RCMD-RS (P=0.07). In terms of other mutations, all 76 study patients were wild-type for JAK2 and MPL, whereas IDH2 R140Q mutations were present in 4 (5%) patients, including 3 with concomitant ASXL1 and IDH2 mutations. IDH1 mutation was seen in only 1 patient. 25 of 43 (58%) patients screened were mutated for SF3B1, including 4 (9%) who were mutated for both ASXL1 and SF3B1. Significant associations were evident between ASXL1 and IDH2 mutations (P =0.02) but not between ASXL1 and SF3B1 mutations (P =0.61). iii) Clinical correlates of ASXL1 mutations: Among all 76 study patients, presence of ASXL1 mutation did not correlate with age (P =0.30), hemoglobin level (P =0.17), platelet count (P =0.53), BM blast percentage (P =0.17), WHO morphologic category (P =0.34), transfusion dependence (P =0.84), IPSS-R cytogenetic categories (P =0.93), or IPSS-R risk group (P =0.33). The results were unchanged when analyzing the 51 patients with MDS-RS (i.e. ≥15% BM RS). Amongst the ASXL1 mutated patients (n =21) IPSS-R cytogenetic categories were as follows: 13 patients with normal karyotype (62%), 2 patients with trisomy 8 (10%), 1 patient each with -Y, del(11q), del(5q), del(20q), isochromosome 17 and monosomy 7 (5% each). iv) Prognostic impact of ASXL1 mutations: Median follow-up was 42.5 months, during which time 69 (91%) deaths and 8 (11%) leukemic transformations were documented. ASXL1 mutated patients had a median survival of 29 months, compared to 45 months in ASXL1 wild-type patients (P =0.04). However, the difference in median survival was no longer significant during multivariable analysis, which instead identified only IPSS-R and transfusion need as being independent predictors of inferior survival. Amongst those patients without excess blasts (n =63), presence of ASXL1 mutation predicted inferior survival in univariate analysis (P =.04) and significance was sustained during multivariable analysis that included IPSS-R cytogenetic categories (P =.04) but became borderline when WHO morphologic category was included (P =0.06); ASXL1 mutated patients had a shortened median survival of 43 months compared to 66 months in ASXL1 wild-type patients (P =.04). In the presence of ≥15% BM RS, ASXL1 mutations did not affect survival (P =0.48); the results were the same for RARS (P =0.9) and RCMD-RS (P =0.64). Conclusions : ASXL1 mutations might not affect survival in MDS patients with ≥15% BM RS, including those with RARS or RCMD-RS. Furthermore, an apparent survival disadvantage seen in ASXL1 -mutated MDS patients with ≥1% BM RS was accounted for by IPSS-R. Disclosures Pardanani: Stemline: Research Funding.

Abstract Background The median survival (OS) of patients with primary myelofibrosis (PMF; 6.5y) is significantly shortened compared to reference population (Cervantes et al, JCO 2012; 30:2981). OS is predicted by the four risk categories of IPSS, Dynamic-IPSS and DIPSS-plus score, nevertheless pts heterogeneity persists within these categories, necessitating improved risk stratification. We recently reported that pts harboring mutations in any one of the prognostically relevant ASXL1, EZH2, IDH1/2 and SRSF2 genes constitute a IPSS-and DIPPS-plus independent Molecular High Risk category (MHR+) characterized by significant reduction of OS and leukemia free survival (LFS) (Vannucchi et al, Leukemia 2013). The aim of this work was to analyze the impact of the number of prognostically relevant mutated genes on OS and LFS in PMF. Patients and methods Two independent cohorts are included, a “test cohort” from Europe, analyzed at diagnosis, and a “validation cohort” from Mayo Clinic, analyzed at any time after diagnosis. Mutation analysis was performed in DNA from whole blood or granulocytes using RTQ-PCR, HRM and direct sequencing; all mutations were confirmed at least twice. The prognostic value of the molecular variables with regard to OS and LFS was analyzed by Cox regression. Test cohort It included 490 pts (median age 61y; males = 301) risk stratified by IPSS into high (n=74, 15%), intermediate-2 (n=93, 19% ), intermediate-1 (n=147, 30%) and low (n=176, 36%). The median follow-up was 3.63y (95% CI, 0.06-28.33); 161 pts died (33.0%), of whom 76 (15.6%) had progressed to acute leukemia after a median of 3.4y (0.04-28.3) from diagnosis. One hundred forty-six pts (29.8%) presented at least one of the four aforementioned mutated genes and were classified as MHR+. The OS of MHR+ pts was significantly reduced compared to patients with no mutations (n=344): 80.7 vs 148.9 mo (HR 2.2, CI95% 1.6-3.03). One hundred twelve (22.8%) pts had 1 mutation and 34 (6.9%) had 2 or more mutations. In univariate analysis presence of 2 or more mutated genes was significantly more detrimental for OS (29.5mo; HR 4.12, IC95% 2.6-6.4) than having 1 mutated gene (84.2mo; HR 1.8, IC95% 1.2-2.5) compared with having no mutations (148.9mo). Multivariate analysis results adjusted for IPSS indicated that having two or more mutations was an independent prognostic factor for OS (HR 2.9, 95% CI 1.8-4.5). Notably, the prognostic relevance of harboring two or more mutations involved both lower (low and intermediate-1; HR 1.87, 95% CI 1.3-2.6) and higher (intermediate-2 and high; HR 1.6, 95% CI 1.2-2.1) categories of IPSS. Also LFS resulted significantly shorter (129mo; HR 3.1, 95%CI 1.9-4.8) in MHR+ pts compared with pts with no adverse mutations (323mo). Having 2 or more mutations correlated with greater reduction of LFS (79.6mo; HR 7.02, CI95% 3.9-12.6) than having one mutation only (133.9mo; HR 2.2, IC95% 1.3-3.7) as compared with no mutations (323.2mo). Multivariate analysis showed that IPSS high risk category (HR 4.5; 95% CI 2.3-8.8) and two or more mutated genes (HR 5.3; 95% CI 2.9-9.8) predicted independently for a significant reduction in LFS. The negative impact on LFS of having 2 or more mutated genes compared to one or no mutations was maintained in the lower (HR 6.4, 95% CI 2.6-15.2) and higher (HR 5.5, 95% CI 2.3-12.8) risk categories. Validation cohort Validation cohort included 262 patients (median age 64 years; males = 167) risk stratified by DIPSS-plus into high (n=89, 34%), intermediate-2 (n=92, 35%), intermediate-1 (n=46, 18%) and low (n=34, 13%). One hundred forty-six pts (56%) displayed none of the four aforementioned mutations, 93 (36%) harbored one mutation whereas 23 (9%) harbored two or more mutations. In univariate analysis, having two or more mutations (HR 3.7; 95% CI 2.2-6.1) was significantly more detrimental for OS than having no mutations, which is more favorable than having one mutation (HR 1.9; 95% CI 1.4-2.7). When adjusted for DIPSS-plus, the presence of two or more mutations retained its significance (HR 2.1; 95% CI 1.3-3.6) and outperformed ASXL1 mutation alone in its prognostic relevance Conclusions Overall, these results show that the number of prognostically relevant mutated genes correlate with OS and LFS in pts with PMF, suggesting that screening for these mutations might help to improve risk stratification. Disclosures: No relevant conflicts of interest to declare.

Abstract Abstract 3068 Background: Abnormal cytokine expression accompanies the disease process in primary myelofibrosis (PMF) and the levels of proinflammatory cytokines have been shown to be down-regulated by JAK inhibitor therapy. The prognostic relevance of increased cytokine levels in PMF has not been systematically studied. The objectives of the current study were to describe the spectrum of abnormal cytokine expression in PMF and determine its prognostic relevance. Methods: A multiplex biometric sandwich immunoassay (Invitrogen multiplex assays, Invitrogen Corporation-Carlsbad CA, USA) was used for simultaneous analysis of 30 cytokine levels. Diluted plasma was mixed with fluorescently addressed microbeads bound with anti-cytokine antibodies. Biotin anticytokine secondary antibodies were then added and allowed to bind to cytokine-bead complexes. Total surface fluorescence for each bead was then measured using a Luminex (Luminex Corporation, Austin, TX, USA) fluorescent flow-based detection system. Results: Cytokine levels at diagnosis were analyzed for a total of 127 treatment naïve patients with PMF, who had stored sera at the time of diagnosis or first referral to the Mayo Clinic. The circulating levels of 30 cytokines in the multiplex assay were compared between PMF patients (n=127) and normal controls (n=31). Significant differences (p<0.05) were evident for 20 of the 30 cytokines: IL-1b, IL-1RA, IL-2R,IL-6, IL-8, IL-10, IL-12, IL-13, IL- 15, TNF-α, G-CSF, INF-α, INF-γ, MIP-1α, MIP-1β, HGF, CXCL10, MIG, MCP-1 and VEGF. On univariate analysis, increased levels of 9 of these 20 cytokines were associated with shortened survival (p<0.05): IL-1RA, IL-2R, IL-6, IL-8, IL-12, IL-15, MIP-1α, CXCL10 and MIG. On multivariable analysis, 4 of these 9 cytokines maintained their prognostic value for survival: IL-2R (p=0.0011), IL-8 (p<0.0001), IL-15 (p=0.0124) and CXCL10 (p=0.0031). The independent prognostic value of these 4 cytokines was sustained during additional multivariable analysis that included the Dynamic International Prognostic System Score and cytogenetic risk category. Next, we divided the 127 patients into three cytokine risk categories based on the serum levels of IL-2R, IL-8, IL-15 and CXCL10: none elevated, one or two elevated and three or four elevated. The respective median survivals were 69, 37 and 19 months (p=0.01; Figure 1) and such cytokine risk categorization was both DIPSS and cytogenetic risk independent; p values for multivariable analysis including cytokine, DIPSS, cytogenetic risk groups were 0.02, 0.0001 and 0.004, respectively. Conclusions: Elevated serum levels of IL-2R, IL-8, IL- 15 and CXCL10 independently predict shortened survival in PMF and such prognostication appears to be independent of DIPSS and cytogenetic risk category. The current study suggests clinical relevance of multiple cytokines in PMF that could be therapeutically targeted. Disclosures: No relevant conflicts of interest to declare.