Abstract: Nearly 20% of patients with myelofibrosis progress to blast phase disease; an aggressive form of acute myeloid leukemia. Although previous studies have implicated loss of TP53 or JARID2 in progression, by and large the genetic events that lead to conversion to blast phase remain unknown. To identify genes whose loss drives progression, we performed a focused CRISPR/Cas9 screen in which murine Jak2V617F bone marrow cells expressing Cas9 were transduced with two separate sgRNA libraries of known tumor suppressor genes and subjected to colony replating assays. Transduction of one of the two libraries led to serial replating and enhanced self-renewal of the Jak2V617F cells. Subsequent DNA sequencing revealed enrichment of all four guides targeting STK11, the gene that encodes LKB1 which regulates a number of key cellular pathways including energy utilization by activation of AMPK. To confirm that loss of Stk11 is the event that leads to increased clonogenicity, we collected cells from Jak2V617F/Vav-Cre+ and control Vav-Cre+ mice and induced Stk11 knockout by electroporating Cas9-Stk11 sgRNA ribonucleoprotein complexes. Consistent with the screening results, only Jak2V617F Vav-Cre+ cells with Cas9-Stk11 sgRNA showed serial replating. To determine whether Stk11 is required for growth of cells with a different driver of enhanced JAK/STAT signaling, we doubly transduced Stk11 homozygous floxed bone marrow cells with MPLW515L-mCherry and Cre-GFP to delete Stk11. As expected, cells with both MPLW515L and Cre recombinase showed enhanced self-renewal, while singly infected control cells failed to replate. These results demonstrate that activation of JAK/STAT signaling can overcome the requirement for Stk11 in normal hematopoiesis and suggest that STK11 loss may be a strong driver of malignant transformation in combination with enhanced JAK-STAT signaling. We next investigated the mechanism by which loss of STK11 cooperates with enhanced JAK/STAT signaling to promote leukemia. RNA-sequencing of wild-type, Stk11+/+/ MPLW515L, and Stk11-/-/MPLW515L hematopoietic cells revealed enrichment of a number of pathways related to hypoxia, oxidative phosphorylation and mitochondrial translation in cells lacking LKB1. Western blot assays confirmed activation of mTOR signaling as well as HIF1a stabilization and pathway activation, both of which have been reported to lie downstream of LKB1 loss. We also performed a number of studies to determine the relevance of reduced LKB1 expression to leukemic progression. First, we induced deletion of Stk11 in mice that were transplanted with HSPCs expressing MPLW515L after development of the MPN phenotype. Loss of Stk11 caused a rapid lethality that was associated with enhanced bone marrow fibrosis and osteosclerosis. We also observed accumulation of leukemic blasts in small clusters consistent with AML transformation arising in the spent phase MPN. Additionally, we deleted STK11 by CRISPR/Cas9 in primary MPN patient samples and monitored their engraftment in immunocompromised mice. We observed enhanced engraftment and increased reticulin fibrosis and osteosclerosis in mice that received the STK11 edited cells compared to those with non-targeted sgRNA. Third, we compared the expression of STK11 in paired blast and chronic phase myelofibrosis patient samples by RT-PCR. Consistent with the hypothesis that loss of STK11 facilitates leukemia, we found that its expression was decreased by more than 50% in five of seven paired post-MPN AML patient samples, with two having STK11 levels below 20%. We further validated downregulation of LKB1 by immunohistochemistry on paired chronic and blast phase MPN specimens and observed little staining in the blast phase specimens. Finally, to further show that the mechanism of in vitro enhanced self-renewal is related to leukemia progression, we stained the paired marrows for HIF1a and saw a dramatic increase in staining at the AML phase. We also analyzed RNA-seq data of paired chronic versus blast phase MPNs specimens and observed that there is a strong congruence of enriched pathways that are associated with the in vitro mouse HSPC phenotype and the human blast phase progression, such as oxidative phosphorylation and hypoxia. Together, our study demonstrates that loss of LKB1/STK11 promotes transformation of cells with activated JAK/STAT signaling and that STK11 is a prominent candidate tumor suppressor gene in post-MPN AML. Disclosures Gurbuxani: UpToDate: Honoraria. Hoffman:Dompe: Research Funding; Protagonist: Consultancy; Abbvie: Membership on an entity's Board of Directors or advisory committees; Forbius: Consultancy; Novartis: Membership on an entity's Board of Directors or advisory committees. Levine:Astellas: Consultancy; Amgen: Honoraria; Gilead: Honoraria; Qiagen: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Morphosys: Consultancy; Novartis: Consultancy; Prelude Therapeutics: Research Funding; Loxo: Current equity holder in private company, Membership on an entity's Board of Directors or advisory committees; Imago: Current equity holder in private company, Membership on an entity's Board of Directors or advisory committees; C4 Therapeutics: Current equity holder in private company, Membership on an entity's Board of Directors or advisory committees; Isoplexis: Current equity holder in private company, Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Honoraria, Research Funding; Roche: Consultancy, Honoraria, Research Funding; Lilly: Consultancy, Honoraria; Janssen: Consultancy. Rampal:Galecto: Consultancy; Incyte: Consultancy, Research Funding; Constellation: Research Funding; Stemline: Consultancy, Research Funding; Celgene: Consultancy; Jazz Pharmaceuticals: Consultancy; CTI Biopharma: Consultancy; Abbvie: Consultancy; Pharmaessentia: Consultancy; Promedior: Consultancy; Blueprint: Consultancy.

Abstract: The Myeloproliferative Neoplasms (MPNs), including Polycythemia Vera (PV), Essential Thrombocythemia (ET), and Primary Myelofibrosis (PMF) are clonal hematopoietic disorders. JAK2V617F mutations are observed in approximately 90-95% of PV cases, but only 40-50% of ET and PMF cases. Although JAK2 exon 12 and LNK mutations are observed in the majority of JAK2V617F-negative PV patients, candidate gene and exome sequencing studies to date have not identified activating oncogenes in the majority of JAK2V617F-negative ET/PMF patients. Thus, further genetic investigations are needed to define the mutational architecture of these JAK2 wildtype MPNs in order to gain insight into the biology of these diseases, the clinical implications of genetic events that do occur, and the elucidation of potential therapeutic targets. Aims To characterize the spectrum of genetic alterations in JAK2 wildtype chronic-phase myeloproliferative neoplasms. Methods We identified 32 patients with a confirmed diagnosis of an MPN (per 2008 WHO criteria), including MF, PV and ET, who were negative for JAK2V617F using a CLIA-certified allele specific assay for the JAK2 disease allele. Genomic DNA and total RNA was isolated from formalin fixed paraffin embedded (FFPE) tissue, blood and bone marrow aspirates. Adaptor ligated sequencing libraries were captured by solution hybridization using two custom baitsets targeting 374 cancer-related genes and 24 genes frequently rearranged for DNA-seq, and 258 genes frequently rearranged for RNA-seq. All captured libraries were sequenced to high depth (Illumina HiSeq), averaging 〉 590X for DNA and 〉 20,000,000 total pairs for RNA, to enable the sensitive and specific detection of genomic alterations. Results High coverage sequencing allowed us to identify JAK2V617F mutations in two patients (allele burden 3-5%) that were below the limit of detection of the CLIA assay. The most common mutations observed in JAK2V617F-negative MPN were in ASXL1 (22% of patients) and in TET2 (9%). Taken together, mutations in known epigenetic modifiers (ASXL1, TET2, DNMT3A, EZH2, MLL) were observed in 43% of samples, including a MLL-PTD mutation in one patient with PMF. We identified mutations in spliceosome components (SRSF2, U2AF1), in a subset of patients, consistent with previous reports, and in each case mutations in spliceosome components were mutually exclusive. We identified mutations in the JAK-STAT pathway (MPL, TYK2) and the RAS pathway components (KRAS, NF1) in 9% of this patient cohort, suggesting that there are alternate disease alleles that activate signaling in JAK2V617F-negative MPN. RNA-sequencing identified a ETV6-ABL1 fusion in one patient, and we identified amplification of PIK3CA in one patient in our cohort; these data suggest fusion genes and amplifications activate signaling in a subset of patients with JAK2V617F-negative MPN. We also identified novel mutations in MPN patients which have not been reported to date, including mutations in DNA repair genes (ATM and BRCA) in 25% of cases and mutations in the Notch signaling pathway (NOTCH1-4) in 31% of cases. The functional implications of these novel mutations remain to be elucidated. In univariate analysis, ASXL1 mutations were found to associate with impaired overall survival (Figure 1, p=0.049). These findings are consistent with data demonstrating an impaired survival in patients with MDS and PMF, and suggest that ASXL1 mutations represent an important biomarker for adverse outcome in JAK2V617F-negative MPN. Conclusions These data demonstrate that the mutational spectra of JAK2V617F-negative MPN includes genes implicated in epigenetic regulation, novel mutations which activate gene signaling, and fusion genes/copy number alterations which provide a novel mechanism of oncogenic activation not previously reported in MPN. ASXL1 mutations occur frequently in JAK2 wildtype Philadelphia-Chromosome negative MPNs, and are associated with impaired overall survival. Collectively, these findings support the importance of ASXL1 mutations in predicting outcome in JAK2V617F-negative MPN, demonstrate that mutations in signaling effectors and in epigenetic regulators are common in MPN, and illustrate the genetic heterogeneity of JAK2V617F-negative MPN. Disclosures: Rampal: Foundation Medicine: Consultancy. Levine:Foundation Medicine, Inc: Consultancy.

Abstract: Myelofibrosis is characterized by bone marrow fibrosis, atypical megakaryocytes, splenomegaly, constitutional symptoms, thrombotic and hemorrhagic complications, and a risk of evolution to acute leukemia. The JAK kinase inhibitor ruxolitinib provides therapeutic benefit, but the effects are limited. The purpose of this study was to determine whether targeting AURKA, which has been shown to increase maturation of atypical megakaryocytes, has potential benefit for patients with myelofibrosis. Patients and Methods: Twenty-four patients with myelofibrosis were enrolled in a phase I study at three centers. The objective of the study was to evaluate the safety and preliminary efficacy of alisertib. Correlative studies involved assessment of the effect of alisertib on the megakaryocyte lineage, allele burden, and fibrosis. Results: In addition to being well tolerated, alisertib reduced splenomegaly and symptom burden in 29% and 32% of patients, respectively, despite not consistently reducing the degree of inflammatory cytokines. Moreover, alisertib normalized megakaryocytes and reduced fibrosis in 5 of 7 patients for whom sequential marrows were available. Alisertib also decreased the mutant allele burden in a subset of patients. Conclusions: Given the limitations of ruxolitinib, novel therapies are needed for myelofibrosis. In this study, alisertib provided clinical benefit and exhibited the expected on-target effect on the megakaryocyte lineage, resulting in normalization of these cells and reduced fibrosis in the majority of patients for which sequential marrows were available. Thus, AURKA inhibition should be further developed as a therapeutic option in myelofibrosis. See related commentary by Piszczatowski and Steidl, p. 4868

Abstract: The Philadelphia chromosome–negative myeloproliferative neoplasms (MPNs), including polycythemia vera (PV), essential thrombocythemia (ET), and the prefibrotic form of primary myelofibrosis (PMF), frequently progress to more overt forms of MF and a type of acute leukemia termed MPN-accelerated phase/blast phase (MPN-AP/BP). Recent evidence indicates that dysregulation of the tumor suppressor tumor protein p53 (TP53) commonly occurs in the MPNs. The proteins MDM2 and MDM4 alter the cellular levels of TP53. We investigated in 1,294 patients whether abnormalities involving chromosomes 1 and 12, which harbor the genes for MDM4 and MDM2, respectively, and chromosome 17, where the gene for TP53 is located, are associated with MPN disease progression. Gain of 1q occurred not only in individuals with MPN-BP but also in patients with PV and ET, who, with further follow-up, eventually evolve to either MF and/or MPN-BP. These gains of 1q were most prevalent in patients with a history of PV and those who possessed the JAK2V617F driver mutation. The gains of 1q were accompanied by increased transcript levels of MDM4. In contrast, 12q chromosomal abnormalities were exclusively detected in patients who presented with MF or MPN-BP, but were not accompanied by further increases in MDM2/MDM4 transcript levels. Furthermore, all patients with a loss of 17p13, which leads to a deletion of TP53, had either MF or MPN-AP/BP. These findings suggest that gain of 1q, as well as deletions of 17p, are associated with perturbations of the TP53 pathway, which contribute to MPN disease progression.

Abstract: Myelofibrosis (MF) is a myeloproliferative disorder that exhibits considerable biological and clinical heterogeneity. At the two ends of the disease spectrum are the myelodepletive or cytopenic phenotype and the myeloproliferative phenotype. The cytopenic phenotype has a high prevalence in primary MF (PMF) and is characterized by low blood counts. The myeloproliferative phenotype is typically associated with secondary MF (SMF), mild anemia, minimal need for transfusion support, and normal to mild thrombocytopenia. Differences in somatic driver mutations and allelic burden, as well as the acquisition of non-driver mutations further influences these phenotypic differences, prognosis, and response to therapies such as JAK2 inhibitors. The outcome of patients with the cytopenic phenotype are comparatively worse and frequently pose a challenge to treat given the inherent exacerbation of cytopenias. Recent data indicate that an innate immune deregulated state that hinges on the myddosome-IRAK-NFκB axis favors the cytopenic myelofibrosis phenotype and offers opportunity for novel treatment approaches. We will review the biological and clinical features of the MF disease spectrum and associated treatment considerations.

Abstract: Hairy cell leukemia (HCL) is a chronic lymphoproliferative disorder recently found to be characterized by somatic BRAFV600E mutations. The malignant cell in HCL exhibits features consistent with a mature B-lymphocyte, including cell-surface expression of the pan-B-cell marker CD19 and monotypic surface immunoglobulins with clonal rearrangements of immunoglobulin heavy and light chains. Despite possessing these stereotypic features, the cell of origin of HCL has been long debated, and no cell type along the continuum of developing B-lymphocytes has been definitively identified as the normal counterpart of HCL cells. We hypothesized that HCL may originate from immature hematopoietic cells, and therefore investigated the hematopoietic-stem/progenitor cell (HSPC) compartment in HCL patients. We found that HCL patients exhibited a significantly increased frequency of immunophenotypically defined long-term hematopoietic stem cells (LT-HSCs; lineage-negative (Lin-neg) CD34+CD38-CD90+CD45RA- cells), pro-B cells (Lin-neg CD10+ cells), and CD34-CD38+ CD10+CD19+ hematogones, as well as a decreased frequency of granulocyte-macrophage progenitor cells (Lin-neg CD34+CD38+CD45RA+CD123+) relative to age-matched normal controls. Sequencing of cDNA from highly pure FACS-sorted cell populations from the bone marrow of HCL patients revealed the presence of the BRAFV600E allele in LT-HSCs and in pro-B cells (Figure). Transplantation of LT-HSCs from the pretreatment bone marrow of HCL patients into NOD/SCID/IL2r-gnull mice resulted in stable human grafts characterized by an expanded B-progenitor population and development of a clonal population of hCD19+hCD103+hCD25+ B cells characteristic of HCL 6 months after transplantation. Together, these data suggest that HCL arises from HSCs that then differentiate into committed B-cells which ultimately give rise to the characteristic clonal B-cell proliferation of HCL. Given the human HSC genetic and functional cell data, we conditionally expressed BRafV600E from its endogenous locus at different stages of hematopoiesis, including in HSPCs and committed B cells. Mice with conditional expression of BRafV600E in Mx1Cre+ BRafV600E knock-in mice died of a lethal hematopoietic malignancy characterized by features of human HCL including splenomegaly, anemia, thrombocytopenia, increased circulating sCD25, and increased clonogenic capacity of B-lineage cells (evidenced by infinite serial replating in the presence of IL-7) (Figure). This disorder was transplantable into lethally-irradiated recipient mice. In contrast, mice with expression of BRafV600E restricted to the B-cell lineage with Cd19 Cre manifested no overt malignant phenotype up to one year of age. Stimulation of these mice with alloantigen through injections of sheep red blood cells resulted in germinal center B-cell hyperplasia, but still did not result in development of a clonal B-cell proliferation. Recent case reports have noted that refractory HCL patients respond to mutant BRAF inhibition with vemurafenib. We investigated the effect of vemurafenib on HSPCs and hematopoiesis in patients treated on a phase II study of the mutant BRAF inhibitor vemurafenib for relapsed/refractory HCL as well as in our in vivo murine models. Flow cytometric analysis of bone marrow cells from vemurafenib treated HCL patients revealed normalization of HSPC frequencies within three months of starting therapy, concomitant with an improvement in peripheral blood counts. Consistent with this, evaluation of the in vitro clonogenic capacity of sorted LT-HSC's from the bone marrow of HCL patients revealed a significant increase in myeloid/erythroid colony formation in HCL patients treated for 3 months with vemurafenib compared to their pretreatment marrows. Likewise, treatment of wildtype mice transplanted with Mx1Cre+ BRafV600E mutant bone marrow cells revealed improvement in anemia and hepatosplenomegaly with in vivo therapy. Overall, these findings link the pathogenesis of HCL to a specific somatic genetic abnormality present in HSCs and provide evidence that mature B-cell malignancies can initiate in the HSC compartment. Moreover, these data suggest that the use of therapies targeting MAP kinase signaling in HCL may lead to durable remissions not only by eliminating the mature leukemic cells but also through targeted inhibition of signaling and survival in HCL initiating cells. Disclosures: No relevant conflicts of interest to declare.

Abstract: Introduction: CMML is a clinically heterogeneous myeloid neoplasm hallmarked by the coexistence of dysplastic and proliferative patho-clinical features, which can include cytopenias, constitutional symptoms, splenomegaly, and leukocytosis. However, according to the FAB classification schema, CMML may be differentiated in dysplastic and proliferative subgroups only by the presence of leukocytosis in the latter (WBC ≥13 x 109/L). We hypothesize that incorporation of other clinically discriminating features may yield a more informative CMML stratification system. To address this, we propose three distinct CMML categories and explore their clinical relevance leveraging our existing international CMML database (Padron E et al. Blood Cancer J. 20151). Method: 1622 WHO-defined CMML cases diagnosed between 1973 and 2014 were collected from eight large cancer centres that include up to 80 discrete data elements as previously described1. Cases were placed into three clinically distinct groups and the Pearson Chi-Square test and the Kruskal-Wallis test were applied respectively to compare categorical and continuous characteristics. The Kaplan-Meier (KM) method was used to estimate median OS and the log rank test was used to compare survival curves.Cox models whereapplied to obtain univariate and adjusted hazard ratios. Identification of optimal cut-off values for continuous variable was supported by graphical inspection of martingale residuals from the null Cox model. Statistical analyses were done in SPSS v23 and R v.3.3.0. Results: We propose three categories to delineate clinically distinct CMML subtypes: (1) Myelodysplastic (MD)-CMML: WBC≤10 x 109/L, PB-immature myeloid cells (IMC) = 0%, no splenomegaly (2) MD/MP-CMML: WBC 10-20 x 109/L or WBC ≤10 x 109/L but PB-IMC 〉 0% and/or splenomegaly (3) Myeloproliferative (MP)-CMML: WBC 〉 20 x 109/L. A recursive partitioning approach was used to identify the WBC cut points, with splenomegaly and IMC added to more accurately depict the MPN aspect of CMML. Numbers of patients included in the MD-, MD/MP-, and MP-CMML subcategories were 319 (19.7%), 789 (48.6%) and 514 (31.7%), respectively (Table 1). According to the FAB criteria, the MD/MP group included 521 (66%) MD- (WBC ≤13 x 109/L) and 268 (34%) MP-CMML (WBC 〉 13 x 109/L) patients suggesting that the proposed classification clinically reclassifies FAB-defined CMML. Within the MD/MP group, 344 patients (21.2%) had a WBC 〈 10 x 109/L but with IMC 〉 0% and/or splenomegaly. Comparison of overall survival (OS) among proposed groups demonstrated that this classification schema was capable of discriminating the CMML natural history (Figure 1). In comparison to MD/MP-CMML, the unadjusted OS Hazard Ratio (HR) was 0.60 (95% CI 0.49-0.73) for MD-CMML and 1.57 (95% CI 1.36-1.81) for MP-CMML (p 〈 0.001 for both). This difference was retained after adjusting for BM-Blasts ( 〈 vs ≥5%), IPSS or CPSS cytogenetics (High-risk vsInterm/low-risk), and elevated levels of LDH (HR was 0.73 with 95% CI 0.57-0.93, p=0.011 for MD-CMML, and 1.33 with 95% CI 1.12-1.58, p=0.001 for MP-CMML, respectively). We next explored whether each proposed group may have distinct variables that uniquely govern its prognosis. Peripheral blood blasts 〉 5%, RBC- and PLT-transfusion-dependence predicted poor OS only in the MD and MD/MP-subcategories (p 〈 0.005), but had no impact in the MP-CMML. Male gender and elevated LDH were only significantly associated with prognosis in the MD/MP group (p 〈 0.05 and p 〈 0.005, respectively) while absolute lymphocyte count 〉 2.5 x 109/L and absolute monocyte count 〉 10 x109/L were uniquely predictive for shorter survival in the MP-CMML subcategory (p 〈 0.005). Notably, we additionally identified gene mutations that uniquely predicted prognosis in each proposed group (Table 2). Conclusions: We demonstrate that our proposed 3-group clinical classification schema is capable of independently stratifying prognosis. Further, our analysis identified clinical and genetic variables that uniquely govern each group's prognosis, suggesting independent clinical behaviour. Further investigations are warranted to validate these groups. Disclosures Jabbour: ARIAD: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding; Novartis: Research Funding; BMS: Consultancy. Fenaux:Celgene, Janssen,Novartis, Astex, Teva: Honoraria, Research Funding. Kantarjian:Bristol-Myers Squibb: Research Funding; ARIAD: Research Funding; Amgen: Research Funding; Pfizer Inc: Research Funding; Delta-Fly Pharma: Research Funding; Novartis: Research Funding. Komrokji:Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding; Novartis: Consultancy, Speakers Bureau.

Abstract: The Philadelphia-chromosome negative myeloproliferative neoplasms (MPNs) Essential Thrombocytosis (ET), Polycythemia Vera (PV), and Myelofibrosis (MF) are characterized by mutations, which drive JAK-STAT pathway activation. Several studies have demonstrated the presence of recurrent somatic mutations outside of the JAK-STAT pathway, which accumulate over time, and may impact disease phenotype and outcome. We sought to determine the influence of somatic mutations on clinical phenotype and prognosis. We sequenced a total of 30 genes recurrently mutated in myeloid malignancies in a cohort of 162 MPN patients (pts) using a next generation sequencing platform. The cohort included 49 pts with ET, 26 PV, 38 Primary Myelofibrosis (MF), 11 Post ET MF, 14 Post PV MF, 12 with leukemic transformations of MPN (LT), 7 with MPN-unclassified (MPN-U) and 5 others. Median age was 59 years and 79 were men. A Total of 288 gene mutations were identified with the most commonly mutated genes being JAK2 (n=121, 74%), TET2 (n=31, 19%), DNMT3A (n=18, 11%), ASXL1 (n=16, 10%), IDH2 (n=10, 6%), RAS (n=12, 7%), TYK2 (n=8, 5%) and TP53 (n=7, 4%). We did not find any mutations in NPM1, CBL, SRSF2 and no FLT3 -ITD. CALR was not assessed in 20 pts and these were excluded from mutation number analysis. Importantly, we identified a relationship between the absolute number of mutations found per pt, disease phenotype, and age (table 1). Pts with/without prior chemotherapy or radiotherapy exposure did not have a difference in mutation number (1.5 vs. 1.9). Cases of ET or PV with fibrotic transformation had more mutations in ASXL1, RAS, EZH2, PHF6 and MPL than pre fibrotic ET or PV suggesting these may be relevant in disease progression and development of fibrosis. Mutations in TET2, RAS and PHF6 were more frequent in cases with LT compared to those with chronic phase MPN. Pts over 40 were more likely to have mutations in TET2 (p=0.026) and JAK2 (p=0.019) and ASXL1 mutations were more common in pts with abnormal cytogenetics than in those with normal cytogenetics (p=0.003). Thrombotic events, which are an important cause of morbidity in MPN patients, negatively correlated with mutations in ASXL1 (p=0.044). Prognosis as measured by DIPPS and DIPSS-Plus scores appeared to correlate with the average number of mutations found in MF patients (table 2). We examined several cases for which serial samples were available, and noted the acquisition of new mutational events despite ongoing therapy. We noted that the most commonly acquired mutations occurred in epigenetic modifying (DNMT3A, TET, IDH, ASXL1) and in growth signaling pathway (RAS, CBL) genes. These occurred despite active therapy and often without an overt change in clinical phenotype. Further details of these serial samples will be presented. We conclude that the number and spectrum of somatic mutations correlate with disease phenotype of MPN. Younger pts have fewer mutations, as do pts with normal cytogenetics. JAK2 and TET2 mutations were more common in older pts. We show that a subset of pts acquire mutations in epigenetic modifiers and in genes involved in growth signaling pathways during disease course, and that mutations in TET2, RAS and PHF6 were enriched at the time of leukemic transformation. Taken together, these results indicate that mutations outside the JAK-STAT pathway influence disease phenotype, and that the acquisition of mutations over time may predict for disease progression. Serial evaluation of mutational burden over time therefore warrants exploration in the clinical setting. Table 1. Average number of mutations appeared to correlate with disease phenotype, age and abnormal cytogenetics. Average Number of Mutations N Mean (SD) P-value Age 〈 40 years 13 1.4 (0.9) 0.026 Age 〉 40 years 12 2 (1) No Thrombosis 113 2 (1) 0.712 Thrombosis 28 1.9 (1) Normal Cytogenetics 64 1.8 (0.9) 0.016 Abnormal Cytogenetics 40 2.3 (1.2) ET/PV/PMF 99 1.8 (0.8) 0.029 LT 10 3 (1.5) ET/PV 66 1.6 (0.7) 0.01 Post ET/PV MF 22 2.3 (1.1) ET 44 1.6 (0.7) 〈 0.001 PV 22 1.5 (0.9) PMF 33 2.2 (0.9) Post ET/PV MF 22 2.3 (1.1) LT 10 3 (1.5) Table 2. Disease prognostic scores in MF appear to correlate with the average number of mutations found per patient. Risk category N Average number mutations DIPSS Low 8 1.5 Intermediate-1 19 2.4 Intermediate-2 9 1.8 High 1 5 DIPSS-Plus Low 6 1.5 Intermediate-1 12 2 Intermediate-2 14 2.2 High 1 5 Figure 1. Comutation map of genomic alterations. Each hash mark on x-axis represents an individual patient. Figure 1. Comutation map of genomic alterations. Each hash mark on x-axis represents an individual patient. Disclosures Levine: CTI BioPharma: Membership on an entity's Board of Directors or advisory committees; Loxo Oncology: Membership on an entity's Board of Directors or advisory committees; Foundation Medicine: Consultancy.

Abstract: Background: The presentation of acute myeloid leukemia (AML) as an extramedullary tumor occurs in up to 9% of cases, making it an uncommon event and one that poses difficulties in clinical decision-making. Current data on treatment and outcomes for myeloid sarcoma are based largely on case reports and clinician experiences involving a small number of patients. Radiation is often added to standard chemotherapy regimens in order to attempt to improve rates of complete remission, however there is a paucity of data to guide the utilization of radiation therapy in addition to chemotherapy. Aims: To determine the impact of chemotherapy, the combination of chemotherapy and radiation therapy, and transplantation on the survival of patients with myeloid sarcoma. Methods: Beginning with a PubMed search from 1985 to 2014, we performed a review of the literature on cases of isolated myeloid sarcomas (without bone marrow involvement) presenting as the initial diagnosis of AML. We identified 7 large case reports or reviews where individual patient characteristics (including treatment course, site of disease, overall survival (OS) and time to development of bone marrow involvement) were defined. We also undertook a chart review of the patients treated at Memorial Sloan Kettering Cancer Center from 1990 to July 2014 who upon initial presentation of AML demonstrated extramedullary involvement. We excluded patients with extramedullary disease at relapse, those presenting with lymph node, splenic disease or leukemia cutis in both data sets. For the MSKCC data, we included those who presented with bone marrow involvement due to the low number of patients available for analysis. Results: The literature search identified 71 analyzable patients and there was no significant difference in OS whether patients with isolated myeloid sarcoma were initially treated with radiation therapy (n= 14, 11.2 months), chemotherapy (n=40, 19 months) or the combination of chemotherapy and radiation (n=17, 24 months, p=0.41) Fig 1. The median OS of these patients was 16 months (n=71). At MSKCC, 27 patients were available for analysis. The median 3 year OS of patients presenting with AML and extramedullary involvement was 41%. There was a significant improved 1 year OS (p=0.002) if patients were treated with chemotherapy alone (n=19), as opposed to the combination of radiation and chemotherapy (n=6), 89% vs. 33%, respectively. There was no difference in OS if patients presented with bone marrow involvement or with isolated disease (35% vs. 55% at 3 years, p= 0.51). Lastly, patients who underwent allogeneic stem cell transplantation (n=13) exhibited a median 44% 3 year OS. Only seven of these patients were in first complete remission. Summary/ Conclusion: Patients with AML who present with extramedullary disease seem to derive no additional survival benefit from the inclusion of radiation therapy to chemotherapy as initial treatment. Also, presentation with isolated myeloid sarcoma does not appear to confer a worse prognosis when compared to myeloid sarcoma with bone marrow involvement and likely represents a different manifestation of the same disease process. For comparison, regarding overall survival, patients enrolled in ECOG 1900 who received standard therapy had a median OS of 15.7 months and those with intermediate risk cytogenetics and an intermediate risk mutational risk profile, as described by Patel et al in NEJM 2012, had a 3 year OS of 42% comparable to the outcomes we noted above. This suggests that patients with myeloid sarcoma do no worse than those with bone marrow only disease. The survival of patients who underwent allogeneic stem cell transplantation in our study had a similar median 3 year OS to patients with AML who undergo allogeneic stem cell transplantation studied by the CIBMT and published in Blood 2012 by Saber et al. The 3 year OS rates for patients who underwent a matched related donor transplant in this publication was 39%, and 37% for those who had a matched unrelated donor transplant. In summary, though the sample size is notably small and acknowledging the limitations of retrospective data, patients with myeloid sarcoma during the initial presentation of AML should undergo risk stratification and treatment in the same manner as patients with AML without extramedullary disease. Figure 1 Figure 1. Disclosures No relevant conflicts of interest to declare.