Search

Background While allogeneic hematopoietic cell transplantation (alloHCT) can be curative for patients with acute myeloid leukemia (AML), relapse remains a significant challenge. Previous work has suggested that disease status at time of transplant and cytogenetics are important predictors of relapse. However, it is unclear if common somatic mutations or dimorphisms of MHC class I chain-related gene A (MICA), a ligand of the natural killer (NKG2D) receptor on immune effector cells that helps mediate NK cell alloreactivity, also contribute. Moreover, the mechanisms of early relapse are an area of ongoing investigation. We assessed risk factors for relapse within 6 and 12 months after alloHCT. Methods We conducted a single center, retrospective analysis of adults with AML who underwent a first alloHCT. Analysis was restricted to patients with T-cell replete HLA-8/8 matched related or unrelated donor. In addition to cytogenetic risk group stratification by European LeukemiaNet criteria (Döhner H, et al, Blood 2016), a subset of patients had a 36-gene somatic mutation panel assessed prior to alloHCT by next-generation sequencing. Dimorphisms at the MICA-129 position have previously been categorized as weaker (valine/valine: V/V), heterozygous (methionine/valine: M/V), or stronger (methionine/methionine: M/M) receptor binding affinity. Risk factors for early relapse were assessed with Fine and Gray competing risk regression with results as hazard ratios (HR) and 95% confidence intervals (CI). Results From 2000 - 2017, 319 adult AML patients were identified meeting inclusion criteria. Median age at transplant was 51 years (range, 18-74), with 95% Caucasian. The distribution of low, intermediate, and high HCT-CI scores was 28%, 28%, and 44%, respectively. 75% of patients were transplanted ≤12 months from diagnosis. Disease status at transplant included 48% in first complete remission (CR1), 19% in second CR (CR2), 33% in third CR or relapsed/refractory or untreated (collectively, In univariable analysis, non-Caucasian race, disease status Conclusion Relapse after alloHCT for AML remains a challenge. In our study, the strongest risk factors for early relapse after alloHCT remains absence of CR1 disease status at transplant and adverse-risk cytogenetics. We observed no prognostic effect of somatic mutations nor MICA dimorphisms prior to transplant on 6 or 12-month relapse post-transplant. Further interrogation of pre-transplant or post-transplant persistence of somatic mutations in a larger series may better risk stratify subjects who may benefit from more intensive or innovative approaches to prevent post-transplant relapse. Disclosures Nazha: MEI: Consultancy. Advani:Amgen: Research Funding; Pfizer: Honoraria, Research Funding; Novartis: Consultancy; Glycomimetics: Consultancy. Carraway:Novartis: Speakers Bureau; Amgen: Membership on an entity's Board of Directors or advisory committees; Balaxa: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; FibroGen: Consultancy; Jazz: Speakers Bureau; Agios: Consultancy, Speakers Bureau. Gerds:Celgene: Consultancy; Apexx Oncology: Consultancy; Incyte: Consultancy; CTI Biopharma: Consultancy. Sekeres:Celgene: Membership on an entity's Board of Directors or advisory committees; Opsona: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees; Opsona: Membership on an entity's Board of Directors or advisory committees. Maciejewski:Apellis Pharmaceuticals: Consultancy; Alexion Pharmaceuticals, Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Apellis Pharmaceuticals: Consultancy; Ra Pharmaceuticals, Inc: Consultancy; Ra Pharmaceuticals, Inc: Consultancy; Alexion Pharmaceuticals, Inc.: Consultancy, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Majhail:Atara: Honoraria; Incyte: Honoraria; Anthem, Inc.: Consultancy.

Approximately 25% of MDS patients (pts) progress to AML; 15% of those with lower-risk, and up to 40-50% of those with higher-risk disease. Predicting those who are likely to progress to AML early in their disease course could directly impact treatment decisions: lower-risk pts with higher risk of AML transformation may be offered transplant, while higher-risk pts may be considered for AML-like therapy. In this study, we developed a genomic model that evaluates mutational patterns and their association with AML progression. Development of this model mimics Netflix or Amazon9s recommender system in which customers who bought products A and B, are likely to buy C: pts who have a mutation in genes A and B, are then likely to progress to AML. Clinical and genomic data ofMDS pts diagnosed between 1/1996 and 9/2016 were analyzed. A panel of 60 gene mutations obtained by next generation targeted deep sequencing was included. Association rules using Apriori algorithm was used to study the relationship between multiple genes/cytogenetic abnormalities and AML progression in an unbiased approach. Association rules are a machine learning method that uncovers relationships between variables in a given dataset. Rules with highest confidence (>90% that an association exists) and highest lift (strength of the association) were chosen. Univariate and multivariate analyses were used to evaluate the impact of mutations on AML progression, adjusting for covariates (age, IPSS-R). A total of 527 pts with lower-/higher-risk MDS were analyzed, of those 105 (20%) progressed to AML, with a median time to progression (TTP) of 13 months (interquartile range, 6.3-29.8). Median age at diagnosis was 67 years (range, 19-99) and 38% were female. Risk categories by IPSS-R included: 78 (15%) very low, 200 (38%) low, 95 (18%) intermediate, 98 (18.5%) high, and 56 (10.5%) very high; IPSS-R cytogenetic risk groups were 15 (3%) very good, 331 (63%) good, 87 (16%) intermediate, 37 (7%) poor, and 57 (11%) very poor. The most commonly mutated genes were SF3B1 (14%), ASXL1 (13%), TET2 (12%), SRSF2 (11%), DNMT3A (10%), STAG2 (9%), and TP53 (8%). Per IPSS-R, 14% of pts with very low/low, 21% with intermediate and 30% with high/very high categories progressed to AML. Univariate analyses identified the following cytogenetic and mutational abnormalities to be associated with AML progression: complex karyotype (CK, P =.019), ASXL1 ( P =.023), DNMT3A ( P =.011), FLT3 ( P =.031), PHF6 ( P =.018), PTPN11 ( P =.037) , RUNX1 ( P =.001) , STAG2 ( P =.023) , TET2 ( P =.047) , U2AF1 ( P =.038), and ≥3 gene mutations( P =.002), whereas SF3B1 was associated with lower risk of AML progression ( P =.025). In multivariate Cox regression analyses, including significant mutations, age ( P =.5) and IPSS-R scores ( P RUNX1 (HR2.8 ; 95% CI 1.6-5.3; P= .001), and ≥3 gene mutations (HR=1.65; 95% CI 1.1-2.5; P= .021) were independent factors that associated with AML progression, while SF3B1 (HR=.417; 95% CI .205-.851, P =.016) was associated with lower risk of progression. Association rules identified the following combinations of cytogenetic/genomic abnormalities for AML progression: (ASXL1, RUNX1, STAG2/BCOR), (ASXL1, STAG2, ZRSR2/SRSF2), (ASXL1, TET2, Trisomy8 ), (RUNX1, STAG2, Trisomy 8 ), ( CK , TP53, TET2) . These combinations were associated with a 5 month reduction in time to AML progression.[Figure 1] Using genomic data to reliably identify patients at highest risk for progression to AML early in their disease course can dramatically alter treatment recommendations. Independent factors such as RUNX1 and ≥3 gene mutations predicted for AML progression, while SF3B1 was associated with a reduced risk. Certain cytogenetic/genomic abnormalities were associated with shorter time to AML progression. This data suggests that MDS pts with ≥3 genomic abnormalities or certain genomic combinations could be offered more aggressive therapy early in their disease course including those with lower risk disease. Disclosures Sekeres: Celgene: Membership on an entity9s Board of Directors or advisory committees. Gerds: Incyte: Consultancy; CTI BioPharma: Consultancy. Advani: Takeda/ Millenium: Research Funding; Pfizer: Consultancy. Maciejewski: Alexion Pharmaceuticals, Inc.: Consultancy, Membership on an entity9s Board of Directors or advisory committees, Other: Speaker Fees; Apellis Pharmaceuticals: Consultancy; Ra Pharma: Consultancy.

Paroxysmal Nocturnal hemoglobinuria (PNH) has been traditionally considered a monogenic disease due to somatic mutations in PIGA gene. While selective immune pressure was implicated in PNH evolution via expansion of a PIGA mutant clone in a privileged environment, history of aplastic anemia (AA) is not always present in a manifest hemolytic PNH. Similarly, expansion of PNH clone after AA therapy does not seem to correlate with the strength of immunosuppression and the quality of response. Our initial studies (J.Clin.Invest.2014; 124 :4529-4538) indicated that additional mutational events may be present in PNH and act as intrinsic factors modifying clinical features and leading to a differential expansion. We hypothesize that subclonal hits may be present in some patients with PNH to explain their expansion dynamics and clinical features. To that end we have collected 319 patients with bone marrow failure of whom 202 had PNH clone including 70 hemolytic PNH cases. For a subgroup of these patients (n=197), deep multi-targeted NGS has been performed to identify mutations in 78 myeloid genes of which 38 genes were found to be mutated (≥1) in this PNH spectrum disease. 116 serial samples for 33 patients were also sequenced. As a control group, we have included AA stably negative for PNH clone (n=113). For analytic purposes, on clinical grounds we have sub grouped patients with PNH clone into AA/PNH (31%) defined as patients with presence of bone marrow failure but no clinical or laboratory evidence of hemolysis and WBC PNH clone 20%. Our analysis focused on driver mutations. In cross-sectional analysis, in addition to PIGA we have identified somatic hits in 10/26 (38%) of AA/PNH pts. (On average 2.1 hits/per positive patient) compared to 27/58(47%) of PNH patients (on average 1.22 hits per positive patients. Subdividing PNH patients, 20/35(57%) had a mutation in primary PNH compared to 7/23(30%) in the sPNH group. In contrast, 42/113(37%) patients had a mutation in the control AA group with a (2.1 mutations per positive patient). The mean VAF for AA is 29% compared to 23% in AA/PNH vs. 34% in PNH vs. 31%in sPNH. In all patients with PNH clone, 44% had a mutation vs. 37% in the AA group. AA/PNH subcohort contained hits in cohesins genes (RAD21-SMC3-STAG2 (8 %), ZRSR2 (8%), BCORs (4%) and TET2 (4%). PNH group had mutations in BCOR/BCORL1 (BCORs) (11%), cohesions, ZRSR2 and TET2 (6% each). In sPNH, a similar trend was observed with the most common hits in EZH2-SUZ12 (9%), ZRSR2 (9%), and BCORs, TET2, LUC7L2 (4% each). In contrast, AA without PNH, the 5 most frequently mutated genes were TET2, ASXL1, EZH2-SUZ12 AND BCORs (5% each) and RUNX 1 (4%). BCORs mutations were more frequent in PNH group compared to AA, (11% vs. 5%). In summary, BCOR/L and ZRSR genes showed more mutations in PNH clone cohort while AA cohort did not show any dominant hits. To determine whether somatic hits observed cross-sectionally led to establishment of stable subclones or were merely transient, we performed a longitudinal analysis. In PNH, 3/4 clones were transient which included EZH2, PHF6 and STAG2. In AA/PNH patients, 9/14 mutational hits appeared in the course of their disease (TET2, CEBPA, CUX1, STAG2, BCOR, SMC3) but 10/14 faded away in subsequent course and thus only 4 were stable clones. In AA cohort, 32 additional mutational hits were noted in 77 serial samples and 34 mutations vanished in subsequent course. In sum, our results show that in PNH, similar to AA, additional somatic hits are common and involve some of the commonly mutated myeloid genes seen in MDS. While likely, many of these subclonal events are transient, seemingly lacking the proper ancestral context, some of the hits resulted in significant expansions (sweeping mutations) and thus may modify the clinical course including speed of the PNH clone evolution. Disclosures No relevant conflicts of interest to declare.

Primary myelofibrosis is a stem cell-derived clonal malignancy characterized by unchecked proliferation of myeloid cells, resulting in bone marrow fibrosis, osteosclerosis, and pathologic angiogenesis. Bone marrow fibrosis (BMF) plays a central role in the pathophysiology of the disease. This review describes current issues regarding BMF in primary myelofibrosis, including the pathophysiology and impact of abnormal deposition of excess collagen and reticulin fibers in bone marrow spaces, the modified Bauermeister and the European Consensus grading systems of BMF, and the prognostic impact of BMF on the overall outcome of patients with primary myelofibrosis. The impact of novel therapeutic strategies, including JAK-STAT inhibitors and allogeneic stem cell transplant, on BMF is discussed. Implications for Practice: Bone marrow fibrosis (BMF) plays an important role in the pathophysiology and the clinical outcomes of patients with primary myelofibrosis. The severity of BMF correlates with the clinical manifestations of the disease and impacts the survival in patients with myelofibrosis. Treatment with ruxolitinib has been shown to reverse BMF and to continue that trend with ongoing treatment. Further studies to fully understand the mechanisms of fibrosis, to further explore the ability of currently available agents (e.g., JAK-STAT inhibitors) to stabilize and/or reverse fibrosis, and to develop additional fibrosis-targeted therapies are warranted.

Several prognostic scoring systems have been developed to risk stratify patients with myelodysplastic syndromes (MDS) [1–7]. The International Prognostic Scoring System (IPSS) has been the most wid...

Myelodysplastic syndromes (MDS) are heterogeneous hematopoietic neoplasms that are driven by somatically acquired genetic mutations and epigenetic alterations. Accurate risk stratification is essential for delivery of risk-adaptive therapeutic interventions. The current prognostic tools sum the impact of clinical, pathologic, and laboratory parameters. Newer technologies with next-generation targeted deep sequencing and whole-genome and -exome sequencing have identified several recurrent mutations that play a vital role in the pathophysiology of MDS and the impact of these genetic changes on disease phenotype. Equally important, well-annotated databases of MDS patients with paired clinicopathologic and genetic data have enabled better understanding of the independent prognostic impact of several molecular mutations on important clinical endpoints such as overall survival and probability of leukemic progression. Cumulative evidence suggests that genomic data can also be used clinically to aid with the diagnosis, prognosis, prediction of response to specific therapies, and the development of novel and rationally targeted therapies. However, the optimal use of this mutational profiling remains a work in progress and currently there is no standard set of genes or techniques that are recommended for routine use in the clinic. In this review, we discuss the genomic revolution and its impact on our understanding of MDS biology and risk stratification. We also discuss the current role and the challenges of the application of genetic mutational data into daily clinical practice and how future research could help improve the prognostication precision and specific therapy selection for patients with MDS. Implications for Practice: Heterogeneity in clinical outcomes of MDS is partly related to interpatient variability of recurrent somatic mutations that drive disease phenotype and progression. Although clinical risk stratification tools have functioned well in prognostication for patients with MDS, their ability to predict clinical benefits of specific MDS therapies is limited. Molecular testing shows promise in aiding diagnosis, risk stratification, and therapy-specific benefit prediction for MDS patients. Nonetheless, logistical issues related to assay performance standardization, validation, interpretation, and development of guidelines for how to use the results to inform clinical decisions are yet to be resolved.

Response to tyrosine kinase inhibitors at three months is a predictor for long-term outcome in chronic myeloid leukemia patients treated with tyrosine kinase inhibitors. We analyzed 456 newly diagnosed chronic myeloid leukemia patients treated with tyrosine kinase inhibitors to determine their outcome based on their response at six months. Forty-four (10%) patients did not achieve major cytogenetic response at three months: 18 of 67 (27%) patients treated with imatinib 400; 18 of 196 (9%) with imatinib 800; and 8 of 193 (4%) with 2nd generation tyrosine kinase inhibitors. Among them, 19 (43%) achieved major cytogenetic response at six months and subsequently had an overall outcome similar to the patients who achieved a major cytogenetic response at three months. In conclusion, the response to tyrosine kinase inhibitors at three months is a static, one-time measure. Assessing the response at six months of patients with poor response at three months may provide a better predictor for long-term outcome.

GvHD-free, relapse-free survival after reduced-intensity allogeneic hematopoietic cell transplantation in older patients with myeloid malignancies

African-American (AA) patients have a younger age at diagnosis and worse outcomes compared to Whites (WTs) across many cancers, including acute myeloid and acute lymphoblastic leukemias.1, 2, 3 Although inferior outcomes for AAs compared to WTs may be related to differential access to medical care or socioeconomic status, disease biology and genetic fingerprint may have a role. A recent study from the Cancer Genome Atlas has shown that WT and AA women with stage I to III breast cancer have differences in the genomic landscape that is correlated with a higher incidence of TP53 mutations, fewer PIK3CA mutations and greater intratumor genetic heterogeneity for AA compared to WTs. This finding suggests that the poor outcome of AA women with breast cancer is not only driven by the disease characteristics such as younger age at diagnosis and a higher incidence of triple-negative disease, but also related to differences in the genomic landscape of their disease.

Acute myeloid leukemia (AML) is a heterogeneous neoplasm characterized by the accumulation of complex genetic alterations responsible for the initiation and progression of the disease. Translating genomic information into clinical practice remained challenging with conflicting results regarding the impact of certain mutations on disease phenotype and overall survival (OS) especially when clinical variables are controlled for when interpreting the result. We sequenced the coding region for 62 genes in 468 patients with secondary AML (sAML) and primary AML (pAML). Overall, mutations in FLT3, DNMT3A, NPM1 and IDH2 were more specific for pAML whereas UTAF1, STAG2, BCORL1, BCOR, EZH2, JAK2, CBL, PRPF8, SF3B1, ASXL1 and DHX29 were more specific for sAML. However, in multivariate analysis that included clinical variables, only FLT3 and DNMT3A remained specific for pAML and EZH2, BCOR, SF3B1 and ASXL1 for sAML. When the impact of mutations on OS was evaluated in the entire cohort, mutations in DNMT3A, PRPF8, ASXL1, CBL EZH2 and TP53 had a negative impact on OS; no mutation impacted OS favorably; however, in a cox multivariate analysis that included clinical data, mutations in DNMT3A, ASXL1, CBL, EZH2 and TP53 became significant. Thus, controlling for clinical variables is important when interpreting genomic data in AML.

Background:Chronic anemia is most often the presenting symptom and major issue for patients (pts) with low-risk myelodysplastic syndrome (LR-MDS). Erythropoietic stimulating agents (ESAs) are the initial therapy for many LR-MDS pts and can predictably yield benefit in pts with low transfusion burden and a low erythropoietin (EPO) level. (Hellstrom-Lindberg, 2013). However, only about 40-50% of pts respond to erythropoietin (ESAs). We posit that improvement of the predictive algorithm for response to ESA therapy might be optimized with inclusion of molecular mutation informatics. If validated, the model could inform those pts with risk of ESA treatment failure and thus favor alternative approaches such as hypomethylating therapy (Thus, sparing unnecessary pt exposure, cost and time to an ineffective agent, as well as delay to delivery of a possibly more effective therapy). Furthermore, the impact of EPO stimulation/treatment on the clonal dynamics of LR-MDS is not yet well understood nor described. Thus, we analyzed molecular profiles of LR-MDS pts in order to identify mutations associated with ESA failure and the subsequent clonal alterations post ESA exposure. Methods:Of 394 pts who had LR-MDS based on International Prognostic Scoring System (IPSS) criteria, and were treated at Cleveland Clinic from 2002-2015, we studied 49 primary cases with clinical and mutational data. Next generation targeted deep sequencing was performed for a panel of 60 genes that are known to be commonly mutated in myeloid malignancies. ESA response was assessed in the first 8 weeks of treatment according to International Working Group (IWG) criteria (Cheson, 2006). We further studied clonal alterations upon ESA treatment in a small subset of patients who had sequential samples during the course of treatment. Results: Median age of 50 pts was 76 yrs (range, 45-87) and 56% were male. 28% had MDS with single lineage dysplasia, 21% had MDS with ringed sideroblasts, 38% had MDS with multilineage dysplasia, 6% had isolated del(5q), and 7% had MDS-unclassifiable. 65% had normal cytogenetics at diagnosis and 44% had low EPO level at baseline. 43% of pts responded to ESA therapy. Mutations commonly found in our cohort included SF3B1 (25% in responders vs 22% in non-responders), U2AF1 and DNMT3A (15% in responders vs 11% in non-responders), ASXL1 (10% in responders vs 15% in non-responders), and TET2 (10% in responders vs 11% in non-responders). In 20 responders, 2 cases had CBL2 mutation, while 2 other cases harbored JAK2 mutation, which were not detected in non-responders. In 27 non-responders, NGS identified EZH2, ETV6, and RUNX1 mutations in 2 individual cases while one case harbored both ETV6 and RUNX1 mutations. None of these mutations were found in responders. Of interest, among 8 patients who had low baseline EPO level without response to ESA, 3 patients harbored RUNX1 mutation. Serial analysis of eight cases revealed genomic diversity that may give insight to prognostic models for EPO response. For example, we found in one case who had a U2AF1, CBL mutations prior to starting ESA, at the time of evolution to AML this pt had acquired additional mutations involving RUNX1, IDH1, and GPR98. Additional cases with high risk mutations at presentation including EZH2, ASXL1, DNMT3A, U2AF1did not respond and evolved to secondary acute myeloid leukemia (sAML). Conclusions: Evaluation of the mutational spectrum in LR-MDS patient, revealed insight into clonal dynamics that may be determinants of EPO responsiveness. We found in our cohort that LR-MDS individuals harboring RUNX1 and ETV6 mutations did not respond to ESA directed therapy and were noted to evolve to high risk MDS or sAML. Notably, pts with SF3B1 mutations can respond to ESA therapy but not all pts appear to reliably do so. These mutational data will inform a future algorithm to help predict ESA responsive patients in LR-MDS that will require prospective validation. Disclosures Carraway: Novartis: Membership on an entity's Board of Directors or advisory committees; Baxalta: Speakers Bureau; Amgen: Membership on an entity's Board of Directors or advisory committees; Celgene: Research Funding, Speakers Bureau; Incyte: Membership on an entity's Board of Directors or advisory committees.

Recurrent somatic nonsense PHF6 mutations have been reported in patients with T-acute lymphocytic leukemia, AML and chronic myeloid leukemia in blast crisis. Germ line (GL) PHF6 mutations are responsible for Borjeson−Forssman−Lehmann syndrome (BFLS), a hereditary X-linked disorder characterized by mental retardation and dysmorphic features. PHF6 is a highly conserved 41kDa protein with ubiquitous expression in hematopoietic cells, including CD34+ cells. We screened patients (N=1166) with myeloid neoplasms by targeted multi-amplicon deep NGS targeting all ORFs of PHF6 to determine the prevalence and distribution and molecular context of PHF6 gene alterations. In total, we identified and verified 52 cases with somatic PHF6 mutations, 32 of which were frameshift or nonsense mutations and with a strong male predominance (76%). Mutations were distributed almost equally between 2 DNA binding domains. Previously, PHF6 has been included in other screening panels (Haferlach et al. 2014 and Papaemmanuil 2013) with somatic mutations found in 24/944 and 21/738 MDS cases, respectively. SNP-array karyotyping showed that microdeletions involving the PHF6 locus were present in about 1.2% of myeloid neoplasms, but affected only female patients. The most frequent chromosomal aberration observed in conjunction with PHF6 mutations was trisomy-8 (P=.018). The most commonly associated somatic mutations included RUNX1 (P=.001) and IDH1 (P=.008) but not IDH2 (P>.1). There was no impact on overall survival with respect to PHF6 mutant status in total or within individual risk groups (low risk (RA,RARS) vs. high-risk (RAEB1/2). Concomitant PHF6 and RUNX1 mutations were associated with particularly poor prognosis. RUNX1 mutational status correlated with PHF6 expression levels and PHF6 expression inversely correlated with RUNX1 mRNA levels. Subsequent analysis of clonal architecture using VAF calculations and serial samples for these cases suggested that PHF6 may function as a founder driver gene in 18% of cases. PHF6 variant allelic frequency (VAF) varied between disease subtypes, with the highest clonal burden found in AML patients (P Recent studies have proposed that PHF6 deficiency leads to impaired cell proliferation, cell cycle arrest at G2/M phase and DNA damage. Following shRNA knockdown, hematopoietic cell lines showed only moderately accelerated growth and increased response to growth factors, while EPO-dependent UT7, did not result in growth factor autonomy. To delineate the possible pathophysiological pathway involving PHF6, we compared transcriptional expression profiles of 5 different cell lines with shPHF6 to WT counterparts. We then studied the consequences of PHF6 knockdown on transcriptional profiles. We have found 1020 transcripts differentially expressed (with at least 1.5x change up/down) in the context of shPHF6 knock down. Concordant results among all 5 cell lines resulted in 354 genes that were upregulated and 766 that were down-regulated. Analyses with primary patient data derived from low PHF6 expressors and mutant cases found a concordance of 71 upregulated genes and 80 genes that were downregulated. The most significant functional group of transcripts that was found to be modulated was a family belonging to ribosomal biogenesis pathway (pFDR In conclusion, our results indicate that PHF6 mutations are generally present in more aggressive types of myeloid neoplasms, frequently associated with RUNX1/IDH1 mutations. Our functional in vitro studies along with recently published reports suggest an association of PHF6 deficiency with transcriptional regulation and thereby provide a basis for a phenotype conveyed by ancestral lesions, consistent with its role as a tumor suppressor gene. Disclosures Sekeres: Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees. Makishima:The Yasuda Medical Foundation: Research Funding.

Background Lower-risk (LR) MDS (Low/Int-1 per International Prognostic Scoring System (IPSS)) are a heterogeneous group of disorders characterized mainly by refractory anemia and transfusion dependency. As survival of this patient (pt) population is measured in years. Goals of therapy focus on decreasing blood transfusions, improving quality of life, while minimizing treatment toxicities. While achieving complete remission (CR) in higher-risk MDS correlates with improved overall survival (OS), its relationship to OS in LR MDS is not well defined. We evaluated the impact of achieving CR on OS in LR MDS and defined the clinical characteristics that may predict for this response in this pt population. Method Included pts were diagnosed with MDS (per 2008 WHO criteria) and had LR disease with clinical and pathologic data collected from MDS Clinical Research Consortium institutions. Only pts with bone marrow blasts of 5-9% who would thus qualify both as having LR MDS and for being eligible to assess CR were included. Responses (including CR, PR, HI, stable disease and progressive disease) were defined per International Working Group 2006 criteria. OS was calculated from the time of achievement of best response to time of death or last follow-up. Cox proportional hazard analysis that included all clinical variables and treatment characteristics was used to identify independent prognostic factors. Results Of 1470 pts included in the database, 999 identified with LR disease, and 174 had bone marrow blasts of 5-9%. The median age was 60 years (range, 22-87), and 37% were female. Median neutrophil count was 1.25 X 109\L (range, .10-51.0), hemoglobin was 9.8 g/dl, platelets were 109 k/ml (range, 18-562), and bone marrow blasts were 6% (range, 5-9%). Best responses to therapy included: 26 pts (15%) with CR, 10 pts (6%) with PR, and 13 pts (7%) with HI. Among pts who achieved CR/PR/HI, 27 received HMA (25 with azacidtine +/- combination and 2 with single agent decitabine), 16 intensive chemotherapy, 2 lenalidomide, and 4 received other therapies. With a median follow up from diagnosis of 31.2 months, the median time from diagnosis to best response was 11.9 months (range, .69-81.0) and was similar in pts who achieved CR compared to PR/HI (11.5 vs. 12.4 months, respectively, p = .74). The median OS from time of CR/PR/HI for the entire cohort was 21.3 months. The median OS for pts who achieved a CR was longer compared to pts with PR/HI (46.5 vs. 18.5 months, respectively, p = .03). In multivariate analyses that included clinical variables and treatment history, achieving CR remained an independent prognostic factor for longer OS (HR .32, p = .03) but no individual demographic, clinical or treatment variables were predictive of CR. Conclusions Similar to pts with higher-risk MDS, LR MDS pts who achieve CR to therapy have improved OS compared to those with PR or HI. CR is thus an important endpoint in LR MDS, though is difficult to predict. As OS is measured in years in LR MDS, CR may be used as a surrogate endpoint for OS in clinical trials in this pt population. Disclosures Komrokji: Novartis: Consultancy, Speakers Bureau; Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding. Roboz:Cellectis: Research Funding; Agios, Amgen, Amphivena, Astex, AstraZeneca, Boehringer Ingelheim, Celator, Celgene, Genoptix, Janssen, Juno, MEI Pharma, MedImmune, Novartis, Onconova, Pfizer, Roche/Genentech, Sunesis, Teva: Consultancy. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees.

Somatic mutations constitute clonal markers now amenable to monitoring by deep NGS. While transient and low frequency clones have been described in AA, their pathophysiologic link to the overtly clonal complication of AA, secondary MDS following AA (sMDS), has not been established. Clarification of this relationship may provide clues as to the genesis of sMDS. Identification of predictive markers for AA patients at risk for this complication is necessary. The etiology of sMDS within AA may include either expansion of a preexisting clone or truly late clonogenic events. In both instances, progression may result in clonal escape. To address these questions we studied 258 AA and 60 PNH patients, identifying 35 patients (11%) who evolved to sMDS. Cytogenetic analysis showed abnormal karyotype in 76% cases; 5% had complex karyotype and -7/del(7q) was present in 67% of cases. The presence of a PNH clone was detected in a similar proportion of cases that transformed to sMDS vs. those that did not (P=.76). For comparison, we have also analyzed primary de novo cases of MDS (pMDS) with (N=19) and without (N=161) -7/del(7q). In contrast to sMDS, -7/del(7q) was present in 12% of cases of pMDS. Using WES on 8 cases and a 60 gene targeted panel on 15 cases, confirmed mutational events and chromosomal aberrations were found in 21/23 patients with sMDS; 18/23 cases of sMDS had at least 1 confirmed somatic mutation. By comparison of mutational profiles ASXL1, RUNX1, PIGA, SETBP1, and CBL were most common in sMDS (26.1%, 21.7%, 18%, 13% and 13%, respectively). Because sMDS included a high proportion of patients with -7/del(7q), we compared sMDS with -7/del(7q) to pMDS with -7/del(7q) for coexisting mutational events. Mutations in RUNX1 appeared to be more frequent in sMDS vs. pMDS (27% vs. 0%, P=.03). In contrast, TP53 was more common in pMDS (7% vs. 32%, P=.1). Similarly, there were several other distinctive differences between all sMDS and pMDS irrespective of cytogenetics: mutations in U2AF1 were common in pMDS, mutations in RUNX1 appeared to be more frequent in sMDS vs. pMDS (22% vs. 5.5%, P=.02). Mutations in PIGA gene constituted a marker for sMDS derivation from AA. To discern a possible biological relationship we have also compared mutational profiles of hypocellular pMDS and sMDS, but no significant differences aside from PIGA prevalence were found. If sMDS is derived from mutations present at the AA stage, one would expect overlaps in the mutational spectrum of AA before and after evolution to MDS. DNMT3A, BCOR/BCORL1, and PDGFR family mutations were found at higher frequency in AA while ASXL1, RUNX1, SETBP1, PIGA, and CBL were higher in sMDS. Thus, cross-sectional analysis suggests that most of the clonal events occurring during the course of AA do not initiate sMDS. To further examine these findings we performed serial sequencing analyses: in 7 patients with sMDS WES was performed and clonal architecture was analyzed. We then queried whether mutations present in MDS were detected in archival samples at presentation using deep targeted NGS (depth 5-10x104 rds. In 4/7 cases the alterations appeared to be ancestral events for sMDS evolution. When an additional 68 AA cases were studied by deep NGS, somatic mutations were present in 31% of AA patients at presentation. Patients with clonal events at presentation tended toward worse progression free survival compared to patients without mutations (P=.1). Mutations found at both initial presentation and upon evolution were suggestive of a slow expansion of previously cryptic clones (ASXL1, CUX1, TET2, CBL, RUNX1, and SETBP1). Patients with these gene mutations (n=21) before immunosuppressive therapies (IST) had worseoverall survival compared to patients without these mutations (n=47; P=.009). To assess the potential impact of IST, we also investigated a subset of 37 patients (25 responders/ 12 refractory) following IST. Clonal somatic events were identified in 42 of them, but there was no association between the response to IST and somatic mutations at presentation (P=.7). Our results demonstrate that while subclonal mutations indicative of oligoclonal hematopoiesis are frequent in AA, the presence of specific permissive ancestral somatic events at the outset of AA predisposes patients to sMDS, a feature that has diagnostic and prognostic implications. Disclosures Makishima: The Yasuda Medical Foundation: Research Funding. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees. Maciejewski:Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees; Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Celgene: Consultancy, Honoraria, Speakers Bureau.

Proteasome inhibitors (PIs) capitalize on the constitutive activation of NF-KB in AML cells and increase chemosensitivity to anthracyclines and cytarabine. We combined the second generation PI, ixazomib, with the standard AML salvage regimen of MEC (mitoxantrone, etoposide, cytarabine). The primary objectives of this study were to determine the dose limiting toxicity (DLT), maximum tolerated dose (MTD), and phase 2 dose of ixazomib in combination with MEC in relapsed/ refractory (R/R) AML. Secondary objectives included evaluating the efficacy of this combination and correlating response to the gene expression profile and CD74 expression, which may identify a subset of leukemias in which NF-KB is operative with increased sensitivity to PI (Attar et al. CCR 2008; 14: 1446-54). Methods: Patients (pts) were treated at Cleveland Clinic and University Hospitals of Cleveland from Oct 2014 to present. An IND was approved by the FDA, and the protocol was approved by each institutional review board. Eligibility: age 18-70 yrs, R/R AML, and cardiac ejection fraction ≥ 45%. The fraction of blasts positive for CD74 was assessed by flow cytometry. Samples were stored for gene expression profiling pre- and post-treatment (at the time of response assessment). Pts received MEC: mitoxantrone (8 mg/ m2), etoposide (80 mg/m2), and cytarabine (1000 mg/m2) intravenous (IV) Days 1-6. Ixazomib, provided by Takeda, was given orally on Days 1, 4, 8, and 11 and was dose escalated using a standard 3x3 design. Dose levels (DLs): 1 (1.0 mg), 2 (2.0 mg), 3 (3.0 mg), 4 (3.7 mg). An additional 18 pts were to be treated at the MTD. One cycle of treatment was administered. Response was assessed by bone marrow aspirate/ biopsy by Day 45 and complete remission (CR) was defined by IWG criteria (Cheson 2006). Toxicities were graded according to NCI CTCAE v 4.03. Toxicities secondary to neutropenia or sepsis were not considered DLTs. DLTs included: (1) ≥ Grade 4 non-hematologic toxicity (NHT) with the exception of nausea, vomiting/ alopecia and drug-related fevers; (2) any ≥ Grade 3 neurologic toxicity; (3) grade 4 platelet or neutrophil count 50 days beyond the start of chemotherapy and not related to leukemia; (4) any Grade 4 NHT > grade 2 by 45 days beyond the start of chemotherapy. Grade 2, 3, and 4 hyperbilirubinemia were redefined as 1.5-< 10x upper limits of normal (ULN), 10-20 x ULN, and > 20 x ULN. Results: Of 23 pts enrolled, 22 are evaluable. The median age was 58 yrs (range 31-70), 12 (52%) were male and the median baseline WBC was 2.56 K/ uL (range 0.1-62.9). The median time from initial diagnosis to registration was 7.1 months (range 1.4-36.8) and 7 pts (30%) had a history of an antecedent hematologic disorder. Thirteen pts were in 1st relapse and 10 pts were refractory to their last therapy. One pt had received a prior allogeneic hematopoietic cell transplant (AHCT), 7 pts had FLT3 ITD mutations and 7/ 21 pts (33%) had adverse cytogenetics per CALGB 8461 criteria at the time of relapse. At DL1, 1 DLT occurred (grade 4 thrombocytopenia), so this DL was expanded to 6 pts. At DL2, 2 pts developed Grade 4 thrombocytopenia; therefore, the MTD of ixazomib was 1.0 mg. The most common grade 3-5 NHTs in the dose escalation phase were febrile neutropenia (100%), hypoalbuminemia (25%), hypokalemia (42%), hypotension (33%), and respiratory failure (33%). No adverse events in the dose escalation phase were attributed to ixazomib alone. The overall response rate was 55% [CR/ CR with incomplete count recovery (CRi)], and 9 pts proceeded to AHCT. Five of these 9 pts remain alive with a median follow-up of 12.8 months. Five pts had CD74 expression performed. Two pts had high levels of CD74 expression (> 80%); and both achieved CRi. Myeloid mutation panel data was available in 14 pts. Previous data has demonstrated the number of mutations in DNTMT3A, TP53, ASXL1, and NRAS (0, 1, >1) is associated with a worse response to salvage therapy (Advani et al, abstract 3825, ASH 2015). Seven pts had at least one of these mutations and 6 of the 7 achieved CR/ CRi. Conclusions: The combination of MEC and ixazomib was well-tolerated and produced an overall response rate of 55% in patients with relapsed/ refractory AML irrespective of molecular mutation status. The combination is safe with a similar toxicity profile to MEC alone. CD74 expression may represent a biomarker for response to this therapy. Results from gene expression profiling will be complete by the time of the meeting and will be presented. Disclosures Mukherjee: Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding; Celgene: Consultancy, Honoraria, Research Funding. Caimi:Genentech: Speakers Bureau; Gilead: Consultancy; Roche: Research Funding; Novartis: Consultancy. Maciejewski:Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Honoraria, Speakers Bureau. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees.

Introduction The major reasons for failure to achieve cure in the majority of AML patients (pts) are primary refractoriness of disease to initial chemotherapy or failure to maintain complete remission (CR) that has been achieved (relapse). There is no uniformly accepted standard treatment for relapsed or refractory (RR) AML, with most available therapies regarded as palliative or as a bridge to allogeneic transplantation. While the past two decades have witnessed trials of several investigational therapies in RR AML, data regarding the effectiveness of these interventions remains unclear. We studied the impact of experimental drugs in RR AML pts by undertaking a comprehensive analysis of all phase 2 and 3 randomized clinical trials (RCTs) reported in the past 3 decades. Methods We searched PubMed, Embase, Cochrane Controlled Trials Register electronic databases, ClinicalTrials.gov and conference abstracts from the American Society of Hematology (ASH), American Society of Clinical Oncology (ASCO) and European Hematology Association (EHA) websites covering a period from 1988 to 2015. Key words used during this search included "refractory" or "relapsed" or "AML" or "phase II" or "phase III" or "randomized". Only double-arm, phase II with a sample size of at least 50 pts and phase III RCTs conducted in RR AML pts were included. Two reviewers independently extracted data on study methods, participants, therapies, and outcomes from all eligible trials: differences in how to classify agents in RCTs were resolved by discussion. The primary outcomes examined in the experimental arms (EAs) and standard arms (SAs) included CR rates, disease-free survival (DFS), refractory disease rates, treatment-related mortality (TRM) rates and overall survival (OS). Odds ratios (OR) were used to summarize differences between EAs and SAs. The DerSimonian and Laird random-effects model was used to compare them and to assess the overall impact of time. Results Of 5500 included pts, 40.5% were treated on 21 double-arm, phase II trials, 51% on 10 phase III trials and 6.6% analyzed through 4 retrospective studies. There was no change in CR rates in either EAs (p=.21) or SAs (p=.15) over time (Figure 1). The CR rates in EAs tended to be higher than in SAs [OR=1.24; 95% CI, 1.02-1.50, p=.03). Rates of disease refractoriness to salvage regimens in both EAs (p=.70) and CAs (p=.31) did not change over time and these rates were not significantly different between treatment arms [OR=0.82; 95% CI, 0.62-1.08, p=.16]. TRM rates tended to decrease over time but the change was not significant in either group [p=.24 for SAs and p=.33 for EAs]. TRM rates were higher in SAs compared to CAs but did not reach statistical significance [OR=1.21; 95% CI, 0.97-1.50, p=.09]. Over time, there was no significant change inDFS in either group (p=.32 for CAs and p=.58 for EAs). DFS rates did not differ between EAs and SAs [OR=1.01; 95% CI, 0.86-1.19, p=.89] (Figure 2). OS tended to remain stable over time in both groups [p=.85 for SAs and p=.66 for EAs]. While OS tended to be higher in SAs, it did not reach statistical significance [OR=0.93; 95% CI, 0.83-1.05, p=.27]. Conclusions: These findings indicate a lack of significant or clinically meaningful improvement in disease outcomes, including OS, in RR AML pts treated within RCTs over the past 3 decades. Greater efforts need to be directed towards designing RCTs using novel statistical approaches and directed agents based on recent discoveries of targetable mutations. Disclosures Carraway: Amgen: Membership on an entity's Board of Directors or advisory committees; Baxalta: Speakers Bureau; Celgene Corporation: Research Funding, Speakers Bureau; Novartis: Membership on an entity's Board of Directors or advisory committees; Incyte: Membership on an entity's Board of Directors or advisory committees. Advani:Pfizer Inc.: Consultancy, Research Funding; Blinatumomab: Research Funding. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees. Mukherjee:Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding; Celgene: Consultancy, Honoraria, Research Funding.

Several population-based cancer registries have consistently reported improvements in overall survival (OS) for patients (pts) with AML treated with intensive chemotherapy (IC) across all age groups extending up to the age of 80 years. While major advances in supportive care and greater numbers of allogeneic transplantations have contributed to improvements in OS in AML, the role of experimental IC regimens has not been comprehensively analyzed. We conducted a systematic review of all phase three randomized clinical trials (RCTs) in AML conducted over the past forty years to evaluate the contribution of experimental IC regimens in clinical outcomes, along with OS in these pts based on the Surveillance, Epidemiology, and End Results (SEER) database. The following electronic databases were searched for eligible non-acute promyelocytic leukemia trials covering the period from 1973 to 2016: Cochrane Controlled Trials Register, PubMed, Embase databases and American Society of Hematology (ASH), American Society of Clinical Oncology (ASCO) and European Hematology Association (EHA) websites for conference abstracts. Analyses were restricted to phase three RCTs using IC regimens. The principal outcomes examined in the experimental arms (EAs) and control arms (CAs) included complete remission (CR) rates, leukemia-free survival (LFS), and OS. Outcomes were stratified by age and cytogenetic risk, in studies where information was available. RCTs were classified as including younger or older AML pts based on whether >65% of the patients in a study were < 60 or ≥ 60 years (yrs) old, respectively. We also evaluated the outcomes of pts diagnosed with AML outside of clinical trials by analyzing survival trends of 61,000 AML pts in SEER registry. A total of 71 studies comprising 36,796 AML pts were included. The majority of pts in RCTs were < 60 years (68.6%); 54.3% were male. Eligibility criteria for study entry across all RCTs included a mean ECOG performance status of 2-3, mean upper limit (UL) of serum creatinine of 2.3 mg/dl and mean UL of serum bilirubin of 2.6 mg/dl. Mean CR (stratified by age and cytogenetics), mean LFS, OS and treatment-related mortality were similar between EA and CA cohorts (Table 1). Over the study period, there was no trend towards improved CR in EAs and CAs in both younger and older AML trials or even when stratified by cytogenetic risk groups. Similarly, no significant change in LFS trend was observed over time in the EAs and CAs in both younger and older pts. However, over time EAs in younger and older AML trials had a significant increase in median OS, at a rate of 29.52 days/yr (d/yr) (6.84-51.84; P=.01), and 13.68 d/yr (1.8-25.92; P=.05), respectively (Figure 1). For those in CAs, the corresponding rates for younger and older pts was 32.76 d/yr (8.64-56.88; P=.009) and 12.24 d/yr (2.88-27.93; P=.10), respectively. By comparison, there was a significant trend towards improvement in median OS over time in the SEER cohort for younger pts, defined as < 60 yrs (18.36 d/yr; 30-212.4, P This systematic review of evidence from phase three RCTs in AML suggests a minimal improvement in OS with use of experimental IC regimens, with gains mostly seen in younger pts. No significant improvements in LFS trends were observed in any age group, demonstrating a lack of aggregate clinical efficacy of experimental drugs over this time period. Greater survival gains in OS in pts enrolled in CAs of RCTs as compared to the SEER population suggests a strong selection bias in RCTs, thereby limiting extrapolation of study findings to the real world population. Disclosures Carraway: Celgene Corporation: Research Funding, Speakers Bureau; Baxalta: Speakers Bureau; Novartis: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Incyte: Membership on an entity's Board of Directors or advisory committees.

Acute undifferentiated leukemia (AUL) and mixed phenotype acute leukemia (MPAL) are acute leukemias of ambiguous lineage representing less than 5% of all adult leukemias. Ontogenetically, they are thought to be caused by precursor lesions in early pluripotent hematopoietic progenitor cells having the ability to differentiate into both lymphoid and myeloid lineages. In contrast, common lymphoid and myeloid leukemias are likely derived from precursor lesions in committed progenitor cells. The molecular characteristics of the cell of origin of these leukemias have not been fully characterized. Given the ambiguous classification of these disorders, we used NGS to evaluate the molecular profiles of this unique subset of leukemias. Our panel tested for the 60 most commonly mutated genes in myeloid malignancies (MM), which can be used to further elucidate driver mutations that play a role in the pathogenesis of this unique entity. Molecular data was available for 18 of 35 patients (pts) (51%), of whom 8 were T/myeloid, 7 were B/myeloid, and 3 were AUL. 14 pts out of the 18 had samples at the time of initial diagnosis, 3 AUL cases harbored mutations commonly seen in MM, including one patient with DNMT3A and a germ line JAK3 mutation, another with U2AF1 and PHF6 mutations, and one pt with SF3B1 and MECOM mutations. Of the 5 B/myeloid cases, 4 had only one mutation detected by the sequencing panel, including TET2, PTPN11, KDM6A, and PHF6, and one had no mutations, but did have complex cytogenetics. Of interest, we identified four patients with T/myeloid MPAL that harbored a FLT3-ITD mutation. Notably, one of these cases had additional DNMT3A and TET2 mutations while another case had a WT1 mutation. We assert that the mutational data for MPAL supports that T-myeloid MPAL encompasses a borderland between early T precursor (ETP)-ALL/AML in addition to more typical cases of T-ALL/AML given its similar immunophenotype to ETP ALL (CD117, CD13 and CD33 positive). The overlap between these two entities is illustrated in the figure below. With this immunophenotyping and molecular profiling we propose a new provisional entity, ETP-myeloid MPAL in addition to already existing ETP ALL. Of 1250 pts with acute leukemia diagnosed from 1997-2016 at Cleveland Clinic, 35 had immunophenotypic characteristics of ambiguous leukemia based on WHO criteria. The median age at diagnosis was 50 years, and 22 (63%) were male. There were 31 MPAL (16 T/myeloid, 15 B/myeloid) and 4 AUL cases. Among the 31 MPAL patients, 2 were mixed lineage chimera (subclones), and 29 had true biphenotypic features. 48% of the cohort had normal cytogenetics at diagnosis, 23% had other cytogenetic abnormalities, 14% had complex karyotype, 9% had MLL gene rearrangement, and 6% had t(9;22). 62% of pts were treated with acute lymphoblastic leukemia (ALL) directed induction chemotherapy (IC), while 38% were treated with acute myelogenous leukemia (AML) directed IC. Rates of complete remission (CR) were significantly higher in pats treated with ALL-IC regimen (82%) versus AML-IC regimen (23%) (p=.003). In our study, two cases with the recurrent lymphoid mutations WT1 and KDM6A, that achieved CR with upfront ALL-IC, while one case with FLT3-ITD, DNMT3A, and TET2 mutations achieved CR with AML-IC treatment. Pts In conclusion, clinical evaluations of ambiguous leukemias demonstrate that those with normal karyotype have the worst outcome of all pts diagnosed with MPAL. This can be partly explained by mutations that are detected at the time of diagnosis such as FLT3-ITD and DNMT3A, which have been associated with poor survival in AML with normal karyotype. Identification of actionable targets like FLT3-ITD can improve clinical outcomes for this subset of patients prior to BMT or for those ineligible for BMT. The existence of a unique subset of pts with immunophenotypic overlap between ETP-ALL/AML, that were classified in routine practice as T/myeloid leukemia, are important to recognize due to their unique actionable phenotype. Molecular profiling of these cases identified gene mutations more akin to MM. We propose that sequencing may play an important role in optimizing the therapeutic approach for patients with ambiguous lineage leukemia. Figure Figure. Disclosures Carraway: Novartis: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Incyte: Membership on an entity's Board of Directors or advisory committees; Baxalta: Speakers Bureau; Celgene: Research Funding, Speakers Bureau.

DNMT3A, a member of the DNA methyltransferases family along with DNMT1 and DNMT3B, is located on chromosome 2p23. Recurrent somatic mutations in DNMT3A are typically heterozygous and found mostly in non-CBF AML, less frequently in MDS and MPN. DNMT3A mutations are reported with other common myeloid mutations including NPM1, FLT3 and IDH1/2. The most canonical DNMT3A mutations are missense alteration in the R882 codon, accounting for >60% of all DNMT3A mutations and they imply dominant negative consequences. Overall, DNMT3A mutations carry a poor prognosis compared to the AML or MDS with wild type (WT) DNMT3A, although data within different subgroups (e.g., incorporating cytogenetic profiles) are conflicting. We hypothesized that molecular consequence of R882 mutations will differ from those of other somatic alterations of DNMT3A and may also result in distinct clinical features and outcomes. To test this theory, we analyzed a cohort of 1174 patients with myeloid neoplasias including 32% AML, 33% MDS, 13% MDS/MPN, 6% MPN and 16% other bone marrow failure disorders. These cases were subjected to multiamplicon targeted deep NGS including all ORFs of DNMT3A and other recurrently mutated genes. After application of various bioanalytic algorithms, confirmatory sequencing and thus stringent exclusion of all artifacts and germline alterations, we identified 140 somatic mutant cases (12% of the cohort), including 89 missense mutations (53 at R882, 19 at R693 and 17 other non-canonical missense alterations) and 51 truncations/frame shifts (all heterozygous). There was an age-related increase in the incidence of DNMT3A mutations, with the peak occurrence at 35-40 yrs. of age. Mutations in DNMT3A were most common in AML (54% in primary (p) AML, 8% in secondary (s) AML) followed by MDS (28%), MDS/MPN (4%), MPN (3%) and other bone marrow failure disorders (3%). Mutation in the R693 codon and truncating mutations were most commonly associated with MDS (p=.013) and sAML (p=.0013) whereas mutation occurring in codon R882 and other non-canonical missense mutations were frequently associated with pAML (p=.00001). For the whole cohort, DNMT3A mutations were most frequently associated with NPM1 (21% vs 8%, p=.014), FLT3 (24% vs. 2%, p=.0001), and IDH1/2 (26% vs. 8%, p=.001), compared to wild type DNMT3A. However, PRC2 complex mutations were less likely to occur in the context of DNMT3A mutations (6% vs. 24%, p=.0006). Canonical R882 mutation was commonly associated with FLT3 (p=.03) mutations, while truncating mutations were not (p=.03). Analyses of clonal hierarchy by ranking of VAF values demonstrated that 53% of DNMT3A mutations were dominant (mean VAF 39%, range 5-93%) (n=74/140). When DNMT3A mutations were dominant, IDH 1/2 (14%), TET2 (9%), ASXL (5%), PRC2 complex (3%) and BCOR (3%) mutations were common secondary events. In subgroup analyses, 55% of mutations in the R693 codon were dominant compared to 45% in R882 and 47% in truncating mutations. TET2 mutations were the most common associated secondary hits in dominant R693 mutations (n=10) compared to truncating (n=24) and R882 mutations (n=23) (40% vs. 8% vs. none, p=.0001). When DNMT3A mutations are secondary (mean VAF 34%, range 1-60%), as in 47% of our cases (n=66/140), then the common first hits were TET2 (10%), U2AF1 (8%) and cohesin complex (RAD21, SMC3, STAG2) mutations (6%). Dominant DNMT3A mutations correlated with MDS/MPN (60%, p=.007), while secondary DNMT3A mutations correlated with sAML (73%, p=.001). DNMT3A mutant myeloid neoplasms showed worse survival (p DNTMT3A mutations often occur as founder lesion in AML. Our study shows that different types of mutations other than canonical R882 alterations may have a differential impact on OS and distinct clinical features. Disclosures Carraway: Celgene Corporation: Research Funding, Speakers Bureau; Novartis: Membership on an entity's Board of Directors or advisory committees; Baxalta: Speakers Bureau; Amgen: Membership on an entity's Board of Directors or advisory committees; Incyte: Membership on an entity's Board of Directors or advisory committees.

Introduction: Approximately 50% of patients with MDS are anemic at initial diagnosis, with many becoming transfusion dependent (TD), necessitating the introduction of therapy with a goal of achieving transfusion independence (TI). In addition to transfusions, treatments to improve hemoglobin levels have historically been limited to erythropoiesis-stimulating agents (ESAs), whereas patients with more advanced disease may require treatment with a hypomethylating agent (HMA) (azacitidine [AZA] or decitabine [DAC]) or lenalidomide (LEN). As there is no predetermined treatment course, these therapies are continued until unacceptable toxicity, lack/loss of response, or disease progression. We examined treatment patterns, clinical outcome(s) of patients with MDS, and the physician's report of reason for treatment discontinuation (previously shown to vary from patients' perspectives) in patients who became TD at or after MDS diagnosis (Gerds et al. Blood 2014;124:abstract 2642). Methods:Data were derived from disease-specific physician surveys and patient charts, which provided information on demographic, treatment, and outcome data on lower-risk TD MDS patients. Patient inclusion criteria were: age ≥ 18 years, diagnosis of Low/Intermediate-risk MDS (International Prognostic Scoring System defined) 2-6 years prior to study entry, and becoming TD at least once during the minimum follow-up of 24 months. TD was defined as having received ≥ 2 transfusions within 8 weeks during the follow-up period. Patients who progressed to higher-risk disease or acute myeloid leukemia prior to becoming TD, had an additional malignancy, or were in an MDS clinical trial were excluded. Demographics, disease history, treatment history, TI, and reasons for treatment discontinuation were collected and reported descriptively. Results: A total of 239 physicians provided information on 1,221 lower-risk TD MDS patients. The median age of patients was 65 years (range 27-95 years), 56.3% were male, and median time since diagnosis was 3.2 years (range 2-6 years). Along with packed red blood cell transfusions, 354 patients (29%) were prescribed LEN [of whom 12.7% had del(5q)], 348 patients (28.5%) were prescribed ESAs, and 32 patients (2.6%) were prescribed HMAs. During the follow-up period, 31.3% of ESA patients discontinued therapy (at a median of 12 months; range 3-47 months), 32% of HMA patients discontinued therapy (at a median of 10 months; range 3-52 months), and 26% of LEN patients discontinued therapy (at a median of 13 months; range 2-30 months). In the LEN-treated group, only 9.8% of patients who discontinued therapy had del(5q). The main reason cited for treatment discontinuation across therapy groups reported by physicians was "patients completing the scheduled course of treatment" (28.2%), which occurred in 32.6%, 20.2%, and 22.6% of the LEN, ESA, and HMA groups, respectively. Other reported reasons for discontinuation included "insufficient initial response" (20.0%), "patients no longer responding to therapy" (19.0%), "disease progression" (18.0%), "death" (13.3%), and "worsening hemoglobin levels" (12.8%). Across therapy groups, "patient preference to stop therapy" was reported less frequently (9.7%), occurring in 7.6%, 8.3%, and 12.9% of the LEN, ESA, and HMA groups, respectively. Conclusions:Over 30% of TD lower-risk MDS patients receiving LEN, ESAs, AZA, or DAC discontinued therapy. Per physician reports, the most frequent reason for discontinuation of therapy was completion of scheduled treatment course, which is in stark contrast to recommendations in consensus guidelines, as MDS is a chronic disease in which treatment should be continued ad infinitum. The significant number of patients who stop therapy for alternative reasons suggests opportunities for further investigation. Education on the expected duration of therapy is essential to help support physicians and inform patients about optimal treatment decisions in the care of all MDS patients. Disclosures Gupta: Kantar Health: Employment; Celgene Corporation: Consultancy, Research Funding. Binder:Celgene Corporation: Employment, Equity Ownership. Carraway:Incyte: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Research Funding, Speakers Bureau; Baxalta: Speakers Bureau; Amgen: Membership on an entity's Board of Directors or advisory committees. Hawthorne:Kantar Health: Employment; Celgene Corporation: Consultancy, Research Funding. King-Concialdi:Kantar Health: Employment; Celgene Corporation: Consultancy, Research Funding. McGuire:Celgene Corporation: Employment, Equity Ownership.

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 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 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 (p2.5 x 109/L and absolute monocyte count >10 x109/L were uniquely predictive for shorter survival in the MP-CMML subcategory (p 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.

Background: While treatment with the hypomethylating agents (HMAs) azacitidine (AZA) and decitabine (DAC) improves cytopenias and prolongs survival in MDS patients (pts), only 30-40% of pts respond. Genomic and/or clinical models that can predict which pts will respond could prevent prolonged exposure to ineffective therapy, avoid toxicities and decrease unnecessary treatment costs. Machine learning (ML), a field of artificial intelligence, is an advanced computational analysis of complex data sets that can overcome some of the limitations of standard statistical methods. ML uses computational algorithms to automatically extract hidden information from a dataset by learning from relationships, patterns, and trends in the data. Thus, ML can produce powerful, reliable and reproducible predictive models based on large and complex datasets. The aim of this project is to build a geno-clinical model that uses ML algorthims to predict responses to HMAs. Methods: We screened a cohort of 433 pts with MDS who received HMAs at multiple academic institutions for the presence of common myeloid somatic mutations in 29 genes. Responses were assessed per International Working Group 2006 criteria. Five popular supervised classification ML algorithms including: random forest (RF), tree ensemble (TE), naive bayes (NB), decision tree (DT), and support vector machine (SVM) algorithms were used individually and in combination to enhance the accuracy of the proposed model (bag of model approach). For each iteration, the dataset was divided randomly into training and validation cohorts. The partition of the dataset was repeated multiple times randomly to minimize biases in pt selection. A 10-fold cross validation was also used on the entire dataset to assure data reproducibility. Important variables were selected using backward feature elimination and tree depth scores. Performance was evaluated according to the area under curve (AUC) and accuracy matrix. All analyses were done using KNIME (an open analytic platform for ML). Results: Among 433 pts, 193 (45%) received AZA, 176 (40%) DAC, and 64 (14%) received HMA +/- combination. The median age was 70 years (range, 31-100) and 28% were females. Responses included: 95 (58%) complete remission (CR), 14 (3%) marrow CR, 16 (4%) partial remission (PR), and 59 (14%) hematologic improvement (HI). For the purpose of this analysis, pts with CR/PR/HI were considered as responders. The most commonly mutated genes were: ASXL1 (31%), TET2 (22%), SRSF2 (17%), RUNX1 (15%), and DNMT3A (14%). In univariate analyses, no single mutation was more prevalent in responders compared to non-responders except NF1 (more common in non-responders, p = .04). A logistic regression multivariate analysis did not produce a reliable and reproducible model. When applying ML algorithms on learner (80% randomly selected pts) and predictor cohorts, the accuracy rate in predicting responses for RF was 64%, for TE 60%, for NB 60%, for DT 66%, and for SVM 51%. When results from each model were combined (a bag of models approach), the accuracy increased to 69%. Backward feature elimination and tree depth scores identified the following factors as predictors of response: hemoglobin 69 years, TP53 with variant allelic frequency (VAF) >15%, CBL VAF >30%, and RUNX1 VAF > 25%. Only ASXL1mutations at any VAF were predictive of HMA resistance. Interestingly, none of the mutations were selected for response or resistance when the models did not include VAF. Neither treatment modality with azacitidine vs. decitabine vs. combination nor treatment center impacted response. When the analysis was restricted to pts with higher-risk disease by IPSS, the accuracy rate in predicting responses improved: for RF it became 71%, for TE 65%, for NB 60%, for DT 64%, and for SVM 76%. When the analysis was focused on pts who achieved CR vs. No CR, the models predicted the response differently. The RF and TE models were able to predict No CR with an accuracy rate of 75% and 76% respectively. Other models were able to predict CR and No CR with lower accuracy. Conclusion: We propose a novel geno-clinical model that uses machine intelligence to predict HMA response/resistance in pts with MDS. The model has a higher accuracy rate in higher-risk MDS pts. ML can open opportunities in translating genomic data into reliable predictive models that can aid physicians in clinical decision making. Disclosures Bejar: Celgene: Consultancy, Honoraria; Foundation Medicine: Consultancy; Genoptix: Consultancy, Honoraria, Patents & Royalties: No royalties.

Objectives: TP53, a tumor suppressor gene, is frequently mutated in myeloid malignancies. The negative impact of TP53 mutations in myelodysplastic syndromes (MDS) has been described previously but there is controversy regarding the prognostic impact of the mutation's characteristics (location, type, passenger vs. driver, and others). Methods: We sequenced DNA samples from 732 patients (pts) with MDS and related myeloid malignancies for the presence of TP53 mutations and 61 other genes that have been described as recurrently mutated in MDS. Overall survival (OS) was measured from the time of diagnosis to time of death or last follows up. Variant allele frequencies (VAFs) adjusted by zygosity were used to define clonal architecture of driver clones. Results: Of 80 mutations detected in 73 (10%) pts, 66 (88%) were missense, 7 (9%) were nonsense, and 7 (9%) were frame shift deletions/insertions. Pts with TP53 mutations had a higher WBC (4.6 vs. 3.9 X 109/L, p = .04), higher bone marrow blast % (median 9 vs. 3, p =0.1), and a higher risk category by revised International prognostic Scoring System (56% vs. 27%, p =.01) compared to pts with TP53 wild type. TP53 mutations were commonly occurred with TET2 (16%), PRPF8 (13%), ASXL1 (11%), DDX54 (8%), DNMT3A (8%), and IDH2 (8%). The mean VAF for TP53 was 41.9 (5-100). TP53 mutations were defined as drivers in 20% of samples, passengers in 40%, and mosaic in 40%. Mutation positions included: 19 (24%) in the DNA binding domain, 2 (3%) in the transactivation domain, 1 (1%) in the tetramerization domain and 58 (72%) other. With a median follow up of 16.4 months, the median OS for the entire group was 8.24 months. Patients with TP53 as driver mutations had a worse OS compared with patients with TP53 as passenger mutations (median, 2.2 vs. 13.0 months, respectively, p =.02). Similarly, OS by TP53 VAF categorized as low (50%) was 12.4, 8.5, and 3.4 months, respectively. Among patients with available treatment data, 29 patients were treated with the hypomethylating agents, azacitidine or decitabine, 17% responded (60% CR, and 40% PR), and OS was similar compared to patients treated with other therapies. Patients treated with Hematopoietic stem cell transplant (HCT) had superior OS compared to pts not receiving HCT (median, 14.9 vs. 8.9 months, respectively, p =.05). Conclusions: TP53 mutations are associated with poor outcome in MDS, but the outcome varies depending on the type of mutation and VAF. Treatment with HCT remains a valid treatment option in a subset of patients but novel treatment strategies are desperately needed. Disclosures Mukherjee: Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding; Celgene: Consultancy, Honoraria, Research Funding.

Background Several prognostic models have been developed to risk stratify patients (pts) with MDS including: the International Prognostic Scoring System (IPSS), Revised IPSS (IPSS-R), World Health Organization classification-based Prognostic Scoring System (WPSS), and MD Anderson Prognostic Scoring System (MDPSS). All except for the MDPSS were developed in treatment-naive pts and, if validated in treated pts, were done so primarily in those receiving one line of therapy. Incorporation of molecular data into the IPSS-R improves its predictive power, and adding molecular data to the IPSS can upstage or downstage some pts. In this study, we compared the prognostic utility of each prognostic model, after adding molecular data, in treated MDS pts. Method Clinical and mutational data from MDS pts diagnosed between 1/2000-1/2013 were analyzed. A panel of 60 gene mutations that were described as commonly mutated in myeloid malignancies was included. Patients who underwent hematopoietic cell transplant (HCT) were censored at the time of transplant. Univariable and multivariable analyses on the training cohort were performed by applying Cox proportional hazards regression analyses that included age, model score, and molecular mutations with an outcome of overall survival (OS). All molecular models were then applied to the validation cohort. The fit of the proposed models to the data was assessed by using Akaike's information criterion (AIC, lower values imply better fit) and concordance (c-) index. Results A total of 610 pts were included and divided into two cohorts, training (404 pts), and validation (206 pts). Median age of the training cohort was 67 years (range, 19-88); 83 pts (20%) had MDS/MPN including chronic myelomonocytic leukemia (CMML). Pts received a median of 2 lines of therapies (range, 0-7) and 15% of pts underwent HCT. First line therapies included: supportive care (22%), growth factors (22%), azacitidine +/- combinations (30%), decitabine +/- combinations (6%), lenalidomide (5%), induction chemotherapy (3%), immunosuppressive therapy (3%), and other therapies /clinical trials (9%). The median OS in the training and validation cohorts per scoring system (SS), i.e. IPSS, WPSS, MDPSS, and IPSS-R, is summarized in Table 1. The most common mutations in the training cohort were: TET2 (19%), ASXL1 (16%), SF3B1 (13%), DNMT3A (9%), STAG2 (9%), RUNX1 (8%), and U2AF1 (7%). In univariate analyses, mutations in EZH2 (HR1.8, p = .02), TP53 (HR2.3, p < .001), RUNX1 (HR 1.5, p = .05), and NPM1 (HR1.49, p = .001) had a negative impact on OS while SF3B1 (HR .34, p < .001) mutations were associated with favorable outcome. In multivariate analyses that included SS, mutations (from the list above), and age (except for the MDPSS, which already includes age), the following independent prognostic factors for OS were identified: age, EZH2, SF3B1, TP53, and each SS. Based on the fitted coefficients of each prognostic factor, a molecular version of each model including IPSSm, WPSSm, MDPSSm, and IPSS-Rm was proposed with median OS in the training and validation cohorts as summarized in Table 1. The addition of molecular data improved (reduced) the AIC and (raised) the C-index for the IPSS (2332.6, .69 vs. 2365.8, .64), WPSS (2322.3, .68 vs. 2356.2, .65), MDPSS (2308.4, .71 vs. 2323.4, .69), and IPSS-R (2305.2, .70 vs. 2330.5, .66), respectively. Further, the addition of molecular data to the IPSS upstaged 37% of pts from lower- to higher-risk disease and downstaged 5% of intermediate-1 to low risk disease. In the WPSS, it upstaged 21% of pts and downstaged 24%; in the MDPSS it upstaged 19% and dowstaged 22% of pts from intermediate-1 to low risk; and in the IPSS-R it upstaged 26% to higher-risk disease and 59% of pts with intermediate risk to a higher risk category. Conclusions The addition of molecular data to established MDS prognostic models can improve their predictive power, even in treated MDS patients. More importantly, adding molecular information can upstage or downstage pts into different risk categories. This study highlights the importance of incorporating molecular data into clinical prognostic systems. Disclosures Mukherjee: Celgene: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding.

Background: U2AF1 forms a heterodimer for the recognition of the 3' splice site during pre-mRNA splicing. Somatic U2AF1 mutations are present in approximately 10% of MDS patients. Most U2AF1 mutations are recurrent at 2 highly conserved hotspots, while non-canonical mutations are rare. U2AF1S34 and U2AF1Q157 mutations map within the zinc finger domains of the protein, resulting in distinct downstream effects. We have previously shown that U2AF1Q157 mutant patients have distinct splicing patterns compared to U2AF1WT with a set of misspliced targeted genes, including ARID2 and EZH2. In contrast, recent work focusing on S34 suggests a distinct subset of misspliced genes, including ATG7 (Park SM, Molecular Cell, 2016). The biological and clinical implications of these 2 distinct mutations are unknown. We investigated the differences between these mutations with respect to clinical outcomes and molecular background, including their impact on clonal architecture. Methods: We first collected molecular and clinical data on a cohort of 1700 patients with myeloid neoplasms (median follow up 1.0 year, range 1-5 years), median age was 65 years (range, 11-93). Targeted deep sequencing of a panel of frequently mutated genes (64) was applied. Our analyses included somatic mutational patterns, clonal hierarchy, and mutational correlation of the cohort of patients with U2AF1S34 and U2AF1Q157 and those without mutations in this gene. U2AF1 mutations were found in 5% (78/1700) of patients, all of them were missense and in a heterozygous configuration. Results: Both mutations were equally distributed in the cohort: U2AF1S34 (45%, 35/78), and U2AF1Q157 (46%, 36/78). Other mutations (Q84, E124, E152, and R156) were detected at a lower frequency (9%). We then dissected the clonal hierarchy of both U2AF1 mutations and found that 44% (34/77) were ancestral while 56% (43/77) were secondary. In MDS, most U2AF1 mutations (77%, P=.002) were dominant, while subclonal U2AF1 mutations were evenly distributed between the subentities. U2AF1S34 or U2AF1Q157 were equally likely to be dominant (21% vs. 27%; ancestral events; P=.09, respectively). Similarly, S34 and Q157 mutant clones had similar median variant allele frequencies (3-52% vs. 8-64%). U2AF1 S34 mutant cases had similar OS to patients carrying U2AF1Q157 (N=35 vs. N=36; 10 vs. 15 months; P=.209; LogR=.65). When we compared the impact of ancestral vs. secondary U2AF1S34 and U2AF1Q157 we found that MDS patients carrying ancestral U2AF1 mutations had a shorter OS compared to MDS patients carrying secondary U2AF1 patients (N=26 vs. N=18; 13 vs. 34 months; LogR=.04). Of note, ancestral U2AF1S34 patients had shorter OS compared to ancestral U2AF1Q157 patients (13 vs.11; 10 vs.15 months; P=.03; LogR=.86). Given these differences, we also investigated the mutational spectrum of U2AF1MUT patients. Cross sectional analysis identified that the top genes mutated in the U2AF1 mutant cohort were: ASXL1 (26%), BCOR/L1 (15%), TET2 (13%), DNMT3A and PHF6 (12%), ETV6 (10%), RUNX1 and STAG2 (9%), and SETBP1 (8%). Transcriptional factor and DNA-methylation genes were predominantly mutated in U2AF1MUT patients (35% and 24%, respectively). Exploring the association between S34/Q157 vs. other gene mutations, S34 co-occurred with BCOR/L1 mutations (P=.007, 24%), while Q157 mutations co-occurred with ASXL1 (P=.003, 44%) irrespective of their rank in the clonal hierarchy. When S34 was the dominant mutation, secondary mutations included ETV6, BCOR, and CUX1. In contrast, when Q157 was the ancestral event, secondary mutations included ASXL1 and DNMT3A. Subclonal S34 occurred in the context of ancestral RUNX1, BCOR/L1, CUX1 and DNMT3A, while subclonal Q157 followed ancestral ASXL1, EZH2, PHF6 and TET2. Conclusion: In sum, U2AF1S34 and U2AF1Q157, consistent with their differential missplicing consequences, create a distinct molecular milieu leading to differences in clinical outcomes. Disclosures Makishima: The Yasuda Medical Foundation: Research Funding. Carraway:Novartis: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Research Funding, Speakers Bureau; Baxalta: Speakers Bureau; Amgen: Membership on an entity's Board of Directors or advisory committees; Incyte: Membership on an entity's Board of Directors or advisory committees. Maciejewski:Celgene: Consultancy, Honoraria, Speakers Bureau; Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees.

Background: Cytogenetic abnormalities have been described in almost 50% of patients with MDS and area strong and independent risk factor for prognosis. The interstitial deletion in the short arm of the chromosome 12 [del(12p)], is a characteristic but rare abnormality in MDS patients. Del(12p) abnormality has been described in approximately 1-5% of patients as a sole anomaly and is also found in up to 4% of patients along with an additional cytogenetic alteration. Isolated del(12p) is classified as a good risk abnormality according to the Revised International Prognostic Scoring Systems (IPSS-R). The commonly deleted region between 12p12.2 and 12p13.1 encompasses the ETV6 gene. To date, besides mutations in the transcriptional factor ETV6 and in the cell signaling KRAS gene, no other molecular mutations have been associated with del(12p). Murine studies have highlighted a role of the transcriptional factors ETV6 and RUNX1 in the impairment of both erythroid and platelets maturation. Here we investigated the presence of alternative molecular factors associated with del(12p) possibly influencing clinical outcomes and disease phenotypes. Methods: We studied the molecular and clinical data of a total of 2834 patients with myeloid neoplasms and found that 3% (93/2834) had alterations in chromosome 12. The median age was 67 years (24-84), with a male: female ratio of 56:37. Del(12p) occurred in 71% of cases (66/93); among them 14% (9/66) had isolated del(12p), 9% (6/66) had del(12p) + 1 additional alteration and 77% (51/66) carried a complex karyotype. The additional alteration to del(12p) included -7/del 7q (N=3), del(5q) (N=1) and t(X;20) (N=1). Cases with del(12p) were also classified according to disease type (MDS=40, AML=16; MDS/MPN=10; P=.057) and according to MDS risk group [lower-risk (33%, 22/66) and higher-risk (45%, 30/66)]. We applied whole exome sequencing and a targeted deep sequencing panel of 64 most frequently mutated genes in myeloid neoplasms. The ETV6 (12p13.2) gene was deleted in 55% (36/66) of cases while the KRAS (12p12) gene was deleted in 32% (21/66) of cases. One-third (32%, 21/66) of patients had deleted both genes. Two patients were hemizygous for KRAS. Results: Comparing patients with del(12p) (isolated, +1 alteration) to patients without alterations in chromosome 12 (n=2741), those with del(12p) had lower hemoglobin levels compared to patients without 12p aberrations (9.2 g/dL (6-16) vs. 9.7 g/dL (3-17); P=.009) and lower platelets counts (47 x109/L (8-577) vs. 73 x109/L (2-2336); P=.04). We noted that patients with isolated del(12p) had a longer median OS compared to patients with del(12p) associated with a complex karyotype [14 months (1-27) vs. 7 months (5-8)] although this difference was not significant. We then analyzed the mutational profile of the del(12p) cohort (isolated, +1 alteration) and compared their mutational spectrum with that of cases diploid for 12p. The most recurrently mutated genes in cases with del(12p) compared to cases diploid for 12p included RUNX1 (27% vs. 7%; P=.01) and DNMT3A (27% vs. 9%; P=.04). When we analyzed all the cases with del(12p) abnormalities (isolated, +1 alteration and complex) the significantly mutated genes were the transcriptional factors TP53 (38% vs. 4%; P=.0001) and RUNX 1 (14% vs. 7%; P=.04) and the histone modifier ASXL1 (21% vs. 10%; P=.01) We then analyzed the gene expression profile of patients carrying the del(12p) abnormality and found that KRAS mRNA expression levels of patients with MDS with del(12p) had a 2-fold reduction compared to the levels of healthy subjects (P=.017). Similarly, we observed also a decrease in ETV6 mRNA expression levels in patients with del(12p) (P=.07). Conclusions: Patients with del(12p) had lower levels of hemoglobin and platelets counts compared to patients without this cytogenetic abnormality. Mutations in transcriptional factors such as RUNX1 were commonly detected in this cohort, suggesting a possible mechanism contributing to the role of ETV6 in the impairment of erythroid and megakaryocytic cell maturation. Disclosures Sole: Celgene: Membership on an entity's Board of Directors or advisory committees. Maciejewski:Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Celgene: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees.

Several familial mutations predisposing to the development of leukemia have been identified, mostly in association with childhood or adolescent presentations of myeloid neoplasms or bone marrow failure states. However, many of these genetic syndromes have variable penetrance and latency. Thus, late presentations of familial diseases are possible as with DDX41 mutations associated with adult myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML). Growing evidence exists to support the notion that germ line (GL) alterations may play a greater than previously known role in otherwise typical spontaneous adult myeloid neoplasms. These alterations may include known predisposition leukemia genes (i.e.,. ETV6, RUNX1, CEBPa, and DDX41), heterozygous alterations known to generate recessive disorders, or new unidentified genes. The search for such predisposition genes is difficult but we hypothesized that genes frequently damaged by somatic mutations may also be affected by GL alterations, a similar principle as seen for CEBPa or TP53. Further, that a collection of rare variants (MAF We analyzed a cohort of 4159 patients with myeloid neoplasms using WES and targeted deep next generation sequencing (NGS). Extensive scoring algorithms, along with mutational databases were used to detect GL variants. Serial tumor samples were examined to discriminate GL from somatic defects. We identified 2518 TET2 SNPs in 1475 patients. Multiple SNPs were seen in the same patient, 28% of patients had 2 SNPs, 11% had 3 and 6% had 4 or more SNPs. Of those 52% were found in proximity to the catalytic domain, and 10% were homozygous. TET2 SNPs irrespective of their population frequency were found in 21% of MDS, 27% of MDS/MPN, 31% of MPN, and 33% of AML patients. We focused our analysis on the top 10 canonical SNPs with the highest odds ratio (OR) when comparing frequencies in our population to that of healthy controls derived from the ExAC base of 60,000 individuals. The SNPs with the highest OR was present in 2% of patients (p.Val1718Leu) vs. 0.5% of controls. Combinations of SNPs (biallelic configuration) were then investigated; SNPs p.Leu1721Trp and p.Pro363Leu co-occurred in 8% of patients, and p.Val218Met, p.Leu34Phe and p.His1778Arg co-occurred in 3%. We also queried whether GL TET2 SNPs create a haplotype (i.e., increased the risk for acquisition of somatic TET2 mutations) or are mutually exclusive. Overall, 32% of cases with any TET2 SNP also acquired a TET2 somatic mutation vs the WT form for these (68%). Conversely, among the somatic mutant carriers, for example SNP p.Met1701Ile, SNP p.Tyr864His and SNP p.Gly355Asp had an OR vs. controls of 1.1. In particular, 55% of all TET2 mutants had SNP p.Leu1721Trp in a biallelic fashion vs. 46% of somatic WT cases. This suggests a predisposition haplotype. We also examined the relationship between unrelated somatic mutations and "top OR" TET2 SNPs. No correlation was seen between SNPs and other somatic mutations. In sum, several GL TET2 variants or their combination may constitute complex, likely low penetrance predisposing factors for myeloid neoplasm that also interact with somatic lesions and lead to cancer decades later. Disclosures Meggendorfer: MLL Munich Leukemia Laboratory: Employment. Nadarajah:MLL Munich Leukemia Laboratory: Employment. Maciejewski:Celgene: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees; Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau.

Background Several recurrent somatic mutations have been identified in MDS and these mutations play an important role in disease pathophysiology and outcome. BCOR and BCORL1 are located on chromosome X and interact with histone deacetylases and other cell functions. The BCOR gene is mutated (BCORMUT) in 4-6% of MDS patients (pts) and is associated with poor outcome. BCORL1 mutations (BCORL1MUT ) are present in Methods We sequenced DNA samples from pts with MDS and related myeloid malignancies (MDS/myeloproliferative neoplasms (MPN) and secondary AML) for the presence of BCOR and BCORL1 mutations and 58 other genes that have been described as recurrently mutated in myeloid malignancies. The Revised International Prognostic Scoring System (IPSS-R) was calculated as descried previously. Overall survival (OS) was measured from the time of diagnosis to time of death or last follow up. Variant allele frequencies (VAFs) adjusted by zygosity were used to define architecture of driver clones. Results Of 621 included MDS pts, 29 (5%) had BCOR mutations and 13 (2%) had BCORL1 mutations. Patients with BCOR mutations were younger (median age 63 vs. 68 years, p= .04), and had a lower platelet counts at diagnosis (63 vs. 93 109/L, p = .01) compared to BCORWT pts. Cytogenetic risk categories per IPSS-R for BCORMUT was similar to BCORWT: very good 3% vs. 2%, good 65% vs. 62%, intermediate 10% vs. 17%, poor 13% vs. 7% and very poor 6% vs. 9%, respectively, p = .62). Risk categories per IPSS-R were also similar for BCORMUTcompared toBCORWT:very low 10 vs. 16%, low 31 vs. 40%, intermediate 24 vs. 18%, high 17 vs. 14%, and very high 13 vs. 10%, respectively, p = .69). BCOR mutations were missense in 9 pts (31%), frameshift insertion/deletion in 10 (34%), stopgain in 8 (28%), and nonsense in 2 (6%). BCOR is commonly co-mutated with ASXL1 (p=.008), RUNX1 (p= .0001), NF1 (p =.002), ETV6 (p =.026), BCORL1 (p =.0001), MECOM (p =.021), RAD21 (p =.021), and CEBPA (p = .0001). Clonal architecture analysis identified BCOR mutations as ancestral, subclonal, and mosaic in 41%, 21% and 38% of cases respectively. The median OS for BCORMUT pts was 24.5 months compared to 17.9 months for BCORWT, p = .23). The impact of BCOR mutations on OS was neutral even after adjustment for age and IPSS-R risk categories. Pts with BCORL1MUT had lower WBC counts (median 2.7 vs. 4 109/L p = .02), and lower platelets counts at diagnosis (72 vs. 91 109/L, p = .02) compared to BCORL1WT pts. In cytogenetic analyses, BCORL1MUT was associated with a higher incidence of very good cytogenetics per IPSS-R (15% vs. 2%, p =.001). Further, BCORL1MUT were more likely to be classified in the intermediate risk group per IPSS-R (46 vs. 18%, p = .01) and less likely to be classified in the low risk group (7 vs. 40%, p = .01) compared to BCORL1WT. BCORL1 mutations were missense in 7 patients (53%), frameshift insertion/deletion in 3 (23%), stopgain in 1 (7%), and nonsense in 2 (15%). BCORL1 is commonly co-mutated with TET2 (p =.001), BCOR (p Conclusion BCOR and BCORL1 mutations occur with low frequency in MDS. These mutations can be ancestral or subclonal. The impact of BCOR mutations on OS was neutral even after adjustment for age and IPSS-R risk scores. Although the median OS for pts with BCORL1MUT was longer compared to BCORL1WT,it was not statistically significant. A larger pt population with BCOL1 may show a positive impact of these mutations on OS. Disclosures Mukherjee: Celgene: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding. Advani:Blinatumomab: Research Funding; Pfizer Inc.: Consultancy, Research Funding. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees.

TET2 is one of the most commonly mutated genes in myeloid neoplasia. The frequency of TET2 mutations (TET2MUT) increases with age and they have been found in aging healthy controls, in whom their presence was associated with a subsequent risk of developing a myeloid neoplasm. The TET2 gene product is an enzyme that uses alpha-ketoglutarate (αKG) and vitamin C to hydroxylate 5-methylcytosine (5MC), leading to both passive demethylation during cell replication and active demethylation via base excision DNA repair enzymes. Despite large studies, there is no consensus opinion as to the clinical impact of TET2 mutations and their mechanistic role in the pathogenesis of MDS. It is likely that the heterogeneity of TET2MUT, their configuration, sub-clonal context and co-associated variables result in biological heterogeneity that precludes proper assessment of clinical impact. To address these issues we analyzed a cohort of 4974 patients with myeloid neoplasms using targeted deep sequencing of a panel of 60 genes found to be most frequently affected by somatic mutations in myeloid neoplasms. A total of 1861 TET2 alterations were identified as somatic in 1238 cases using various bioanalytic algorithms and sequencing of germline DNA where possible. Of these mutations, 80% were frame shift/stop codons (fs/sc) likely leading to various truncations, while 20% were missense (ms) mutations. While no hotspots for mutations were found, 53% were located in the proximity of the catalytic domain. The most recurrent ms alteration was p.Ile1873Thr, found in 2% of all TET2MUT cases. Biallelic TE2MUT were present in 45% of mutant cases; of which fs/sc alteration combinations were the most common (333/557; 60%), and an additional 9% were homozygous fs/sc variants. Biallelic ms mutations were present in 5% of biallelic cases (3% of homozygous ms configurations). Clinically, TET2MUT were found in 17% of MDS patients, 65% of MDS/MPN, 12% of MPN, 21% of non-core binding factor pAML, and 26% of sAML, and were most frequently associated with normal cytogenetics (p TET2MUT tend to display a higher number of additional mutations than TETWT (p Analysis of clonal architecture including co-associated events revealed that 40% of TET2MUT were dominant/possibly ancestral within the clonal hierarchy. In cases with a dominant TET2MUT, the most commonly associated somatic events involved TET2 (18%), followed by SRSF2 (8%), ASXL1 (7%) and DNMT3A (6%). In those cases (26%) in which TET2 was not the dominant clone, SRSF2 (11%), ASXL1 (10%), DNMT3A (9%), and EZH2 (9%) were the most common early events. In 33% of TET2MUTcases, the TET2MUTclones were co-dominant with another TET2 (29%), SRSF2 (15%), DNMT3A (11%), or ASXL1 (9%). TET2MUT confers worse survival than WT cases (p Our analyses demonstrate that TET2 lesions are strongly selected for in myeloid neoplasia and a gene dose effect may dictate biological and molecular features. While ancestral TET2 ancestral events do not determine the subsequent phenotype, subtypes of myeloid neoplasms appear to be associated with specific sets of secondary or tertiary sub-clonal events. Disclosures Nadarajah: MLL Munich Leukemia Laboratory: Employment. Maciejewski:Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Celgene: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees. Meggendorfer:MLL Munich Leukemia Laboratory: Employment.

Targeted and unbiased next generation sequencing (NGS) has contributed to a better understanding of the molecular pathogenesis of myeloid neoplasms, including MDS. Discovery efforts have identified novel classes of mutated genes, while deep NGS approaches have yielded a better appreciation of clonal hierarchy, inter-case variability and intra-tumor heterogeneity. MDS is a disease continuum characterized by a wide spectrum of often overlapping lesions that determine phenotype, while also serving as initiation and progression events. In addition to somatic lesions, germ line (GL) alterations can serve as bona fidenon-clonal ancestral events that play an underappreciated role in MDS pathogenesis. While some of these lesions are associated with childhood familial leukemia syndromes, others are unknown, and are likely characterized by a low/variable penetrance and delayed disease manifestation. To delineate clonal dynamics in MDS, we sequenced whole exomes of 262 cases with primary MDS and related disorders. For validation and confirmation we also deep sequenced a cohort of 1,686 additional cases with a various type of myeloid malignancies. An extensive bioanalytic pipeline and confirmatory sequencing, including GL DNA analysis, was used to discriminate somatic vs. GL lesions and exclude sequencing artifacts. Initially we focused on driver somatic events in significantly mutated genes. All somatic mutations were subjected to clonal hierarchy analysis using variant allele frequencies (VAFs). In selected cases (n = 180), serial analyses were performed. Using VAF rankings of each event, a position within the clonal hierarchy was assigned; while each patient has a single dominant clone, some may have a founding chromosomal abnormality and others may have VAFs too close to distinguish, i.e. have co-dominant events. In general, multiple subclonal events are detected in each patient. For the purpose of this analysis we distinguished between 2 types of ancestral events: 1) driver non-clonal mutations (e.g., GL TP53, RUNX1, ETV6) and 2) predisposition non-clonal events (FA genes, telomerase genes, BRCA1/2). The latter do not influence the clonal architecture. Based on average sequencing depth, 5,474 somatic mutations were identified: 241 (92%) were clonal dominant and 234 (89%) were sub-clonal (secondary) events. The median number of mutations in subclonal events per case was 13. The number of mutations in subclonal events was higher than that in events that were clonal dominant (4,881 vs 593). No genes were mutated in a purely dominant fashion and some genes were almost entirely subclonal, e.g., RAS and FLT3. For each dominant event, there is a frequent secondary lesion, e.g., dominant TET2 mutations are followed by subclonal second TET2 events, SRSF2 and ASXL1 lesions. Thus, novel relationships between dominant and subclone events were found, indicating the presence of invariant functional interactions among different mutations in MDS pathogenesis. In a confirmatory cohort studied by NGS targeted to a selected panel of significantly mutated genes, the number of subclonal events increased due to greater coverage and thus sensitivity. The spectrum of dominant events, however, should not differ as they are inherently associated with a high clonal burden. For examples, TP53 clonal mutations frequently co-occur with TP53 subclonal mutations (12%, p=.004), but are exclusive of STAG2 subclonal mutations. EZH2 clonal and ASXL1 secondary mutations also co-occur. Classifications of clonal and secondary events may have prognostic and diagnostic implications. We identified a spectrum of novel predisposition and non-clonal driver variants by comparing to ethnically weighted control populations. Eight mutations (3%, 8/262 cases) in 3 genes (DDX41, TP53, and ELANE) were identified as driver non-clonal mutations because identical mutations were reported in familial leukemia syndromes, while 16 mutations (6%) in 3 genes (CSF3R, BRCA1, and RPL5) were identified as non-clonal predisposition events. Detailed understanding of such clonal dynamics and complexity of clonal hierarchical complexity may have clinical significance, both for somatic mutations and for germline events. Increasing clonal burden of extracted genes associated with predictive prognostic impact should be prospectively validated in a more uniform and larger cohort of MDS cases. Disclosures Makishima: The Yasuda Medical Foundation: Research Funding. Mukherjee:Celgene: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Ariad: Consultancy, Honoraria, Research Funding. Sole:Celgene: Membership on an entity's Board of Directors or advisory committees. Carraway:Celgene: Research Funding, Speakers Bureau; Baxalta: Speakers Bureau; Incyte: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees. Sekeres:Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees; Celgene: Membership on an entity's Board of Directors or advisory committees. Ogawa:Kan research institute: Consultancy, Research Funding; Sumitomo Dainippon Pharma: Research Funding; Takeda Pharmaceuticals: Consultancy, Research Funding. Maciejewski:Alexion Pharmaceuticals Inc: Consultancy, Honoraria, Speakers Bureau; Apellis Pharmaceuticals Inc: Membership on an entity's Board of Directors or advisory committees; Celgene: Consultancy, Honoraria, Speakers Bureau.

230 ASXL1 Mutations in AML: Molecular Biomarker for Secondary AML? Jingmei Hsu, MD/PhD, Anita Kumar, MD/MS, Martin Carroll, MD, Noelle Frey, MD, Nirav Shah, MD, Jianhua Zhao, PhD, Elizabeth Hexner, MD, Alison Loren, MD/PhD, Alexander Perl, MD, David Porter, MD, Jennifer Morrissette, PhD, Selina Luger, MD Hematology/Oncology, Abramson Cancer Center of the University of Pennsylvania; Department of Pathology and Laboratory Medicine, Abramson Cancer Center of the University of Pennsylvania

Patients with newly diagnosed AML (n = 360) including 137 (38%) with normal karyotype (NK) were evaluated. Overall, 60 (16.6%) patients, including 46 of the 137 (33.5%) NK patients, had NPM1 mutation at baseline. Thirty-nine patients (30 NK) had available NPM1 status at the time of complete remission (CR) and all (100%) were negative for mutated NPM1. Among the patients with mutated NPM1 at baseline, 10/39 overall (25%) and 7/30 NK (23%) patients relapsed. NPM1 status was available for eight patients (six with NK) at the time of relapse. Among them, 7/8 overall (87%) and 5/6 NK (83%) patients had mutated NPM1, while 1/8 overall (12%) and 1/6 NK (16%) patients remained NPM1 wild type. Among the 300 patients (including 91 with NK) with wild type NPM1 at diagnosis, none acquired a mutated NPM1 clone, either at CR or at relapse. We conclude that mutated NPM1 is a stable and reliable prognostic marker in AML and can be used to assess MRD.

Background: Previous studies have suggested that NPM1 mutations may be a marker for response to all-trans retinoic acid ( ATRA) given as an adjunct to intensive chemotherapy in older patients with acute myeloid leukemia (AML). Patients and Methods: we examined the impact of the addition of ATRA among patients with diploid cytogenetics treated on a randomized phase II study of fludarabine + cytarabine + idarubicine +/- GCSF +/- ATRA with available data on their NPM1 mutation status. Between September 1995 and November 1997, 215 patients were enrolled in the study. Among them 70 patients had diploid cytogenetic and are the subjects of this analysis. Results: The median age of the 70 patients was 66 years (range 23-87). Twenty (29%) of patients had NPM1 mutations. Among them 7 (35%) did and 13 (65%) did not receive ATRA in combination with chemotherapy. Complete remission (CR) was achieved in 71 % of patients treated with ATRA as compared to 69% without ATRA (P = 0.62). With median follow up of 12.5 years, the overall survival (OS), event-free survival (EFS), and relapse-free survival (RFS) were similar among patients who received ATRA compared to no ATRA regardless of NPM1 mutation status. Conclusion: The addition of ATRA to intensive chemotherapy did not affect the overall outcome of patients with AML regardless of NPM1 mutation status.

Acute myeloid leukemia (AML) in the elderly is associated with several distinctive biological and clinical features compared to younger patients. Despite the advances in supportive care and the introduction of less intensive chemotherapy regimens, the overall outcome for this population remains poor. More importantly, the decision making process for choosing the appropriate treatment for individual patient, based on their comorbidities and the biological features of their disease, continues to be challenging for the treating physicians. Currently, a significant number of elderly patients with AML do not receive treatment above and beyond supportive care; several studies have suggested that patients who receive any therapy have a better outcome than patients who receive palliation alone. Furthermore, the development of novel, targeted, and less intensive therapies is providing new options suitable for older patients with multiple comorbidities and with high risk disease features. In this review, we will highlight the challenges that face the treating physicians when encountering elderly patients with AML and will describe some of the potential strategies under development for treating older AML patients and the available data from recent clinical trials.

Omacetaxine mepesuccinate (formerly known as homoharringtonine [HHT]) is a natural alkaloid with significant anticancer activity partly through inhibition of protein synthesis and induction of apoptosis. Prior to the development of tyrosine kinase inhibitors (TKIs), HHT was the most active therapy in chronic myeloid leukemia (CML) after interferon failure. Subsequent trials showed that HHT and omacetaxine are active in patients failing several TKIs or carrying the T315I mutation.This review will discuss the preclinical development of HHT and omacetaxine mepesuccinate in CML and the clinical studies leading to its approval by the Food and Drug Administration (FDA).A sizable number of patients with CML will develop TKI resistance, frequently through the acquisition of BCR-ABL1 kinase domain mutations. Omacetaxine is active in patients with CML after failure to multiple TKIs and in those carrying the T315I mutation, which is highly resistant to all FDA-approved TKIs except for ponatinib. Both ponatinib and omacetaxine have been recently approved by the FDA and represent useful treatment options for patients with CML who failed several TKIs and/or acquired the T315I mutation. The development of an oral formulation of omacetaxine would greatly facilitate its use and provide an attractive option for TKI-based combinatorial strategies.

Introduction Trisomy 21 is frequently noted in patients with AML. In adults, +21 has traditionally been considered an intermediate-risk cytogenetic aberration. Patients and Methods We analyzed 90 patients with newly diagnosed AML harboring +21. Four cytogenetic subgroups were defined based on associated cytogenetic abnormalities: +21 alone, +21 with favorable, +21 with intermediate, and +21 with unfavorable cytogenetics. Results Fifty-four percent of patients with +21 AML achieved a complete remission (CR) or CR with incomplete platelet recovery (CRp) after induction therapy with a trend toward improved CR/CRp rates in patients with +21 alone/+21 with favorable cytogenetics compared with patients with +21 with intermediate/+21 with unfavorable cytogenetics (76% vs. 50%; P = .057). Time to progression (TTP) was 12 months (range, 5-19) and overall survival (OS) was 9 months (range, 7-11) for the entire group. TTP was longer for patients with +21 alone (not reached) or with +21 with favorable cytogenetics (101 months) compared with those with +21 with intermediate cytogenetics (2 months) or +21 with unfavorable cytogenetics (11 months) ( P = .02). Similarly, OS was improved in patients with +21 with favorable cytogenetics (not reached) or +21 alone (107 months), compared with +21 with unfavorable cytogenetics (9 months) or +21 with intermediate cytogenetics (8 months) ( P P = .04 and P = .001; respectively). Conclusion Isolated +21 hitherto classified as intermediate-risk cytogenetics might actually behave as a favorable-risk cytogenetics in adult AML patients.

To the editor: Myelofibrosis (MF) is a heterogeneous, hematopoietic stem cell malignancy characterized by abnormal proliferation of myeloid cells of variable maturity and function [1]. The most common clinical manifestations of the disease are anemia, leukocytosis, leukopenia, thrombocytosis, constitutional symptoms, and marked splenomegaly with extramedullary hematopoiesis [1]. Bone marrow fibrosis (BMF), which results from abnormal deposition of reticulin and collagen fibers in the bone marrow plays a major role in the pathophysiology and clinical manifestation of the disease. Various scoring systems have been developed to evaluate the amount of reticulin and collagen fibers in BMF, the most widely accepted is the 4-grade scoring system (MF 0–3) recommended by the European consensus [2]. The prognostic effect of BMF in MF remains controversial; however, it has been suggested that BMF may affect overall survival (OS) in patients with MF [3–5]. We retrospectively analyzed the medical records of 537 patients who were diagnosed with MF according to World Health Organization criteria [6] and were referred to our institution between February 2005 and December 2009. Project was based on a chart review protocol approved by the Institutional Review Board. Paraffin-embedded trephine bone marrow biopsies obtained from patients at their first presentation to MD Anderson were reviewed. Grades from 0 to 3 were documented, with grade 3 representing the most severe grade of fibrosis [2]. Twenty-five patients were excluded from the final analysis because of incomplete documentation of BMF grading in their reviewed biopsies. The OS rate was calculated from the time of first biopsy at MD Anderson to the time of death or last follow-up; the EFS rate was calculated from the time of first biopsy at MD Anderson to time of death, leukemia transformation, or last follow-up. The IPSS and the DIPSS were calculated as described previously [7,8]. Differences among variables were evaluated by the chi-square for categorical variables and by Kruskal-Wallis and Mann-Whitney U tests for continuous variables. A Cox proportional hazard regression module was used to determine the independent predictors of severity of BMF. Time-to-event analyses were performed by the Kaplan-Meier method, and survival curves were compared with the log rank test. The final analysis included 512 patients (352 patients with primary MF, 88 with post-PV MF, and 72 with post-ET MF). The patient’s demographics and disease characteristics are summarized in Table 1. The median time from diagnosis of MF to the first bone marrow biopsy done at MD Anderson was 4 months (range, 0–388 months), and the median follow-up was 16.5 months (range, 0–79 months). Higher grades of BMF were associated with a lower hemoglobin level (P < 0.001), lower white blood cell count (P = 0.03), and lower platelet count (P = 0.02), higher percentage of blasts in peripheral blood (P = 0.001), as well as, significantly larger spleen (P = 0.01) and liver (P = 0.04). In addition, abnormal cytogenetic findings were associated with higher grades of BMF (P = 0.01); but, neither the presence of the JAK2 V617F mutation nor the allele burden correlated with the severity of BMF (P = 0.28). However, in a multivariate analysis, lower hemoglobin level (P = 0.003), higher percentage of blasts in the peripheral blood (P = 0.015), and larger spleen (P = 0.006) were the only independent prognostic parameters for the severity of BMF. In addition, survival risk assessment of patients using the IPSS and DIPSS scoring systems correlated well with the severity of BMF (Table 1). Vener and colleagues [3] reported the prognostic significance of BMF grading in MF. One hundred thirteen patients were included in their study, 29% had MF-0, 24% MF-1, 23% MF-2, 11% MF-3, and 13% secondary MF. Patients with MF-3 had inferior OS compared to patients with MF-0 (P = 0.001), and MF-1+2 groups (P = 0.003). In addition, Thiele and colleagues [5] had also reported inferior OS among patients with grades MF-2 or MF-3 compared with those with grades MF-0 or MF-1 (statistical significance was not reported) suggesting that BMF grading may play an important role in the prognosis of patients with MF. Recently, Barosi and colleagues [4] compared the clinical features of prefibrotic myelofibrosis (MF-0) to primary myelofibrosis fibrotic type (MF-1,2,and 3) and found that patients with prefibrotic myelofibrosis are predominantly females with younger age, higher hemoglobin, higher platelet count, lower white blood cell count and smaller spleen. Furthermore, with median follow up of 43 months (range, 1–375 months), the median OS for prefibrotic mylofibrosis patients was not reached and was significantly higher than patients with primary myelofibrosis fibrotic type (16.6 years, P

7032Background: Population-based registries have reported improvements in overall survival (OS) for patients (pts) with acute myeloid leukemia (AML) receiving intensive chemotherapy (IC). In this s...

Background The effect of prognostically important gene mutations (MUTs), nucleophosmin (nucleolar phosphoprotein B23, numatrin) (NPM1) and fms-related tyrosine kinase 3 (FLT3), in elderly patients with acute myeloid leukemia (AML) is not well defined. Patients and Methods We analyzed 557 patients, 65 years of age or older with newly diagnosed AML, treated at our institution between 2000 and 2010 with cytotoxic chemotherapy. NPM1 and FLT3 analysis were available in 146 patients (26%) and 388 patients (70%), respectively. Results NPM1 and FLT3 MUTs occurred in 16% and 12% of patients, respectively. No difference in median overall survival was observed between FLT3-MUT and NPM1-MUT patients who received cytotoxic chemotherapy. Outcome was significantly better among patients with NPM1-MUT/FLT3-wild type (WT) genotype (n = 14) compared with patients carrying FLT3/NPM1 genotypes other than NPM1-MUT/FLT3-WT (n = 125). The complete remission rates were 71% and 49%, respectively ( P = .11). The median survival was 21.5 months vs. 9.0 months and estimated 2-year survival rates were 51% vs. 38%, respectively ( P = .003). NPM1 and FLT3 MUTs appear to occur less frequently in elderly AML patients. The prognostic effect of isolated NPM1- or isolated FLT3-MUT is minimal. Conclusion Elderly AML patients with NPM1-MUT/FLT3-WT genotype have significantly improved outcomes compared with patients with other NPM1/FLT3 genotypes when treated with cytotoxic chemotherapy.

FMS-like tyrosine kinase 3 internal tandem duplication (FLT3-ITD) mutations are among the most frequent molecular aberrations in patients with acute myeloid leukemia. We retrospectively analyzed 324 patients with acute myeloid leukemia treated with front-line induction chemotherapy between October 2004 and March 2010. Fifty-six patients had FLT3-ITD mutation at diagnosis. Fifty-one (91%) patients with FLT3-ITD achieved complete remission. Thirteen patients had FLT3 analysis at complete remission. None had FLT3-ITD. Twenty-five (49%) patients with FLT3-ITD relapsed. Of these, 13 (52%) had FLT3-ITD at relapse (3 negative and 9 not done). Among the 201 patients without FLT3-ITD at diagnosis who achieved complete remission, 77 (38%) relapsed among whom 8 (10%) patients acquired FLT3-ITD clone. We conclude that FLT3-ITD mutations are unstable at follow up and may occur for the first time at relapse. Therefore, FLT3-ITD is not a reliable marker for minimal residual disease in acute myeloid leukemia.

To validate the recently reported European Treatment and Outcomes Study (EUTOS) score, we applied it to 465 patients with early chronic phase chronic myeloid leukemia treated with standard-dose imatinib (n = 71), high-dose imatinib (n = 208), or second-generation tyrosine kinase inhibitors (n = 186), and assessed its ability to predict event-free survival (EFS), transformation-free survival (TFS), and overall survival (OS). The median follow-up was 69 months. The overall complete cytogenetic response and major molecular response rates were 92% and 85%, respectively. The 3-year EFS, TFS, and OS rates were 86%, 95%, and 97%, respectively. Of the 465 patients, 427 (92%) were in low EUTOS score category. There was no difference in the major molecular response, TFS, EFS, and OS rates between patients with low and high EUTOS score, overall and within specific therapies. In conclusion, 8% of patients with chronic phase chronic myeloid leukemia treated at our institution are in the high EUTOS score; in this population, the EUTOS score was not predictive for outcome.

Approximately 1 in 1000 pregnancies in the United States are plagued by concurrent neoplastic diseases. These account for the second leading cause of maternal mortality in the United States.1 The majority of these cases are related to solid tumors with only 25% of the cases related to hematologic malignancies.2 Hairy cell leukemia (HCL) is a relatively uncommon lymphoproliferative disorder.3 The outcome of patients with HCL has significantly improved with the introduction of purine nucleoside analogues including cladribine and monoclonal antibodies including rituximab. Remission rates of > 90% have been documented after just 1 course of therapy.4,5 The management of the pregnant patient with hematologic malignancies represents a diagnostic, therapeutic, and social challenge, requiring a multidisciplinary team approach. There is scant published literature regarding treatment of HCL in pregnancy. The therapeutic efficacy, teratogenicity, and maternal toxicities of purine analogues and rituximab in pregnancy are poorly defined.6,7 Orlowski published the first report of successful pregnancy outcome after treatment of HCL with cladribine, suggesting that fertility may be preserved in some female patients exposed to purine analogues.8 However, to our knowledge this is the first report of a favorable outcome in a pregnant HCL patient treated sequentially with a monoclonal antibody (rituximab) followed by a purine analogue (cladribine).

Background: Several validated prognostic models exist for patients (pts) with myelodysplastic syndromes (MDS), including the International Prognostic Scoring System (IPSS), the Revised IPSS (IPSS-R), the World Health Organization (WHO) classification-based Prognostic Scoring System (WPSS), and the MD Anderson Prognostic Scoring System (MDAPSS). All were developed in pts with newly diagnosed MDS, and their prognostic value in subsequent stages of disease, such as at the time of hypomethylating agents failure (HMAs, azacitidine (AZA) and decitabine (DAC), has not been established. Despite this, the IPSS and IPSS-R is often used to determine clinical trial eligibility for pts who fail HMA and has been used by the FDA for drug labeling in this setting. Here in we developed a new prognostic model that predicts outcome post HMAs failure (HMAF). Methods Included patients were diagnosed with higher-risk MDS (per 2008 WHO criteria, higher-risk defined as IPSS Intermediate-2/High) with clinical and pathologic data entered into the MDS Clinical Research Consortium database. The IPSS, IPSS-R, WPSS and MDAPSS were calculated at the time of diagnosis and HMAF. HMAF was defined as no response to AZA or DAC following >4 cycles, loss of response, or progression to acute myeloid leukemia (AML) at any time after starting therapy. Responses were defined per International Working Group criteria (IWG 2006). Overall survival (OS) was calculated from the time of diagnosis to time of death or last follow up when the models were applied at diagnosis and from HMAF date to time of death or last follows up when the models were applied at the time of HMAF. Cox proportional hazard analysis within the multivariable model-building with fractional polynomials (MFP) approach, which automatically select from all factors at the time of HMAF, was used to build the new model. Akaike information criterion (AIC) was used to compare fits from Cox proportional hazards models. Results Of 450 higher-risk MDS pts who failed HMAs, 311 (69.1%) were treated with AZA and 139 (30.9%) with DAC. The median age at diagnosis was 70 years (range: 35-91). Best responses (BR) to HMA were: 96 (21.3%) with complete remission, 40 (8.9%) partial remission, 46 (10.2%) hematologic improvement, 180 (40.0%) stable disease, and 88 (19.6%) with progressive disease. The median number of cycles received during treatment was 6 (range, 2-51). With a median follow up of 17.4 months (IQ range, 16.1, 18.7), the median OS from diagnosis for the entire group was 18.5 months (IQ range, 17.2, 19.8). Median OS from diagnosis was similar for patients treated with AZA compared to DAC (18.0 months vs. 20.3 months, p = .36). The median OS after HMAF was 7.3 months (IQ range, 6.3, 8.4). Survival plots for each prognostic scoring system at diagnosis and HMAF are shown in Figure 1. Comparing the predictive power of these scoring systems at the time of HMAF, the AICc for each model was: MDASS (3541.1); IPSS-R (3562.0), IPSS (3570.0), and WPSS with AICc of (3572.2) (lower AICc indicates better fit of the model). Given the lower predictive power of the current prognostic models at the time of HMAF, we developed a new prognostic model specific for this patient population. Our MFP modeling approach selected 6 factors that have significant association with OS at the time of HMAF in the final Cox multivariate model (Table 1). The new model identified two risk groups: Low: score < 2.25, median OS 11.0 months (95% CI 8.8-13.6) and a high risk group with score of > 2.25 and median OS 4.5 months (95% CI 3.9-5.3). Using the internal model validation assessment, the estimated AICc for the new model was 3520.4 (lowest AICc). When the new model was applied at time of diagnosis, the AICc decreased to 3515.1, a much smaller decrease compared to the existing prognostic systems built at diagnosis: MDASS (3515.7), IPSS-R (3528.2), WPSS (3537.5) and IPSS (3537.7). Conclusion Currently available MDS prognostic scoring systems should be used cautiously in pts at the time of HMAF and, given their inconsistent reliability, should be avoided for clinical trial eligibility or drug labeling. A new prognostic model was developed specific for this patient population. Table 1. The Post-HMA model Table 1. The Post-HMA model Figure 1. Overall survival by scoring systems at diagnosis and at the time of HMA failure Figure 1. Overall survival by scoring systems at diagnosis and at the time of HMA failure Disclosures Komrokji: Celgene: Consultancy, Research Funding; Incite: Consultancy; Novartis: Speakers Bureau; GSK: Research Funding. Steensma:Celgene: Consultancy; Amgen: Consultancy; Incyte: Consultancy; Onconova: Consultancy. Padron:Novartis: Speakers Bureau; Incyte: Research Funding. List:Celgene Corporation: Honoraria, Research Funding. Sekeres:TetraLogic: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees.

Recently, rare somatic nonsense PHF6 mutations and deletions have been reported in patients with T-ALL, AML and blast crisis CML. Germ line PHF6 mutations have been described in Borjeson−Forssman−Lehmann syndrome (BFLS), a hereditary X-linked disorder characterized by mental retardation and somatic deformities. Patients with BFLS have been also reported to develop leukemia, suggesting PHF6 mutations may predispose to cancer. PHF6 is a highly conserved 41kDa protein showing ubiquitous expression in a variety of tissues, including bone marrow, CD34+ cells and leukocytes. The function and molecular pathogenesis in hematological disorders is unknown. PHF6 has been suggested to be a tumor suppressor gene (TSG) involved in the control of rRNA synthesis. Recent CHIPseq experiments showed that PHF6 binds upstream of the regulatory sequence of RUNX1. In an index case of a young adult female patient with proliferative CMML with dysmorphic features, we have identified remarkable GL mosaicism for PHF6 mutation (p.K44fs), confirmed by deep sequencing of bone marrow, CD3+ cells, spleen and skin tissue. Subsequently, we screened patients with myeloid neoplasms by targeted multi-amplicon sequencing to determine the prevalence and distribution of PHF6 gene alterations. Sequencing results from 1122 cases were analyzed (778 by targeted deep sequencing and 344 by whole exome sequencing). In total, we identified 45 cases with PHF6 mutations, 32 of which were frameshift or nonsense mutations. Previously, PHF6 have been included in screening panels by Haferlach et al., (Leukemia 2014) and Papaemmanuil et al., (Blood 2013) and somatic mutations were found in 24/944 and 21/738 cases of MDS, respectively. The somatic nature of these defects was confirmed by analysis of non-clonal CD3+ lymphocytes, Thus the incidence of PHF6 mutations ranges from 4.3% in current study to 2.8% and 2.5% reported by others and are most frequently observed among patients with secondary AML (33%), suggesting that PHF6 mutations are not uncommon driver events in myeloid neoplasia. Gender distribution showed a strong male predominance (76%), indicating that retention of a single copy of PHF6 may be protective. There was no significant sex difference in the transcriptional expression of PHF6 itself. The most frequent chromosomal aberration observed in conjunction with PHF6 mutations was trisomy-8 (p=.08). The most commonly associated somatic mutations were in RUNX1 (p=.001) and IDH2 (p=.008). Concomitant PHF6 and RUNX1 mutations are associated with a poor prognosis in AML, and occur predominantly in males. There was no association observed between low expressors of PHF6 and RUNX1 mutations or RUNX1 expression levels. Conversely, RUNX1 mutant cases without somatic PHF6 mutations were not observed to have low transcriptional PHF6 levels. Subsequent analysis of clonal architecture using variant allelic frequency calculations and serial sampling suggested that mutated PHF6 may function as a founder driver gene in proportion of cases, while RUNX1 mutations are acquired as secondary events. Recent studies proposed that PHF6 deficiency leads to impaired cell proliferation, cell cycle arrest at the G2/M phase and an increase in DNA damage. To delineate a possible pathophysiological pathway involving PHF6 we compared transcriptional expression profiles of low expressors to those with normal levels of PHF6. The most notably deregulated group of genes were clustered to a functionally related group of ribosomal RNA proteins (p In conclusion, our results indicate that PHF6 mutations are generally present in more aggressive types of myeloid neoplasms and arefrequently associated with RUNX1/IDH2 mutations. Our functional in vitro studies, along with recently published reports, suggest an association of PHF6 deficiency with transcriptional regulation and thereby provide a basis for a transcriptional repressor phenotype conveyed by ancestral lesions, consistent with a role for PHF6 as a TSG. Disclosures Levine: Foundation Medicine: Consultancy; 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. Sekeres:Celgene Corporation: Membership on an entity's Board of Directors or advisory committees.

Background Achieving a complete remission (CR) in patients with newly diagnosed acute myeloid leukemia (AML) after induction chemotherapy with cytarabine and an anthracycline (7+3) remains an important treatment goal associated with better overall survival (OS). Approximately 25-30% of younger, and up to 50% of older patients (pts) fail to achieve CR. AML pts with residual leukemia at day 14 receive a second cycle of the same regimen; whether these pts have worse survival than pts not requiring re-induction is unclear. Information on pts with primary refractory AML and the best treatment strategy in this setting are limited. Methods Pts with newly diagnosed AML treated at our institution between 1/2000 and 1/2015 were included. Pts received standard induction chemotherapy with cytarabine for 7 days and an anthracycline for 3 days (7+3). Bone marrow biopsies were obtained at day 14 and a second cycle of the same regimen (7+3 for younger adults, 5+2 for older adults) was given to pts with residual leukemia (blasts > 5%). All responses were assessed at day 30 +/- 5 days post induction. Response was defined as CR and CR with incomplete hematologic recovery (CRi) or platelet recovery (CRp) per International Working Group (IWG) 2003 response criteria. Cytogenetic risk stratifications were based on CALGB/Alliance criteria. OS was calculated from the time of diagnosis to time of death or last follow up. A panel of 62 gene mutations that have been described as recurrent mutations in myeloid malignancies was used to evaluate whether genomic data can be used to predict response. Results: Among 227 pts with AML, 123 received 7+3 and had clinical and mutational data available. Median age was 60 years (range, 23-82). Median baseline WBC was 8.2 X 109/L (range, 0.3-227), hemoglobin 8.9 g/L (range, 4.7-13.8), platelets 47 X 109/L (range, 9-326), and BM blasts 46% (range, 20-95). Cytogenetic risk groups were: favorable in 12 (10%), intermediate in 68 (56%) [normal karyotype in 44 (36%)], and unfavorable in 42 (34%). A total of 93 pts (76%) responded, 69 (74%) received 1 cycle of induction and 24 (26%) required re-induction at day 14 due to residual leukemia. A total of 39 pts (32%) received allogeneic stem cell transplant (ASCT): 18 (46%) from a matched sibling donor, 16 (41%) from a matched unrelated donor and 5 (13%) had an umbilical cord transplant. With a median follow up of 13.5 months, the median OS for the entire group was 13 months (m, range, 0.1-120). The median OS for pts who failed 1-2 cycles of 7+3 was significantly worse than pts who responded (median 2.6 vs 16.9 m, p = 0.002). When pts undergoing ASCT were censored, the median OS was 2.3 vs 9.9 m, p= 0.003, respectively. Overall, 33 pts (27%) had residual leukemia at day 14 and received re-induction, 24 (72%) achieved a response at day 30+/- 5 days. The median OS for pts who received re-induction was inferior compared to pts who did not (10.1 vs. 16.1 months, p= 0.02). When pts who received ASCT were censored, the OS was similar (8.5 vs. 7.4 months, p = 0.49, respectively). Among the 30 pts with persistent disease following induction therapy at day 30, 11 (37%) died from induction complications, 6 (20%) received salvage therapy with mitoxantrone/etoposide/cytarabine, 3 (10%) received high dose cytarabine, 2 (7%) received azacitidine, and 8 (27%) received best supportive care. Among pts who received salvage chemotherapy 56% achieved CR and proceeded with ASCT. Two pts had ASCT with residual leukemia and relapsed within 3 m of ASCT. Pts who received ASCT after induction failure had a significantly better OS compared to non-transplant pts (median OS 22.0 vs. 1.4 months, p < 0.001, respectively); however, this benefit was only seen in pts who had ASCT in CR. We then investigated if genomic mutations can predict response or resistance to chemotherapy. Out of the 62 genes tested, only a TP53 mutation was associated with resistance, p = 0.02. Further, pts with TP53 mutations had significantly inferior OS compared to TP53 wild type regardless of ASCT status (1.4 vs 14.8 m, p< 0.001) Conclusion: Pts with newly diagnosed AML who fail induction chemotherapy with a 7+3 regimen have a poor outcome. Re-induction with the same regimen at day 14 for residual leukemia converted most non-responders to responders, but was associated with worse OS. ASCT improves outcome only in pts who achieve CR with salvage therapy. TP53 mutations predicted resistance to chemotherapy with 7+3. Disclosures Carew: Boehringer Ingelheim: Research Funding. Sekeres:TetraLogic: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees.

Deletion of the 9q21.32 locus is observed in some patients with AML, suggesting an undescribed tumor suppressor resides in this region. HNRNPK is an attractive candidate, as it is one of six genes mapped to this region and is thought to drive AML progression when mutated. Mechanistically, hnRNP K functions as a DNA and RNA binding protein that transcriptionally and translationally governs gene expression. Together, these findings suggest hnRNP K may serve as a currently uncharacterized tumor suppressor and that its haploinsufficiency plays a pivotal role in AML progression. To interrogate a potential relationship between HNRNPK haploinsufficiency and AML, we performed fluorescence in situ hybridization (FISH) using bone marrow aspirates from newly diagnosed AML patients. These analyses revealed the HNRNPK gene is specifically lost in a subset of AML patients and results in a significant decrease in HNRNP K expression in CD34+ cells from patients harboring this deletion (Fig. 1A). To directly examine the potential tumor-suppressive activities of hnRNP K in vivo, we generated a haploinsufficient hnRNP K (hnRNP K+/-) mouse model. hnRNP K haploinsufficiency resulted in reduced survival and hematologic neoplasms with marked genomic instability (Fig. 1B). Critically, hnRNP K+/- hematopoietic stem progenitor cells (HSPCs) were transplantable and had the capacity to develop hematopoietic neoplasms in recipient mice. To examine the mechanism driving these tumor phenotypes, we analyzed the cytokine profile of hnRNP K+/- mice. This examination revealed that several critical pro-proliferation and myeloid differentiation cytokines were significantly overexpressed (e.g.; IL-3, IL-6, GM-CSF, and G-CSF), resulting in activation of the JAK-STAT pathway (Fig. 1C). Next, we analyzed the proliferation potential of hnRNP K+/- HSPCs and mouse embryo fibroblast (MEFs) using cell based studies. Here, we observed a significant increase in the number of colony forming units in HSPCs and proliferation potential in both HSPC and MEFs between cells from hnRNP K+/- and wild type mice. Critical to the proliferative advantaged observed in hnRNP K+/- cells, activation of the p53/p21 pathway was significantly reduced in hnRNP K+/- MEFs following exposure to ionizing radiation. To more fully explore the molecular mechanisms responsible for these phenotypes, we performed whole-transcriptome analyses. Pathway analysis revealed that hnRNP K haploinsufficiency resulted in the attenuation of critical p53- and C/EBP pathways (Fig. 1D). We next validated expression changes in critical genes known to govern tumorigenesis, myeloid differentiation, and proliferation. Here, we observed a significantly down regulation in C/EBPa, C/EBPβ, and p21 at both the transcript and protein level (Fig. 1E-F). To assess the relationship between hnRNP K expression and regulation of these genes, we performed Chromatin Immunoprecipitation (ChIP) and RNA immunoprecipitation (RIP) analyses. ChIP analysis revealed the transcriptional activities of hnRNP K through its direct interaction with the promoter of these genes. Critically, reduced hnRNP K expression significantly diminished these interactions, resulting in their reduced expression. Interestingly, RIP analyses revealed that hnRNP K may also specifically regulate the expression of the tumor suppressive p42 isoform of C/EBPα through its translation regulation of the C/EBPα transcript. Together, these results suggest hnRNP K is a bona fide tumor suppressor and that its loss contributes to the development of hematologic malignancies. Given its critical regulation of the p53 pathway, drug therapies targeting the p53 pathway may serve as efficacious treatment options for patients with 9q deletions. To address this, we are currently examining the efficacy of Mdm2-p53 antagonist (e.g.; Nutlin-3 and AMG-232) using primary patient samples and primary cells from hnRNP K+/- mice with hematologic malignancies. Figure 1. Figure 1. Disclosures No relevant conflicts of interest to declare.

Background Acute myeloid leukemia (AML) is a complex, heterogeneous neoplasm characterized by the accumulation of complex genetic alterations that are responsible for the initiation and progression of the disease. Secondary AML (sAML) represents a progression from antecedent hematologic disorders such as myelodysplastic syndromes (MDS) or myeloprolifrative neoplasms (MPN). Certain acquired mutations have been reported to be specific for sAML when compared to primary AML (pAML), but many limitations exist when cytogenetic grouping or other parameters are taken into account. In addition, some mutations have been shown to impact survival in some studies, but not others. Methods We performed targeted deep sequencing on samples from bone marrow and peripheral blood of pts diagnosed with sAML and pAML and treated at our institution between 1/2003-1/2013. Additional data on pAML was added from The Cancer Genome Atlas (TCGA). A panel of 62 gene mutations described as frequently recurrent mutations in myeloid malignancies were assessed. Cytogenetic grouping was defined by CALGB/Alliance criteria. Differences were compared using Fisher's exact test and the Mann-Whitney U test for categorical and continuous variables, respectively. Overall survival (OS) was calculated from the time of diagnosis to last follow up or death. Results: A total of 496 pts included: 273 with pAML and 223 with sAML. Comparing pAML to sAML, pts were younger (median age 59 vs. 68 years, p 3 abnormalities) compared to 11% and 16% for pAML (p< .001, .009), respectively. Mutations in ASXL1 (p We then evaluated whether the mutations that were specific to each AML phenotype had an impact on OS. We observed different mutations that impacted OS in each phenotype: DNMT3A (HR 1.81, 95% CI 1.28-2.57, p Conclusion Clear genomic variations exist between sAML and pAML. Although some of these genomic changes are more specific to each phenotype in general, this specificity and the impact on OS differed for each cytogenetic subgroup, highlighting the complexity of interpreting genomic information in pts with AML and the need to incorporate both cytogenetic and molecular data in prognosis-driven treatment decisions. Disclosures Sekeres: TetraLogic: Membership on an entity's Board of Directors or advisory committees; Celgene Corporation: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees.

Despite documented success of immunosuppressive therapy (IST) in the treatment of aplastic anemia (AA), a significant minority of patients remain refractory, most responses are incomplete, and allogeneic stem cell transplantation is not available for older patients or those with significant comorbidities. Until introduction of the cMpl agonist eltrombopag, anabolic steroids were the most commonly used salvage drugs. At least theoretically, engaging growth factor receptors with eltrombopag has the potential to promote the evolution or expansion of mutant clones and thereby increase the rate of progression to secondary MDS, a feared complication of AA occurring in 10-20% of patients. Recently we and others reported detection of clonogenic somatic mutations typical of MDS in patients with AA and PNH. Subsequent study demonstrated that mutations characteristic of sMDS can be found in some patients at presentation of AA and may constitute risk for subsequent progression to MDS. As the risk of MDS evolution was a prominent concern when filgrastim was more widely used in management of AA, now similar questions have been raised regarding use of eltrombopag, be it as salvage therapy or to complement IST. Recently, one of our primary refractory patients receiving eltrombopag progressed to AML. This clinical observation led to investigation of the impact of eltrombopag on evolution and clonal expansion using deep sequencing of a cohort of patients with AA. DNA from bone marrow cells was sequenced before and after initiation of eltrombopag to evaluate clonal expansion or evolution using a targeted multi-amplicon deep sequencing panel of the top 60 most commonly mutated genes in MDS. Among 208 AA patients treated at Cleveland Clinic, we identified 13 patients (median age 68 yrs.) who were treated with eltrombopag for IST-refractory AA; median duration of treatment was 85 wks. The overall response rate, defined as sustained improvement in blood counts and transfusion independence after 12 weeks of therapy, was 46% (6/13), while 38% (5/13) of patients showed stable disease with intermittent transfusions (one of whom underwent HSCT). Among the two non-responders, one patient developed a PNH clone and another progressed to AML (see below). Expansion of PNH granulocytes after eltrombopag treatment was observed in 2 patients. Two patients had chromosomal abnormalities at initial diagnosis, one with t (10; 18) in 2 metaphases, and one with an extranumeral Y chromosome. Use of next generation sequencing (NGS) allows for the quantitative detection of clonal events. We hypothesized that serial analysis by NGS before and after eltrombopag therapy may provide clues as to potential effects of this drug on clonal evolution. Sequencing analysis before eltrombopag treatment revealed the presence of a sole clonal mutational event in 3/13 cases, including CEBPA, EZH2, and BCOR. In the patient with a CEBPA mutation, the mutation persisted during treatment with minimal clonal expansion evidenced by a change in VAF from 53% to 65%. In the second patient, NGS results revealed the initial presence of an EZH2 mutation. A post eltrombopag sample clearly identified acquisition of additional clonal events in genes highly associated with advanced disease and clonal evolution (RUNX1 and U2AF1), as well as slight expansion of a persistent EZH2 clone from 2 to 8%. The third patient harbored a BCOR mutation which expanded markedly, increasing from 8% to 21%, and was accompanied by a hematological response. Sequencing results after eltrombopag treatment revealed the acquisition of new somatic mutations in 5/13 (38%) cases: 2 new CEBPA mutations, 1 new BCOR mutation, and, as discussed, one case with an initial EZH2 mutation in which RUNX1 and U2AF1 mutations were later discovered. In the 5th patient, evolution to AML was observed and accompanied by a large DNMT3A and U2AF1 clone that was absent on initial evaluation. In conclusion, we did observe occasional expansion of clones with potentially leukemogenic mutations during treatment with eltrombopag. At our institution a case control study of patients with refractory aplastic anemia without treatment with eltrombopag is ongoing; ideally a prospective trial would be needed to confirm results. Our results suggest that the initial detection of certain somatic mutations (CBL, SETBP1 and RUNX1) associated with post-AA MDS may contraindicate use of eltrombopag in AA. Disclosures Sekeres: Celgene Corporation: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; TetraLogic: Membership on an entity's Board of Directors or advisory committees.