Post ASH speaker abstracts 2018


Abstracts to support presentations from Austin Kulasekararaj

Mutational Signatures Associated with Intensity and Duration of Smoking in Myelodysplastic Syndromes (MDS)

Result Type: Paper
Number: 424
Presenter: Yazan Madanat
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Inflammation, Immunity, and Iron

Yazan F. Madanat, MD1, Tomas Radivoyevitch, PhD2*, Aziz Nazha, MD1, Yazeed Sawalha, MD1, Karam Al-Issa, MD3, Nour Abuhadra, MD3*, Teodora Kuzmanovic, BA4*, Bartlomiej Przychodzen5*, Cassandra M. Hirsch, BS5*, Bhumika J. Patel, MD3, David Seastone, PhD, DO6, Yogenthiran Saunthararajah, MD5, Sudipto Mukherjee, MD, PhD, MPH3, Anjali S. Advani, MD3, Hetty E. Carraway, MD, MBA1,3, Aaron T. Gerds, MD, MS1, Jaroslaw P. Maciejewski, PhD, MD, FACP5and Mikkael A. Sekeres, MD, MS3

1Taussig Cancer Institute, Department of Hematology & Medical Oncology, Cleveland Clinic, Cleveland, OH
2Quantitative Health Sciences, Cleveland Clinic, Cleveland, OH
3Leukemia Program, Taussig Cancer Institute, Department of Hematology and Medical Oncology, Cleveland Clinic, Cleveland, OH
4Department of Translational Hematology and Oncology, Taussig Cancer Institute, Cleveland Clinic, Cleveland, OH
5Translational Hematology and Oncology, Cleveland Clinic, Cleveland, OH
6UPMC Regional Cancer Center, Erie, PA


Background: MDS pathogenesis is a multi-step molecular process resulting from somatic mutations acquired over time. Smoking is a known risk factor for MDS which may accelerate this process. To further investigate the contribution of tobacco use in MDS development, we studied the association between smoking and specific genomic alterations in MDS patients (pts).

Methods: MDS and CMML pts diagnosed between 1996 and 2016 with both smoking histories and sequencing data were included. The Center for Disease Control definition defined current, ex- and never smokers, the latter defined as pts with zero smoking duration/packs per day (PPD). Next generation targeted sequencing was performed for a panel of 60 genes known to be commonly mutated in myeloid malignancies. Fisher’s exact test and Wilcoxon’s test compared proportions and medians, respectively. Poisson regression modeled total numbers of mutations per pt. The probability of a mutation in a particular gene was modeled using conditional logistic regression. Adjustment was done for both sex and age, fits were stratified using age cut points of 0, 40, 50, 60, 65, 70, 75, 80, 85 and 90 years (y). Odds ratios (OR) were found as exponentials of gene status coefficients of such fits.

Results: In 672 pts, the median age at diagnosis was 68y (range, 20-103), for pt characteristics, see Table 1. Median OS 34 mo (95% CI 30-39) and median follow-up time was 21.5 mo (IQR, 10-45). Median tobacco exposures were 30 pack-years (py) in current smokers (range, 6-80) and 25 in ex-smokers (range, 0.3-175), P=.2. More men than women ever smoked (74% v 45%; P<.0001). Longer duration of smoking associated with higher IPSS-R categories (P=.012) and cytogenetic risk scores, P=.037. Most pts (502, 75%) had at least one mutation; 23% had a single gene mutation, 19% had 2, 15% had 3, and 18% had >3. The most frequently mutated genes were TET2 17%, ASXL1 16%, SRSF2 14%, SF3B1 12%, DNMT3A 10%, RUNX1 9%, STAG2 9%, and TP53 7%.

Poisson regression analysis was used to assess the impact of smoking duration (y) or PPD on mutation acquisition. A positive correlation of greater smoking intensity (> 2 PPD v <0.5, relative increase (RI) =1.46, P=.011) and longer duration of smoking (>40y v <10, RI = 1.31, P =.011) was seen with increased number of mutations [Figure 1].

Impact of duration of smoking was evaluated using logistic regression. Duration of smoking >40y v <1 was significantly associated with the following molecular abnormalities: FLT3 (P=.036), EZH2/del7 (P=.038) and NRAS (P=.046). Comparing 20-40y v <1 yielded SETBP1 (P=.005), DNMT3A (P=.010) and EZH2/del7 (P=.036). Comparing 10-20y v <1 yielded STAG2 (P=.039) and NOTCH1 (P=.046). Finally, comparing durations of 1-10y v <1 yielded PHF6 (P=.016). OR per mutation, see Table 2.

Focusing on intensity of smoking, >2 PPD v <0.5 was significantly associated with higher rates of mutations in FLT3 (P=.013), EZH2/del7 (P=.024), IDH1 (P=.041), and NPM1 (P=.044). Even low intensity smoking, comparing 0.5-1 PPD v <0.5, was significantly associated with mutations in SETBP1 (P=.002), DDX54 (P=.042) and DNMT3A (P=.046). For the combination of smoking duration and intensity, >40py v <10 was associated with IDH1 mutation (P=.046), while 20-30 py v <10 significantly associated with RAD21 (P=.016), DDX54 (P=.018), SRSF2 (P=.041), del7/EZH2 (P=.042), and NF1 (P=.045) mutations. Comparing 1-10py v <1, significant associations were found for IDH1 (P=.004), PRPF8 (P=.011), SETBP1 (P=.022), ETV6 (P=.040) and TET2 (P=.047) mutations [Table 2]. Analyses were repeated after excluding CMML and MDS/MPN pts to control for the common genomic phenotype seen in these disorders, with identical results.

Conclusions: Tobacco smoking predisposes to either acquisition or selection of distinct somatic mutational patterns from those seen in “spontaneous/non-tobacco exposed” MDS. EZH2/del7 abnormalities associated with longer durations and heavier smoking patterns, while SETBP1 mutation associated with lighter smoking for prolonged durations. Lesions associated with light smoking for years may be related to chronic inflammatory changes, whereas those associated with higher intensity may reflect direct DNA damage. Validation of these findings warrants earlier sequencing in pts with pre-MDS conditions such as clonal hematopoiesis of indeterminate potential to better identify the onset of somatic mutations and their role in MDS pathogenesis.

The Use of Immunosuppressive Therapy (IST) in Patients with the Myelodysplastic Syndromes (MDS): Clinical Outcomes and Their Predictors in a Large International Patient Cohort

Result Type: Paper
Number: 422
Presenter: Maximilian Stahl
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Inflammation, Immunity, and Iron

Maximilian Stahl, MD1, Michelle Deveaux2*, T. M. M de Witte, MD, PhD3, Judith Neukirchen, MD4*, Mikkael A. Sekeres, MD, MS5, Andrew M. Brunner, MD6, Gail J. Roboz, MD7, David P. Steensma, MD8, Vijaya R. Bhatt, MD9, Uwe Platzbecker, MD10, Thomas Cluzeau, MD, PhD11, Pedro H Prata, MD12, Pierre Fenaux, MD, PhD13, Amir T. Fathi, MD, MD, BS14, Alex Smith, PhD15*, Ulrich Germing, MD16*, Ellen K Ritchie, MD17, Vivek Verma, MD18*, Nikolai A. Podoltsev, MD, PhD1, Leonor Arenillas, MD19*, Thomas Prebet, MD, PhD1, Valeria Santini, MD 20, Steven D. Gore, MD1, Rami S. Komrokji, MD21 and Amer M. Zeidan, MD22

1Yale Cancer Center, New Haven, CT
2Department of Biostatistics, Yale School of Public Health, New Haven, CT
3Radboud University Medical Centre, Nijmegen, Netherlands
4Dept. of Hematology, Oncology and Clinical Immunology, Heinrich Heine University Düsseldorf, Duesseldorf, Germany
5Taussig Cancer Institute, Department of Hematology & Medical Oncology, Cleveland Clinic Taussig Cancer Institute, Cleveland, OH
6Hematology-Oncology, Massachusetts General Hospital, Boston, MA
7Weill Cornell Medical College, New York
8Dana-Farber Cancer Institute, Boston, MA
9University of Nebraska Medical Center, Omaha, NE
10Department of Medicine I, University Hospital Dresden, Dresden, Germany
11CHU de Nice, Nice, France
12Hématologie Clinique, Hôpital Saint Louis, Paris, France
13Hopital Saint-Louis, Paris, France
14Massachusetts General Hospital, Harvard Medical School, Cambridge, MA
15Epidemiology and Cancer Statistics Group, University of York, York, United Kingdom
16Dept. of Hematology, Oncology and clinical Immunology, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany
17Weill Cornell Medicine, New York, NY
18University of Nebraska Medical Center, Omaha
19Laboratorio Citología Hematológica. Servicio Patología, GRETNHE, IMIM Hospital del Mar Research Institute, Barcelona, Spain
20Azienda Ospedaliero Universitaria Careggi, Florence, Italy
21Malignant Hematology, H. Lee Moffitt Cancer Center, Tampa, FL
22Section of Hematology, Department of Internal Medicine, Yale Cancer Center, East Haven, CT


Introduction: All patients with lower-risk (LR)-MDS experience primary or secondary failure of erythropoiesis-stimulating agents (ESAs), lenalidomide, or hypomethylating agents (HMAs), and there is an unmet need for other active therapies. IST has activity in patients with LR-MDS, but is infrequently used. We studied the clinical outcomes and predictors of response in patients with MDS treated with IST in a large multi-center international cohort.

Methods: Data was collected from 11 centers in the United States and in Europe. Responses and red blood cell (RBC) transfusion independence (TI) were defined by the modified 2006 MDS IWG criteria. Kaplan-Meier methods estimated overall survival (OS) from initiation of IST to death or end of follow-up. Multivariable logistic regression models estimated odds for response and TI, and multivariable Cox Proportional Hazard models estimated hazards ratios (HR) for OS. Covariates included IST received, age at diagnosis, MDS classification, MDS risk category per International Prognostic scoring system (IPSS) and revised IPSS (IPSS-R), WHO-based Prognostic scoring system (WPSS), LR-MDS Prognostic Scoring System (LR-PSS), transfusion dependence, presence of paroxysmal nocturnal hemoglobinuria (PNH) and large granular lymphocytes (LGL) clones, HLA-DR15 positivity, IPSS-defined cytogenetic risk groups, and genetic mutations in TP53, IDH1/2, ASXL1 and SF3B1 genes.

Results: A total of 367 patients received IST; of these, disease risk according to the IPSS was low (23%), int-1 (68%), or either int-2 or high risk (9%). Median age at diagnosis was 65 years (range [R], 15-95), and 63% were male (Table 1). Median number of therapies prior to IST was 1 (R, 0-7) with 249 patients receiving ≥ 1 prior therapy. Most common prior therapies were ESAs (43%) and HMAs (29%). In 149 (40.6%) patients only prednisone was used; these patients were excluded from response and survival analysis. Of the other 198 patients, IST used from most to least frequently was ATG (46%: Rabbit/Horse 62/38%), cyclosporine (11%), tacrolimus (5%) and others (8%). ATG was combined with cyclosporine in 25% of patients. 114 of these 198 patients had response criteria reported: 13.2 % (95%CI, 7.8%-21.1%) had a complete remission (CR), 5.3% (95%CI, 2.2%- 11.6%), had a partial response (PR) and 31.6% (95%CI, 23.4%-41.0%) achieved hematologic improvement (HI) resulting in an overall response rate (ORR) of 45%. In contrast, 38.6% (95%CI, 29.8%-48.2%) of patients had stable disease (SD) and 11.4% (95%CI, 6.5%-19.1%) had progressive disease (PD). TI was achieved in 38.3% of patients. Median OS from time of IST initiation for all patients was 47.2 months (95%CI, 35-69.4 months). For patients who achieved a response (CR+PR+HI) to IST, the median OS was not reached (95%CI, 47 months-un-reached) whereas patients without a response had a median OS of 26 months (95%CI, 19.8-46.6 months) (p=0.0004) [Figure 1A]. Similarly, for patients who achieved TI with IST, median OS was not reached, whereas patients who remained transfusion dependent had a median OS of 52 months (95%CI, 20.7 months-un-reached) [Figure 1B]. In univariate analyses, female sex (female vs. male, HR 0.6, p=.04) predicted better OS whereas higher-risk IPSS (HR 1.9, p=.01) and IPSS-R score (HR1.3, p=.02) predicted worse OS [Figure 2A]. In multivariate analyses, only bone marrow blast >5% remained a statistically significant predictor of OS (>5% vs. ≤ 5%, HR 6.6, p<0.0001). In a univariate analysis of response, SF3B1 mutational status was associated with lower response rates (mutated vs. non-mutated, OR 0.2, p=.04); in multivariate analysis no predictive factors for response were identified. Importantly and in comparison to prior reports, age, prior transfusion dependence, MDS risk scores, type of IST used (ATG+cyclosporine vs. others), presence of a PNH or LGL clone and HLA DR15 positivity, and mutations were not predictive of response to IST [Figure 2B].

Conclusions: To our knowledge, this is largest reported cohort of patients with MDS treated with IST. While we confirmed that IST is associated with an ORR of 45% and leads to TI in 39% of patients, we could not confirm the predictive value of several previously described biomarkers of response. Apart from SF3B1 mutations, which appear to negatively affect response to IST, no other factors were identified that predicted response to IST. OS was better for patients who achieved an objective response or TI.

Immune Checkpoint Profiling of TP53 Mutant and Wild-Type Myeloid Malignancies: TP53 Mutations Direct Immune Tolerance Via an Immunosuppressive Phenotype

Result Type: Paper
Number: 423
Presenter: David Sallman
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Inflammation, Immunity, and Iron

David A Sallman, MD1, McLemore Amy1*, Rami S. Komrokji, MD1, Kathy McGraw, PhD1, Susan M. Geyer, PhD2*, Erika Eksioglu, PhD, MS1, Najla Al Ali, M.Sc1*, Jeffrey E. Lancet, MD3, Sheng Wei, MD1, Eric Padron, MD1 and Alan F. List, MD1

1Malignant Hematology, H. Lee Moffitt Cancer Center, Tampa, FL
2Health Informatics Institute, University of South Florida, Tampa, FL
3Malignant Hematology, H. Lee Moffitt Cancer Ctr. & Rsrch. Institute, Tampa, FL


Background Augmenting adaptive immunity through immune checkpoint blockade introduced a paradigm shift in the treatment of cancer patients. Although aberrant mRNA upregulation of several immune checkpoints (e.g. PDL1) has been described in CD34+ cells of myelodysplastic syndrome (MDS) patients compared to normal controls (Yang et al., 2014), clinical investigations of checkpoint inhibitors in myeloid malignancies to-date have shown low response rates to monotherapy with a lack of response biomarkers. MDS is genetically heterogeneous evidenced by the array of somatic gene mutations that may impact differential checkpoint display. In concordance, our interrogation of the AML TCGA dataset identified TP53 mutant (MT) AML as an immunologically distinct subset with checkpoint upregulation. Therefore, our goal was to comprehensively characterize checkpoint molecules by lineage and stage of maturation in MDS to discern relationships with immune cell subsets in molecularly defined cohorts.

Patients and Methods Utilizing the Moffitt Cancer Center tissue core and MDS database, bone marrow mononuclear cells were obtained from genetically profiled MDS (n=23), CMML (n=5) and secondary AML patients (n=22). NGS of up to 54 genes was performed at the time of sample acquisition with 5% VAF threshold and 500x minimum depth of coverage. Multi-color flow cytometry acquisition of immune checkpoints with a myeloid and lymphoid panel (Table 1) was performed using a BD LSRII, and analysis was conducted using FlowJo v10.3. Clinical variables and outcomes of patients were characterized at the time of sample procurement. Mann-Whitney and Kruskal-Wallis tests were used for comparative analyses. Median values of each checkpoint were calculated to compare outcomes of patients with increased versus decreased expression. Kaplan-Meier curves were used to estimate OS and analyzed from the date of sample collection. Multivariate Cox regression models were created to adjust for clinical characteristics.

Results A total of 50 patients were analyzed (25 TP53 MT and wildtype (WT), respectively) with a median age of 72 years and male predominance (58%). Overall 38% of patients (n=19) were treatment naïve while 40% were HMA refractory (n=20). TP53 MT patients had significantly higher percentage of PDL1+ hematopoietic stem cells (HSC; Lin/CD34+/CD38,15.5% v 5.7%; P=0.008) with a non-significant increase in progenitor cells (HPC; Lin/CD34+/CD38+, 15.4% v 9.2%; P=0.3). Additionally, PD1+/CD71+ erythroid progenitors (6.5% v 28.1%; P<0.0001)) and granulocytic myeloid derived suppressor cells (MDSC-G (Lin/CD33+/ CD38/CD11b+/CD14); 10 v 23.7%; P=0.001) were significantly decreased in TP53 MT versus WT patients. The percentage of PDL1HSC or HPC were similar in HMA-naïve versus HMA-failure or patients actively on HMA (P=0.61 and P=0.84, respectively). More importantly, TP53 MT patients were enriched for an immune suppressive/anergic profile with increased % of TIM3+ MDSC-G (4% v 2%, P=0.04) and decreased % of OX40 cytotoxic T-cells (2.2 v 2.9%; P=0.006). As TP53 MT strongly predicts for inferior OS (P<0.0001), survival was similarly predictive with checkpoint alterations associated with TP53 MT.

Although there was no difference in % of T-regulatory cells (T-regs; CD3+/CD4+/ CD25+/CD127dim/) in TP53 MT versus WT patients (P=0.75), there was a trend for increased ICOS+ T-regs in TP53 MT patients (16.9% v 13%; P=0.09). Increased ICOS+ T-regs was associated with inferior OS in the total cohort (5.7 v 13.3 months, P=0.003)) and TP53 WT cohort (5.5 months vs NR, P=0.01)). In multivariable analysis incorporating age, BMT status, HMA failure, and revised international prognostic scoring system (IPSS-R) category, increased ICOS+ T-regs remained an independent co-variate for inferior OS in both the total cohort (HR 2.28, 95% CI 1.13 to 4.62; P = 0.02) and TP53WT cohort (HR 4.69, 95% CI 1.33 to 16.56; P = 0.02)).

Conclusion In patients with MDS and sAML, immune checkpoint expression is heterogeneous in HSCs and immune cell subsets. TP53 mutation portends an immunologically distinct, immune evasive phenotype that may be the predominant driver of their poor prognosis. Increased % of tumor infiltrating ICOS+ T-regs is highly predictive of outcome, independent of TP53 mutation status. We conclude that comprehensive immune checkpoint profiling may offer rationale for targeted immunomodulatory therapeutic strategies.

Potential Familial Predisposition to Hematologic Neoplasm Detected Incidentally in Adult Patients Presenting with Myeloid Neoplasms

Result Type: Paper
Number: 585
Presenter: Vincent Funari
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Genes Inform Genes, Morphology, and Clinical Outcomes

Vincent Funari, PhD1*, Maya Thangavelu, PhD2, Wanlong Ma, MS1*, Ivan De Dios, BS1*, Sally Agersborg, MD, PhD1*, Forrest Blocker, PhD1* and Maher Albitar, MD3

1NeoGenomics, Inc., Aliso Viejo, CA
231 Columbia, NeoGenomics, Inc., Aliso Viejo, CA
3NeoGenomics, Valley Center, CA

Introduction: Germline mutations in CEBPA, ETV6, RUNX1, and GATA2 have been reported to be associated with predisposition to hematologic malignancies. Next Generation Sequencing (NGS), used to identify somatic mutations to aid in targeted therapies, may incidentally uncover inherited germline mutations. Recognizing familial predisposition is important as it may impact management and surveillance. We assessed the prevalence of potentially clinically significant germline mutations detected incidentally by NGS in samples from patients investigated for myeloid disorders.

Methods: NGS was performed to identify somatic mutations in 54 genes, on 6100 consecutive bone marrow aspirates or peripheral blood specimens between January 2014 and mid-2017. The average depth of sequencing was 10,000x. DNA was extracted using the QIAamp DNA Mini Kit and subjected to the TruSight Myeloid panel (Illumina, San Diego, CA). Most (92%) of the common dbSNP variants (>10% allele frequency in dbSNP) were found between 45-55% variant allele frequency (VAF) or 98-100% VAF for hetero- and homozygous states. This suggests that common SNPs are mostly found between those frequencies; hence, alleles with these frequencies were defined as candidate germline variants. However, to further enrich for germline variants, we only analyzed those cases with at least one other somatic variant with VAF of 5-40%; the rational being that it is unlikely that a sample would have two true independent somatic mutation events in different subclones. Following this enrichment process, 2951 specimens were eligible for further analysis for potential germline mutations. The ClinVar and dbSNP databases were used to assess allele frequencies and clinical significance.

Results: In 2951 of 6100 specimens, we reported at least one somatic mutation with a variant allele frequency in the range of 5-40%. A total of 5322 variants were identified in this subset of specimens. Among the candidate germline mutations in 4 genes implicated in myeloid neoplasm predisposition, we identified 93 unique putative loss of function variants that met our germline criteria. Nine of these mutations were identified in more than 3 cases and more than half of these were represented in ClinVar as germline loss of function variants. One of them, RUNX1: p.Arg201Gln, identified in 5 patients, is described as a germline pathogenic variant. In contrast, the other 84 variants identified in the 4 genes were rare (observed in 1 or 2 cases) and few (7/84) had been characterized in dbSNP. Most loss of function germline variants are unique and rare with few hotspots, which is consistent with our data.

Analysis of germline mutations in genes, which have not yet been implicated in predisposition to myeloid neoplasm, revealed 4858 candidate mutations in 23 genes: 15.4% (750/4858) were unique, 13.6% (102/750) were more common (> 3 cases), while most (86.4%; 648/750) were found in only 1 or 2 cases. Some of the more common variants were identified in ClinVar as pathogenic germline variants, including IDH2 (p.Arg140Gln; MIM: 613657), identified in 84 cases, and a TP53 loss of function mutation (p.Pro72Arg) identified in 424 cases.

Conclusion: Germline mutations, GATA2, RUNX1, CEBPA, and ETV6, known to predispose to myeloid malignancies, may be identified incidentally during testing for somatic mutations in myeloid malignancies.

In addition, if the NGS panel includes genes other than those known to predispose towards myeloid malignancies, pathogenic variants of wide ranging clinical implication may be uncovered, as observed in the cases with mutation in IDH2 or TP53. Although the specific mutation observed in IDH2 has been identified most often as a somatic mutation, it has been reported to be inherited in one instance. Inheriting this allele can confer D-2-hydroxyglutaric aciduria-2 (D2HGA2). The TP53 mutation observed is known to be involved in other hereditary cancers (e.g. Li-Fraumeni syndrome) and could affect the recommended therapy of antineoplastic agents.

Rare incidentally identified potential germline mutations remain a challenge as their clinical significance may be unknown. These mutations require additional investigation to determine if they could contribute to clinical management.

Hypoplastic Myelodysplastic Syndrome: Clinical, Histopathological and Molecular Characterization

Result Type: Paper
Number: 588
Presenter: ELISA Bono
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Genes Inform Genes, Morphology, and Clinical Outcomes

ELISA Bono, MD1*, Donal McLornan2*, Erica Travaglino3*, Shreyans Gandhi, MBBS, MD, MNAMS2*, Anna Gallì3*, Alesia Abigael Khan4*, Austin G. Kulasekararaj, MBBS, MD, MRCP, FRCPath2*, Emanuela Boveri5*, Kavita Raj2*, Chiara Elena1*, Robin M. Ireland, FRCP2, Antonio Bianchessi1*, Jie Jiang2*, Gabriele Todisco1*, Luca Malcovati, MD1, Judith CW Marsh, MD6, Mario Cazzola, MD1 and Ghulam J Mufti, DM, FRCP, FRCPath2*

1Department of Molecular Medicine & Hematology Oncology, University of Pavia & IRCCS Policlinico S. Matteo Foundation, Pavia, Italy
2Department of Haematological Medicine, King’s College Hospital NHS Foundation Trust, London, United Kingdom
3Department of Hematology Oncology, IRCCS Policlinico S. Matteo Foundation, Pavia, Italy
4Guy’s and St Thomas’ NHS Foundation Trust, London, United Kingdom
5Unit of Pathology, Fondazione IRCCS Policlinico San Matteo, Pavia, Italy
6Department of Haematological Medicine, Kings College Hospital, London, United Kingdom


Introduction and Aims : Hypoplastic Myelodysplasic Syndrome (h-MDS) has distinct features compared to normo- or hypercellular MDS, with a higher response rate to immunosuppressive therapy and more favorable prognosis. However, to date, reproducible diagnostic criteria for h-MDS have not been clearly defined, making differential diagnosis from other bone marrow failure (BMF) syndromes challenging. Based on a large and well-annotated consecutive cohort of patients with BM hypocellularity from two tertiary centres, we aimed to delineate the clinical, histopathological and molecular features of h-MDS.

Methods: Of 1262 consecutive adult patients analyzed at King’s College Hospital, London, UK and IRCCS Policlinico San Matteo & University of Pavia, Italy: 533 patients had a hypocellular BM, including 205 MDS patients with BM cellularity ≤25%, 77 MDS with reduced age-adjusted BM cellularity, 139 with Aplastic Anemia (AA), 97 with a diagnosis of Idiopathic Cytopenia of Undetermined Significance (ICUS) and 15 with congenital BMF (c-BMF) and 729 with normo- or hypercellular MDS (n-MDS).

Results: Comparison of clinical features and outcome of patients with h-MDS as defined by a BM cellularity ≤25% (n=205) and those with a reduced age-adjusted cellularity (n=77) did not detect any significant differences. Therefore, a reduced age-adjusted BM cellularity was adopted as a criterion to define h-MDS (n=282). Compared with patients with n-MDS, those with h-MDS were significantly younger, displayed higher levels of hemoglobin (p=.01), lower neutrophil and platelet counts and lower percentage of BM blasts, ring sideroblasts (RS) and CD34+cells as determined by immunohistochemistry (all p<.001). No significant difference was observed in the distribution of chromosomal abnormalities. A significantly higher prevalence of PNH clones was observed in h-MDS (24% vs 2%, p<.001). Those patients with h-MDS showed a significantly lower risk of leukemic evolution compared with n-MDS (4-year cumulative incidence 14% vs 26%, p=.003). Comparing h-MDS with AA, h-MDS showed a significantly higher percentage of RS (p=.004), BM blasts, dysmegakaryopoiesis and dysgranulopoiesis (p<.001), BM fibrosis (p=.002), clusters of CD34+ cells and chromosomal abnormalities (p<.001). H-MDS patients showed a significantly higher risk of leukemic evolution compared with AA (4-year cumulative incidence 14% vs 0%, p<.001).

Targeted sequencing of 24 commonly mutated myeloid genes was performed on 358 patients (73 h-MDS, 45 AA and 240 n-MDS). Twenty-five patients with h-MDS (34%) showed one or more somatic mutation (median 1; range 1-4). Comparing h-MDS to n-MDS revealed a significantly lower number of mutations per subject and variant allele frequency (VAF) (p<.001). Moreover, a significantly lower number of mutations per patient and VAF were observed in AA compared to h-MDS (p=.031 and p=.003, respectively). Thirty-seven patients with a hypocellular BM had a peripheral blood telomere length (TL) ≤1st percentile and/or a germline mutation consistent with a diagnosis of c-BMF, leading to an estimated prevalence of unidentified c-BMF of 13% in cases without robust morphological evidence of MDS.

Based on these findings, we defined a diagnostic score (h-score), including variables with the highest specificity for MDS: dysmegakaryopoiesis, dysgranulopoiesis, BM fibrosis, clusters of CD34+ cells, RS 2-14%, BM blasts 2-4% (score 1); RS≥15% and BM blasts ≥5% (score 2). By applying ROC analysis, a cut off value of 2 was associated with the highest specificity for MDS (0.98). Notably, 71% of patients with a hypoplastic BM and low h-score (0-1) had no evidence of clonal disease by cytogenetic or mutation analyses or had a mutation pattern consistent with age-related clonal hematopoiesis. These patients showed a significantly better OS and lower risk of leukemic evolution (P<.001).

Conclusions: Integration of cyto-histological and genetic features in patients with a hypocellular BM identifies two distinct groups, one with clinical and genetic features highly consistent with a clonal disease with high risk of leukemic evolution, and one with features more consistent with a non-malignant bone marrow failure. Furthermore, we identified relevant enrichment of patients with previously undiagnosed c-BMF.

Promising Results of a Phase 1/2 Clinical Trial of Ruxolitinib in Patients with Chronic Myelomonocytic Leukemia

Result Type: Paper
Number: 162
Presenter: Eric Padron
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Predicting Drug Response Using Novel Genomic Algorithms

Eric Padron, MD1, Amy E. DeZern, MD2, Sandrine Niyongere, MD3, Markus Christian Ball, PhD1*, Maria Balasis1*, Hanadi Ramadan, BA1*, Jeffrey E. Lancet, MD4, Alan F. List, MD1, Ruben A. Mesa, MD5, Gail J. Roboz, MD6, David P. Steensma, MD7, Mikkael A. Sekeres, MD, MS8 and Rami S. Komrokji, MD1

1Malignant Hematology, H. Lee Moffitt Cancer Center, Tampa, FL
2Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD
3H. Lee Moffitt Cancer Center, Tampa, FL
4Malignant Hematology, H. Lee Moffitt Cancer Ctr. & Rsrch. Institute, Tampa, FL
5UT Health San Antonio Cancer Center, San Antonio, TX
6Weill Cornell Medical College, New York
7Dana-Farber Cancer Institute, Boston, MA
8Taussig Cancer Institute, Department of Hematology & Medical Oncology, Cleveland Clinic Taussig Cancer Institute, Cleveland, OH


Introduction: Chronic myelomonocytic leukemia (CMML) is a myeloid neoplasm characterized by GM-CSF hypersensitivity, marrow dysplasia, blood monocytosis, and a propensity for acute myeloid leukemia transformation. The overall survival is estimated at 34 months with no drug therapies known to improve survival. We previously reported results of a phase 1 study demonstrating that ruxolitinib, a JAK1/2 inhibitor, is safe and has clinical activity in CMML at doses up to 20mg BID. To confirm efficacy we performed a phase 2 clinical trial and here report the combined results of the phase 1 and 2 clinical studies.

Methods: All CMML WHO subtypes were included without regard to previous therapy. Key exclusion criteria included an ANC < 0.25x103 c/dL and a platelet count < 35x103c/dL. The use of GM-CSF analogs was prohibited. The Kaplan-Meier method was used to estimate survival and log-rank test to determine statistical significance. Somatic mutations were identified at baseline using a 49 gene targeted next generation sequencing panel (RainDance Technologies) and a custom 46-plex Luminex assay was used to profile baseline cytokine levels (R&D). The study was registered at NCT01776723

Results: Between April 2015 and October 2016, 29 patients were enrolled in the phase 2 study resulting in a combined 49 CMML patients treated with ruxolitinib. In phase 2, 56 percent of patients had the proliferative subtype (MPN-CMML) defined by WBC >13x109/L and 28% had higher-risk disease by the Global MD Anderson Scoring System. Six patients (20%) received a prior hypomethylating agent and 85% of patients were WHO subtype CMML-1. Therapy-emergent toxicities were similar to those reported in phase 1. The mean duration of therapy was 219 days (range, 6-675 days). Three hematologic responses, including bi- or trilineage responses (2 HI-E, 2 HI-P, 1 HI-N), one marrow CR, and one partial bone marrow response per 2006 IWG criteria were observed. Six of 13 patients (46%) with splenomegaly had a 50% or greater reduction in spleen size by physical exam. When combining IWG and spleen responses, the overall response rate was 46% (n=11 of 24), including clinical benefit by MDS/MPN IWG response criteria. Using the Myeloproliferative Symptom Assessment Form (MPN-SAF) administered electronically daily, a reduction in total symptom score was noted. The median OS (mOS) from start of treatment was 28 months (95% CI 18-38), and, for CMML-MDS and CMML-MPN, was 32 months and 28 months respectively, (p=0.3). To evaluate whether ruxolitinib may have disease-modifying properties in CMML, analysis of mOS from time of diagnosis was performed comparing all ruxolitinib treated patients (n=49) to an historical dataset with over 1800 CMML patients (Padron et al 2015). The mOS at the time of diagnosis in CMML ruxolitinib-treated CMML patients was 69 months compared to 31 months in the historical CMML cohort (p=0.03). Adjusting for age, Global MD Anderson model, CMML WHO subtype, CMML-MPN, and the interval from diagnosis, treatment with ruxolitinib was associated with improved OS (HR 0.42, p <0.005). Although no mutation or class of mutations was associated with response, peripheral blood plasma levels of CXCL9 (p=0.02) and decreased levels of CCL5 (p=0.018) and RAGE (p=0.016) were statistically associated with higher likelihood of response.

Conclusions: Ruxolitinib has promising activity in CMML patients with particular benefit in those with proliferative symptoms and may have disease-modifying activity. The MPN-SAF may be a useful tool to measure symptom burden in the context of ruxolitinib therapy for CMML. Overexpression of cytokines and soluble innate immune receptors is associated with a higher likelihood of response to ruxolitinib therapy. Further clinical study is warranted to validate the biomarkers of response and disease modifying capacity of ruxolitinib in CMML.

Eltrombopag for Refractory Severe Aplastic Anemia: Dosing Regimens, Long-Term Follow-up, Clonal Evolution and Somatic Mutation Profiling

Result Type: Paper
Number: 777
Presenter: Thomas Winkler
Program: Oral and Poster Abstracts
Session: 508. Bone Marrow Failure: Biology and Therapy

Thomas Winkler, MD1, James N. Cooper, MD2*, Danielle M. Townsley, MD, MSc2, Phillip Scheinberg, MD3, Sophia Grasmeder, RN2*, Colin Wu, Ph.D.4*, Jennifer Lotter2*, Marie J Desierto, MS5*, Ronan Desmond, MD6, Matthew J. Olnes, MD, PhD7, Katherine R Calvo, MD, PhD8, Neal S. Young, MD2 and Cynthia E. Dunbar, MD2

1National Heart, Lung, and Blood Institute, NHI, Bethesda, MD
2National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD
3Clinical Hematology, Antônio Ermírio de Moraes Cancer Center, Hospital Sao Jose and Beneficencia Portuguesa, Sao Paulo, Brazil
4Office of Biostatistics Research, National Institutes for Health, Bethesda, MD
5Department of Medicine, Division of Hematology, University of Washington, Seattle, WA
6Consultative Hematologist, Dublin, IRL
7Alaska Native Tribal Health Consortium, Anchorage, AK
8NIH Clinical Center, Bethesda, MD


Eltrombopag (EPAG) received FDA approval for treatment of refractory severe aplastic anemia (rSAA) in 2014, based on our phase I/II dose escalation trial of single agent EPAG at 50-150 mg daily over a period of 12 weeks (Olnes NEJM 2012; Desmond Blood 2014).

Two observations warranted further investigation of EPAG in this unique patient population. First, cell count kinetics and lack of acute toxicities suggested that extended administration of EPAG at a fixed dose of 150mg could speed and improve response rates. Second, 19% of patients developed new cytogenetic abnormalities on EPAG, raising concerns that EPAG might promote progression to MDS/AML.

We conducted a subsequent phase II study of EPAG given at a fixed daily dose of 150mg for 6 months in patients with rSAA (NCT01891994). Thirty-nine patients enrolled between July 2013 and April 2017. Primary endpoint was hematologic response at 6 months. Responding participants could continue EPAG treatment. Secondary endpoints included response at 3 months and the rate of clonal cytogenetic evolution.

Nineteen of 39 (49%) patients met criteria for hematologic response at 6 months. Of these, 5/19 (26%) patients would have been deemed non-responders at 3 months of treatment. EPAG was continued in 18 patients on the extension arm. EPAG was discontinued for robust response in 13/18 (72%) after a median duration of drug administration of 12 months (6-27.5 months). EPAG was re-initiated for relapse in 3/13 patients, and all 3 recovered response.

At median follow up of 6 months (range 2 – 39 m), 6/39 patients (15%) developed marrow cytogenetic abnormalities, a rate comparable to our previous cohort.

Given the similar rates of response and clonal evolution in our two consecutive studies, we analyzed the relationship between outcomes and cytogenetic progression for all patients (n=83) at up to 8 years of follow-up. Sixteen of 83 (18%) patients clonally evolved (Table 1). Clonal evolution was an early event after EPAG initiation. Evolution occurred within 6 months in 13/16 evolvers (81%), and in 6/6 evolvers with high risk chromosome 7 abnormalities (5/6 within 3 months). The frequency of high risk clonal evolution 24 months post intervention is comparable to historic controls with rSAA. However, direct temporal comparisons of evolution events are limited by differences in the sequence of cytogenetic tests. Non-chromosome 7 cytogenetic abnormalities were often transient, and not associated with dysplasia. Two evolvers continued EPAG off the rSAA protocol and cytogenetics normalized (UPN 71,14).

The acquisition and selection of somatic mutations, particularly of myeloid candidate genes recurrently mutated in MDS/AML, has been proposed to be an initiating step in clonal evolution. We performed whole exome sequencing (WES) on samples obtained pre-EPAG treatment and at the primary response endpoint and/or time of clonal evolution in 21 responding patients and in 11 patients with cytogenetic evolution. Candidate gene mutations were detected in patients who responded (6/21) and in those with cytogenetic evolution (4/11). Clonal hematopoiesis without an identifiable driver mutation was common. In post-EPAG samples, additional myeloid candidate gene somatic mutations were detected in 2 cytogenetic evolvers and in 4 responding patients, all at low variant allele frequencies (VAF) that were close to the 2.5% detection threshold. There was no significant change in VAF in either candidate or in non-candidate genes in responders and in cytogenetic evolvers. Only one early cytogenetic evolver (UPN 01) showed an expansion of mutated clones (SETBP1 and RUNX1), from VAF of 5% to 40%, when 80% of bone marrow metaphases showed chromosome 7q deletion.

In summary, extended administration of EPAG at a fixed dose of 150 mg daily for 6 instead of 3 months induces additional responses in a subgroup of patients with rSAA. After EPAG was discontinued, most patients maintained durable robust responses. The temporal relationship between clonal evolution and drug exposure suggests that in a subgroup of patients, EPAG may promote expansion of dormant pre-existing clones with an aberrant karyotype. No clinical or laboratory findings prior to therapy, including WES, predicted response or risk of clonal evolution. Careful monitoring of refSAA patients treated with EPAG is indicated, particularly in the first 6 months of treatment.

Heterodimerization of TPO and IFNγ Impairs Human Hematopoietic Stem/Progenitor Cell Signaling and Survival in Chronic Inflammation


Number : 4
Program: General Sessions
Session: Plenary Scientific Session

Luigi J. Alvarado, PhD*, Alessio Andreoni, PhD*, Heather D. Huntsman, PhD*, Hai Cheng, MD*, Jay R. Knutson, PhD*and Andre Larochelle, MD, PhD

National Institutes of Health, Bethesda, MD

Thrombopoietin (TPO) is the main regulator of hematopoietic stem and progenitor cell (HSPC) self-renewal and survival. Upon binding to its receptor, c-MPL, TPO activates cell signaling through JAK-STAT and other pathways. In recent clinical trials, eltrombopag, a small molecule c-MPL agonist, improved trilineage hematopoiesis in subjects with severe aplastic anemia (SAA). Paradoxically, TPO levels are already markedly elevated in these patients. To explain this paradox, we previously showed that IFNγ, a key proinflammatory cytokine implicated in the destruction of HSPCs in SAA, inhibited TPO signaling in human CD34+ HSPCs cultured in the presence of both cytokines. In contrast, eltrombopag could evade this inhibition in vitro, resulting in improved maintenance of progenitors in clonogenic CFU assays, and long-term repopulating cells in NSG transplantation models compared with TPO-containing cultures (Cheng et al., ASH abstract 2016). In this study, we sought to characterize the mechanisms by which eltrombopag evades IFNγ blockade of c-MPL signaling. Because activation of both c-MPL and IFNγ receptor by their respective ligands induces negative regulatory feedback mechanisms from the SOCS family, we first measured expression of SOCS proteins in CD34+ cells exposed to IFNγ and TPO or eltrombopag. SOCS expression was equally upregulated by TPO and eltrombopag in the presence of IFNγ, suggesting an alternative explanation for eltrombopag’s ability to escape IFNγ-induced perturbation of TPO signaling. Because TPO and eltrombopag distinctively bind to the extracellular or juxtamembrane domain of c-MPL, respectively, we hypothesized that IFNγ may decrease binding affinity of TPO, but not eltrombopag, to c-MPL. We used microscale thermophoresis (MST) to assess the impact of IFNγ on TPO binding to its receptor. As previously reported, we showed that TPO and c-MPL interact via both a high affinity (KD, app<0.11 ± 0.04 nM) and a low affinity (KD, app=1100 ± 130 nM) binding site. In contrast, IFNγ had no interaction with c-MPL. Remarkably, addition of 100-molar excess IFNγ prevented binding of TPO to c-MPL at the low affinity site, but had no impact on the high affinity binding site (KD, app<0.20 ± 0.04 nM) (Panel A). To explain this result, we hypothesized that TPO and IFNγ may directly interact and form heteromeric complexes that hinder binding to c-MPL, whereas the non-peptide small molecule eltrombopag could conceivably evade that process. Strikingly, MST assays revealed a specific, one-site heterodimeric interaction between TPO and IFNγ (KD, app=540 ± 30 nM), while no heterodimer formation was observed between either TPO or IFNγ and other early-acting cytokines (SCF and Flt3L) (Panel B). These data suggest that TPO:IFNγ heterodimers may be responsible for the hindered TPO:c-MPL low-affinity interaction observed in the presence of IFNγ. To confirm this finding, we investigated whether TPO-induced c-MPL dimerization was interrupted in the presence of IFNγ. Using raster image correlation spectroscopy (RICS) to determine average diffusion coefficients of GFP-tagged c-MPL in live cells, we have preliminarily found that IFNγ negatively affects c-MPL dimerization in cells cultured with TPO, but not with eltrombopag. Further studies are underway. Taken together, our data provide a new and provocative paradigm to explain the observed negative impact of the proinflammatory cytokine IFNγ on human HSPC maintenance in bone marrow failure syndromes, and the ability of a small molecule (eltrombopag) to evade this inhibition. We propose that, under chronic inflammatory conditions, IFNγ specifically heterodimerizes with TPO, resulting in (i) occlusion of the low-affinity binding site of TPO to c-MPL, (ii) impaired receptor dimerization, (iii) perturbation of TPO-induced signaling pathways, and (iv) decreased survival of human HSPCs (Panel C). This new understanding could have far-reaching clinical implications for other disorders of chronic inflammation.

Efficacy and Safety of Imetelstat in RBC Transfusion-Dependent (TD) IPSS Low/Int-1 MDS Relapsed/Refractory to Erythropoiesis-Stimulating Agents (ESA) (IMerge)

Result Type: Paper
Number: 4256
Presenter: Pierre Fenaux
Program: Oral and Poster Abstracts
Session: 637. Myelodysplastic Syndromes—Clinical Studies: Poster III

Pierre Fenaux, MD, PhD1, Azra Raza, MD2, Edo Vellenga, MD3, Uwe Platzbecker, MD4, Valeria Santini, MD 5, Irina Samarina, MD6*, Koen Van Eygen, MD7*, Maria Díez-Campelo, MD, PhD8*, Mrinal M. Patnaik, MD, MBBS9, Laurie Sherman, BSN10, Libo Sun, PhD11*, Helen Varsos, MS, RPh11*, Esther Rose, MD11*, Aleksandra Rizo, MD, PhD11 and David P. Steensma, MD12

1Hôpital St Louis, Paris, France
2Columbia Presbyterian, New York, NY
3UMCG, Groningen, Netherlands
4Universitätsklinikum Carl Gustav Carus Dresden, Dresden, Germany
5AOU Careggi-University of Florence, Firenze, Italy
6Emergency Hospital of Dzerzhinsk, Nizhny Novgorod, Russian Federation
7AZ Groeninge – Oncology Centre, Kortrijk, Belgium
8Hosp. Clinico Univ. De Salamanca, De Salamanca, Spain
9Division of Hematology, Mayo Clinic Rochester, Rochester, MN
10Janssen Research & Development, LLC, Spring House, PA
11Janssen Research & Development, LLC, Raritan, NJ
12Dana-Farber Cancer Institute, Boston, MA


Background: There are limited treatment options for anemic patients with lower-risk MDS who are relapsed after or are refractory to ESA. Imetelstat is a novel, first-in-class telomerase inhibitor that targets cells with short telomere lengths and highly active telomerase, and has clinical activity in myeloid malignancies (Cancer Res 2014;4:362; NEJM 2015;373:920; NEJM 2015;373:908; BJC 2016;6:e405). Targeting MDS clones with imetelstat may improve outcomes including anemia in MDS patients relapsed/refractory to ESA. We explored the safety and efficacy of imetelstat in RBC TD patients with IPSS Low or Intermediate-1 (Int-1) risk MDS relapsed/refractory to ESA.

Methods: IMerge is an ongoing 2-part, global, multicenter study in TD patients (planned enrollment 200) with Low or Int-1 IPSS risk MDS relapsed/refractory to ESAs and with transfusion requirement of ≥4 units over 8 weeks prior to study entry. ESA-naïve patients with sEPO >500 mU/mL are also eligible. Part 1 is an open-label, single-arm study assessing the efficacy and safety of imetelstat administered IV every 4 weeks at a dose of 7.5 mg/kg. Part 2 is a double-blind, randomized design comparing the efficacy of imetelstat to placebo. The primary endpoint is the rate of RBC transfusion-independence (TI) lasting ≥8 weeks. Key secondary endpoints include safety, rate of ≥24-week TI, time to and duration of TI, and hematologic improvement (HI) rate. Here we report findings from the 32 patients enrolled in Part 1 with a median follow up of 48 weeks.

Results: 29 (91%) patients had ECOG 0–1, median age was 68.5 years. 59% of patients were IPSS Low and 41% Int-1. 13 patients (43%) had sEPO>500 mU/mL. 34% had a cytogenetic abnormality, including 22% with del(5q). Prior treatments for MDS included ESAs (88%), lenalidomide (38%), and decitabine or azacitidine (HMAs) (25%); 56% of patients were both lenalidomide and HMA naïve. Baseline median RBC transfusion burden was 6 units per 8 weeks (range: 4–14). As of May 2017, RBC-TI ≥8-week was achieved in 34% of patients. Median time from treatment initiation to onset of TI was 8 weeks with a median duration of TI of 19 weeks, while 16% of patients achieved ≥24-week TI. 63% of patients achieved erythroid HI; transient platelet improvement was reported for one patient. Of the 13 lenalidomide and HMA naïve patients without del(5q), 54% achieved ≥8-week TI, 31% had ≥ 24-week TI and 69% achieved erythroid HI. TI response was 35% (6/17) and 38% (5/13) in patients with sEPO level ≤500 mU/L and >500 mU/L, respectively. The most frequently reported adverse events were cytopenias, including grade ≥3 and 4 neutropenia in 66% and 41% of patients, respectively and grade ≥3 and 4 thrombocytopenia in 50% and 19% of patients, respectively. Dose reductions or cycle delays due to adverse events were required for 59% of patients. One patient experienced neutropenic fever and two patients had grade 3 thrombocytopenia concurrent with grade 1 bleeding events both considered related to imetelstat; all recovered without sequelae. 28 patients (88%) had reversible LFT elevations by at least one grade; 4 of these were by 2 grades. Reversible grade 3 elevations were reported in 3 patients.

Conclusions: In IPSS Low/Int-1 RBC transfusion dependent MDS patients relapsed/refractory to ESA, TI was observed in 34% and erythroid HI in 63% with imetelstat therapy. TI response appeared to be independent of sEPO level. Reversible cytopenias were the most frequent adverse events, which were generally manageable with dose reduction or delays. Elevated LFTS were reversible. Patients who were naïve to lenalidomide and HMAs and who lacked del(5q) had a higher rate of durable TI (54%) than other groups, which should be further explored.