ALBERT

Description: INTRODUCTION Anemia is the most frequent cytopenia in lower-risk MDS. Erythropoietic-stimulating agents (ESAs) are commonly used in these patients. The use of ÒclassicalÓ parameters (EPO and ferritin levels) and the revised IPSS (IPSS-R) has been proposed1 (SantiniÕs score) to predict response to ESAs and overall survival (OS) among patients with lower risk MDS by IPSS and a favorable Nordic group score2. OBJECTIVES The main objective of the study was to evaluate overall response rate (ORR) to ESAs and OS according to the proposed SantiniÕs score in an independent and large cohort of anemic lower risk MDS patients receiving treatment with ESAs. METHODS Data from 530 anemic patients with low/int1 risk IPSS de novo MDS (according to FAB and WHO criteria) and sufficient follow-up data available were recorded in Spresas3 (SPanish Registry of Erythropoietic Stimulating Agents Study from GESMD). Two hundred and twenty six patients (42.6% of the patients) were selected according to specific criteria regarding the published SantiniÕs score1: Hb level 350 ng/mL(=1) and IPSS-R very low=0, low=1, intermediate=2 and high=3) yielded a score ranging from 0 to 5. ESAs response rate and overall survival were analysed according to these score. Response to treatment was evaluated according to IWG 2006 response criteria and a multivariate logistic regression analysis was used to identify independent predictors of erythroid response (ER). OS were defined as the time between diagnosis and the corresponding event or last follow up (Feb 2015) and were analyzed using univariable and multivariable Cox proportional hazards regression methods. RESULTS Median age was 77 years (interquartile range [IQR] 25%-75%: 71-83 y), median Hb level at start of treatment was 10 g/dL (IQR25-75: 9-10), median EPO level was 90 (IQR25-75: 27,25-108) and median ferritin level was 338,5 (IQR25-75: 146,5-568,75). Among 139 patients with this data available, 85 patients (61,1%) were RBC transfusion dependent before ESAs treatment. Median time from diagnosis to ESAs treatment was 82 (IQR25-75: 27-353) days. According to the IPSS, 68.6% (N=155) and 31.4% (N=71) were in low and Int-1 risk groups, respectively. Regarding IPSS-R, 23% (N=52), 66.8% (N=151), 9.7% (N=22) and 0.4% (N=1) were in very low, low, intermediate and high risk, respectively. ORR to ESA treatment was 71.2% (N=161), with a median duration of response of 2.06 years. Prognosis factors of ER showed a trend toward to a higher ER among patients in the lower IPSS-R (P〉0.05), low IPSS (p=0.039) and lower EPO levels (p

Description: Abstract 1750 Poster Board I-776 Chronic myelomonocytic leukemia (CMML) is a clonal disorder sharing features of both myelodysplastic syndromes (MDS) and chronic myeloproliferative disorders (MPD). The natural course of CMML is highly variable. Several small series have suggested the prognostic importance of different characteristics but a widely accepted prognostic scoring system for CMML is not available. The main aims of the study were to identify prognostic factors, including cytogenetic findings, for overall survival (OS) and acute leukemic (AL) transformation in a large series of patients with CMML and to develop an easily applicable prognostic scoring index for estimating outcome and planning treatment in the individual patient. Five hundred and seventy-two patients diagnosed of CMML according to FAB and WHO criteria in 25 centers belonging to the Spanish Registry of MDS were included in the study. Actuarial curves of OS and risk of AL evolution were built by Kaplan-Meier method and differences between curves compared with log-rank tests. Multivariate analyses of OS and risk of AL evolution were performed by Cox proportional hazards regression method. The weights assigned to the variables included in the final prognostic scoring system were based on the regression coefficients from the proportional hazards models. Median age was 73 yr and 397 (69%) were males. According to FAB criteria 61% of the patients had MDS-CMML (absolute WBC count '13 × 109/L) and 39% MPD-CMML and by WHO classification 86% were CMML-1 (blasts

Description: Background and Aim Although new agents have been approved for the treatment of MDS, the only curative approach for these patients is allogeneic hematopoietic stem cell transplantation (HSCT). Nevertheless, in these patients this approach has only obtained 40-60% of overall survival. Somatic mutations in MDS have recently been analyzed in order to confirm clonally and also prognostic impact in MDS patients. In this regard, TP 53 mutated gene is present in MDS in less than 10% of patients and is associated with advanced disease and high-risk features. Recent studies confirms poor outcomes in patients with TP 53 mutated receiving allogeneic stem cell transplantation1,2. The present study try to analyze if the development of chronic graft versus host disease (cGVHD) could modify, due to graft versus leukemia effect, the adverse prognosis of these high-risk patients (TP53 mutated patients). Design and Methods Results of HSCT in 92 MDS patients from 5 centers in Spain were retrospectively studied. Samples were collected 1 month prior to transplant. 280ng of the genomic DNA from BM cells was screened for somatic mutations in TP53 gene. The study was done by NGS on a GS Junior Instrument (Roche) according to an amplicon sequencing design. For each sample, eight exons (4-11) were amplified with preconfigured primer plates provided within the IRON II study network. Data analysis, were carried out using the Sequence Pilot software version 3.5.2 (JSI Medical Systems) and GS Amplicon Variant Analyzer software, versions 2.7 and 2.9 (Roche Applied Science). Minimum coverage of sequenced exons was 100 reads and the sensitivity of variant detection was set to a lower limit of 〉2% for bidirectional reads. Only those variants that resulted in amino acid change in the protein sequence were considered. OS and RFS were calculated using the Kaplan-Meier method. The log-rank test was used for comparisons. All calculations were done using SPSS 18.0. Cumulative incidence of relapse was also calculated by xlstat version 2014 program. Results Median age was 54 years (17-69), 71.7% were "de novo" MDS and regarding IPSS, 53% were in the int-2/high-risk category. Other characteristics were in Table 1. In the pre-transplant evaluation, 15 patients out of 92 (16,3%) were TP 53 mutated. The mutations were located in exons 5, 6, 7, 8 and 10. These variations were present in a variable percentageof the cell population (3 to 84%). All mutations were specific nucleotide changes except for two cases. At the time of the last update, 16 patients had relapsed (17.4%) and 40 had died (43.5%). After a median follow up of 15.5 months, OS was 56.5%. Median OS for patients with mutated TP53 trend a toward to be shorter than survival for patients without mutated TP53 (median of 7 mo vs median not reached, respectively, p=0.156). Multivariate analysis for OS confirmed complex karyotype (HR 5,588, 95CI 1,794-17,407, p=0.003) and no developement of cGVHD (HR 3,531, 95IC 1,634-7,632, p=0.001) as predictors for poor outcome. Cumulative incidence of relapse was 20.3% (+/-4.3%) at 1 years. Mutational status of TP53 significantly influenced on relapse (53.3% +/-12.9% vs 13.7% +/-4% at 1 year for patients with vs without TP 53 mutation (Gray test=0.001, Figure 2). Regarding Relapse Free Survival (RFS), after a median of follow up of 17 months, RFS was 67.9% and as previously suggested, the presence of TP 53 mutation had an impact on RFS (41.7% for mutated (median RFS of 6 months) and 75% for non mutated patients (median RFS not reached), p=0.009). Multivariate analysis for RFS confirmed age (HR 1.054, 95CI 1.005-1.106, p=0.032) and TP 53 mutated (HR 3.054, 95IC 1.145-8.149, p=0.026) as predictors for lower RFS. Regarding 15 patients with mutated TP 53, 7 did relapsed and 9 had died. Developement of cGVHD showed a trend toward to improve outcome among TP 53 mutated patients, with a better OS and RFS for those developing cGVHD as compared to those who did not (OS of 55% vs 17% for patients with and without cGVHD, p=0.039, Figure 2 and RFS of 71% vs 50%, respectively, p=0.3). Conclusions Mutated TP53 pre-allo patients presents poor outcome as compared to not mutated, as previously described Bejar1 and Kim2. Nevertheless, the developement of cGVHD could overcome the adverse impact of this factor due to the developement of graft versus tumor efect, improving survival curves (OS and RFS) as compared to previous published results. Study supported by GRS-1033/A/14 P53. 1.-BŽjar, JCO 2014, 32(25). 2.-Kim, BBMT 2015, Epub ahead of print. Figure 2. Figure 2. Figure 3. Figure 3. Disclosures Sanz: JANSSEN CILAG: Honoraria, Research Funding, Speakers Bureau. Valcarcel:AMGEN: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; NOVARTIS: Honoraria, Membership on an entity's Board of Directors or advisory committees; GSK: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; CELGENE: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Díez-Campelo:CELGENE: Research Funding, Speakers Bureau; JANSSEN: Research Funding; NOVARTIS: Research Funding, Speakers Bureau.

Description: Numerous recurrently mutated genes have been described in the last years for acute myeloid leukemia (AML). Nevertheless, only a few of them (NPM1, CEBPA, FLT3), and usually individually analyzed, have been systematically incorporated into clinical laboratories to stratify prognosis and guide risk-adapted therapy. So far, technology made impossible their simultaneous study, however, development of high throughput techniques as "next generation sequencing" (NGS) allows a parallel assay. Recently, our group designed a NGS panel (custom panel) interrogating the complete coding sequence of 87 genes potentially involved in AML. Interactome analysis revealed that 13 genes were highly implicated in AML pathogenesis. Although NGS has been broadly utilized in the investigational scope, its application to the routine diagnosis is still far from become a standard. The aim of this study is to validate the clinical applicability to routine laboratories of a hotspot NGS panel that includes these 13 genes and other potentially actionable targets. We included 62 "de novo" AML patients in which we tested the Ion Ampliseq AML community panel (Life Technologies) (IAAC panel) in the Ion PGM/Proton platforms. This panel includes hotspots of ASXL1 (exon 12), BRAF (V600E), CBL (exons 8-9), FLT3 (codons 676, 830-850), IDH1 (exon 4), IDH2 (exon 4), JAK 2 (exon 14), KIT (exons 8, 10, 11 and 17), KRAS (exons 2-4), NRAS (exons 2-4), PTPN11 (exons 3,7,8,13), RUNX1 (exons 3-8) y WT1 (exons 7 and 9), and the entire coding sequence of CEBPA, DNMT3A, GATA, TET2 and TP53. The design does not include the FLT3-ITD region. This panel requires only a total of 40 ng of DNA, far less than the custom panel (250 ng), which is very convenient in a clinical laboratory, where the input sample can be limited. The hotspot NGS panel detected 153 variants in 62 patients (2.47 mutations/patient). Mean read depth and uniformity were 1580 and 94.89%, respectively. The IAAC panel detected 25 NPM1 mutations, 20 TET2 and DNMT3A mutations, 14 in CEBPA, 12 in TP53, 11 in RUNX1, 8 in FLT3, 7 in IDH2, 6 in ASXL1, GATA2 and NRAS, 5 in PTPN11 and KRAS, 3 in WT1, 2 in CBL and 1 mutation in IDH1, KIT, and BRAF. Only 4 patients (6%) remained wild-type for these genes after the analysis with the IAAC panel. These samples had previously been analyzed by conventional molecular biology techniques (CMBT) for FLT3-D835, NPM1-T288, DNMT3A-R882 and CEBPA. The IAAC panel found 100% of these previously known mutations plus 5 extra mutations that were negative by CMBT. These mutations were reconfirmed by Sanger sequencing. Therefore, the IAAC panel detects more mutations than CMBT. To further assess the suitability of the hotspot panel, we selected a subset of 25 patients that had been also analyzed by the custom panel (Sure Desing Tool (Agilent) and an Illumina platform). This allowed us to analyze the complete coding sequence of the genes included in the IAAC panel and thus look for mutations outside the hotspots. A total of 53 variants were found with the custom panel. Fifty of these variants (95%) were also detected with the IAAC panel. The remaining 3 variants (5%) were located outside the hotspot regions. Additionally, the IAAC panel detected 13 variants in the overlapping regions that were not found with the custom panel. This could be explained because when focusing on recurrent regions of particular genes, it is possible to increase the mean read depth and therefore reach higher sensibility, which can be achieved with the IAAC panel but not with the custom panel. In conclusion, the Ion Ampliseq AML community panel detects mutations currently analyzed in most of clinical laboratories with validated prognostic relevance (NPM1 and CEBPA) in one assay, with low sample input requirement and with 100% sensibility compared with CMBT. In addition, this panel is able to find out alterations in these genes that are lost by CMBT. Moreover, other mutations with probable diagnostic or prognostic value and/or potential therapeutic targets are also studied and identified with high sensibility. Only a few changes are excluded of the covered regions. Therefore, IAAC panel is useful for routine diagnosis; however, detection of FLT3 internal tandem duplications is not possible, which limits its clinical utility. Disclosures No relevant conflicts of interest to declare.

Description: Key Points Risk assessment is crucial in patients with CMML because survival may range from a few months to several years. Integrating clinical features, morphology, and genetic lesions significantly improves risk stratification in CMML.

Description: Transfusion dependency seems to have a major prognostic impact in patients with myelodysplastic syndrome (MDS) (Malcovati L et al. J Clin Oncol2007;25:3503). Preliminary data also suggest that the development of iron overload could influence outcome (Malcovati L et al. J Clin Oncol2005;23:7594 and Garcia-Manero G et al. Leukemia2008;22:538), but small numbers have precluded a meaningful analysis of the prognostic value of this characteristic. The main aim of this study was to evaluate the independent prognostic value of transfusion dependency (as defined in WHO-based Prognostic Scoring System [WPSS]) and iron overload (defined as serum ferritin level 〉1,000 ng/mL) in a large series of 2,994 patients (median age, 74 yr) with de novo MDS according to FAB criteria (2,107 MDS according to WHO criteria). Complete transfusional history was available in 2,241 patients (835 transfusion dependent [TD] at diagnosis, 526 TD during follow-up, and 880 non-TD) and serum ferritin levels in 1,634. Karyotyping was successfully performed in 2,074 patients, who could then be classified by the International Prognostic Scoring System (IPSS) as low (861 patients), intermediate-1 (748), intermediate-2 (311), and high-risk (154). The numbers of patients in the five risk categories defined by the WPSS (available for 1,228 patients) were 257 (21%) in very low, 385 (31%) in low, 217 (18%) in intermediate, 271 (22%) in high, and 98 (8%) in very high, closely similar to those reported in the original WPSS series. Actuarial curves of overall survival (OS) and risk of evolution to acute myeloblastic leukemia (AML) were built by Kaplan-Meier method and differences between curves compared with log-rank tests. Multivariate analyses of OS and risk of evolution to AML were performed by Cox proportional hazards regression method, with development of transfusion dependency and iron overload entered as time-dependent covariates. Other variables included in the prognostic factor analyses were age, gender, hemoglobin level, absolute WBC, PMN, and platelet counts, proportion of blasts in blood and marrow, percentage of dysplastic features in the three different hematopoietic cell lines, cytogenetics according to IPSS cytogenetic risk subgroups, FAB and WHO classifications, ferritin, beta-2 microglobulin, erythropoietin, and LDH levels at diagnosis, and IPSS and WPSS risk categories. All the previous variables showed a statistically significant relationship with OS and/or AML risk on univariante analyses. Median OS for TD patients at diagnosis, TD patients during evolution, and non-TD patients was 19, 60, and 96 months, respectively (P

Description: Deletion of the long arm of chromosome 5 is the most frequent chromosomal abnormality in MDS (10–15% of MDS cases). Patients with del(5q), particularly those with the ‘5q-syndrome’ have a much better prognosis than other MDS subtypes. Although the presence of additional chromosome abnormalities (ACA), apart from 5q-, has been suggested to negatively influence this favourable outcome, the exact prognostic impact of ACA remains unknown. The aim of the present study was to analyse the prognostic value of ACA in a large series of patients with MDS with 5q- abnormality, treated with supportive care. Three-hundred and five MDS patients with del(5q) were selected from a 3128 cases database that included 1004 patients from the Spanish Haematological Cytogenetics Working Group (GCECGH) (Solé et al., 2005) and 2124 patients from the German-Austrian MDS Study Group (Haase et al., 2007). Patients were separated into two groups: group A (n=204), all del(5q) cases as a single anomaly and group B (n=101) with additional cytogenetic anomalies. Patients in Group B were subdivided according to: the number of additional anomalies (1 to 3 5 anomalies); and the type of additional cytogenetic aberrations: chromosomes 1 and 3, monosomy 7, 7q-, trisomy 8, trisomy 11, trisomy 13, 12p-, involvement of chromosome 17, -18/18q-, 20q-, trisomy 21, loss of X/Y chromosome, and unrelated clones. The series includes 90 males (29.5%) and 215 females (70.5%) with a median age of 66 years (range: 3–92 yr). Using FAB criteria (n=294): 52% had RA, 9% RARS, 30% RAEB, 8% RAEB-t and 1% CMML. WHO classification was available for 217 patients: 52% had ‘5q- syndrome’, 1% RA, 0% RARS, 2% RCMD, 2% RSCMD, 13% RAEB-1, 20% RAEB-2, 1% CMML, 8% AML and 1% were unclassifiable. Overall, 204 (67%) of the patients presented 5q- isolated, 52 (17%) 5q- with one additional abnormality, 10 (3%), 6 (2%), 7 (2%) and 26 (9%) with 2, 3, 4 and 5 or more additional abnormalities, respectively. Follow-up data were available for 273 patients (89.5%). Median survival was 48 months for all. Median survival for patients with isolated del(5q), with one additional abnormality and with two or more additional abnormalities (complex karyotypes) was 69, 55 and 8 months, respectively (P

Description: Background: Lenalidomide (LEN) is an immunomodulatory drug which binds cereblon through a glutarimide ring modulating the substrate specificity of CRL4CRBNE3 ubiquitin ligase complex, resulting in the proteosomal degradation of specific disease-related proteins. LEN is approved for the treatment of RBC transfusion-dependent (TD) low and int-1 risk MDS patients with del(5q), 70% of whom reach RBC transfusion-independence (TI) and 50% complete cytogenetic remission. It is also under investigation in RBC -TD low and int-1 risk MDS without del(5q) resistant to erythropoietin stimulating agents and with 20-30% of patients reaching RBC-TI. Herein we aimed to study whether molecular mutations in MDS patients with and without del(5q) influenced the response to LEN treatment. Methods: We collected 95 MDS patients treated with LEN, 72 with del(5q) and 23 without del(5q) (23/23: normal karyotype, presence of ringed sideroblasts and SF3B1MUT). Retrospective clinical information was available for 65 patients. To characterize the mutational spectrum of patients with del(5q) and non del(5q), we combined results from multi amplicon targeted sequencing Ion Torrent (44 cases) and from captured-based targeted deep sequencing (51 cases). The Ion Torrent panel included 39 of the most frequently mutated genes in MDS (ASXL1, BCOR, BRAF, CBL, CDKN2A, CEBPA, DNMT3A, ETV6, EZH2, FLT3, GNAS, IDH1, IDH2, JAK2, KIT, KRAS, LUC7L2, MPL, NF1, NPM1, NRAS, PHF6, PTPN11, RAD21, RPS14, RUNX1, SETBP1, SF1, SF3A1, SF3B1, SMC3, SPARC, SRSF2, STAG2, TET2, TP53, U2AF1, WT1, ZRSR2). The Ion Torrent panel was then extended for the captured-base sequencing and updated with the addition of 43 genes including BCORL1, CALR, CSNK1A1 and KDM6A. The average depth per gene was 780x per sample. Results: Patients with del(5q) vs. non del(5q) had an equal risk distribution (LR: 46/55 vs. 21/22; HR: 9/55 vs. 1/22). We restricted our analyses to LR patients. Median age was 69 yrs (34-90) and M:F ratio was 15:50. A significantly higher median of hemoglobin (Hb) levels were found in del(5q) in comparison to non del(5q) patients (9 g/dL (6-13) vs. 8 g/dL (7-9); P=.001). Molecularly we found that a slightly lower number of mutations occurred in del(5q) [2 mutations (0-6)] compared to the number of mutations in non del(5q) [3 mutations (1-7); P=.001]. WHO 2008 distribution was significantly (P

Description: Background and Aim:It is increasingly recognized that patients with a de novomyelodysplastic syndrome (MDS) onset as young adults, lacking any other feature of a congenital disorder, may share a pathogenic overlap due to the presence of both germline and somatic variants. Identifying an inherited pathogenic variant has important therapeutic implications beyond family counselling: adapting the selection of sibling donor, the use of highly cytotoxic therapy and the monitoring for other cancer development. However, most studies have focused on patients with suspected inherited disorders based on the presence of physical abnormalities and/or family history. In addition, a mixture of pediatric and adult cases is usually reported. The aim of this study is to characterize the germline and tumor variants in a group of adult MDS patients without accompanying congenital physical anomalies and or family antecedent of bone marrow failure. Methods: We included 72 patients from 15 Spanish centers with a diagnosis of MDS between 18 and 60 years old (y.o). Patients with a previously diagnosed or suspected (one physical anomaly or family history) congenital syndrome were excluded. Diagnoses were made in accordance with the WHO 2016 classification. Whole-exome sequencing (WES) libraries were prepared using SureSelectXT Target Enrichment and sequenced on a HiSeq4000 platform (IlluminaInc.). Mean number of reads per sample was 138,726,017 with a Phred Quality Score 〉30 in 95.05% of bases. Read mapping sequence alignment and variant calling were performed using Biomedical Workbench (Qiagen). WES was performed on 72 tumor and 32 paired germinal DNA (buccal swab). To identify potential germline-causal mutations, a selection tool was implemented incorporating 239 genes associated with cause or predisposition to bone marrow failure or cancer. Variants with an ExAC, TOPMed and/or European 1000 Genomes minor allele frequency ≥0.01 were discarded. Results: The median age at diagnosis was 49 y.o. The cohort was categorised into two groups, less or equal 50 y.o. (62.5%) and between 50 and 60 y.o. (37.5%). In the first group, the frequency according to the WHO classification were 12% MDS with single lineage dyplasia (MDS-SLD), 8% MDS with ring sideroblasts (MDS-RS), 11% MDS with multilineage dyplasia (MDS-MLD), 24% MDS with excess blasts(MD-EB), 4% MDS with isolated del(5q)(MDS-del5q), 4% MDS unclassifiable and 4% chronic myelomonocytic leukemia (CMML). Meanwhile, in the group with age more than 50 y.o., the subtypes were 3.7% MDS-SLD, 7.4% MDS-RS, 29.6% MDS-MLD, 40.7% MD-EB, 3.7% MDS-del5q, and 14.8% CMML.Patients less or equal 50 y.o. were stratified based on IPSS-R as very low (4%), low (64%), intermediate (20%), high (12%) and very high (0%); and the group of more than 50 y.o. as very low (14.8%), low (33.3%), intermediate (29.6%), high (11.1%) and very high (11.1%).The mean number of somatic mutations was 0.68 in patients with less or equal 50 y.o. and 1.37 in those between 50 and 60 y.o., p=0.033 (U Mann-Whitney); and regarding germline variants, the first group mean number was 2.44 (p25-75, 1-3) and the second group showed a mean of 1.85 (QI 25-75, 1-3), p= 0,331.In the whole cohort, germline variants were found in 62 out of 72 patients, with the following frequencies: ATR(N=5, 6.9%), followed by BARD1(N=5, 6.9%), ERCC6L2(N=4, 5.6%), MSH6(N=4, 5.6%), TCIRG1(N=4, 5.6%), NBEAL2(N=4, 5.6%), ASXL1(N=3, 4.2%), ATM(N=3, 4.2%), MPL(N=3, 4.2%), NF1(N=3, 4.2%), RECQL4(N=3, 4.2%), SAMD9L(N=3, 4.2%), WRN(N=3, 4.2%).Among germline variants, those reported previously as pathogenic or likely pathogenic, or involving genes associated with familial MDS/AML included: ERCC6L2(N=4, 5.6%), SAMD9L(N=3, 4.2%), and one case mutated for DDX41, FANCC, JAK2, MSH6, SETBP1, MUTYH, BRCA1and RECQL4. In the whole cohort, somatic variants were found in 38 patients, with the following frequencies: TP53(N=7, 9.7%), ASXL1(N=7, 9.7%), SETBP1(N=5, 6.9%), NF1(N=5, 6.9%), SRSF2(N=4, 5.5%). Conclusion:In this subset of young adults with de novo MDS without congenital anomalies and/or familial history suggesting the presence of an undiagnosed congenital syndrome, 18% of the cohort harbored a likely causative germline variant. In addition, we noted a predominance of variants affecting genes with a cancer predisposition limited to the hematopoietic system, rather than classical telomere, DNA damage genes with an established mendelian link. Table. Table. Disclosures Díez-Campelo: Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau.