ALBERT

Description: Key Points IDH1 promotes leukemogenesis in vivo in cooperation with HoxA9. Pharmacologic inhibition of mutant IDH1 efficiently inhibits AML cells of IDH1-mutated patients but not of normal CD34+ bone marrow cells.

Description: Abstract 1709 Introduction: Allogeneic hematopoietic stem cell transplantation (HSCT) is the only curative therapy for patients with myelodysplastic syndromes (MDS), but it is associated with high mortality and morbidity. Predictors for treatment outcome after HSCT are limited. Recently, mutations in ASXL1 have been described as an independent adverse prognostic marker for MDS patients not undergoing HSCT. The aim of this study was to investigate the prognostic impact of ASXL1 in a large cohort of patients with high risk MDS or secondary AML following MDS (sAML) undergoing HSCT. Patients and Methods: Patients (n=105) with a diagnosis of MDS (34.3%) or sAML (65.7%) who received an allogeneic HSCT at Hannover Medical School between 1996 and 2010 and for whom genomic DNA was available from a time when the disease was active, were evaluated for the presence of ASXL1 mutations by direct sequencing. Patients gave their informed consent in accordance with the Declaration of Helsinki, and the study was approved by the institutional review board of Hannover Medical School. Overall survival (OS) and cumulative incidence of non-relapse mortality endpoints, measured from HSCT, were death (failure) and alive at last follow-up (censored). A Cox proportional hazards model was constructed for multivariate analysis and the two-sided level of significance was set at P

Description: Monitoring of measurable residual disease (MRD) provides prognostic information in patients with Nucleophosmin1 mutated (NPM1mut) acute myeloid leukemia (AML) and represents a powerful tool to evaluate treatment effects within clinical trials. We determined NPM1mut transcript levels (TL) by RQ-PCR and evaluated the prognostic impact of NPM1mut MRD and the effect of gemtuzumab ozogamicin (GO) on NPM1mut TL and the cumulative incidence of relapse (CIR) in patients with NPM1mut AML enrolled in the randomized phase III AMLSG 09-09 trial. 3733 bone marrow (BM) and 3793 peripheral blood (PB) samples from 469 patients were analyzed. NPM1mut TL log10 reduction ≥3 and achievement of MRD negativity in BM and PB were significantly associated with a lower CIR rate, after two treatment cycles and at end of treatment (EOT). In multivariate analyses, MRD positivity consistently revealed as poor prognostic factor in BM and PB. With regard to treatment effect, the median NPM1mut TL were significantly lower in the GO-Arm across all treatment cycles, resulting in a significantly higher proportion of patients achieving MRD negativity at EOT (56% vs 41%; P=.01). The betterreduction of NPM1mut TL after two treatment cycles in MRD-positive patients by the addition of GO led to a significantly lower CIR rate (4-year CIR 29.3% vs 45.7%, P=.009). In conclusion, the addition of GO to intensive chemotherapy in NPM1mut AML resulted in a significantly better reduction of NPM1mut TL across all treatment cycles leading to a significantly lower relapse rate. The AMLSG 09-09 trial was registered at www.clinicaltrials.gov as #NCT00893399.

Description: Background: Internal tandem duplications (ITD) in the receptor tyrosine kinase FLT3 occur in roughly 25% of younger adult patients (pts) with acute myeloid leukemia (AML), implicating FLT3 as a potential target for kinase inhibitor therapy. The multi-targeted kinase inhibitor midostaurin shows potent activity against FLT3 as a single agent but also in combination with intensive chemotherapy. Aims: To evaluate the feasibility and efficacy of midostaurin in combination with intensive induction therapy and as single agent maintenance therapy after allogeneic hematopoietic stem cell transplantation (alloHSCT) or high-dose cytarabine (HIDAC). Methods: The study includes adult pts (age 18-70 years (yrs)) with newly diagnosed FLT3-ITD positive AML enrolled in the ongoing single-arm phase-II AMLSG 16-10 trial (NCT: NCT01477606). Pts with acute promyelocytic leukemia are not eligible. The presence of FLT3-ITD is analyzed within our diagnostic study AMLSG-BiO (NCT01252485) by Genescan-based fragment-length analysis (allelic ratio 〉0.05 required to be FLT3-ITD positive). Induction therapy consists of daunorubicin (60 mg/m², d1-3) and cytarabine (200 mg/m², continuously, d1-7); midostaurin 50 mg bid is applied from day 8 onwards until 48h before start of the next treatment cycle. A second cycle is optional. For consolidation therapy, pts proceed to alloHSCT as first priority; if alloHSCT is not feasible, pts receive three cycles of age-adapted HIDAC in combination with midostaurin from day 6 onwards. In all pts maintenance therapy for one year is intended. This report focuses on the first cohort of the study (n=149) recruited between June 2012 and April 2014 prior to the amendment increasing the sample size; the amendment to the study is active since October 2014. Results: At study entry patient characteristics were median age 54 years (range, 20-70, 34% ≥ 60 yrs); median white cell count (WBC) 48.4G/l (range 1.1-178G/l); karyotype, n=103 normal, n=3 t(6;9), n=2 t(9;11), n=20 intermediate-2 and n=7 high-risk according to ELN recommendations, n=14 missing; mutated NPM1 n=92 (62%). Data on response to first induction therapy were available in 147 pts; complete remission (CR) 58.5%, partial remission (PR) 20.4%, refractory disease (RD) 15% and death 6.1%. A second induction cycle was given in 34 pts. Overall response after induction therapy was CR 75% and death 7.5%. Adverse events 3°/4° reported during the first induction cycle were most frequently gastrointestinal (n=34) and infections (n=81). During induction therapy midostaurin was interrupted, dose-reduced or stopped in 55% of the pts. Overall 94 pts received an alloHSCT, 85 in first CR (n=65 age

Description: Background: T effector cells (Teff) within the stem cell graft in allogeneic hematopoietic stem cell transplantation (HSCT) can elicit disabling acute graft-versus-host disease (aGvHD) contributing to transplant-related mortality. Teff as donor lymphocyte infusion (DLI) are a therapeutic option to re-induce complete remission (CR) after leukemia relapse. Usually DLIs are given in dose escalating regimens until CR is achieved or first signs of aGvHD develop. To monitor the DLI induced allo-immune response and the efficacy of aGvHD treatment is a clinical challenge, since no established biomarkers are available. T regulatory cells (Tregs) are thought to play an important role in balancing immune responses and studies have shown effects of adoptive Treg transfer as a therapeutic option in GvHD. T cell receptor (TCR) sequencing of T cell subsets, such as Tregs, after DLI might allow the identification of clones inducing control of GvHD. Here we report data of 29 DLI patients with a median follow-up of 〉1 year allowing us to thoroughly analyze differences of the TCR repertoire in patients with or without aGvHD. Aims: We aimed to analyze Treg TCR diversity and clonality changes over time as a potential biomarker for development and treatment response of aGvHD following DLI. Patients and Methods: The study cohort consisted of 29 leukemia patients after HSCT who received DLI treatment for recurrence of disease, molecular relapse or high risk phenotype. Blood samples were taken before and after DLI. A median of 4 (range 2-6) blood samples per patient were available for TCR-sequencing. The last sample was taken at a median of 123 (27-530) days post DLI. After generation of PBMCs CD4+CD127-CD25+ Tregs were FACS-sorted for cDNA-based CDR3-region amplification of the TCR-β chain. CDR3 amplicons were sequenced on the Illumina MiSeq platform and annotated using the IMGT.org database. Further bioinformatic analyses were based on VDJtools and the tcR R-package. Results: In 18/29 patients we observed clinical symptoms of aGvHD, with blood samples available of the acute onset in 12/18 cases. Treg frequencies and absolute numbers did not differ between aGvHD and noGvHD samples. Treg TCR diversity, assessed via inverse Simpson's diversity index (1/D) increased on average by 322% at the first occurrence of aGvHD compared to the previous sample (Figure 1A, B). Stratifying for aGvHD severity (total grade 1-2 vs. 3-4) did not reveal any group differences. However, stratifying by organ involvement (skin vs. GI/liver) showed a more pronounced increase of 1/D in patients with aGvHD of the GI tract and/or the liver. Moreover, in 11 subjects blood samples were available a median of 7 (3-14) days prior to aGvHD diagnosis. Already at this preclinical time point we detected an increased 1/D of +361% (Figure 1A, B). Again, aGvHD organ involvement significantly affected the result, with GI/liver involvement mainly driving this effect (+567% vs. +1% 1/D). In contrast, patients who did not develop aGvHD at any time after DLI showed on average a slight decrease of -8% 1/D compared to the previous time point. Next, we analyzed all aGvHD patients with at least one sample available after initiation of treatment (local and/or systemic steroids). Control or remission of aGvHD symptoms was accompanied by decreased 1/D of on average -58% compared to the previous samples (Figure 2). In 9/11 patients we saw a focusing of the Treg TCR repertoire; in the other two (Figure 2, red lines) no or only partial clinical response to systemic steroids was reported. Conclusion: Our data describe detailed changes in the Treg compartment on the TCR level following development and treatment of aGvHD. Currently, aGvHD diagnosis following DLI treatment relies solely on clinical symptoms, deciding whether further dose increments of DLI can be administered. As our data suggest, Treg TCR sequencing may support transplant specialists with (1) detection of patients at risk for aGvHD, even prior to clinical symptoms, (2) identification of patients eligible for further dose increments (compare Figure 1B), and (3) assessment of aGvHD treatment with guidance whether higher dosage of steroids and/or alternative immunosuppressive treatment might be required (compare Figure 2, red lines). Future prospective studies are needed to replicate these findings in a large cohort, potentially enabling identification of specific Treg TCR clones controlling the allo-immune response in aGvHD. Disclosures Ganser: Novartis: Membership on an entity's Board of Directors or advisory committees. Koenecke:BMS: Consultancy; abbvie: Consultancy; Amgen: Consultancy; Roche: Consultancy.

Description: Background:Acute myeloid leukemia (AML) is the second most common form of leukemia and the most frequent cause of leukemia-related deaths in the US. While complete response (CR) rates can be as high as 80% in patients (pts) undergoing initial induction chemotherapy, the majority of pts relapse. Pts who relapse or who fail to achieve CR after first cycle of induction therapy have a bleak prognosis. Selinexor inhibits the major nuclear transport protein exportin (XPO1), which is overexpressed in many cancers. Hematological neoplasms are particularly susceptible to inhibition of XPO1 and undergo apoptosis, while normal hematopoietic cells are largely spared. A Phase I study with selinexor monotherapy shows encouraging results in AML pts (NCT01607892). Selinexor's mechanism of action provides a novel approach and is a suitable agent for combination therapy with Ara-C and idarubicin. Methods: Pts with relapsed/refractory AML received Ara-C (100mg/m2, d1-7) and idarubicin (10mg/m2, d1, 3, 5) every 4 weeks.Two dose schedules of selinexor were administered in combination with chemotherapy: Cohort 1 received 40mg/m2 orally, twice weekly in 4 week cycles while cohort 2 received a selinexor flat dose of 60mg orally twice weekly for 3 weeks. The rationale for employing two dose levels of selinexor was to investigate differences in toxicity and efficacy and to determine the maximum tolerated dose/recommended phase II dose. The data cutoff date was July 28, 2016. Results: Forty-two pts were enrolled into the study between 09/2014 and 06/2016 in 3 German sites. Cohort 1 comprised 27 pats: 16 M, 11 F, median age 58 years (range 22-78), who had received a median of 2 prior treatment regimens (PTR) (range 1-5) including 10 pts (37%) with prior stem cell transplantation (SCT), 33% falling into the unfavourable cytogenetic risk group. Cohort 2 comprised 15 pts: 9 M, 6 F, median age 60 years (range 29-77), who had received a median of 1 PTR (range 1-2). Six patients (40%) had prior SCT and 40% falling into the unfavourable cytogenetic risk group; all 15 pts were evaluable for safety; 11 pts for efficacy (4 pts did not have a follow-up examination due to early death). Only one third of pts had late relapse (〉12 months) in both cohorts. In cohort 1 all 27 pts received induction cycle 1 (IC 1) and one pt received a second induction cycle (IC 2). Two pts had consolidation therapy. For cohort 2 the values were 15, 0 and 2 respectively. The most frequent grade 3/4 toxicities attributable to selinexor for cohort 1 were vomiting (4%, for a median of 22 days), nausea (11%, for a median of 9 days) and diarrhea (52%, for a median of 6 days). For cohort 2 the respective values were vomiting (7%, one day), nausea (7%, 21 days) and diarrhea (40% median of 2 days). The median recovery time to neutrophils ≥ 0.5/nl and platelets ≥50/nl after IC 1 in pts who experienced a CR/CRi was 40 vs. 32 days and 37 vs. 29 days, for cohorts 1 and 2 respectively. One possibly drug related death occurred in cohort 1 (systemic inflammatory response syndrome) and one in cohort 2 (hemophagocytosis syndrome). In cohort 2 three cases (sepsis, pneumonia, asystole) resulted in fatal, although unrelated, outcomes. The overall response rate (ORR) for IC 1 in cohort 1 was 55.5% (22.2% CR, 33.3% CRi). 40% of these received SCT or donor lymphocyte infusions (DLIs), achieving a median relapse free survival (RFS) of 454 days and median overall survival (OS) of 465 days. Of the cohort 1 pts achieving CR/CRi after IC 1 and not being transplanted (N=9), three relapsed resulting in a median RFS of 241 days. Median OS for non-transplanted pts was 339 days. ORR in cohort 2 (N=11) was 54.5% (27.3% CR, 18.2% CRi; 9.1% morphologic leukemia-free state). Of cohort 2 pts that achieved CR/CRi after IC 1, none received SCT or DLI and no relapses had occurred as of the data cutoff date, resulting in a median RFS and OS of 96 days; the observation period is still on-going. As of the data cutoff date, 44.4% of all pts in cohort 1 (N=27) were alive with a median OS of 328 days and a median observation period of 330 days, and 90.9% of pts in cohort 2 (N=11) with a median OS and median observation period of 96 days. Conclusion: The prognosis of relapsed/refractory AML patients is remarkably poor. Our results suggest that combined treatment of Ara-C, idarubicin and selinexor might be an effective and tolerable treatment option and serve as a bridge to transplant. The lower selinexor dose shows better tolerability and should be further explored in this setting. Disclosures Fiedler: Gilead: Other: Travel reimbursement, Gilead; Ariad/Incyte: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees, Other: Travel reimbursement, Amgen, Patents & Royalties: Amgen, Research Funding; Novartis: Membership on an entity's Board of Directors or advisory committees; Koltan: Research Funding; Teva: Other: Travel reimbursement Teva. Heuser:Pfizer: Research Funding; Karyopharm Therapeutics Inc: Research Funding; Bayer Pharma AG: Research Funding; Celgene: Honoraria; Novartis: Consultancy, Research Funding; Tetralogic: Research Funding; BerGenBio: Research Funding. Bokemeyer:Karyopharm: Research Funding.

Description: To integrate available clinical and molecular information for cytogenetically normal acute myeloid leukemia (CN-AML) patients into one risk score, 275 CN-AML patients from multicenter treatment trials AML SHG Hannover 0199 and 0295 and 131 patients from HOVON/SAKK protocols as external controls were evaluated for mutations/polymorphisms in NPM1, FLT3, CEBPA, MLL, NRAS, IDH1/2, and WT1. Transcript levels were quantified for BAALC, ERG, EVI1, ID1, MN1, PRAME, and WT1. Integrative prognostic risk score (IPRS) was modeled in 181 patients based on age, white blood cell count, mutation status of NPM1, FLT3-ITD, CEBPA, single nucleotide polymorphism rs16754, and expression levels of BAALC, ERG, MN1, and WT1 to represent low, intermediate, and high risk of death. Complete remission (P = .005), relapse-free survival (RFS, P 〈 .001), and overall survival (OS, P 〈 .001) were significantly different for the 3 risk groups. In 2 independent validation cohorts of 94 and 131 patients, the IPRS predicted different OS (P 〈 .001) and RFS (P 〈 .001). High-risk patients with related donors had longer OS (P = .016) and RFS (P = .026) compared with patients without related donors. In contrast, intermediate-risk group patients with related donors had shorter OS (P = .003) and RFS (P = .05). Donor availability had no impact on outcome of patients in the low-risk group. Thus, the IPRS may improve consolidation treatment stratification in CN-AML patients. Study registered at www.clinicaltrials.gov as #NCT00209833.

Description: Background: Activating mutations in receptor tyrosine kinases like FLT3 (FLT3mut) lead to an aberrant signal transduction thereby causing an increased proliferation of hematopoietic cells. Internal tandem duplications (FLT3-ITD) or mutations in the tyrosine kinase domain (FLT3-TKD) occur in about 25% of younger adult patients (pts) with acute myeloid leukemia (AML), with FLT3 -ITD being associated with an unfavourable outcome. FLT3mut present an excellent target for small molecule tyrosine kinase inhibitors (TKI). The multi-targeted kinase inhibitor midostaurin (PKC412) is currently under investigation as a FLT3-inhibitor in combination with intensive chemotherapy. Monitoring of the efficacy of such a targeted therapy and correlation of the results with clinical outcome will be of major importance. The plasma inhibitor activity (PIA) assay allows the visualization of the level of dephosphorylation of the target under TKI therapy. Preliminary data suggest a correlation between the grade of dephosphorylation, as a marker for the activity of the TKI, and clinical outcome. Aims: To individually measure the level of FLT3 dephosphorylation by PIA analysis in a large cohort of FLT3-ITD AML pts treated within our AMLSG16-10 trial (NCT: NCT01477606) which combines midostaurin with intensive chemotherapy, and to correlate the results with clinical outcome. Methods: Plasma samples from pts (age 18-70 years) with newly diagnosed FLT3-ITD AML were obtained at different time points for PIA analysis. All pts were enrolled on the ongoing AMLSG 16-10 trial applying intensive therapy in combination with midostaurin (50mg twice a day). For consolidation therapy, pts proceeded to allogeneic hematopoietic stem cell transplantation (alloHSCT) as first priority; pts not eligible for alloHSCT were intended to receive 3 cycles of age-adapted high-dose cytarabine (HiDAC) in combination with midostaurin from day 6 onwards. In all pts one year of maintenance therapy with midostaurin was intended. PIA analyses were performed at defined time points (day 15 of induction, each consolidation cycle, at the end of each treatment cycle, every 3 months during maintenance therapy) as previously described (Levis MJ, et al. Blood 2006; 108:3477-83). Results: So far, PIA analyses were performed in 63 pts (median age, 51.6 years; range, 20-70 years) during (n=63) and after (n=73) first and second induction cycle, during (n=40) and after (n=53) consolidation therapy with HiDAC as well as during maintenance therapy (n=82). During and after induction therapy median levels of phosphorylated FLT3 (p-FLT3) were 46.6% (4.5-100%,

Description: Background The cohesin complex presents a ring structure that regulates chromosome segregation during meiosis and mitosis and is thus essential for cell division. Genes that belong to the cohesin complex in somatic vertebrate cells are SMC1A, SMC3, RAD21 (SCC1), STAG2 (SA-2) and STAG1 (SA-1). Recently, mutations in the cohesin complex have been reported to occur in AML. The aim of this study was to evaluate the clinical and prognostic implications of cohesin mutations in 389 uniformly treated younger AML patients. Patients and Methods Diagnostic bone marrow or peripheral blood samples were analyzed from 389 younger adult patients with de novo (n=348) or secondary AML (n=41) with French-American-British (FAB) classification M0-M2 or M4-M7, who were entered into the multicenter treatment trials AML SHG 0199 (n=276) or AML SHG 0295 (n=113). Patient samples were assessed for frequently occurring mutations including FLT3-ITD, NPM1, DNMT3A, IDH1, IDH2. In the subgroup of cytogenetically normal AML (CN-AML), additional mutation analyses were performed for CEBPA, MLL-PTD, WT1 and WT1 SNP rs16754, NRAS, and expression levels of BAALC, ERG, EVI1, MN1, MLL and WT1. Sequencing of all coding exons of the cohesin genes was performed with the SOLiDTM system (Life Technologie, Darmstadt, Germany). Sequences were analyzed twice separately using the DNAnexus software and a separate pipeline of bioinformatics software. Identified mutations were validated by Sanger sequencing. Results 27 patients (7%) were identified to carry a mutation in one of the cohesin genes. 10 patients harboured a mutation in STAG1. The second most frequently mutated gene in the cohesin complex was SMC3 with 7 patients showing this mutation. 5 patients harboured a mutation in RAD21, 3 patients in STAG2 and 2 patients in SMC1A. All mutations were mutually exclusive. The majority of mutations were missense mutations, apart from 2 frameshift mutations and 3 nonsense mutations. All evaluable patients lost their cohesin mutation during remission and with the exception of one patient regained their mutation during relapse. Patients with mutations in the cohesin genes were older (47 vs 44 years, P=.04), had a higher white blood cell count (WBC) (P=.046) and a higher peripheral blast count (P=.048). We observed a strong correlation between cohesin gene and NPM1 mutations (P=.007). BAALC expression was lower in patients with a mutation compared to patients without a mutation in the cohesin genes (P=.026). In order to get a better understanding of when mutations in the cohesin complex occur during clonal evolution we evaluated the allelic burden of mutations in the cohesin complex. Because of the strong association between NPM1 and cohesin mutations, the allelic ratio of mutated and wildtype NPM1 was compared to the allelic ratio of mutated and wildtype cohesin genes. Interestingly, we found a similar mutation burden between NPM1 and genes of the cohesin complex (Pearson correlation coefficient R=0.67), suggesting that cohesin gene mutations occurred in the same clone as NPM1 mutations. When considering all mutations in the cohesin complex as one group, overall survival (OS), relapse free survival (RFS) and complete remission rate (CR) were not influenced by the presence of cohesin mutations (OS: HR 1.23; 95%CI 0.77-1.99; P=.4; RFS: HR 0.92; 95%CI 0.51-1.65; P=.77; CR: mutated 77% vs wildtype 76%, P=.83). We next evaluated the prognostic influence of each gene in the cohesin complex separately in all patients. STAG1, SMC3 and RAD21 mutations had no influence on OS and RFS. A trend for a reduced OS was observed for STAG2 mutated patients (OS: HR 3.02; 95%CI 0.97-9.46; P=.057), although the analysis is limited by the small number of mutated patients. In the subgroup of patients with cytogenetically normal AML we did not identify a difference in OS, RFS or CR rates for patients with (n=18) or without mutations (283) in the cohesin complex (OS: HR 1.00; 95%CI 0.52-1.92; P=.99; RFS: HR 0.64; 95%CI 0.30-1.39; P=.26; CR rates: mutated 80% vs wildtype 73%, P=.64). Conclusion Mutations in the cohesin genes represent novel, recurrent mutations in AML. We found a strong association between these mutations and mutations in NPM1. Cohesin complex mutations as a group had no prognostic impact in all and in cytogenetically normal AML patients. Disclosures: Schlenk: Amgen: Research Funding; Pfizer: Research Funding; Novartis: Research Funding; Chugai: Research Funding; Ambit: Honoraria.

Description: Background: CBF-AML is defined by recurrent genetic abnormalities which encompass t(8;21)(q22;q22), inv(16)(p13.1q22) or less frequently t(16;16)(p13.1;q22). Most frequent secondary chromosome aberrations in t(8;21) AML are del(9q) or loss of a sex chromosome, and in inv(16)/t(16;16) AML trisomy 22 or trisomy 8. At the molecular level mutations involving KIT, FLT3, or NRAS were identified as recurrent lesions in CBF-AML. However, the underlying genetic alterations which might trigger relapse in CBF-AML are not well delineated. Thus, the aim of our study was to characterize the clonal architecture of relapsed CBF-AML. Methods: We performed mutational profiling (KIT, FLT3-ITD, FLT3-TKD, NRAS, ASXL1) in paired samples obtained at diagnosis and at relapse from 66 adults with CBF-AML [inv(16), n=43; t(8;21), n=23] who all were treated within the AMLSG studies. Results: In inv(16) AML, the following mutation pattern was identified at diagnosis: KIT 13/40 (33%; exon 8, n=6; exon 17, n=5; exon 8+17, n=1; exon 11, n=1; missing data, n=3), NRAS 18/43 (42%), FLT3-TKD 4/43 (9%); none of the pts harboured FLT3-ITD or ASXL1 mutations. At the time of relapse, there was a shift in the mutation pattern in 26 pts (60%): KIT mutations (exon 8, n=5; exon 17, n=2; exon 8+17, n=1) were lost in 8 pts and 1 pt acquired an exon 17 KIT mutation; similarly, 15 pts lost and 1 pt gained NRAS mutation, respectively. Of note, all FLT3-TKD mutations were lost at the time of relapse, and only one pt gained a FLT3-ITD mutation. Based on these findings we calculated the stability in inv(16) AML for KIT, NRAS and FLT3-TKD mutations as 38%, 17%, and 0%, respectively. AML with t(8;21) presented a different diagnostic mutation profile: KIT 9/23 (39%; exon 17, n=8; exon 11, n=1), FLT3 -ITD 3/23 (13%), NRAS 2/23 (9%), and ASXL1 1/23 (4%); there were no FLT3-TKD mutations. At the time of relapse, the mutation pattern changed in 9 pts (39%); KIT mutations were lost in 4 pts (exon 17, n=3; exon 11, n=1), but acquired in 2 pts with both of them located in exon 17; only 1 pt lost the NRAS mutation. FLT3-ITD was lost in 2 and gained in 3 pts. There was no change in the ASXL1 mutation status. Thus, the stability for KIT, NRAS, FLT3-ITD and ASXL1 mutations in t(8;21) AML was calculated as 56%, 50%, 33% and 100%, respectively. Of note, mutations affecting the KIT and NRAS gene were almost mutually exclusive; there were only 3 pts with concurrent KIT and NRAS mutations at diagnosis [inv(16), n=2; t(8;21), n=1]. Conclusion: CBF-AML cases display a high degree of molecular heterogeneity with shift of the mutation pattern at relapse in both CBF-AML subtypes. The frequent loss of KIT and NRAS mutations at relapse suggests that there might be other important secondary lesions driving relapse. Ongoing high-resolution genome-wide profiling will further unravel the clonal hierarchy and genomic landscape in CBF-AML. Disclosures Götze: Novartis: Honoraria; Celgene Corp.: Honoraria. Greil:Celgene: Consultancy; Ratiopharm: Research Funding; Sanofi Aventis: Honoraria; Pfizer: Honoraria, Research Funding; Boehringer-Ingelheim: Honoraria; Astra-Zeneca: Honoraria; GSK: Research Funding; Novartis: Honoraria; Genentech: Honoraria, Research Funding; Janssen-Cilag: Honoraria; Merck: Honoraria; Mundipharma: Honoraria, Research Funding; Eisai: Honoraria; Amgen: Honoraria, Research Funding; Cephalon: Consultancy, Honoraria, Research Funding; Bristol-Myers-Squibb: Consultancy, Honoraria; AOP Orphan: Research Funding; Roche, Celgene: Honoraria, Research Funding. Schlenk:Boehringer-Ingelheim: Honoraria; Teva: Honoraria, Research Funding; Janssen: Membership on an entity's Board of Directors or advisory committees; Novartis: Honoraria, Research Funding; Daiichi Sankyo: Membership on an entity's Board of Directors or advisory committees; Pfizer: Honoraria, Research Funding; Arog: Honoraria, Research Funding.