No abstract

2792 Introduction: Somatic mutations of key candidate genes have gained interest as biomarkers predicting poor survival in myelodysplastic/myeloproliferative neoplasms (MPN) and myelodysplastic syndromes (MDS). RUNX1 (runt-related transcription factor 1) deregulations constitute such a disease-defining molecular aberration and are usually tested applying a combination of denaturing high-performance liquid chromatography and direct Sanger sequencing. Patient-specific RUNX1 mutations were proposed to represent clinically useful molecular alterations to follow disease progression from MDS to s-AML. Study design: Using genomic DNA obtained from mononuclear cells a next-generation amplicon deep-sequencing (NGS) assay targeting the complete coding region of RUNX1 was developed on a longitudinal series of 116 retrospective samples obtained from 25 patients collected between 11/2005 and 6/2010 (454 Life Sciences, Branford, CT). Subsequently, this assay was applied to characterize an unselected prospectively collected MPN/MDS patient cohort during their course of disease. Results: Here, we present analyses on a cohort of 534 patients (females: 200; males 334). The median age was 72.0 years (25.2–95.7 years). The cohort included 149 chronic myelomonocytic leukemias (CMML), 11 cases with 5q- syndrome, 10 cases with refractory cytopenia with unilineage dysplasia (RCUD), 15 cases with refractory anemia with ring sideroblasts (RARS), 105 cases with refractory cytopenia with multilineage dysplasia (RCMD), 135 cases with refractory anemia with excess blasts-1 (RAEB-1), 87 cases with refractory anemia with excess blasts-2 (RAEB-2), and 22 cases with t-MDS, respectively. In total, 130 RUNX1 mutations were observed in 17.8% (95/534) of these patients. The mutational clone size ranged from 1.7% to 94% and amounted to a median of 31%. In comparison to our data from an AML cohort, i.e. 460 patients at diagnosis with 112 (24.3%) cases mutated, the median clone size was about 10% lower in MPN/MDS. In detail, 74.7% (71/95) of patients harbored one mutation, whereas 25.3% (24/95) of cases harbored two (17.9%; 17/95) or >=3 (7.4%; 7/95) mutations. The 130 RUNX1 mutations were characterized as follows: 29% frame-shift mutations, 42% missense, 14% nonsense, 13% exon-skipping, and 2% in-frame insertion/deletion alterations, respectively. The following codons were recurrently mutated: Arg174 (8/95 patients; 9.4%), Arg177 (6/95 patients; 7.0%), and Arg135 (5/95 patients; 5.3%). The mutations were predominantly located in the RHD domain (55%) and TAD domain (13%) and in cases with 2 or more alterations only 15% (4/24) harbored mutations outside of these regions. In all cases with 3 concomitant mutations both domains were affected (4/4 patients). Further, cases with >1 RUNX1 mutation were more frequently observed in CMML (33.3%; 8/24 mutated), RAEB-1 (17.2%; 5/29 mutated) and RAEB-2 (34.5%, 10/29 mutated) as compared to other disease groups, respectively. In subsequent serial analyses including 56 samples from 22 cases the amplicons carrying the respective known alteration were analyzed with increased coverage for disease status monitoring (in median 833 reads/amplicon were sequenced; leading to a sensitivity of ∼1:800). With a median time span of 2.5 months between the molecular analyses a total of 2 to 4 samples per patient were analyzed. In 5/22 patients, this assay then allowed to monitor the treatment success of allogeneic stem cell transplantation: in 3 cases the mutations known before transplantation became undetectable; in 2 cases the same mutated clones still remained detectable at a level of 0.2% and 23%, respectively. Further, in 17 patients quantitative assessment of mutated RUNX1 read counts was used to monitor stable disease (n=12) or allowed to follow an increasing clone size in 3 patients that progressed into s-AML (39% -> 53% increase; 31% -> 42% increase; 7% -> 37% increase). Summary: Unbiased techniques such as deep-sequencing provide the required diagnostic specificity and sensitivity to enable classification and individualized monitoring of disease progression. We here demonstrate that amplicon-based NGS is a suitable method to accurately detect and quantify the broad spectrum of molecular RUNX1 aberrations with high sensitivity. It is therefore suitable for therapy guidance. Disclosures: Kohlmann: MLL Munich Leukemia Laboratory: Employment; Roche Diagnostics: Honoraria. Grossmann:MLL Munich Leukemia Laboratory: Employment. Harbich:MLL Munich Leukemia Laboratory: Employment. Dicker:MLL Munich Leukemia Laboratory: Employment. Nadarajah:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

747 Introduction: RUNX1 (runt-related transcription factor 1) deregulations constitute a disease-defining aberration in AML. RUNX1 mutations were proposed as clinically useful biomarkers to follow disease progression from MDS to s-AML, as well as to monitor minimal residual disease (MRD). Study design: First, a next-generation amplicon deep-sequencing (NGS) assay was developed and a validation study was performed on genomic DNA obtained from mononuclear cells on a longitudinal series of 116 retrospective samples obtained from 25 patients. These samples were collected between 11/2005 and 6/2010 and were characterized for RUNX1 mutations by DHPLC and Sanger sequencing (conventional methods). In median, 3,346 reads per amplicon were generated and in all cases NGS analyses concordantly detected the mutations known from conventional methods. Furthermore, in 2/25 (8%) cases, NGS detected additional low-level mutations with 0.9% and 3.2% of reads mutated that were not observed by standard Sanger technique. Concerning MRD monitoring, in 7/25 (28%) cases an increasing clone size, i.e. mutations as low as 0.2% - 7.0%, was detectable up to 9 months earlier than by conventional methods. This established assay then was applied to characterize an unselected prospectively collected cohort during the subsequent 12-months routine diagnostics period starting 07/2010. Results: In total, 2,705 NGS RUNX1 mutation analyses were performed on a variety of hematological malignancies. We report on analyses on 460 AML cases at diagnosis including 369 de novo AML, 57 s-AML, and 34 t-AML cases (median age: 68 years; females: 204; males 256). 51% of cases presented with a normal karyotype, 38% harbored non-complex cytogenetic alterations, 10% carried a complex aberrant karyotype, and 1% of patients were characterized by favorable cytogenetics. Overall, 141 RUNX1 mutations were observed in 24.3% (112/460) of cases. At diagnosis, the clone size ranged from 2% to 95% (median: 40%). 82% (92/112) of mutated patients carried one, whereas 18% (20/112) harbored two (n=17) or more (n=3) mutations. The 141 mutations were characterized as follows: 43% (60/141) frame-shift mutations, 34% (49/141) missense, 15% (21/141) nonsense, 5% (7/141) exon-skipping, and 3% (4/141) in-frame insertion/deletion alterations, respectively. The mutations were predominantly located in the RHD domain (54%) or TAD domain (20%). In subsequent serial NGS analyses 31/112 evaluable RUNX1 mutated cases were studied and in 88 individual samples the alterations detected at diagnosis were specifically investigated with high coverage. With a median sampling interval of 50 days for the NGS analyses between 2 and 9 samples per patient were analyzed during the first year of treatment. In this cohort, three categories of patients were detectable: (i) 55% (17/31) of patients responded to therapy and were characterized by a total clearance of the mutated clone at the first time point of follow-up (804-fold median sequencing coverage; sensitivity ∼1:800). (ii) A second group consisted of 10% (3/31) of patients with refractory disease that stayed mutated, but were excluded from further analyses since they underwent transplantation. (iii) The third group comprised 35% (11/31) of patients: None of these patients demonstrated a clone size reduction below 0.7% of reads at the first follow-up analysis (reduction to a median of 21% mutated reads; range 0.7% - 41%). Also, at the second time point (in median 108 days after initial diagnosis), mutated clones were still detectable (reduction to a median of 8% mutated reads; range 4% - 15%). Most of these cases (10/11) had refractory disease as assessed by cytomorphology or molecular analyses. 10/11 cases did harbor a normal karyotype; n=1 with del(7q). Further, 6 of these 11 patients with refractory disease, as defined using NGS, were found to carry RUNX1 double mutations. Finally, in all (3/3) cases with double mutations in the same domain and refractory disease a changing antiparallel distribution of the clone size from initial diagnosis to first follow-up was observed. Conclusions: NGS accurately detects and quantifies RUNX1 mutations in AML with high sensitivity. The technique of deep-sequencing was observed to be superior to current routine methods, in particular during follow-up and in detecting MRD and thus has the potential to enable an individualized monitoring of disease progression and treatment efficacy. Disclosures: Kohlmann: MLL Munich Leukemia Laboratory: Employment; Roche Diagnostics: Honoraria. Grossmann:MLL Munich Leukemia Laboratory: Employment. Harbich:MLL Munich Leukemia Laboratory: Employment. Dicker:MLL Munich Leukemia Laboratory: Employment. Alpermann:MLL Munich Leukemia Laboratory: Employment. Nadarajah:MLL Munich Leukemia Laboratory: Employment. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.