Search

Abstract 2535 Background: Minimal residual disease is well established as an important prognostic parameter in a number of hematologic diseases and is used for stratification and treatment decisions in adult and childhood acute lymphoblastic leukaemia (ALL). DNA- based PCR analysis of Ig-/TCR-rearrangements have been well standardised during the last decade, resulting in robust systems for MRD analysis with an intra- and inter-assay variability of less than half a log and low false positivity. In Ph+ALL, MRD analysis relies on RNA-based RT-PCR techniques that are highly sensitive but as yet lack standardisation between laboratories. Since laboratories differ substantially in both their methodology and analysis strategy, interpretation and meaningful comparison of results between laboratories and from different studies is difficult or impossible. Moreover, and in contrast to CML, there are no generally accepted definitions of molecular response that could identify thresholds on which to base therapeutic decisions. Aims: To assess i) the variability of BCR-ABL quantification between laboratories with recognized expertise in bcr-abl analysis, ii) identify optimal laboratory methods and iii) to standardize analysis and interpretation of RT-PCR results for Ph+ALL, the EWALL and ESG-MRD-ALL consortia initiated a multinational project for quality assurance involving more than 30 laboratories from 14 countries worldwide. We here report the results of the first six m-BCR-ABL laboratory control rounds, performed between march 2008 and august 2011, that focussed on inter-intraassay variability of rt-PR techniques, comparison of housekeeping genes, plasmids, platforms and reagents. Methods: Serial dilutions of the BCR-ABL positive cell line Sup B15 with the BCR-ABL negative cell line Nalm 6 covering 5 logs were produced. In the 6 lab rounds, 1355 aliquots, each generated from with a total amount of 5 × 10E+06 cells in a final volume of 1 ml were produced, stabilized in TRIZOL or buffer RLT and frozen at −20°C until shipment or centralized isolation of RNA and reverse transcribed. Participants were asked to process the material using predefined procedures. Results: In the first QC round a major finding was the high variability in RNA-yield between laboratories despite using the same extraction method, with an up to 3 log difference in ABL copy numbers. In addition, there was rare false-positivity and false negativity (specifity 94.8%). With centrally produced cDNA the variability improved and there was a lower frequency of false-positivity reaching 100% specifity in QC 2. By using standardized assays, in 12 of 33 laboratories no conversion factor is needed, 20 of 33 laboratories are at the moment within one log difference when comparing the BCR-ABL/ABL ratios. Comparison of housekeeping genes (ABL/GUS) revealed non-linearity at high transcript levels resulting in an underestimation of intraindividual log reductions and may be relevant when assessing disease prognosis. The quantitative range was generally above 10E-04. Conclusion: Initiation of laboratory control rounds and implementation of standardized methodology improved the variability and reduced the frequency of false-positive results. Sensitivity of current techniques is still unsatisfactory for diseases with aggressive growth kinetics such as Ph*ALL. Further standardisation of technical and analytical methodology will provide the basis for defining levels of molecular remission and relapse and help to ensure comparability between laboratories and clinical trials at the European level. Disclosures: Goekbuget: Micromet: Consultancy. Ottmann:Novartis Corporation: Consultancy; Bristol-Myers Squibb: Consultancy, Research Funding.

Abstract 894 Various techniques have been employed to detect BCR-ABL kinase domain mutations in patients with chronic myeloid leukemia (CML) resistant to imatinib or second generation tyrosine kinase inhibitors. This variation may at least partially explain the different frequencies of mutations that have been reported. Furthermore, the pattern of individual mutations reported seems to depend on the specific method used for mutation detection. Standardized techniques and protocols for the detection of BCR-ABL mutations will be necessary to obtain comparable mutation results within clinical studies. The first objective of this study conducted within the EUTOS (European Treatment and Outcome Study) for CML program was to record the mutation analysis techniques and protocols that are used for routine diagnostics by 28 national reference laboratories in 23 European countries, 9 of whom perform regular mutation analyses as a central laboratory for national or international clinical trials. The second objective was to evaluate the techniques by analysis of blinded samples containing various BCR-ABL kinase domain mutations. Initially a web-based survey was conducted with a total of 39 technical and PCR-specific questions. Most laboratories (n=17; 61%) perform 1–10 mutation tests per month (range 100) and use published protocols for nested RT-PCR from peripheral blood or bone marrow leukocytes. Failure to amplify BCR-ABL mRNA at low BCR-ABL transcript levels was reported to be the most common PCR related difficulty in 16 (57%) laboratories. Sanger sequencing is applied for routine BCR-ABL mutation analysis in 27 (96%) laboratories. Additional screening techniques (e.g. denaturing high-performance liquid chromatography, high-resolution melting) or more sensitive detection methods (e.g. allele-specific oligonucleotide PCR, pyrosequencing) are used routinely by eight laboratories each, respectively. The application of positive or negative mutation control samples was reported by 13 (46%) or 17 (61%) laboratories, respectively. Quantitative mutation analysis is performed routinely in nine (32%) laboratories. To evaluate laboratory performance we prepared a series of control samples that were distributed in a blinded fashion to testing laboratories. Seventeen Ba/F3BCR-ABL cell lines harboring various BCR-ABL kinase domain mutations were mixed with non-mutated Ba/F3BCR-ABL to produce dilutions ranging from 1% to 100% of mutant alleles. Three samples were non-mutated Ba/F3BCR-ABL only and were used as negative controls. Mutated and non-mutated Ba/F3BCR-ABL cell lines were diluted into HL60 cells to simulate a BCR-ABL level of 10% on the International Scale. Twenty blinded cDNA samples were sent out on dry ice to each participating laboratory (total of 560 samples). Twenty-three labs have reported their results so far (460 samples). The three non-mutated samples were identified correctly in 68/69 (99%) tests. For the 20 mutated samples we distinguished those with high (≥20%; n=11) and low (≤10%; n=9) levels of mutation. For the high level samples 218/253 (86%) tests were reported correctly. Of the incorrect results, 22 were scored as negative and 13 were false positives (T315I, n=7; D276G, K357E, D363Y, N374I, S386S, E505K, n=1 each). For the low level samples only 15/138 (11%) tests correctly identified the mutation. Most (122/138; 88%) were reported as undetectable and a false positive result (L323H) was reported in one case. We conclude that Sanger sequencing is the most frequently applied technique for routine analysis of BCR-ABL kinase domain mutations in CML in Europe. In general it reliably identifies mutations when the proportion of mutant alleles comprise 20% or more. Nevertheless, false negative and false positive results were reported in a substantial proportion of samples with ≥20% mutation level (35/253, 14%). For mutations that are present at 10% or less mutant alleles, routine methods mainly failed to identify mutations. This study provides a basis for further comparisons and standardization efforts comparable with the introduction of the international scale for quantification of BCR-ABL transcripts. Disclosures: No relevant conflicts of interest to declare.