Stability Of Conversion Factors For Bcr-Abl Monitoring - Implications For The Frequency Of Validation Rounds

Abstract Abstract 893 Introduction: A European collaborative harmonization study involving 61 laboratories is being conducted under the auspices of the European Treatment and Outcome Study (EUTOS) for CML that aims to facilitate reporting of molecular BCR-ABL quantification results according to the International Scale (IS). The aim of this analysis was to investigate the effectiveness of this process and specifically the stability of conversion factors (CF) over time. Methods: The currently accepted way of adopting the IS is to establish and validate a laboratory-specific CF which is then used to convert local results to the IS. For round 1, preliminary CFs were calculated by centrally distributing standard samples containing 10–20 million WBC approximating to 10%, 1%, 0.1%, and 0.01% BCR-ABL IS. Rounds 2 and 3 were employed to refine the CF calculations using 25–30 CML patient samples from each participating laboratories covering a range of BCR-ABL levels between 0.01% and 10%. Log BCR-ABL values for the same samples were compared between reference and local laboratories applying the Bland-Altman bias plot. In order to judge the stability of each laboratory`s methodology, a CF index (ratio of round 3 CF divided by round 2 CF) was calculated and evaluated according to its capability to achieve optimum concordance of results. Results: Of the 61 laboratories participating in round 1, evaluable patient samples have been provided to date by 56 and 30 laboratories in rounds 2 and 3, respectively. Of the 30 laboratories with complete data, 12 had stable CFs (defined as a CF index within 0.75–1.33) whereas 18 laboratories were outside this range. Comparison of the CFs derived from round 2 with those derived from round 3 revealed better and more consistent concordance between laboratories with stable CFs compared to those with unstable CFs. For the 12 stable laboratories, 79% (round 3 CF) vs 79% (round 2 CF) of the samples were within a 2-fold range (0.5–2.0) and 93% vs 89% were within a 3-fold range (0.33–3.0). For the 18 unstable laboratories, 74% vs 55% of the samples were within a 2-fold range (0.5–2.0), p=0.0005 and 92% vs 77% were within a 3-fold range (0.33–3.0), p=0.0005. 2 of 12 laboratories with stable CFs and 8 of 18 laboratories with unstable CFs indicated changes in either one or more components of their procedures (cDNA synthesis, PCR platform, RQ-PCR protocol) that may have impacted on their CFs. Conclusion: These data indicate that CFs may be unstable in some laboratories even in the absence of significant changes to laboratory protocols. Further, it supports the need for continuous revalidation of CFs. In laboratories with unstable CFs we suggest revalidation within 3 to 6 months whereas those with stable CFs should be assessed on a yearly basis. We also suggest that laboratories with unstable CFs need to rigorously examine their internal processes to identify potential sources of variation. Disclosures: Müller: Novartis: Honoraria, Research Funding. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.