Xpert^® BCR-ABL Ultra^¥ Demonstrated Good Inter-Laboratory Reproducibility in Three Consecutive GBMHM External Quality Assessments

Introduction: Monitoring BCR-ABL transcript levels in peripheral whole blood of patients using real-time quantitative PCR is standard of care in the management of Chronic Myeloid Leukemia. Therefore, it is essential to ensure that variability in testing methodology is tightly controlled and compatible with accurate assessments of treatment response. Xpert^® BCR-ABL Ultra^¥ (Ultra) is an automated cartridge-based assay developed by Cepheid for sensitive assessment of BCR-ABL p210 transcript levels. Every Ultra assay is standardized onto the WHO International Scale (IS) during production. The GBMHM (Groupe des Biologistes Moléculaires des Hémopathies Malignes) program has previously validated a sample preparation protocol designed for Ultra External Quality Assessment (EQA). We report here results from 3 consecutive EQA rounds for Ultra.

Methods: Three Ultra EQA panels have been produced at GBMHM, containing a total of 11 members. Each member was produced from two contrived cell lysates, one BCR-ABL negative and one BCR-ABL e14a2/b3a2 positive, prepared according to the two-step Ultra sample preparation protocol per the package insert, but starting from 4 mL of a cell suspension titrating 2.5x10⁶ cells/mL instead of blood (Table 1). Typically, 30 ml of each panel member was produced and aliquoted into 1mL samples before shipping on dry-ice to the participating laboratories. Upon receipt, individual laboratories further prepared the samples for testing following the 2^nd lysis procedures.

Results: A total of 209 samples were sent to up to 21 different labs and tested in 15 different Ultra lots. The %BCR-ABL/ABL IS ratios in 207 of 209 samples were returned to the GBMHM EQA program. The calculated average ratios were between 0.0066% and 24% IS with coefficients of variation (CVs) ranging from 16% to 77% (Table 1.). The data showed lower CVs at higher %BCR-ABL/ABL ratios. Nighty-five percent confidence intervals (95% CIs) for distribution of the average ratios ranged from ±4.8 fold for sample UC in EQA Round 19.2 to ±1.4 fold for sample UB in Round 19.2. Inter-lab reproducibility was acceptable, with less than 50% CV in 5 out of 8 panel members. The highest CV was observed in the panel member with an average ratio less than 0.01%IS. There was a trend towards an improvement of the CVs in the last EQA round compatible with an ongoing learning process.

Conclusions: Overall, inter-laboratory reproducibility data obtained here are compatible with accurate measurements of BCR-ABL mRNA levels using Ultra, demonstrating equivalent or better results compared with non-automated methods. Since 15 different Ultra lots were used, this study verified that the Ultra standardization process is fully operative. To our knowledge, this is the first assessment of Ultra inter-laboratory reproducibility.

^¥For Research Use Only. Not for use in diagnostic procedures. Not reviewed by any regulatory body.

View largeDownload slide

View largeDownload slide

Close modal