Pb1981: experience of four laboratories of the italian cml labnet network in the use of the cepheid cartridge system

Topic: 8. Chronic myeloid leukemia - Clinical Background: Chronic myeloid leukemia is characterized by rearrangement of the BCR::ABL1 oncogene resulting in the production of p210 and p190 fusion proteins with deregulated tyrosine kinase activity. Monitoring BCR::ABL1 transcript levels in PB patients on TKI therapy using RT-qPCR is the gold standard for management of CML. Xpert BCR-ABL Ultra (named Xpert® BCR-ABL Ultra p210) is a cartridge-based assay that automates all quantitative process integrating all steps directly from clinical samples in less than 3 hours. We need to understand the differences in measurement to ensure that variability between the two methods is controlled for treatment response assessments. Aims: The aim of the study was to compare two different methodological approaches. This was done by measuring the agreement between the values obtained from the analysis of PB samples from CML patients. These analysis were performed in parallel with the use of the Cepheid cartridge and with the individual assay used in 4 reference laboratories of the Italian Labnet Network. These assays have been validated and described in the Italian Laboratory Recommendations (RIL). Furthermore, the methodological comparison was done through the use of a reference material such as a well-known ACROMETRIX BCR::ABL1 Reference Panel and the analysis of two samples received from the UK NEQAS Control Panel. Methods: 25-30 CML peripheral blood samples were evaluated using the Xpert BCR-ABL Ultra test and each of the 4 comparator assays of 4 study sites. The ACROMETRIX material is a panel of 5 vials with a known of concentration of BCR::ABL1 aligned with the International scale (I.S). The two samples supplied by the UK NEQAS panel consist of lyophilized samples with BCR-ABL1 IS levels ranging from 10% to 0.0032. Results: The analysis included a total of 193 measurements, which were quantified by both the Cepheid and the Standard in-house method. The measurements were divided by level of disease, with 31 for MR1, 36 for MR2, 54 for MR3, 26 for MR4, 28 for MR4.5, and 18 for MR5. For each method and level of disease, the median, interquartile range, minimum, and maximum were reported in the table. Bias and 95% Limit of Agreements from Bland-Altman analysis were also computed. The agreement on MR assignment was evaluated using Cohen’s Kappa (squared weights) and more conservative equal weights, which yielded values of 0.88 (p<0.001) and 0.75 (p<0.001), respectively. These results indicate substantial to almost perfect agreement (McHugh ML. Interrater reliability: the kappa statistic. Biochem Med (Zagreb). 2012;22(3):276-82. PMID: 23092060; PMCID: PMC3900052.) TABLEFinally, we evaluated the agreement between the standard and Cepheid results according to Branford et al. (Blood 2008; 112: 3330–3338) by applying three criteria. The criteria take into account the ratio of the standard to the Cepheid result. Overall, 193 observations gave the following results: N=109 (56.5%) between 0.5 and 2.0 fold difference N=155 (80.3%) between 0.33 and 3.0 fold difference N=182(94.3%) between 0.2 and 5.0 fold difference The data obtained fall within the required ranges and confirm the level of agreement between the two analysis systems. Summary/Conclusion: Statistical analysis of the data obtained with the two systems and on the three different types of material demonstrated a good correlation at all different levels of disease. The CEPHEID cartridge system could therefore be considered as an alternative to the analysis systems currently validated within the Italian network CML LabNet. Acknowledgements Thanks to Elena Bergatto and Marco Vigliano for the support of the project Keywords: Chronic myeloid leukemia, Peripheral blood