Ps1177 prognostic Factors for Molecular Relapse‐free Survival After Discontinuation of Tyrosine Kinase Inhibitors in Chronic Myeloid Leukemia Based on Russian Multicenter Prospective Trial (Ru‐ski)

Abstract

Results: Fig. 1a: Data from samples containing BCR-ABL1 transcripts detected in all 6 wells in the In-house assay (i.e. those samples with least stochastic variation). The reported IS between the two methods are similar, with an average fold change of 1.2 between the assays (ranging from 0.4 to 2.1). Fig. 1b: A plot of the data, from fig. 1a, gives a regression equation of Y = 1.043 * X + 0.81, and R 2 of 0.85. Fig 1c: Data from samples with detectable low-level disease in the In-house assay where one or more BCR-ABL1 wells contain under 3 transcripts. As the IS from Asuragens assay nears the LOD (sample 33 and onward), we see higher IS with our in-house method. The recent EUTOS guidelines recommend using 3 as the lowest number when scoring a positive well, and this may in some cases result in IS overestimation. Transcript number using the commercial kit was much larger due to the single-well multiplex setup, thereby reducing the effect of stochastic variation. Two samples with detectable disease with the In-house assay were not detectable with the commercial assay, these were also below this assays LOD. Fig 1d: Data from samples with undetectable disease in the In-house assay. 4 samples show detectable disease when using the commercial kit.Inspection of the reference gene data for these 4 cases indicate a higher sensitivity in the commercial assay, thereby allowing an increased chance for detecting the few transcripts that may be present. Summary/Conclusion: In our hands, data obtained from the commercial kit was essentially equal to our in-house assay, conforming to EAC/EU-TOS standards, even when run on equipment as yet not IVD approved by Asuragen. The largest variations were seen for samples close to the limit of detection, where our In-house assay had several BCR-ABL1 wells containing less than 3 transcripts.