The ability to consistently detect low levels of BCR-ABL transcripts in patients with chronic myeloid leukemia (CML) is important in the assessment of treatment outcomes in patients on tyrosine kinase inhibitor (TKI) therapy. Particularly, BCR-ABL assays that are sensitive in the measurement of deep level response may aid in the identification of potential candidates for treatment discontinuation. Xpert® BCR-ABL Ultra detects the most common BCR-ABL transcripts below MR4.5 (Molecular Response at 4.5-log reduction) or 0.0032%, which is widely accepted as the clinical threshold that defines candidates who can safely discontinue TKI therapy. The present studies were designed to verify the limit of detection (LoD) for the Xpert® BCR-ABL Ultra assay below MR4.5 on the international Scale (IS) in clinical samples for both the b3a2 and b2a2 transcripts.

To overcome the challenge of testing numerous replicates requiring large volumes of patient samples,, serial dilutions ranging from BCR-ABL/ABL levels of 10% to <0.001% (IS) were prepared as contrived samples using CML patient blood with initial BCR-ABL level >10% (IS) and pooled blood from CML negative patients, ranged from 10% to <0.001% (IS). Twenty-one replicates of each dilution were measured for %BCR-ABL/ABL (IS). Determination of the LoD was performed by the statistical analysis to identify the lowest concentration of %BCR-ABL/ABL (IS) per test that can be reproducibly distinguished from negative samples with 95% confidence. The acceptable precision for %BCR-ABL/ABL (IS) is defined as the ability to detect at least a 3-fold difference for all concentrations tested. In addition, analytical LoD studies were performed using spike-in CML cell lines and cell-line derived RNAs, carrying either b3a2 or b2a2 transcripts. Furthermore, the clinical sensitivity study was conducted using blood from twelve low BCR-ABL transcripts level CML patients on TKI therapy, who had achieved and maintained MMR (Major Molecular Response) [0.1% (IS)] with reporting below 0.05% (IS).

Consistent results were observed in the both the diluted CML patient blood and spike-in CML cell lines or cell-line derived RNA studies for both the b3a2 and b2a2 transcripts, demonstrating an assay LoD of MR4.5 and below with a less than 2-fold difference at the LoD levels. With the clinical sensitivity study, eleven out of twelve low CML subjects were detected in at least 19 out of 20 replicates tested per subject over a range of 0.038% (IS) (SD=~0.17 Log) to 0.0011% (IS) (SD=~0.4 Log). The overall ABL copy number present in clinical samples in each study was at least 5-10 times the required minimal ABL copy number of ≥32,000 to support a claim of MR4.5 and ≥100,000 for MR5.0.

These LoD evaluations demonstrate that the Xpert® BCR-ABL Ultra assay complies with the international guidelines for assay sensitivity achieving MR4.5 with 5-10 times more than the required ABL copies to confidently identify candidate patients that may benefit from the discontinuation of TKI therapy.


Levitas: Cepheid: Employment. Uy: Novartis: Consultancy, Other: Travel Suppport; Boehringer Ingelheim: Consultancy; GlycoMimetics: Consultancy. Mims: Novartis: Honoraria. Shridhar: Cepheid: Employment. Wu: Cepheid: Employment. Huang: Cepheid: Employment. Wei: Cepheid: Employment. Lykke: Cepheid: Employment. Bates: Cepheid: Employment. Wong: Cepheid: Employment. Day: Cepheid: Employment.

Author notes


Asterisk with author names denotes non-ASH members.