Abstract

Assays used to monitor BCR-ABL levels in CML patients on TKI therapy should be sensitive enough to reliably detect BCR-ABL transcript levels at 4.5 logs below baseline for optimal clinical utility. Here we describe the clinical comparison of Xpert® BCR-ABL Ultra (V2) and Xpert® BCR-ABL Monitor (V1). V1 uses 200uL of peripheral blood. V2 uses 4mLs to enhance sensitivity. Both assays are currently standardized to the IS.

V2 was compared to V1 using prospectively acquired CML patient specimens (N=165) in EDTA, NaCitrate, and PAXgene for a potential 495 paired comparisons (Phase I). Linear regression, Bland-Altman, and concordance at MMR and CMR4.0 analyses were performed. Using blinded precision materials (Ph+ cultured cells containing BCR-ABL e14a2 diluted into Ph- normal donor blood to produce specimens at targeted BCR-ABL1 transcript levels spanning the dynamic range from 10% to 0.001% IS), V1 and V2 were repeat tested by two operators at each of 4 sites 8 times on different days (N = 2 x 8 x 4 = 64 replicates per day; 5 sites participated on Day 1, N=80), for a total N = 272 (Phase II).

96.2% of V1, 93.7% of V2, and 99.6% of external control specimens were successful on the first attempt. Across all types of anticoagulant, linear regression shows a slope of 1.083 +/- 0.028 (95% CI) with intercept of 0.31 +/- 0.032 (95% CI) and an R2 = 94.5% for V2 log10 BCR-ABL/ABL% IS (Y) versus V1 log10 BCR-ABL/ABL% IS (X). Bland-Altman analyses reveal a mean bias across the dataset of 0.27. Positive Percent Agreement (PPA) for V2 vs. V1 was 96.4% (298/309) and Negative Percent Agreement (NPA) for V2 vs V1 was 66.9% (87/130). 39/43 specimens detected by V2 but not V1 were below 0.01% (IS). Concordance studies for MMR showed an Overall Percent Agreement (OPA = [# comparisons with both assays <0.01%] plus [# with both assays >0.01%]/total comparisons) > 97.0% for all comparisons and 98.0%, 93.9%, and 89.4%, respectively, for EDTA, NaCitrate, and PAXgene for CMR4.0. In Phase II, 100% (N=203) of samples at 1%-0.01% were positive by V2; 50.8% (32/63) at 0.001% were positive by V2.

In conclusion, the Xpert BCR-ABL Ultra (V2) assay demonstrates concordance with version 1 of the Xpert BCR-ABL assay across a wide dynamic range in both clinical and contrived specimens. Higher discordance is noted in the range 0.01-0.001% (IS) likely due to the enhanced sensitivity of BCR-ABL Ultra (V2).

Disclosures

Mauro:Novartis Pharmaceutical Corporation: Consultancy, Research Funding; Bristol-Myers Squibb: Consultancy; Ariad: Consultancy; Pfizer: Consultancy. Pinilla-Ibarz:Pfizer: Consultancy, Other: Consulting & Advisory Role, Research Funding, Speakers Bureau; ARIAD Pharmaceuticals, Inc.: Consultancy, Other: Consulting & Advisory Role, Research Funding; Novartis: Consultancy, Other: Consulting & Advisory Role, Research Funding; BMS: Consultancy, Honoraria, Other: Consulting & Advisory Role, Speakers Bureau; Teva: Consultancy, Speakers Bureau. Uy:Novartis: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract