Equivalence of BCR-ABL transcript levels with complete cytogenetic remission in patients with chronic myeloid leukemia in chronic phase

Abstract

Purpose: Chronic myeloid leukemia (CML) patients are monitored by both cytogenetic and molecular assessments, although present guidelines appear to switch from cytogenetic to molecular criteria. Due to the increasing use of molecular measurements, it was the aim of this work to identify a BCR-ABL level according to the international scale (BCR-ABL(IS)) as an equivalent substitute for complete cytogenetic remission (CCyR).

Methods: In total, 1,329 paired data from 557 patients of the German CML-Study IV were evaluated. The data set was divided into a learning set and a validation set. The best cutoff was determined applying a minimal p value approach to the Fisher test.

Results: In the learning set, we found BCR-ABL(IS) values between 0.2 and 1.1 % were well suited for predicting a CCyR. In the validation set, the cutoff level of 1 % led to a mean concordance rate of 90.1 %.

Conclusions: Our results suggest that there is no one-to-one cutoff for BCR-ABL(IS) representing CCyR, but we advise to use the 1 % BCR-ABL(IS) in order to avoid misclassification of CCyR patients.

Fig. 1

Kernel density estimate of the best cutoff points within the 10,000 subsets. The vertical lines denote the modal value (0.35 %) and the median (0.7 %). The highlighted area depicts the area where all cutoff levels were similarly well suited and covers the optimal cutoff levels of 68 % of the 10,000 subsets