Objectives: Acquired resistance to imatinib (IM) in patients with chronic myeloid leukaemia (CML) is frequently due to mutations of the BCR-ABL kinase domain (KD). Current thinking holds that this results from the selection of pre-existing mutant clones on IM. As new BCR-ABL inhibitors with differential activity against KD mutant BCR-ABL have become available, precise quantification of key mutations even at a low level is required to adequately monitor responses. Here we report that the occurrence of G250E, Q252H, Y253F/H, E255K/V, T315I, F317L, F359V mutations at the time of IM resistance is independently associated with a high maximum IM dose.

Patients and Methods: We have developed highly specific allele-specific ligation-PCR assays (L-PCR) to accurately quantify a panel of frequent KD mutations, including G250E, Q252H, Y253F/H, E255K/V, T315I, F317L, F359V. In limiting dilution experiments the L-PCR assays routinely detect between 0.05 and 0.1% of mutant allele in total BCR-ABL, and their average dynamic range is in the range of 4.5 log. Forty-three patients with imatinib failure were analyzed. The median age was 60 (range 20 – 75) years and the median disease duration 64 (range 3–213) months. Eleven patients were in blast crisis, 20 in accelerated phase and 12 in chronic phase.

Results: Patients were treated with chemotherapy (48%) and/or interferon alpha (76%) prior to IM. The median maximal IM dose was 600 (range 500–800) mg and the median duration of total therapy until resistance was 15.5 (range 1 to 75) months. Eighteen patients (42%) had dose reductions due to toxicity. At the time of IM resistance, clonal evolution was present in 21/41 (47%) of the patients. L-PCR identified 50 mutations in 29/43 patients (67%) (table1). One, 2, 3 or 4 different mutations were identified in 14, 11, 2 and 2 patients, respectively. The T315I and E255K/V mutations accounted for 32/50 (64%) of all mutations, while Q252H and Y253H were not detected. Twelve mutations (24%) were confirmed by direct sequencing (DS) and an additional M315T mutation and a K247R polymorphism were detected. Thirty eight (76%) mutations were negative by DS for the corresponding mutation from the L-PCR panel but additional mutations (L248V, M351T, H396Rx2, L298Vx2) were identified. All 38 L-PCR positive mutations were confirmed in a second independent experiment. The mutated clone was significantly smaller in mutations with no confirmation by DS (median 0.2, range 0.05–12.63%) compared to positive by DS (median 40.76, range 13.58–100%, p=0.003, Wilcoxon test). The detection of one or more mutations was significantly more frequent in patients with a maximum IM dose of 800mg (n=15/17, 88%) compared to less than 800mg (14/26, 53%, p= 0.02 Fisher’s exact test). Multivariate analysis (Wald forward) confirmed that a maximum IM peak dose of 800mg is an independent prognostic parameter to detect a mutation from the L-PCR panel at the time of IM resistance.


  1. High sensitive testing with L-PCR detects mutations with 4-fold increased frequency compared to direct sequencing.

  2. Mutations, including p-loop and T315I, may be selected by exposure to higher drug levels.

  3. The predominance of T315I and E255K/V mutations at a low level is consistent with the findings in newly diagnosed patients with Ph+ALL (Pfeifer et al. Blood 2007).

Author notes

Disclosure:Research Funding: Research support from Novartis Pharma GmbH. Honoraria Information: Honoraria from Novartis Pharma GmbH.