Finding of kinase domain mutations in patients with chronic phase chronic myeloid leukemia responding to imatinib may identify those at high risk of disease progression

Abstract

Purpose: Kinase domain (KD) mutations in the BCR-ABL gene are associated with resistance to imatinib in chronic myeloid leukemia (CML) but their incidence and prognostic significance in chronic phase (CP) patients without resistance are unclear.

Patients and methods: We analyzed outcome for 319 patients with CML-CP who were treated with imatinib; 171 were in early CP (ECP) and 148 were in late CP (LCP). Patients were screened routinely for mutations using direct sequencing regardless of response status. The 5-year cumulative incidence of mutations was 6.6% for ECP and 17% for LCP patients.

Results: Of the 319 patients, 214 (67%) achieved complete cytogenetic responses (CCyR). The identification of a mutation without other evidence of imatinib resistance was highly predictive for loss of CCyR (RR, 3.8; P = .005) and for progression to advanced phase (RR, 2.3; P = .01), though the intervals from first identification to loss of CCyR and disease progression were relatively long (median, 21 and 16 months, respectively). Mutations in the P-loop (excluding residue 244) were associated with a higher risk of progression than mutations elsewhere.

Conclusion: We conclude that routine mutation screening of patients who appear to be responding to imatinib may identify those at high risk of disease progression.

Fig 1.

Cumulative incidence of kinase domain (KD) mutations. (A) Five-year cumulative incidence of KD mutations (13.9%). (B) Five-year cumulative incidence of KD mutations comparing patients whose best response was complete cytogenetic responses (CCyR; 5.5%) with patients who failed to achieve CCyR (29.1%; P < .0001). (C) Five-year cumulative incidence of KD mutations comparing early chronic phase (ECP) patients (6.6%) with late chronic phase (LCP) patients (17%, P = .01). (D) Five-year cumulative incidence of KD mutations comparing ECP (5.6%) and LCP patients (5.3%, P = .89) but restricted only to those patients whose best response was a CCyR.

Fig 2.

Progression-free survival according to the presence of kinase domain (KD) mutations at 2 years: a landmark analysis.

Fig 3.

Progression-free survival (PFS) in a 2-year landmark analysis according to the position of the amino acid substitution and the level of in vitro resistance of the kinase domain (KD) mutation. (A) PFS in the 2-year landmark analysis according to the type of mutation. The 5-year PFS was 86% for patients with no mutation, 100% for patients with the mutation M244V, 20% for patients with P-loop mutations, and 47% for patients with non-P-loop mutations (P < .0001). Patients harboring the M244V mutation had a PFS no different from that of patients with no mutation (P = .9), but patients with a P-loop mutation and patients harboring non-P-loop mutations had a significantly different PFS when compared with patients with no mutation (P = .01 and P < .0001, respectively). The difference in PFS between patients harboring a P-loop mutation and non-P-loop mutations was also significant (P = .02). (B) PFS in the 2-year landmark analysis according to the degree of in vitro resistance of the KD mutation. The 5-year PFS was 86% for patients with no mutation, 0% for patients highly resistant mutations, and 64.7% for patients with other mutations (P < .0001). Patients with a highly resistant mutations and patients harboring other mutations had a significantly different PFS when compared with patients with no mutations (P = .008 and P < .0001, respectively). The difference in PFS for patients harboring a highly resistant mutation and patients with other mutations was also significant (P = .03).

Fig 4.

Evolution of total BCR-ABL transcript levels and percentages of the mutated transcripts in two selected patients. The figure shows the evolution of the total transcript levels the percentages of mutated subclones and in two representative patients. In both cases the detection of the kinase domain (KD) mutation clearly antedated any noticeable change in the transcript levels (see text). In 16 patients, exemplified by the patient in (A), the proportion of the mutant BCR-ABL clone (E459K) increased progressively until it became the predominant clone (more than 50%). In 14 of these patients the dose of imatinib was increased higher than 400 mg; in the patient represented in (A) the dose was increased to 800 mg at the moment of losing the complete cytogenetic response (CCyR). Seven of these 16 patients progressed to advanced phase. In the remaining 21 patients, exemplified by the patient in (B), the proportion of mutant transcripts (M351T) remained at or below 50% during follow-up (10 cases) or transiently increased higher than 50% (11 cases). In 20 patients, the dose of imatinib was increased above 400 mg. (In the patient represented in (B) the dose was increased to 600 mg at the moment of losing the CCyR and later changed to dasatinib). Ten of 21 patients progressed to advanced phase. BP, blastic phase.