Abstract

We analysed outcome for 211 CML patients treated with imatinib in chronic phase (CP) (99 newly diagnosed and 112 late chronic phase) who were screened for BCR-ABL kinase domain (KD) mutations using direct sequencing regardless of the response status. When a mutation was found all available previous cDNA samples were analysed by pyrosequencing to establish the date of its first occurrence and subsequent kinetics. The median age of patients was 47.4 years. The Sokal risk score was ‘low’ in 57 patients, ‘intermediate’ in 82 and ‘high’ in 72. The median follow up from starting imatinib was 45 months (rage 6 to 89 months). A mutation was detected in 34 of the 211 patients (16%) at a median time of 27 months from starting imatinib. Twenty-two different mutations were identified, the most frequent being M244V (n=6) and F359V (n=3). When studied serially by pyrosequencing the size of the mutant subclone never exceeded 50% of total BCR-ABL transcripts in 8 patients, while in 17 patients it exceeded 90% on at least one occasion. 48 patients discontinued imatinib while still in CP and received either dasatinib, nilotinib or an allograft. The overall progression-free survival (absence of advanced phase) at 5 years was 73%. Major (MCyR) and complete (CCyR) cytogenetic responses were achieved by 153 and 123 patients respectively; 56 patients achieved major molecular response. 24% of the patient with up front cytogenetic resistance had a mutation while 40% of the patients with acquired cytogenetic resistance develop a mutation. In an-intention-to-treat analysis, patients harboring a mutant clone had a poorer PFS at 4 years (78% versus 57%, p=0.0014). The various mutations had no differential effects based on their known imatinib IC50. By multivariate analysis, factors associated with worse PFS were the presence of a KD mutation and failure to achieve CCyR (relative risks for PFS 2.6 and 8.7 respectively, p=0.002). Interestingly, the adverse effect of the presence of a KD mutation was restricted to the patients who achieved a MCyR (PFS 91% versus 62% at 5 years, p = 0.0006); it had no adverse impact on patients who failed to achieve a MCyR (PFS 42% and 49%, p=0.73). Similar results were found when the analysis was repeated according to the achievement of CCyR (data not shown). Surprisingly patients with a continuously low percentage (≤50%) of mutated vs wild type (>50%) clones fared worse than patients in whom the mutated clone became the predominant population (PFS 14% vs 69% respectively, p=0.0005). Comparable results were obtained when the patients were censored at the point of discontinuing imatinib, correcting for the effects of subsequent treatment, ie allografting (data not shown). The fact that the adverse effect of a mutation seems to be restricted to patients who had achieved cytogenetic response, the fact that mutations present at low level seemed to have a remarkable adverse effect and the fact that the in-vitro level of resistance to imatinib of the specific mutation did not affect the PFS could all be explained if the development of a mutation is only a reflection of the genomic instability of the disease that leads to secondary resistance to imatinib and eventually to transformation. Thus genomic instability may be less important in explaining primary resistance to imatinib and eventual transformation in patients with up-front resistance.

Author notes

Disclosure: No relevant conflicts of interest to declare.