Kinase domain mutations in BCR-ABL1 are associated with resistance to tyrosine kinase inhibitors in patients with chronic myeloid leukaemia. Next-generation sequencing (NGS) allows detection of low-level kinase domain mutations, but its relevance in clinical practice remains debated. We aimed to examine the clinical effects of low-level kinase domain mutations identified using NGS in patients with chronic myeloid leukaemia. In this population-based study, we included consecutive patients newly diagnosed with chronic myeloid leukaemia treated with first-line tyrosine kinase inhibitors, and patients identified at the time of resistance to first-line treatment with imatinib at six institutions (teaching hospitals and district hospitals) in southeast England. We screened patients for BCR-ABL1 kinase domain mutations using NGS, irrespective of patient response to tyrosine kinase inhibitor therapy. When we detected a mutation with NGS, we retrospectively analysed all previous samples to establish the date of first occurrence and subsequent kinetics of the mutant subclone (or subclones). The primary endpoints of this study were progression-free and event-free survival at 5 years. Between Feb 1, 2007, and Dec 31, 2014, we screened 121 patients with chronic myeloid leukaemia for BCR-ABL1 kinase domain mutation. 99 consecutive patients were newly diagnosed, with available sequential RNA stored. The remaining 22 patients were diagnosed between June 1, 1999, and June 30, 2006, and were screened at the time of resistance to first-line treatment with imatinib. Imatinib was the first-line treatment for 111 patients, nilotinib for seven patients, and dasatinib for three patients. We detected a kinase domain mutation in 25 (21%) of 121 patients. Low-level kinase domain mutations were first identified in 17 (68%) of 25 patients with mutation. For patients with a complete cytogenetic response, 13 (14%) of 93 patients screened had a mutation. Five (71%) of the seven patients with a clinically relevant mutation lost complete cytogenetic response compared with 15 (17%) of 86 patients without a clinically relevant mutation (80 patients without mutation and six patients with a tyrosine kinase inhibitor-sensitive mutation, p=0·0031). Patients harbouring a mutant clone had poorer 5-year progression-free survival (65·3% [95% CI 40·5-81·8] vs 86·9% [75·8-93·2]; p=0·0161) and poorer 5-year event-free survival (22·2% [CI 5·6-45·9] vs 62·0% [50·4-71·6]; p<0·0001) than did patients without a mutation. We identified a kinase domain mutation in four (10%) of 41 patients with samples available at 3 months after starting first-line tyrosine kinase inhibitor treatment; all four subsequently progressed to accelerated phase disease compared with only three (8%) of 37 without a mutation (p<0·0001). NGS reliably and consistently detected early appearance of kinase domain mutations that would not otherwise be detected by Sanger sequencing. For the first time, to our knowledge, we report the presence of kinase domain mutations after only 3 months of therapy, which could have substantial clinical implications. NGS will allow early clinical intervention and our findings will contribute to the establishment of new recommendations on the frequency of kinase domain mutation analysis to improve patient clinical care. None.
Read full abstract