BackgroundChronic myeloid leukemia (CML) is characterized by the BCR-ABL1 fusion gene. Despite the dramatic improvement of its prognosis in recent years by the development of tyrosine kinase inhibitors (TKIs), a minority of chronic phase (CP) CML patients fail to respond to TKI therapies and progress to blast crisis (BC), showing dismal clinical outcomes. While acquired mutations in ABL1 kinase have been identified as a common mechanism for TKI resistance, recent genetic studies have revealed that patients with BC frequently harbor one or more genetic alterations implicated in myeloid malignancies, suggesting additional mutations other than ABL1 mutations might drive disease progression. However, our knowledge about the mechanism of TKI resistance and progression to BC is largely limited by the scarcity of matched CP and BC samples, which were investigated for genetic alterations in relatively small number of genes. Here, we performed comprehensive genomic studies of CML-BC using paired CP and BC samples to investigate the mutation profiles associated with BC.MethodWe performed whole-exome sequencing of 53 patients with CML-BC, including 40 myeloid and 13 lymphoid crisis cases, as well as corresponding CP controls to investigate acquired mutations during disease progression from CP to BC. We also performed targeted-capture sequencing of known and putative driver genes in an additional 15 CML-BC samples. Combined, a total of 68 CML-BC samples were analyzed for somatic mutations, copy number abnormalities, and structural variations.ResultsCommonly affecting ASXL1, GATA2, and IKZF1, mutations were found only in a minority of CP cases (10/53 [19%]). However, most cases acquired somatic mutations during disease evolution from CP to BC; in whole-exome sequencing, an average of 17 additional non-synonymous mutations were newly acquired per case during evolution from CP to BC. Mutations in CML-BC frequently involved known driver genes, such as ASXL1, RUNX1, ABL1, TP53, BCOR/BCORL1, and WT1. In addition, we identified novel targets of recurrent mutations, including UBE2A, NBEAL2 and KLC2. Of note, most these driver mutations were not detected in corresponding CP samples and newly acquired, whereas ASXL1 mutations were often found in corresponding CP samples in a minor population, suggesting that ASXL1 mutations at CP might play an important role in the disease progression to BC. Mutational profiles were similar between cases with and without a history of TKI therapy before BC, except for frequent ABL1 mutations among TKI-treated cases, mostly affecting the kinase domain. Compared with lymphoid BC, myeloid BC showed a higher number of somatic mutations, which was more prominent for ASXL1, TP53, and WT1 mutations. Copy number abnormalities were rarely found in CML-CP cases (8/53), but were common and newly acquired in 29 (55%) cases with CML-BC, 18 of which showed complex karyotype-like (≥3) abnormalities. Amplification of chromosome 6 and/or 8 were characteristics of myeloid BC, while deletion of chromosome 7 was more characteristic of lymphoid BC. In some cases, structural variations other than BCR-ABL1 translocation were newly acquired in CML-BC, which frequently involved genes implicated in myeloid malignancies such as RUNX1, CBFB, and MECOM. When mutations, copy number abnormalities, and structural variants were combined, most BC cases had at least one driver alterations, which might be involved in CML-BC.ConclusionThrough a comprehensive sequencing analysis using paired samples of CP and BC, we demonstrate a role of additional driver events during the clonal evolution to BC. Additional mutations were common even in CML-CP, some of which might contribute to the progression to BC. DisclosuresTakaori-Kondo:Celgene: Honoraria, Research Funding; Novartis: Honoraria; Janssen Pharmaceuticals: Honoraria; Bristol-Myers Squibb: Honoraria; Pfizer: Honoraria.