Abstract
Background
Allogeneic HCT remains the only curative option for patients (pts) with high-risk or relapsed AML. Genome sequencing technologies such as next generation targeted deep sequencing (NGS) or whole exome sequencing have identified several recurrent somatic mutations in AML that may be associated with HCT outcomes. Genomic abnormalities may persist in pts who achieve complete remission (CR) after induction chemotherapy; however, the impact of HCT on the mutational burden before and after transplant in such pts has not been described. We explore the dynamic changes of the clonal architecture in serial samples obtained from AML pts before and after HCT.
Methods
NGS data on bone marrow and peripheral blood samples obtained at diagnosis, prior to transplant, and at day 30 and day 100 after transplant from pts with AML (according to 2016 WHO criteria) who were treated at our institution between 11/2015 - 1/2017 were analyzed. A total of 170 somatic and germline gene mutations that have been described in hematologic malignancies were sequenced at each time point. The mean depth of the targeted sequencing was 300 fold for the entire cohort. Clonal architecture was defined by using variant allelic frequency (VAF) adjusted for zygosity at each time point.
Results
A total of 28 pts with serial samples were included. The median age at transplant was 57 years (range, 25-74). Cytogenetic analysis at diagnosis included normal karyotype in 15 pts (54%), 9 (32%) with unfavorable risk per Alliance criteria (including 5 pts with complex karyotype), 1 (3%) with t(8,21), and 3 pts (11%) with other (intermediate risk) abnormalities. All pts received induction chemotherapy at diagnosis with cytarabine and anthracycline (7+3) +/- combination except for one pt who received azacitidine. All pts achieved complete remission (CR/CRi) per 2003 IWG criteria prior to transplant. Conditioning regimens were myeloablative (N= 17, 61%) and reduced-intensity (N=11, 39%). Graft sources were bone marrow (N = 17, 61%), and peripheral blood (N =11, 39%). Donor source was matched sibling (N= 9, 32 %), matched unrelated (N= 14, 50 %) or haplo-identical (N= 5, 18%). Among pts with samples available at diagnosis, the most frequently mutated genes were: DNMT3A (46%), NPM1 (21%), FLT3 (ITD/TKD, 20%), TET2 (13%), and WT1 (13%). All pts with abnormal cytogenetics (13 pts) achieved complete cytogenetic remission prior to transplant. Common leukemia-related mutations such as DNMT3A, ASXL1, and TET2 persisted among others prior to transplant even in pts with CR and complete cytogenetic response. Among pts who achieved full engraftment at day 30 and/ or day 100 post-transplant, persistent leukemia-related mutations were still identified. In 2 pts with heterozygous BRCA2 and 1 pt with a heterozygous BRCA1 mutation at diagnosis, mutations persisted at CR, prior to transplant and at day 30 and 100 post-transplant, with a VAF of 50%, suggesting that these likely represented germline mutations. None of these pts had a history of breast or gynecological cancers. The median follow up after transplant was 9.97 months (range, 6.07- 20.17). Among pts with persistent mutations at day 100, 37% relapsed and the rest continued to be in remission until their last follow up.
Conclusion
Genomic analysis showed persistent leukemia-associated mutations in pts with AML in morphological and cytogenetic CR prior to HCT. Regardless of the conditioning regimen, a significant proportion of pts had persistent mutations at day100 after transplant even in pts with full engraftment. Persistent mutations predicted relapse though not all pts relapsed during this study period. These data suggest that deep genome sequencing may help in identifying higher risk pts of relapse post HCT, though longer follow up is needed to monitor the persistent mutations in pts who did not relapse during study time.
Sekeres: Celgene: Membership on an entity's Board of Directors or advisory committees. Gerds: CTI BioPharma: Consultancy; Incyte: Consultancy. Majhail: Sanofi: Honoraria; Anthem, Inc.: Consultancy.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal