PF218 DDX11 MUTATIONS AND CLONAL TRAJECTORY PATTERNS IN ADULTS WITH CYTOGENETICALLY-NORMAL ACUTE MYELOID LEUKEMIA
Y.-L. Zhou,L.-X. Wu,Z.-L. Wang,R. Gale,R.-Q. Lu,J.-L. Li,H. Jiang,Q. Jiang,B. Jiang,S.-B. Cao,F. Lou,Y. Sun,X.-X. Zhang,C. Cui,Y.-R. Liu,Y. Wang,L.-P. Xu,K.-Y. Liu,X.-H. Zhang,X.-J. Huang,G.-R. Ruan
DOI: https://doi.org/10.1097/01.hs9.0000559088.07802.86
2019-01-01
HemaSphere
Abstract:Background: Adults with acute myeloid leukemia with normal cytogenetics (AML-CN) have variable relapse risks. New mutations correlated with relapse risk are needed. Understanding clonal trajectory of relapse in this setting may improve understanding of leukaemia development. Aims: Interrogate genomic landscape and clone trajectory data in adults with CN-AML and identify new mutations associated with relapse risk. Methods: 333 consecutive, newly-diagnosed adults with CN-AML seen in Peking University Peoples Hospital March, 2006 and April, 2018 were studied. Whole-exome sequencing (WES) was done in 23 subjects with diagnosis, remission and relapse samples. Using these and published data we developed a customized panel of 322 AML-related genes. Diagnostic DNA samples of the remaining 310 subjects were studied and the coding regions or hotspot areas of the 322 genes were sequenced by deep target regional sequencing (TRS). Average effective sequencing depths were 200X (145–265X) for WES and 1500X (1000–1800X) for TRS. Median coverages were 99.9% (77.6, 99.9%) and 99.0% (98.0, 99.5%). Results: We identified 3672 non-synonymous somatic variants in 309 genes by TRS. In 305 subjects ≥5 mutations were identified (98%; 95% confidence interval [CI], 97, 100%), with a median of 13 mutations per subject (range, 2–27; Figure). We identified mutations at 4 new loci including DDX11 in 23 subjects (7%), CYP2F1(4%), NPIPA5 (3%) and MED12 (2%). We then interrogated the impact of DDX11 mutations by matching these subjects with 115 controls according to age (<60/≥60 y), NPM1(Y/N) and FLT3-ITD(Y/N) mutations and WBC (<100/≥100x10E+9/L). In multivariable analyses DDX11 mutation was independently-associated with cumulative incidence of relapse (CIR; Hazard Ratio [HR] = 2.15 [1.12, 4.13]; P = 0.021) and relapse-free survival (RFS; HR = 2.12 [1.03, 4.38];P = 0.042). Other mutations identified included driver genes in the RAS/RTK signaling pathway (27%), chromatin modifier pathway(10%), DNA-methylation pathway (8%), transcription factor pathway (8%), histone methylation pathway (5%), cohesin complex pathway (3%) and the spliceosome pathway (3%). NPM1 was the most frequent mutated gene in our cohort (N = 75; 24% [19, 29%]), followed by FLT3-ITD (N = 64; 21% [16, 25%]) and DNMT3A (N = 59; 19% [15, 23%]). Using triplicate samples from the 23 subjects we identified 3 clonal trajectory patterns: (1) Dominant clone at diagnosis and relapse without new mutations (N = 2); (2) Dominant clone at diagnosis with ≥1 new mutation (N = 13); and (3) evolution from a minor clone detected at diagnosis (N = 8).The clonal trajectory of relapse after chemotherapy was like that posttransplant but the latter was associated with more relapse-specific sub-clones possibly related to intensive pretransplant conditioning.Summary/Conclusion: We identified 4 new mutated genes in AML-CN and show DDX11 mutations correlate with a higher CIR and worse RFS in multivariable analyses. We also detected 3 distinct clonal trajectories which provide insight into mechanisms of relapse after chemotherapy and transplantation and the pathogenesis of AML-CN.