The World Health Organization (WHO) recently included germline predisposition to myeloid malignancies as a new entity, recognizing it is clinically important to identify these individuals. Using DISCOVeRY-BMT cohorts, a study of almost 3,000 patient-HLA matched unrelated donor pairs reported to the Center for International Blood and Marrow Transplant (BMT) from 2000-11, we identified pathogenic (P) and likely pathogenic (LP) variants in cancer or myeloid malignancy related genes in leukemia patients and healthy blood or marrow donors, performed gene-based association with myeloid disease and determined if variants and/or genes impact overall survival after transplant.
In 1684 acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) patients and 2246 unrelated healthy donors from DISCOVeRY-BMT ClinVar P/LP variants taken from Illumina HumanExome BeadChip genotype data and frequencies were calculated across both disease groups. Nonsense and missense variants were used for sequence kernel-based association testing as implemented in SKAT-O and likelihood ratio tests were used for single variant association and survival analyses.
We identified 45 pathogenic/likely pathogenic variants in DISCOVeRY-BMT AML or MDS patients and/or donors; 12 autosomal dominant variants in 9 genes (3.1% of patients and 2.6% of donors) and 34 P/LP variants in 24 genes for autosomal recessive conditions identified in 203 patients (10.2%) and 243 donors (10.8%). The start codon, c.3G>A, DEAD/H-box helicase (DDX41) was the most frequent LP/P variant seen patients (11 de novo AML, 6 MDS patients) but no donors. The CHEK2 c.470T>C was the most frequent LP/P variant seen in donors (n=19); this variant was also patients (6 AML, 1 t-AML and 3 MDS). Survival analyses identified a germline non-internal tandem duplication/tyrosine kinase domain FLT3 variant, c.2383G>A associated with 1 year transplant related mortality in AML and MDS patients (P=6.6 x 10-5).
Despite the small amount of exome variation captured by the research Illumina HumanExome BeadChip (2%), we identified biologically relevant germline genomic variants and genes correlated with AML and MDS as well as in the donor population. These results, using a decade of data on US transplant donors, highlight three important areas of focus for genetic counselors and genomic scientists. First the characterization of germline variation in AML and MDS patients receiving a transplant as curative therapy for AML and MDS and the conveyance of this information to the patient and family. Second, AML and MDS patient outcomes who receive a transplant from donors with mutations and what this means for donor selection, as well as how genetic testing impacts the family. Third, better characterization of the role of germline variation in genes with known somatic mutations in survival after transplant as curative therapy for hematologic malignancies.