Germline Risks and Clinical Impacts of DDX41 Mutations in Myeloid Malignancies

Background Pathogenic/likely-pathogenic (P/LP) germline DDX41 variants have been implicated in late-onset myeloid neoplasms (MNs). Despite an increasing number of publications regarding this gene, many important features of DDX41-mutated (DDX41mut) MNs remain to be elucidated, including disease penetrance. Methods Here, we enrolled 9,082 cases with various MNs of divergent ethnicity. Tumor-derived DNA was available for all cases at one or more time points. In particular, diagnostic or treatment-naive tumor samples were obtained in 7,697 cases, of which 2,359 had survival data. Germline DNA was available for 1,350 patients. Bone marrow (BM)- or blood-derived tumor DNA was subjected to whole genome, exome and/or targeted sequencing using different driver gene panels, which included DDX41 and an additional 36 genes in common. We also enrolled 20,238 individuals from different Japanese biobanks for population controls. Results We identified a total of 346 (3.8%) DDX41mut MNs among 9,082 MN cases. Among these, 293 (84.7%) had a germline variant, whereas 53 (15.3%) carried somatic variants alone of which biallelic lesions were suspected for 52%. As many as 66.6% of germline variants are truncating, which was rarely true of somatic variants. Based on the genotyping of known leukemia predisposition genes in consecutively enrolled 1,039 Japanese cases with MNs, DDX41 variants explains ~80% of all known predisposition to adult MNs in Japan. We determined the penetrance of MNs for P/LP DDX41 alleles through a kin-cohort analysis involving a total of 288 and 237 first-degree relatives of patients with MNs with and without P/LP DDX41 germline variants, respectively. The risk is almost negligible (~0%) under the age of 40 years, but rapidly increases to 49.3% (95%CI: 29.8-68.2) by the age of 90. In population analysis, five DDX41 alleles showed a significant enrichment in MNs in Japanese populations and therefore represent P/LP alleles (Odds ratio (OR) of ~10; 95%CI: 4.5-22.8, on average), which was more prominent in female than in male (20.7 vs. 5.0 in OR). Accordingly, the penetrance in males (52.5%) almost double that in females (28.7%). DDX41mut patients with MDS showed a significantly faster progression to AML, which however, confined to those cases having truncating variants, while those with non-truncating variants showed a rather slower progression compared with DDX41WT patients. Co-mutation patterns at diagnosis and at progression to AML are substantially different between DDX41mut and DDX41WT patients. Particularly, CUX1 mutations were overrepresented in DDX41mut cases. Conspicuously, unlike other MDS cases, DDX41mut MDS were not significantly enriched for mutations in AML-related genes, such as RAS and other signaling pathway genes. In contrast to DDX41WT cases, none of the co-mutations significantly affected clinical outcomes. Even biallelic TP53 mutations did not affect overall survival (OS) and as expected, neither IPSS-R nor IPSS-M scores do so. DDX41mut patients showed a significantly better OS than DDX41WT cases with higher-risk MDS and sAML patients and a survival comparable to that of DDX41WT lower-risk MDS cases. OS in patients who received hypomethylating agent (HMA) was significantly better for DDX41mut than DDX41WT patients and so in those who received hematopoietic stem cell transplantation (HSCT), which however, confined to those who received HMA prior to HSCT, highlighting a role of HMA treatment in DDX41mut cases. Conclusion P/LP germline DDX41variants explain the largest risk of adult MNs with very high penetrance. DDX41mut MNs have distinct genetic and clinical/hematological features, representing a unique subset of MNs. IPSS-R/IPSS-M prognostication may not be applied to DDX41mut MDS.