Abstract 2070: The prevalence of DICER1 pathogenic variation in cancer population databases

Abstract

Abstract DICER1 is an RNaseIII endonuclease that is crucial for processing pre-microRNA into mature microRNA. The DICER1 syndrome is a rare, autosomal dominant tumor-predisposition disorder associated with pleuropulmonary blastoma, a rare pediatric lung cancer. Somatic missense variation in “hotspot” codons in the RNaseIIIb domain (E1705, D1709, G1809, D1810, E1813) is observed in DICER1-associated tumors. Previously, we found in the general population (non-TCGA ExAC) that the prevalence of pathogenic DICER1 variation is approximately 1:10,000. In this study, we investigated the prevalence of pathogenic DICER1 variation in common adult and pediatric cancers. We examined two publicly-available cancer datasets: TCGA (adult; 32 cancer types; 9,173 exomes) and TARGET (pediatric; 3 cancer types; 236 exomes). TCGA bam files were downloaded and germline variants were called. Germline VCF files were obtained from TARGET investigators. All datasets were annotated with snpEff and ANNOVAR. Variants were binned into 4 categories: likely benign (LB), variant of unknown significance (VUS), likely pathogenic (LP), or pathogenic (P). We reviewed publicly-available pathology images for the TCGA tumors. The prevalence of DICER1 P/LP variants (in all cancers combined) was 1:834 and 1:118 in TCGA and TARGET, respectively. To reduce potential prevalence inflation, we conducted a more stringent prevalence calculation by only considering loss-of-function and hotspot variants. The prevalence of P variants was approximately 1:4586 in TCGA and 1:236 in TARGET. With these more stringent criteria, a single participant with a uterine corpus endometrial carcinoma (UCEC in TCGA) harbored a two pathogenic germline DICER1 (hotspot and splice donor) variants, and a second UCEC harbored a somatic DICER1 (hotspot) variant. In TARGET, a single participant with a neuroblastoma harbored a germline DICER1 canonical splice variant. There were no atypical or unusual features noted on pathology review of publicly-available slides from TCGA tumors. In TCGA, across many types of cancer, the prevalence of pathogenic DICER1 variation was higher compared with the general population (~1:4,600 vs. ~1:10,000), a finding driven by one cancer type (UCEC). In the smaller TARGET dataset, we observed a single pathogenic germline variant in neuroblastoma. This is the first comprehensive analysis of DICER1 pathogenic variation in large adult and pediatric cancer populations. The novel association of DICER1 variation with UCEC merits additional investigation. Citation Format: Jung Kim, Kris Ann P. Schultz, D. Ashley A. Hill, Douglas R. Stewart. The prevalence of DICER1 pathogenic variation in cancer population databases [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2070.