Comprehensive genomic characterization of upper tract urothelial carcinoma (UTUC).

Abstract

375 Background: Integrated analysis of UTUC was performed to characterize the genomic landscape of UTUC and provide insights into biology. Methods: 31 untreated samples underwent WES, RNAseq, and RPPA. Consensus mutation calls from independent pipelines of 2 centers identified gene expression clusters using unsupervised consensus hierarchical clustering (UCHC). Results: Clinical data are shown in the Table. WES identified mutations in FGFR3 (74.1%; 92% low grade, 60% high grade), KMT2D (44.4%), PIK3CA (25.9%), TP53 (22.2%). APOBEC and CpG signatures were identified. Expressed marker genes from bladder TCGA were not associated with molecular subtypes in this study. UCHC of RNAseq segregated samples into 4 molecular subtypes. Cluster 1: no PIK3CA mutations; enriched for nonsmokers, high grade non-muscle invasive tumors, high recurrence rates and favorable survival. Cluster 2: 100% FGFR3 mutations; enriched for low grade, noninvasive disease, no bladder recurrences. Cluster 3: 100% FGFR3 mutations, 71% PIK3CA, no TP53 mutations; high tobacco use and bladder recurrence (62.5%); tumors all non-muscle invasive. Cluster 4: KMT2D (62.5%), FGFR3 (50%), TP53 (50%) mutations, no PIK3CA mutations; enriched for high grade, muscle-invasive disease, tobacco use, CIS, reduced survival; novel fusion of SH3KBP1-CNTNAP5 was identified, with high expression levels in the sample. Conclusions: We show mutations in UTUC occurring at differing frequencies and subtypes than bladder cancer. Novel fusion SH3KBP1 regulates RTK signaling and acts to recycle TGFβ receptors. Further studies are needed to validate the described subtypes, explore their responses to therapy, and better define the novel fusion mutation. [Table: see text]