Subtype-specific Patterns of Evolution and Clinically Relevant Genomic Instability in Wilms Tumour

Understanding cancer evolution is fundamental to predicting cancer progression. However, the evolution of paediatric cancers is still under-researched. Large cohorts of patients are required to determine consistent evolutionary trajectories that shed light on key steps in cancer development and reveal underlying biology, especially in rare cancers. Additionally, well annotated clinical data is necessary for determining if evolutionary biomarkers are predictive of patient outcome. We performed detailed evolutionary analysis of 64 paediatric kidney cancers, including 60 Wilms tumours (WT), using DNA microarrays and, in a subset of 30 patients, a WT-specific targeted sequencing assay, to detect copy number alterations (CNA) and mutations, respectively. By analysing multiple tissue samples in the majority of cases we could detect mutation heterogeneity in each tumour. We reconstructed clones across the cohort and described their phylogenetic histories, in addition to detecting mirrored subclonal allelic imbalance. Our results highlight pervasive evidence of parallel evolution in WTs affecting CNAs, and CTNNB1 and TP53 mutations. Furthermore, we demonstrate that stromal-type WTs often evolve from a consistent series of events (WT1 mutation, 11p uniparental disomy and CTNNB1 mutation) and we suggest that 19q uniparental disomy is an important ancestral event in both epithelial and diffuse anaplastic WTs. Finally, we propose the total number of evolutionary CNA events as a prognostic biomarker in WTs for event-free survival, particularly in high-risk WT. Overall, this study sheds light on the evolution of the most common paediatric kidney cancer and links evolutionary analysis to fundamental clinical and biological questions in a large cohort of WTs. We conclude that histological subtypes of WT are often defined by consistent evolutionary sequences. We present evidence that a key marker of evolvability, namely CNA diversity measured phylogenetically across multiple tumour sites, is prognostic of patient outcome and should be considered for clinical use to detect the most aggressive blastemal and diffuse anaplastic type WTs.