Genomic markers in differentiated thyroid cancer

Abstract

This thesis explores the potential use of genomic markers to assist the management of differentiated thyroid cancer (DTC) patients throughout the course of their treatment pathway. We explore the use of these markers to prognosticate and risk stratify patients, whilst also investigating their role as sensitive and specific markers of disease burden. The first part of the thesis investigates a genomic signature capable of predicting the presence of high risk histopathological features. These features commonly upstage DTCs from low to higher risk tumours and potentially influence the decision for clinicians to recommend completion thyroidectomy after initial hema-thyroidectomy. We hypothesised that predictive genomic biomarkers obtained from tumour samples provide additional prognostic stratification and assist in surgical decision making. We conducted a retrospective, observational cohort study in adult patients with stage I-IV DTC. The index tumours from these patients underwent multi-mutational analysis using next generation sequencing as well as microRNA (miRNA) differential expression analysis using Nanostring and quantatitive polymerase chain reaction (qPCR). We identified mutational patterns that were significantly associated with aggressive pathological features such as disease recurrence, vascular invasion and distant metastases. By combining the dysregulation statuses of various miRNA species with the somatic mutation status of various oncogenes, we proposed a predictive genomic signature that correctly identified DTC samples with high risk histopathological features with high sensitivity and specificity. These findings could be translated clinically into adjunctive molecular testing that can complement our existing prognostic investigations for better risk stratification. The second part of the thesis addresses the novel use of bespoke patient specific sequencing panels, to increase the limit of detection of circulating tumour DNA (ctDNA) in DTC. We hypothesised that existing sensitive PCR based techniques were limited by the inability to perform large target multiplexing and that the use of Whole Exome Sequencing (WES) to inform the production of patient tumour specific sequencing panels combined with new sequencing techniques, could improve detection rates of ctDNA from peri-treatment blood samples. In this prospective pilot study, we demonstrated that WES was more effective at identifying putative mutational targets with which ctDNA can be identified. We showed that traditional targeted high depth sequencing with singleplex droplet digital PCR(ddPCR) assays detected ctDNA in only a minority of our patient cohort, whilst the rate of detection significantly increased when using patient specific bespoke sequencing techniques for the detection of minimal residual disease. We found that ctDNA detection and tracking analysis is most effective in AJCC stage 2 disease and above, although clinical valididty and utility will need to be confirmed in larger prospective trials.