Discovery and Clinical Utility of Novel Prognostic Biomarkers in Gastroesophageal Cancer

<p dir="ltr">Gastric, gastroesophageal junction, and oesophageal adenocarcinomas are amongst the most common and lethal cancers. Biomarkers of cancer are vital for prognostication of patient survival, prediction of response to treatment, and allow for a better understanding of tumour progression through a treatment course. Understanding clinical biomarkers could allow for novel treatment targets to be identified, or re-utilisation of current therapeutics. This thesis investigates known and potential new molecular and clinical biomarkers of gastroesophageal cancers and further assesses their clinical relevance.</p><p dir="ltr"><b>Chapter 1</b> introduces gastric, gastroesophageal junction and oesophageal adenocarcinomas (GOCs), focusing on current clinically relevant and experimental biomarkers explored in this thesis. Background information is provided on the urokinase plasminogen activator system (uPAS), epithelial to mesenchymal transition (EMT), human epidermal growth factor receptor 2 (HER2), tumour budding (TB) and the effects of patient nutrition status on GOCs.</p><p dir="ltr"><b>Chapter 2</b> details the uPAS, a proteolytic pathway resulting in the remodelling of extracellular matrices facilitating tumour invasions and metastasis. This includes an up-to-date review of the clinical relevance of uPAS in GOCs. uPAS activity is demonstrated widely in GOCs and is associated with adverse tumour features, poorer patient survival and has a potential role as both a circulating biomarker and therapeutic target.</p><p dir="ltr">Tumour budding (TB) is a novel pathological predictive biomarker. <b>Chapter 3</b> describes TB in a local cohort of GOC patients by histological and immunohistochemical investigation. Patients with GOC and high levels of TB are associated with a shorter disease-free survival, confirming previous findings with other GOC patient cohorts globally. For the first time, the expression of the urokinase plasminogen activator receptor (uPAR) and E-cadherin was explored on tumour buds (TBs) directly and each was independently shown to have prognostic potential. A novel score combining uPAR and E-cadherin identifying TBs with a mesenchymal-like phenotype (denoted as TBmes) was shown to be an independent GOC prognostic biomarker, better at detecting high risk patients than standard TB enumeration, or sole evaluation of uPAR or E-cadherin TB expression. The fact that TBmes are strongly associated with poor prognosis also enhances our understanding of the underlying mechanism of TB formation and the role of TBs in GOC metastasis. This paradigm is then explored more broadly within the tumour invasion front (<b>Chapter 3 appendix</b>), where the EMT status of tumour cells is also a strong prognostic factor of GOCs.</p><p dir="ltr">Cancers demonstrate heterogeneity, changing over the patient’s cancer journey. Novel methods are required to more readily identify cancer progression, tumour treatment resistance and patient prognosis. In <b>Chapter 4 </b>human peripheral blood mononuclear cells (PBMCs) from patients with GOC and from healthy individuals were analysed for uPAR expression by comprehensive multicolour flow cytometry. Confirming earlier findings in other pathologies, the proportion of uPAR positive cells was greater than 90% of all monocytes (with significant differences seen between monocyte subsets from GOC) and less than 1% of all CD3+ T-lymphocytes across both cohorts. However, a significantly increased proportion of uPAR positive T-lymphocytes was observed in GOC patients compared to those from healthy individuals. This novel finding was associated with shorter time to radiological progression than low uPAR expressing T-lymphocytes in patients with metastatic GOC. PBMC uPAR expression was also explored in a cohort of head and neck cancer patients undergoing curative treatment (<b>Chapter 4 appendix</b>) for comparison to the GOC cohort.</p><p dir="ltr">HER2 is a well-established biomarker of GOC, offering both prognostic and predictive information, including for standard therapeutic options. In <b>Chapter 5</b>, the association between HER2 expressing GOCs and the development of brain metastasis is established, demonstrating that patients with HER2 positive tumours have a higher rate of brain metastasis. This strongly suggests that patients with HER2 expression should be considered for additional brain imaging and an aggressive approach to their brain metastasis management.</p><p dir="ltr"><b>Chapter 6</b> explores the impact of the nutritional status of GOC patients in the peri-operative setting. There is often a significant delay in the diagnosis of GOCs, which adversely affects the baseline nutrition of patients. Pre-operative malnutrition is multifactorial, including but not limited to cancer related cachexia and mechanical obstruction, and can be further compounded by anti-cancer treatments. We demonstrate that changes in patient nutrition over the pre-operative period result in poorer patient survival. Two methods of assessing GOC patient malnutrition, reduction in weight loss and the clinically validated patient generated subjective global assessment, are associated with worse survival outcomes. The study demonstrates the important need to improve peri-operative nutrition for the long-term survival of patients.</p><p dir="ltr"><b>Chapter 7 </b>addresses the main finding established in each chapter, synthesises this information within the current literature and identifies potential future research opportunities.</p>