Novel strategies for the identification of biomarkers of non-Hodgkin lymphoma: evidence from the European Prospective Investigation into Cancer (EPIC)

Abstract

Non-Hodgkin’s Lymphomas (NHL) represent the eighth most common cancer in Western Europe. Yet despite their widespread prevalence and high mortality rate relatively little is known about the aetiology of these hematological malignancies. Consequently NHL represents an ideal candidate for the discovery of biomarkers lying along the causal pathway. Such biomarkers would allow the improved identification of risk factors and high risk individuals, as well as an enhanced understanding of lymphomageneisis. However, to date there has been little progress in determining validated predictive biomarkers of NHL. This thesis attempts to address some of the issues that have previously hampered the study of NHL through novel strategies of biomarker identification utilising novel methodologies, technologies and statistical techniques. The thesis comprises a nested case-control study within the European Prospective investigation into Cancer (EPIC) cohort and is split into two parts: the ‘validation of biomarkers’ and the ‘integration of biomarkers’. The most exciting finding was the identification of a novel biomarker for Follicular lymphoma based on the t(14;18) translocation which comprises a previously unknown pre-disease condition. Although no other predictive biomarkers were identified this work represents a ‘proof-of-principle’ for the use profile regression in the study of highly dimensional complex datasets, and the possibility of using mass-spectrometry derived metabolic profiles in the study of lymphoma. Part two of the thesis confirmed that the use of the ‘meet-in-the-middle’ approach was a valuable and feasible method for studying the complete causal pathway from risk factor to disease. Together these results highlight potential avenues for further study of NHL and confirm the utility of a number of novel strategies that can aid such work. Additionally it informs on some of the likely challenges that will be involved.