An investigation into the clinicopathological characteristics of radiotherapy response in locally advanced rectal cancer

Abstract

The standard management of Locally advanced rectal cancer (LARC) is neoadjuvant therapy (NAT) by administering radiotherapy (RT), with or without chemotherapy, prior to resectional surgery. Response to NAT is an important prognosticator but is heterogenous, even in patients of the same disease stage. Pathological (pCR), with no specimen evidence of tumour or positive lymph nodes, or Clinical complete response (cCR), with no endoscopic or radiological evidence of residual tumour, are the epitomes of good response. Understanding the mechanisms that govern NAT response, as well as the means by which to assess and predict it, is key to overcoming radio-resistance and increase the number of patients achieving a complete response. Using in-depth analysis of tumours through medical imaging and endoscopy, in addition to analysis of the patient as a host through systemic inflammation and body composition studies, the present thesis examines the clinicopathological factors that characterise response. The present thesis aims to explore the relationships between the tumour and host, validate past studies on NAT response in a large western cohort, and advance the understanding of how to categorise response by developing novel assessments. Chapter 1 provides a broad overview of LARC, including the diagnosis, staging, multimodal and surgical management, and predictors of NAT response. In Chapter 2, the aims of this thesis are outlined. Retrospective analysis Chapters 3 – 10 retrospectively analyse a dataset of patients from NHS Greater Glasgow and Clyde who have undergone curative intent NAT at The Beatson West of Scotland Cancer Centre. These patients have either progressed to organ preservation if a cCR has been achieved or surgical resection. In Chapter 3, the role of MRI for staging and re-staging of LARC is explored. In this chapter, several MRI features are examined and reported to be predictors of recurrence and survival outcomes in LARC. A combination score of these features better stratifies patient prognosis after NAT over more established MRI assessments of response / regression. Chapter 4 reviews the Neoadjuvant Rectal (NAR) Score – a validated surrogate endpoint score of long-term outcomes used to assess response in LARC. The NAR score is validated within the current cohort and is then adapted to produce a novel surrogate score that utilises MRI staging and re-staging. This novel MRI NAR score (mrNAR) has stage independent prognostic value and, unlike the NAR score, can be used to stratify patients who do not progress to resection. Chapter 5 begins to explore the role of the host in NAT response through haematological markers of systemic inflammation. An elevated systemic inflammatory response (SIR) is known to consistently associate with worse prognosis in colorectal cancers. In this chapter I demonstrate that patients who have an elevated SIR at a time point after NAT are more likely to have an incomplete response to NAT. I hypothesise that this elevated SIR, represented by the validated modified Glasgow Prognostic Score (mGPS), results in a degree of radio-resistance. Chapter 6 begins to explore dimensional tumour measurements that are not a part of standard staging criteria. The role of dimensional tumour measurements in NAT response and recurrence remains unclear. In this chapter I explore this and also work to establish the relationship of tumour size to the SIR. Dimensional tumour size measurements provide additional predictive information for NAT response beyond clinical staging criteria and larger tumours are associated with elevated SIR markers. Chapter 7 builds on the study of imaging modalities of the tumour beyond that of standard features or dimensional measurements from diagnostic CT and MRI. This chapter utilises radiomics, which is an emerging field that involves the extraction of numerous quantitative features from medical images using advanced data algorithms. In this chapter, an exploratory radiomics study of