Introduction:
Prognostic assessment of rectal cancers pre-&post-chemoradiotherapy (CRT) relies on precise definitions, reproducibility of regression-assessments and effective application in clinical contexts. Achieving confidence in regression assessment allows more accurate sub-division of known poor prognosticators, which may yield cohorts with better prognosis. Novel analysis methods examining circulating tumour DNA(ctDNA)may identify variations in tumour burden at a microscopic level.
Methods:
A range of methods were employed in this thesis including literature and systematic reviews, cohort studies, interobserver agreements and a multicenter prospective pilot study.
Results:
Histopathological good response is defined as Mandard1&2/Dworak3&4; poor response as Mandard3,4&5/Dworak0,1&2 according to published survival outcomes. A good and poor response on MRI was mrTRG1,2&3 and mrTRG4,5 respectively. The single best individual agreement between pTRG and mrTRG for good-response was 83% achieved using the RCPath, RCRG and Wotherspoon Scales and for poor-response was 89% using RCPath but interobserver agreement was 0.45. Median interobserver agreement for mrTRG was 0.57 (0.14-0.82). Proformas improved MRI report quality (98% of key metrics reported). Tumour regressed by a mean of 36% in the extramural component and 51% in the intraluminal part. T3 tumours (<5mm beyond muscularis) on baseline/post CRT MRI scans and final histopathologic specimens have better survival than T3cd(>5mm). mrEMVI tumours have a 5-fold increased rate of synchronous tumours and almost 4-fold ongoing risk of developing metastases in follow up after surgery. ctDNA has prognostic implications and intraoperative tumour manipulation could lead to a surge of tumour DNA into the circulation but detection may vary according to tumour height.
New knowledge contributed to the management of rectal cancer:
This thesis identified precise definitions for good and poor response for histopathology and MRI tumour regression scales, and identified the pTRG scale most closely correlated with mrTRG. pTRG scales have low interobserver agreements but mrTRG can be taught to inexperienced radiologists with higher K. Proformas help improve the quality of MRI reports. T3 tumour extension is prognostic on histopathology, MRI at baseline/after neoadjuvant therapy. Tumours regress heterogeneously. The metastatic risk associated with mrEMVI and circulating tumour DNA was quantified. ctDNA was identified in mrEMVI positive patients and a surge from intraoperative manipulation was seen, depending upon tumour height.