Use this URL to cite or link to this record in EThOS:
Title: Identifying risk factors and biomarkers for colorectal cancer screening
Author: Thomas, Darren S.
ISNI:       0000 0004 7970 7219
Awarding Body: UCL (University College London)
Current Institution: University College London (University of London)
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Background. An inefficient selection of screenees and low uptake to stool-based screening undermines the potential benefits of the BCSP. The present thesis, through a series of cohort and case-control studies nested within the UKCTOCS involving 202,638 postmenopausal women, studies the viability of risk-stratified and blood-based screening for detecting asymptomatic CRCs. Chapter 3. CRCs diagnosed during 2001-11 following randomisation to the UKCTOCS were identified via EHRs and self-report and later verified with their treating clinician. Notifications for CRC diagnoses were received for 1,085 women by May 2011, in which 514 has CRC, 24 had benign polyp, and 103 had neither diagnosis. Cancer & death registrations and Hospital Episode Statistics used in adjunct had a sensitivity of 98% and a positive predictive value of 92%. EHRs together are a reliable means for ascertaining events of CRC. Chapter 4. A NGRS for risk-stratified screening was derived from a Cox proportional- hazards model of 202,323 women (4,134 CRCs) and 16 shortlisted covariates recorded in EHRs. Reclassifying the risk of a bootstrap resample (n 138,900 (2,678 CRCs)), 5.6% more potentially harmful screens can be undertaken to potentially detect the same proportion of cancers than criterium based on age alone (50-74 years). The top decile had a three-fold increased hazards over the geometric mean. Chapter 5. CA125, CEA, CYFRA21-1, FAP, TIMP-1, and VEGFA were shortlisted from the literature and later evaluated as screening markers on immunoassay measurements of 386 longitudinal preclinical sera drawn 0-4 years before the diagnosis of benign adenoma and CRC and matched controls who remained cancer-free. CEA is elevated up to a year and four years in advance of diagnosis in 23 and 12% of cases, respectively. Other markers had sensitivities of 5-12 and 4-10% at the same lead times. Chapter 6. Novel markers were sought through analyses of pools of preclinical sera using LC-MS/MS with quantitative Tandem Mass Tags. DCD and S100A8 were shortlisted on their differential and longitudinal expression in CRCs and controls and were evaluated using individual measurements obtained with immunoassays. DCD and S100A8 were elevated up to one year before diagnosis for only 10% of CRCs. Chapter 7. The thesis concludes with an exploration of all multi-marker permutations. The top-ranked-a panel of CA125, CEA, FAP, TIMP-1 and VEGFA- had, at 95% specificity, a sensitivity of 23% for detecting CRCs up to one-year and 10% up to four years before diagnosis. The panel performed no better than CEA alone. Afterword. The use of EHRs to ascertain events of CRC in BigData is justified. The NGRS requires refinements before informing risk-stratified screening. CEA is elevated up to one year before diagnosis in 23% of future CRC patients. No circulating proteins are viable as standalone tools for screening for CRC.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available