Use this URL to cite or link to this record in EThOS:
Title: Raman spectroscopy and colorectal cancer : towards early diagnosis and personalised medicine
Author: Jenkins, Cerys A.
ISNI:       0000 0004 7967 4533
Awarding Body: Swansea University
Current Institution: Swansea University
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
The development of healthcare technologies to streamline patient referral systems and diagnose the early onset of disease is of great importance for improving cancer survival and is the basis of this work. This thesis details the development of Raman spectroscopy as a triage tool for urgent suspected colorectal cancer referrals. In this work the development of high-throughput, cost effective standardised platforms for the analysis of biofluids with Raman spectroscopy has been shown. The platforms developed allow the analysis of both dry and liquid biofluid samples. The optimal liquid biopsy for colorectal cancer applications was found to be serum due to its ability to be stored and transported without the formation of precipitates within the samples. Serum samples were then used to optimise dry and liquid HT-platforms for reproducible spectral collection. Principal component analysis (PCA) was used to investigate and optimise inter-user measurements to ensure a robust measurement platform. PCA analysis showed that patient fasting status and sex could have potential effects on spectral reproducibility and diagnostic capability. The liquid HT platform developed had less sensitivity for colorectal cancer detection than a dry platform. However, it showed lower inter-user spectral variations and the overall analysis time for each sample was faster. It was also less susceptible to freeze-thaw sampling effects in terms of diagnostic capability. This made it the method of choice when considering a translatable technology. The limits of the liquid HT platform were investigated with random forest based machine learning to develop diagnostic models for serum spectra. It was established that the technique could be used for the detection of precursor cancer lesions when tested against healthy control patients with a positive predictive value (PPV) of 40.00% and a negative predictive value (NPV) of 88.89%. The technique could also detect CRC in a large cohort of test patients against healthy controls with a NPV of 94.44%. This approaches the NPV of approximately 98% for the gold standard diagnostic test (colonoscopy) for colorectal cancer. The thesis concludes by discussing the clinical translation of the technique as an effective diagnostic based upon the results presented.
Supervisor: Thornton, Catherine A. ; Harris, Dean A. ; Dunstan, Peter R. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral