Understanding the response to pre-operative therapies in colorectal cancer through phenotypic and Raman spectroscopic analysis

Colorectal cancer (CRC) is common and in the UK alone, approximately 42,000 people are diagnosed annually. Treatments offered to CRC patients may vary, but typically high-risk CRC patients may be offered pre-operative radiotherapy (RT), chemotherapy or both. However, CRC is morphologically and molecularly heterogeneous and the degree of tumour response to pre-operative therapies is highly variable – some patients may show a complete pathological response whereby no viable tumour remains and others will show no evidence of tumour regression.
Despite the clinical benefit of pre-operative therapies, they are associated with significant adverse effects, ranging from mild to severe, and in some cases even fatal. Despite this, there are no clinically accepted methods to predict patient response to the therapies and for some patients, the adverse effects may outweigh the potential benefit of receiving them. The overall aim of the work in this thesis was to address the current lack of a reliable, clinically approved method to predict the response to pre-operative therapy in CRC patients using Raman spectroscopy (RS).
Raman spectra were collected from the pre-treatment formalin-fixed paraffin-embedded (FFPE) biopsy specimens of 20 rectal cancer patients that showed a varying degree of tumour response to pre-operative RT, as determined by tumour cell density scoring. Using a PCA-LDA algorithm, RS was able to stratify patient response to pre-operative RT into good and poor with approximately 86 % accuracy, as determined using five-fold cross-validation. Further analysis of the spectra showed the Raman signature of good responders demonstrated increased protein contributions, whereas poor responders showed elevated lipid contributions. In addition, single cell trapping microfluidic devices were employed to investigate the response to common CRC chemotherapies at the single cell level. The results of which demonstrated the feasibility of collecting Raman data from single cells on-chip pre- and post-exposure to chemotherapeutics.
The work detailed in this thesis demonstrates the ability of RS to reliably classify tumour response to pre-operative therapies from routine pre-treatment FFPE biopsies, and the potential to extend this to work to the analysis of single cells on microfluidic devices, highlighting the potential of RS to guide personalised CRC treatment.