| dc.description.abstract | Oesophageal cancer, one of the most aggressive cancer types is considered the seventh most common cancer in terms of incidence and the sixth most common cause of cancer deaths worldwide due to late diagnosis. In the UK, the oesophageal cancer incidence rate has increased by approximately 10% since the 1990s. At present, histopathology is the gold standard method for the diagnosis of oesophageal cancer, which rely on biopsy collection using an endoscopy procedure followed by the histological sample’s preparation. This method is invasive, time-consuming, and largely based on the pathologist's experience of diagnosis. Therefore, new diagnostic techniques are required to provide non-invasive methods for early and rapid diagnosis. Raman scattering has the potential to replace histopathology as the gold standard for diagnosis for a wide range of diseases. Raman scattering provides stain-free imaging with chemical-specificity derived from the intrinsic vibrational signatures of biomolecules. However, the low scattering cross-section severely limits the image acquisition speeds and like conventional histopathology, requires tissue sectioning to provide morphological imaging below the surface of tissue biopsies. Stimulated Raman scattering (SRS) has recently appeared as a powerful technique for (near)real-time Raman imaging in intact tissue samples. Thework in this thesis aimed to develop the stimulated Raman scattering (SRS) for rapid wavelength tuning and chemical imaging of clinical samples, such as cancer biopsies. This was achieved by making modification to a laser cavity to reduce the time of the wavelength tuning by approximately 35 times compared to the original cavity design. Furthermore, the cavity modification led to the spectra being separated efficiently and the wavelength tuning controlled by cavity length changes only. The improved design was applied to image frozen oesophageal tissues, which have four major pathology groups, normal, inflammation, columnar-lined (Barrett's) oesophagus (CLO) and low-grade dysplasia. A large area imaging was performed using the SRS technique at 2930 cm-1 for four different oesophageal tissues, which presented the morphological and structural information. However, histopathological diagnosis depends on the visualisation of the cell nucleus in the tissue. This component was not highlighted until the stimulated Raman histology approach was developed for small regions of interest in the CLO and the low-grade dysplasia sample, which required two different frequencies at 2840 cm-1 and 2930 cm-1. All SRS images were compared to haematoxlin and eosin (H&E) stained sections. Further comparisons were made between SRS and Raman imaging techniques, with SRS offering faster acquisition times and a higher spatial resolution. The spectral signature for the different pathological groups in the oesophageal tissues were explored in the high wavenumber (2800 – 2930 cm-1) region using hyperspectral SRS and compared with the spectra from the Raman. K-means clustering analysis was used to explore the morphochemical information using the CLO and low-grade dysplasia sections. Both techniques were able to demonstrate unique information such as the epithelial cells that form the oesophagus glands and surrounding connective tissue. It is concluded that SRS has the power to be one of the ideal imaging modalities to gather the molecular information in biological samples. However, it still needs more development due to the complexity of the system. | en_GB |
