AbstractThis work describes a preliminary study to evaluate the use of Optical Project Tomography in the diagnosis of colorectal polyps. Colonic polypoid cancers are the earliest detectable form of colorectal cancer and can potentially be cured if completely excised. The diagnosis using conventional 2D histopathology can be difficult due to epithelial displacement and 3D imaging could improve this. Optical Projection Tomography is an in vitro 2D and 3D imaging technique for small biological specimens, which produces 2D virtual sections in three orthogonal planes as well as 3D rendered images.
Important diagnostic features of colorectal polyps such as dysplasia and villous morphology have been studied on OPT. Differentiating between high and low grade dysplasia was challenging due the resolution of the images but a better correlation with H&E sections was found when diagnosing cancer polyps. The 3D properties of OPT enable the surface morphology to be seen in far greater detail than we have seen before and this, understandably, shows a poor association with the 2D classification method of tubular, tubulovillous and villous adenomas which we used in this study. The 3D endoscopic Pit Pattern Classification was demonstrated well on OPT and highlighted that pit patterns frequently co-exist although the clinical relevance of this needs to be explored further.
Novel findings include the detection of specific vasculature patterns in benign and cancer polyps which may help us understand why some polyps bleed and some do not leading to the growth of interval colorectal cancers. A multispectral analysis of images has further explored the potential of OPT showing enhanced images of polyps when different ultraviolet light filters are used. Inter-observer variation exists when diagnosing colorectal polyps on H&E and we have demonstrated that this is even more marked when using OPT.
This is the first comprehensive study of OPT using human tissue and has brought with it many challenges for this reason. OPT has several limitations that prevent it from being marketed commercially as an independent diagnostic tool although it has potential as an adjunct to current histopathological techniques and it requires further validation.
|Date of Award||2016|
|Supervisor||Bob Steele (Supervisor), Francis Carey (Supervisor) & Sarah Wedden (Supervisor)|