Novel endoscopic techniques in early gastrointestinal neoplasia
Student thesis: Doctoral Thesis
Advances in endoscopic practice are challenging traditional views on how neoplasia should be managed in the gastrointestinal system. Identifying pre-malignant changes in the gut is central to successful early management of most cancers. There has been rapid growth in the equipment and technology available for evaluating potentially neoplastic lesions in the gastrointestinal tract. However, the evidence base for how they should be used is limited. This thesis investigates the use of chromoendoscopy and ‘virtual computed chromoendoscopy’ in the gastrointestinal tract for the identification, assessment and management of early gastrointestinal neoplasia. The first part of the thesis reviews data on the health burden of gastrointestinal neoplasia. It reviews the background research into advanced endoscopic techniques in the colon and oesophagus and outlines the deficiencies in the published literature. Chapters 7-9 describe the development and validation of a new tool (N.A.C.) for invivo histology prediction of colonic polyps <10mm in size using Flexible Spectral Imaging Colour Enhancement (FICE) and indigo carmine (IC) chromoendoscopy without optical magnification. Chapter 7 describes a study to determine the optimum FICE setting for assessment of polyps. Setting 4 came out as the best setting for this purpose. Chapters 8 and 9 describe two studies to define the key surface features associated with neoplastic and non-neoplastic polyps using first a picture based study and then in an in-vivo study. N.A.C. was then used in a prospective cohort study of polyp assessment in a Bowel cancer Screening Population, described in Chapter 10. A prospective study of 232 polyps <10mm in size was performed. FICE and IC significantly improved the accuracy of the in-vivo diagnosis of polyps as compared to WLI endoscopy. In-vivo prediction of polyp histology allowed the endoscopist to set the correct surveillance interval in 83% of cases using WLI alone and in 97% of cases using either FICE or IC based on BSG guidelines. Chapter 11 describes a study which attempts to identify the role of high resolution (HD) endoscopes in lesion characterization. HD endoscopes significantly improved the ability of the endoscopist to make an in vivo diagnosis using FICE compared to standard resolution endoscopes with FICE but had no effect on the WLI or IC predicted diagnosis. Chapter 12 describes the use of acetic acid in a cohort study for the identification of neoplasia within Barrett’s oesophagus to establish the sensitivity and specificity of the technique. Acetic acid chromoendoscopy had a sensitivity of 95.5% and specificity of 80% for the detection of neoplasia. There was a correlation between lesions predicted to be neoplasic by acetic acid and those predicted by histological analysis (r=0.98). There was a significant improvement in the detection of neoplasia using acetic acid compared with white light endoscopy (P=0.001). This method is then refined in a second study by attempting to exploit the aceto-whitening reaction. This is described in Chapter 13. Data from 146 areas of Barrett’s was collected from 121 patients. 84% were male. Mean age 69. In total 72/86 metaplasia, 6/14 LGD, 26/27 HGD, 15/15 IMC and 12/12 areas of invasive cancer were recognised correctly by the endoscopist. A ROC curve was produced for the identification of high risk neoplasia (HGD, IMC and invasive cancer) using the aceto-whitening timings. The area under the curve was 0.93 (95% C.I.0.89-0.97), demonstrating a low probability that a randomly chosen positive case will exceed the value for a randomly chosen negative case. Using a cut off threshold of 142 seconds a sensitivity for neoplasia of 98% (95% C.I. 89-100) and specificity of 84% (95% C.I. 74-91) was achieved. The final chapter of the thesis discusses the clinical impact of the research and how it could be introduced to clinical practice. It reflects on where deficiencies in the published literature still exist where the future research needs are.
3.37 MB, PDF document