Prospective Study for Evaluating Colon Polyp Histology With in Vivo Probe Based Confocal Laser Endomicroscopy

Overview

Chromoendoscopy (that involves spraying of dyes over the colonic mucosa) combined with magnification has been utilized for polyp histology identification. Pit patterns on the surface of polyps described by Kudo et al have been shown to have a high diagnostic accuracy in differentiating the polyp types (18, 19). NBI, that is also referred to as "electronic chromoendoscopy" is another technique that has been evaluated for polyp histology identification by highlighting the superficial mucosal and vascular architecture (15, 20, 21). pCLE is another novel addition to the technologies aiming to accomplish in vivo histologic diagnosis with a high degree of accuracy. The pCLE system has three major components (Mauna Kea Technologies, Paris, France). The first is the confocal miniprobe made of approximately thirty thousand optical fibers bundled together and terminated by a distal microsystem. The images obtained have a lateral resolution of 1µm, an axial resolution of 10 µm and a maximum field of view of 240 µm. The depth of observation is from 55 to 65 µm. The miniprobe tip diameter is 2.5 mm and can be passed through the accessory channel of any standard endoscope. The second is the laser scanning unit (excitation wavelength - 488 nm) that combines the functions of laser light illumination and rapid laser scanning, enabling a frame rate up to 12 images per second and signal detection. The third is the control and acquisition software for real time image reconstruction, immediate sequences display and post-procedure analysis and editing tools. Once an area of interest (e.g. a polyp) is identified, 5 ml of 10% fluorescein sodium is injected intravenously; the confocal probe is passed through the accessory channel of the endoscope and placed against the lesion to obtain several high-quality images and video sequences. In a study by Buchner et al from the Mayo Clinic, Jacksonville, (22) this system was used to evaluate confocal images of 37 polyps from 25 patients in a blinded fashion without the knowledge of their histologic diagnosis or endoscopic appearance. The investigators developed the following criteria that were suggestive of neoplastic polyps: villiform pattern, nuclear characteristics - oval/irregular nuclear shape and increased number of nuclei. These features had a sensitivity of 82.6%, specificity of 92.9% and accuracy of 86.5% for the characterization of neoplastic polyps. Similarly, Meining et al (23) have also evaluated criteria for differentiating neoplastic from benign lesions in the colon with encouraging results. The investigators hypothesize that pCLE will have a high rate for accurate characterization of polyp histology real time during colonoscopy

This study will be conducted at the Kansas City VA Medical Center, Kansas City, MO. The investigators have the pCLE system that the investigators are currently using in a multicenter study evaluating patients with Barrett's esophagus. Patients referred for screening and/or surveillance colonoscopy will be prospectively enrolled in this study. Inclusion criteria - referral for screening and/or surveillance colonoscopy and the ability to provide informed consent. Exclusion criteria: prior surgical resection of any portion of colon, prior history of colon cancer, history of inflammatory bowel disease, use of anti-platelet agents or anticoagulants that precludes removal of polyps during the procedure, poor general condition or any other reason to avoid prolonged procedure time, history of polyposis syndrome or HNPCC, inability to give informed consent, inadequate bowel preparation, allergy to fluorescein, pregnancy, and renal insufficiency. Moderate sedation for the procedure will be administered in a standard fashion with intravenous midazolam, meperidine or fentanyl. This will be a single arm study with all procedures being performed with a standard white light colonoscope (CF - H180AL, Olympus America). After cecal intubation, the colonic mucosa will be carefully visualized during withdrawal under high-definition white light and if a polyp is detected, its characteristics will be documented: size, location, and morphology \[Polypoid (sessile, pedunculated) or Non-polypoid: (superficial elevated, completely flat, depressed)\]. Then 5ml of 10% fluorescein sodium will be injected intravenously. Following this the pCLE probe will be passed through the biopsy channel of the colonoscope and placed on the polyp to obtain confocal images and video sequences. During the initial phase unblinded comparisons of the confocal images and videos with the polyp histology (20 adenomas, 20 hyperplastic) will be performed in order to evaluate the criteria described by the Mayo Jacksonville group (22) that differentiate between neoplastic and non neoplastic polyps. Following this initial phase, these criteria will be applied in a prospective manner to predict polyp histology (100 polyps) real time during the procedure prior to their removal. Multiple confocal images and videos of each polyp will be saved with appropriate labeling. Pathologists will be blinded to the endoscopic and pCLE findings of the polyps. Polyp size will be assessed by comparing with sheath of a polypectomy snare or open span of a biopsy forceps. Photo documentation of the polyps will be performed. Polyps will then be removed in the standard fashion with a biopsy forceps or snare and sent for histopathological evaluation. Each polyp will be sent in a separate jar and labeled accordingly. The bowel preparation will be evaluated and graded according to previously reported criteria (24): excellent, good, fair, and inadequate. Patients with inadequate bowel preparation will be excluded. Complications including gastrointestinal bleeding (requiring intervention) and perforation will be recorded. Outcomes: The predicted histology of polyps based on the confocal images and videos and the correlation with actual histology for accuracy of prediction will be the primary outcome. Secondary outcome will be the inter-observer variability in polyp histology prediction based on the confocal images and videos of polyps (by kappa statistics).