A Multicenter, Prospective, Randomized Controlled Trial for the Detection of Colon Polyps and Predicting Histology

This study is a multi-center randomized controlled trial that will be conducted at the Kansas City VA Medical Center, Kansas City, MO and The Barnes Jewish Hospital, St Louis, MO. Patients referred for screening and surveillance colonoscopy will be enrolled. They will be randomized to one of the following three groups: standard white light colonoscopy, high-definition white light colonoscopy or NBI.

All colonoscopies will be performed at the Kansas City VA Medical Center and The Barnes Jewish Hospital, St Louis, MO. Moderate sedation will be administered for the procedure with intravenous midazolam, demerol or fentanyl and Propofol. In patients randomized to the NBI arm, the colonoscope will be inserted to the cecum under white light and will switch to NBI mode at the time of withdrawal. After cecal intubation, the colonic mucosa will be carefully visualized during withdrawal of the scope with one of the three imaging modalities to which the patient has been randomized. No effort will be made to look for polyps during the insertion of the colonoscope. All polyps detected will be documented for their size, morphology (Polypoid: sessile, pedunculated or Non polypoid: superficially elevated, completely flat, depressed), and location. The surface of the polyp will be assessed during the procedure and the histology will be predicted real time by evaluating the surface mucosal and vascular patterns. The patterns would be classified into one of the four patterns described in our pilot study (tubular and gyrus patterns will be combined to tubulogyrus pattern). Any other patterns detected would be described and kept in the miscellaneous category. If a pattern is not identified on a polyp with either white light or NBI, it will be recorded as such and no attempts will be made to predict histology. Photo documentation of the surface patterns of the polyps will be performed. Polyps will then be removed either with a biopsy forcep or snare and sent for histopathological evaluation. Each polyp removed will have one or more corresponding picture of its surface patterns and these will be labeled accordingly. The total time spent in inspecting the colonic mucosa (excluding the time spent in removal of polyps or clearing the colon of liquid and solid debris) will be measured with a stop watch. Representative images of the various polyp patterns will be reviewed, discussed and standardized amongst all the investigators at both sites. In addition, images of all the polyp patterns will be printed and posted in the endoscopy suites at both hospitals.

The bowel preparation will be evaluated and graded as follow:

Excellent (score=1) - >90% mucosa seen, mostly liquid colonic contents, minimal suctioning needed for adequate visualization.

Good (score=2) - >90% of mucosa seen, mostly liquid colonic contents, significant suctioning needed for adequate visualization Fair (score=3) - > 90% mucosa seen, mixture of liquid and semisolid colonic contents, could be suctioned and/or washed Inadequate (score=4) - <90% mucosa seen, mixture of semisolid ands solid colonic contents, which could not be suctioned or washed