localization of the colonoscope tip may not be reliable. We aimed to assess accuracy of colonoscopist localization of polyps at the time of colonoscopy and determine if patient, procedure or colonoscopist variables influence the accuracy of localization. Methods: We conducted a prospective observational study in consecutive patients presenting for elective, outpatient colonoscopy at an academic center. All procedures were performed by board-certified colonoscopists. Polyps found during colonoscope withdrawal were included. The endoscopic location of each polyp was reported by the colonoscopist at the time of resection and prospectively recorded. Magnetic endoscope imaging (ScopeGuide, Olympus, Japan) was used to generate a reference standard for polyp location. For each procedure, a ScopeGuide image of the instrument location on insertion, caecal intubation and during polypectomy was prospectively obtained. Colonoscopists were blinded to the ScopeGuide image. Three separate expert colonoscopists, blinded to the endoscopist’s assessment of polyp location, independently reviewed the ScopeGuide images to obtain a reference standard for polyp location (Cronbach alpha 0.98). We assessed the accuracy of colonoscopist polyp localisation using the ScopeGuide-derived reference standard. We assessed extent of deviation between colonoscopist location and ScopeGuide location, by dividing the colon in 18 segments from caecum to distal rectum. Results: A total of 282 polyps were resected in 155 patients by 16 colonoscopists. Overall accuracy of polyp localization was 85%. Accuracy varied significantly (p!0.001) by colonic segment: caecum 100%, ascending 77%, transverse 82%, descending 56%, sigmoid 90%, rectum 95%. There was no significant deviation between ScopeGuide location and colonoscopist location for polyps located in the distal sigmoid and rectum (mean deviation 0.3 segments, range 0-3 segments, pZ0.32). There was significant deviation between ScopeGuide location and colonoscopist location from caecum to mid-sigmoid colon (mean deviation 0.8, range 0-6 segments, p!0.001). Extent of deviation between actual and predicted location was not influenced by patient sex, age, body mass index, surgical history, bowel preparation or years of colonoscopy experience. Conclusion: Accuracy of localization of polyps is imprecise, and affected by position within the colon. Methods to improve lesion localization at colonoscopy should be studied.