When a polyp is found, it woulf be recorded by a white-light and OE1 model with magnification for assessing and predicting its histology. After that, its really histology will be reported by an pathologist.

When a polyp is found, OE mode 1 with magnification was first used with near focus the polyp and an endoscopist made a real-time prediction of polyp pathology. After that, high-definition mode and OE mode 1 without magnification were used to observe polyp sequencely. The video of all procedure was recorded.

After bowel cleansing with 2L polyethylene glycol, patients received colonoscopy under propofol intravenous anesthesia. When a polyp was found, we began to wash it and record the video. OE mode 1 with magnification was first used with near focus the polyp and an endoscopist made a real-time prediction of polyp pathology. After that, high-definition mode and OE mode 1 without magnification were used to observe polyp sequencely. Biopsy of the polyps were taken after observation, and the size of polyps was estimated by the biopsy forceps. The information of patient age, gender polyp morphology and location were recorded. After a video transcoding and clipping, the pathology type of each video clip was predicted by an endoscopist. The sensitivity, specificity, positive predicting value (PPV), negative predicting value (NPV) and accuracy were calculated. The pathology types of polyps were classified by non-neoplastic polyps and neoplastic polyps according to Vienna Classification. Statistics were performed by NCSS 11 and R software (Version 3.3.2).

Inclusion Criteria:

• Patients will be included if they are found polyps during endoscopy procesure.

Exclusion Criteria:

• Inflammatory bowel disease
• Biopsies were not available.
• Unable to provide informed consent.

Patients who experence an examation of coloscopy found polyps.