Colorectal polyp size is related to the risk of exhibiting advanced histological features. Moreover, polyps larger than 10 mm are associated with an elevated risk of metachronous advanced neoplasia and colorectal cancer (CRC). Consequently, accurate measurement of polyp size, especially at the 10 mm threshold is critical for risk stratification and surveillance intervals. Furthermore, polyp size is also important for the choice of the appropriate resection procedures. Underestimation may lead to delayed diagnosis, thereby increasing the risk of colorectal cancer, while overestimation may result in unnecessary surveillance endoscopies.

Concerns regarding polyp size measurement are increasing. In daily practice, polyp size is mostly estimated based on visual size estimation by the endoscopist, which has proven to be inaccurate, with reported accuracies ranging from 54% to 65%. Furthermore, a prevalent terminal digit preference has been observed, with clustering of size measurements at 5-mm intervals. Additionally, inappropriate surveillance recommendations based on endoscopists' visual assessment have been documented for up to 35% of polyps. Some endoscopists routinely use non-calibrated instruments such as biopsy forceps or polypectomy snares to improve polyp size estimation. However, this approach has proven to be costly and time-consuming. Besides, endoscopes' fisheye lens can lead to size distortions depending on placement of an instrument in the center or periphery of the endoscope image.

AI-assisted systems, such as Virtual Scale Endoscopy (Fujifilm, Tokyo, Japan) and ENDOANGEL-CPS (Renmin Hospital of Wuhan University, Wuhan, China), have been developed to increase the accuracy of polyp size measurement. However, these systems have primarily been tested in video-based studies, which may not accurately reflect real-time conditions. Recently, a new artificial intelligence system (EndoAIM) has been developed that allows polyp size measurement during real-time endoscopy. This system can automatically estimate the diameter of polyps. Preclinical studies have shown promising performance for EndoAIM, with an overall accuracy of 0.98 in discriminating polyp size at the 10 mm threshold. However, its measurement accuracy has not been tested in real patients during real-time endoscopy.

In this research project, we aim to evaluate the measurement accuracy of EndoAIM for estimating colorectal polyp size in real patients during real-time endoscopy.

Inclusion Criteria:

Subjects are eligible if:

1. They have received colonoscopy for screening, surveillance or symptom investigation;
2. Aged 18 years older or above;
3. Written informed consent obtained.

Exclusion Criteria:

Subjects are excluded if they have:

1. Contraindication to colonoscopy (e.g. intestinal obstruction or perforation)
2. Contraindication or conditions precluding polyp resection (e.g. active gastrointestinal bleeding, uninterrupted anticoagulation or dual antiplatelets)
3. Advanced comorbid conditions (defined as American Society of Anesthesiologists grade 4 or above)

Individual participant data will not be shared.

Endoscopists will be blinded from EndoAIM's AI generated results.