Novel image enhancement technology that helps find bleeding points during endoscopic submucosal dissection of gastric neoplasms.

Background And Aim: The management of bleeding during endoscopic submucosal dissection (ESD) is critical and related to the procedure time. We collaborated on a new image enhancement algorithm with parameter optimization for clinical use being developed by FUJIFILM Co. and processed white light image data offline to evaluate the effectiveness of this technology. This study aims to evaluate the clinical usefulness of this technology.

Methods: Eighteen video scenes of bleeding points from five gastric ESDs were selected and processed by the new image enhancement algorithm. The time until a bleeding point was found, visibility of a bleeding point, and color abnormality of the submucosal layer were evaluated by ESD experts, ESD trainees, and endoscopy trainees. The color differences between the bleeding point and the surroundings in CIE-L*a*b* color space were calculated in the original and enhanced images.

Results: The time until a bleeding point was found in the enhanced videos was significantly shorter than that in the original videos (11.10 s vs 13.85 s) (P = 0.017). On a 5-point (-2 to +2) Likert scale of visibility, the enhanced image was slightly superior to the original (+0.45), and the appearance of the submucosa was comparable between images (+0.14). The color difference among the bleeding areas on the enhanced images was significantly larger than that on the original images (10.93 vs 8.36).

Conclusion: This novel image enhancement algorithm emphasizes the color difference between a bleeding point and the surrounding area, which would help find bleeding points faster during ESD for the less experienced endoscopists.