News Releases & Research Results Approval of deep learning-based software (EndoBRAIN®-EYE) as a medical device to support endoscopic image diagnosis

The results of collaborative R&D conducted by Specially Appointed Professor Shin-ei Kudo of the Digestive Disease Center, Showa University Northern Yokohama Hospital and Professor Kensaku Mori of the Graduate School of Informatics, Nagoya University.

• EndoBRAIN®-EYE, software for predicting the presence or absence of lesions on endoscopic images by image analysis using artificial intelligence (AI), was developed and approved as a Class II medical device according to the Pharmaceutical and Medical Device Act, on January 24.
• The deep learning-based software was developed by learning about 3.95 million endoscopic images extracted from videos collected from five domestic institutions (Showa University Northern Yokohama Hospital, National Cancer Center Hospital, Shizuoka Cancer Center, Tokyo Medical and Dental University Hospital, and Cancer Institute Ariake Hospital). A clinical performance test showed that lesions could be accurately detected with 95% sensitivity and 89% specificity, demonstrating sufficient accuracy to support endoscopists.

This R&D project was conducted with the support of "Project on Utilizing High Definition Medical Imaging Data up to 8K quality" by the Japan Medical Research and Development Agency (AMED).