INTRODUCTION

To share the first experience of using the CAD EYE system in children and determine its potential application in pediatric practice.

OBJECTIVE

In 2019, Kudo et al. published the first work dedicated to artificial intelligence (AI)-assisted colonoscopy. This system was developed as an assistant for endoscopists to enhance the efficiency of detecting lesions such as hyperplastic polyps, serrated adenomas, and colorectal cancer. These conditions are classified under BASIC4, representing the most common lesions in adult patients. Currently, the system is not intended for use by children. Pediatric practice predominantly encounters pseudopolyps and hamartomatous polyps, with diseases from the BASIC4 group being exceptionally rare.

CASE REPORTS

We conducted 10 diagnostic colonoscopies on children with various diagnoses. Two patients were undergoing treatment in the rheumatology department and were referred for colonoscopy to rule out inflammatory bowel diseases (IBD). Six children under long-term treatment for IBD underwent surveillance studies. Two patients were in the surgical department, one suspected of a rectal polyp and the other with a confirmed diagnosis of juvenile polyposis. Among the four patients with identified polyps, attempts were made to examine them using AI. The CAD-EYE system mostly identified these formations as non-neoplastic.

CONCLUSION

In our clinical case series, the CAD EYE system proved most useful in the primary search mode in white light and LCI mode, where it assesses the overall landscape and helps focus attention on lesions. In BLI mode, a reliable assessment of the formations was not achieved.