Application of Artificial Intelligence for Early Diagnosis of Gastric Cancer During Optical Enhancement Magnifying Endoscopy

Shandong University

Status

Unknown

Conditions

Magnifying Endoscopy

Artificial Intelligence

Optical Enhancement Endoscopy

Study type

Observational

Funder types

Other

Identifiers

NCT04563416

2018SDU-QILU-3

Details and patient eligibility

About

Previous prospective randomized controlled study demonstrated higher accuracy rate of diagnosing early gastric cancers by Magnifying image-enhanced endoscopy than conventional white-light endoscopy. Nevertheless, it is difficult to differentiate early gastric cancer from noncancerous lesions for beginner. we developed a new computer-aided system to assist endoscopists in identifying early gastric cancers in magnifying optical enhancement images.

Full description

Gastric cancer is the third most common cause of cancer-associated deaths worldwide especially in Asia.Early detection and treatment would cure the disease with 5-year survival rate greater than 90%.However, the sensitivity of conventional endoscopy with white-light imaging (C-WLI) in diagnosis of early gastric cancers (EGCs) is merely 40%. Magnifying image-enhanced endoscopy (IEE) techniques such as magnifying narrow band imaging (M-NBI) have been developed and 2 RCT report that white-light imaging combine with M-NBI can increase the sensitivity to 95%. The strategy that using white-light imaging to detect the suspicious lesion and using M-IEE techniques to make a diagnosis of early gastric cancer is recommend in screening endoscopy.

Optical enhancement (OE) which is one of the M-IEE techniques was developed by HOYA Co. (Tokyo, Japan) . This technology combines digital signal processing and optical filterers to clear display of mucosal microsurface (MS) and microvessel (MV). The advantage of OE is to overcome the darkness of NBI which leads to less usefulness for detect-ability in the full extended gastrointestinal lumen.Nevertheless, it is difficult to differentiate early gastric cancer from noncancerous lesions for beginner, and expertise with sub-optimal inter-observer agreement is essential for the use of M-IEE.

Nowadays, Artificial intelligence (AI) using deep machine learning has made a major breakthrough in gastroenterology, which using gradient descent method and backpropagation to automatically extract specific images features. The diagnostic accuracy in identifying upper gastrointestinal cancer was 0.955 in C-WLI . Polyps can be identified in real time with 96% accuracy in screening colonoscopy. AI show an outstanding application in detection and diagnosis.

This study aims to develop a M-OE assistance model in the diagnosis of EGCs by distinguishing cancer or not.

Enrollment

80 estimated patients

Sex

All

Ages

18+ years old

Volunteers

No Healthy Volunteers

Inclusion criteria