Scientists at Okayama University in Japan have developed an artificial intelligence (AI)-based endoscopic diagnosis system for the early identification of gastric cancer.
Early-stage gastric cancer can be treated using surgical gastrectomy procedures and endoscopic surgery (ESD), which can save the stomach.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
The use of endoscopy treatment or surgery is decided based the depth of cancer within the stomach wall. The treatment plan is decided after analysis of endoscopic images, said the researchers.
To help in early detection of the cancer, the team developed a prototype of the AI endoscope using GoogLeNet to match purpose via the image identification capability of Convolutional Neural Network (CNN) published by Google on the MATLAB numerical analysis software.
The researchers then used a 152-layer convolutional neural network, called ResNet, for intramucosal endoscopic resection in patients who were treated for early gastric cancer at Okayama University Hospital.
Moreover, they developed the AI system and validated its accuracy using endoscopic images of 100 cancers and 50 submucosal invasion cancers and 50 submucosal invasion cancers.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataFor intramucosal cancer, the diagnostic accuracy of the system was observed to be 82.7% sensitivity, 63% specificity, with a 69.1% positive predictive value and a 78.4% negative predictive rate in image unit.
In case unit, the values were 82%, 71%, 73.9% and 79.8% respectively.
A correct diagnosis rate of 72.8% in image units and 76.5% in case units was demonstrated with deep-seated diagnosis of early intramucosal cancer and submucosal invasion cancer combination.
A statement from the researchers read: “If automatic diagnosis of digestive tract endoscope images by AI is realised, then ‘automated diagnosis logic’ will be added to endoscope technology for real time diagnosis.
“Automatic diagnosis will be possible and it will greatly improve the current status of relying on the diagnostic ability of individual endoscopy physicians.”
