To the Editor:
In recent articles published by Gastrointestinal Endoscopy,
1
, 2
, 3
artificial intelligence (AI) has been used to predict the invasion depth of gastric cancer (GC) on the basis of nonmagnifying endoscopic images with high accuracy. The accuracy of white-light imaging (WLI) with AI is from 78.6% to 94.5% when distinguishing GC within the mucosa, the superficial and deeper submucosa, and beyond the submucosa.The diagnosis and invasion depth prediction of GI cancer is always a prerequisite for the treatment choices (eg, endoscopic or surgical resection). In current clinical practice, the invasion depth of GC mainly relies on the macroscopic features from WLI, such as Paris classification, size, color, and ulceration. The diagnostic ability is subjective and changeable. These results may indicate the feasibility of the clinical application of AI based on WLI in determining GC invasion depth to improve accuracy and reduce the workload after the optimization of AI models and algorithms in the future.
For esophageal squamous cell cancer and colorectal cancer, intrapapillary capillary loops (IPCL), the National Institute for Health and Care Excellence guidelines, and the Japan Narrow-Band Imaging Expert Team classification can give us common diagnostic criteria for invasion depth. Specific criteria of GC invasion depth are lacking, which may be a limitation of endoscopic evaluation. Therefore, besides the clinical application to endoscopic diagnosis that is considered the dominant goal of AI, does the high accuracy indicate a criterion of GC invasion depth hidden in WLI? Furthermore, can AI be used to find a specific structure and the corresponding change rule responsible for GC invasion depth, similar to IPCL and pit pattern? If so, it means that image-based AI may promote the development of endoscopic diagnostic criteria of GC invasion depth and also may be applicable to esophageal adenocarcinoma and the mechanism of gastric carcinogenesis behind the structure.
Disclosure
All authors disclosed no financial relationships.
References
- Application of convolutional neural network in the diagnosis of the invasion depth of gastric cancer based on conventional endoscopy.Gastrointest Endosc. 2019; 89: 806-815.e1
- Highly accurate artificial intelligence systems to predict the invasion depth of gastric cancer: efficacy of conventional white-light imaging, nonmagnifying narrow-band imaging, and indigo-carmine dye contrast imaging.Gastrointest Endosc. 2020; 92: 866-873.e1
- Deep learning system compared with expert endoscopists in predicting early gastric cancer and its invasion depth and differentiation status (with videos).Gastrointest Endosc. 2022; 95: 92-104.e3
Article info
Footnotes
Drs Yang and Guan contributed equally to this article.
Identification
Copyright
© 2022 by the American Society for Gastrointestinal Endoscopy
