Artificial Intelligence in Gastrointestinal Endoscopic Ultrasound
Author Information
Author(s): Wu Yizhong, Ramai Daryl, Smith Eric R., Mega Paulo F., Qatomah Abdulrahman, Spadaccini Marco, Maida Marcello, Papaefthymiou Apostolis
Hypothesis
Can artificial intelligence enhance the diagnostic capabilities of endoscopic ultrasound (EUS) in detecting gastrointestinal lesions?
Conclusion
AI has shown promising results in improving the sensitivity and accuracy of EUS for diagnosing gastrointestinal lesions.
Supporting Evidence
- AI models have shown high accuracy in predicting malignancy in gastric gastrointestinal stromal tumors.
- AI can enhance the practice of ultrasound imaging by analyzing and highlighting fleeting, real-time image features.
- AI-assisted novices had significantly lower adenoma miss rates compared to those without AI assistance.
Takeaway
This study looks at how computers can help doctors find problems in the stomach and intestines using special ultrasound pictures. It shows that these computer tools can be really good at spotting issues.
Methodology
The review evaluates various studies on AI applications in EUS, focusing on diagnostic accuracy and training improvements.
Potential Biases
AI models may carry biases based on the data used for training, which may not reflect diverse patient populations.
Limitations
The generalizability of AI models is limited by the specific populations from which training data is obtained.
Statistical Information
P-Value
0.007
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website