Comparing AI and Endoscopy for Diagnosing Stomach Cancer
Author Information
Author(s): Soong Tse Kiat, Kim Guo Wei, Chia Daryl Kai Ann, So Jimmy Bok Yan, Lee Jonathan Wei Jie, Shabbbir Asim, Lum Jeffrey Huey Yew, Soon Gwyneth Shook Ting, Ho Khek Yu
Primary Institution: National University Hospital, Singapore
Hypothesis
Can Raman spectroscopy-based AI provide comparable diagnostic performance to high-definition white light endoscopy for gastric lesions?
Conclusion
The AI-enabled Raman spectroscopy performed similarly to expert endoscopists using high-definition white light endoscopy in diagnosing gastric lesions.
Supporting Evidence
- The AI system achieved 100% sensitivity and 80% specificity in diagnosing gastric lesions.
- Both diagnostic methods showed similar accuracy rates of around 90%.
- The study was conducted in a real-world clinical setting, enhancing its relevance.
Takeaway
This study shows that a new AI tool can help doctors quickly and accurately identify stomach problems, just like experienced doctors do.
Methodology
A randomized double-arm feasibility proof-of-concept trial comparing AI-enabled Raman spectroscopy with high-definition white light endoscopy in patients with suspected gastric neoplasia.
Potential Biases
Potential bias due to the non-blinding of participants and endoscopists to the assigned intervention.
Limitations
The study had a small sample size and was conducted during the pandemic, which limited participant recruitment.
Participant Demographics
Mean age of 71.7 years; 57.1% male; 90.5% Chinese.
Statistical Information
P-Value
1.00
Confidence Interval
95% CI: 40–100%
Statistical Significance
p = 1.00
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website