Comparison of Two Computer-Aided Detection Systems in Colonoscopy
Author Information
Author(s): Goh Eun Chung, Jooyoung Lee, Seon Hee Lim, Hae Yeon Kang, Jung Kim, Ji Hyun Song, Sun Young Yang, Ji Min Choi, Ji Yeon Seo, Jung Ho Bae
Primary Institution: Seoul National University Hospital
Hypothesis
Frequent false alarms from CADe systems might lead to distraction or fatigue during routine clinical practice, potentially causing significant alarms to be overlooked.
Conclusion
The study found that a computer-aided detection system with lower false positive rates improved adenoma detection rates without increasing unnecessary polypectomies.
Supporting Evidence
- The adenoma detection rate (ADR) was significantly higher in the B system compared to the control group.
- The mean adenomas per colonoscopy (APC) was also higher in the B system.
- System B had the lowest non-true lesion resection rates among the groups.
- High-performing endoscopists showed no improvement in ADR across both systems.
Takeaway
This study looked at two computer programs that help doctors find polyps during colonoscopy. One program had fewer mistakes and helped doctors find more polyps without taking out unnecessary ones.
Methodology
This non-randomized controlled prospective study compared two CADe systems with different false positive rates in a clinical setting.
Potential Biases
Potential researcher bias due to the non-randomized nature of the study and the lack of blinding.
Limitations
The study was non-randomized and lacked blinding among participating endoscopists, which could introduce bias.
Participant Demographics
The study included 3,047 subjects aged 45 years or older undergoing colonoscopy for CRC screening or surveillance.
Statistical Information
P-Value
0.012
Confidence Interval
95% CI, 1.02–1.22
Statistical Significance
p=0.012
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website