New Generation Express View: An Artificial Intelligence Software Effectively Reduces Capsule Endoscopy Reading Times.

Area of Science:

• Gastroenterology
• Medical Imaging

Background:

• Capsule endoscopy (CE) reading is a lengthy process.
• Express View (EV) software aims to expedite CE video analysis.
• Evaluating EV's diagnostic performance for small-bowel (SB) lesions is crucial.

Purpose of the Study:

• To assess the diagnostic accuracy of EV software in identifying significant SB lesions.
• To compare the time efficiency of EV mode versus standard reading (SR) for CE videos.

Main Methods:

• Prospective enrollment of 126 patients with suspected SB bleeding or neoplasia undergoing SB CE.
• CE evaluation conducted using both standard reading (SR) and Express View (EV) modes.
• Consensus review for discrepancies between SR and EV readings to re-classify findings.

Main Results:

• EV demonstrated high diagnostic accuracy, with post-consensus metrics reaching 97% sensitivity, 100% specificity, and 98% overall accuracy.
• Median reading time was significantly reduced from 71 minutes with SR to 13 minutes with EV (p < 0.001).
• Small-bowel capsule endoscopy completion rate was 86.5% with no device retention.

Conclusions:

• New-generation Express View (EV) software exhibits high diagnostic accuracy for detecting small-bowel lesions.
• EV significantly decreases the time required for capsule endoscopy video interpretation.
• EV offers a more efficient approach to capsule endoscopy analysis without compromising diagnostic quality.