Use of rapid reading software to reduce capsule endoscopy reading times while maintaining accuracy

Area of Science:

• Gastroenterology
• Medical Imaging
• Software Algorithms

Background:

• Capsule endoscopy (CE) generates vast amounts of image data, making abnormality detection time-consuming.
• Identifying rare lesions within thousands of images presents a significant challenge in routine CE analysis.

Purpose of the Study:

• To evaluate the efficacy of Omni Mode, a novel software algorithm for capsule endoscopy.
• To determine if Omni Mode can reduce image reading times without compromising lesion detection accuracy.

Main Methods:

• A prospective multicenter study involving 9 European centers.
• Consecutive capsule endoscopy cases were analyzed using conventional mode and Omni Mode.
• Lesion detection rates, reading times, and clinical significance (P classification) were compared between the two modes.

Main Results:

• Omni Mode significantly reduced reading times by an average of 24.6 minutes (40% time saving, P < .0001).
• The accuracy of Omni Mode (0.75) was comparable to conventional mode (0.70), with no significant difference in clinical conclusions.
• A total of 2127 lesions were identified and classified, with Omni Mode showing a slight improvement in missed lesions.

Conclusions:

• Omni Mode offers a substantial reduction in capsule endoscopy reading time.
• The software algorithm effectively maintains diagnostic accuracy for lesion detection in CE.
• Omni Mode represents a significant advancement in streamlining CE data analysis.