Document resource
Introduction TOP100 is a feature within capsule endoscopy (CE) software that selects the 100 most significant images from a full CE video, potentially reducing physician reporting time. Previous studies 1 2 have shown good diagnostic performance and agreement with standard human reading (SR), especially for active bleeding and inflammatory lesions; this study evaluates the inter- and intra-observer agreement in findings and patient management between SR and TOP100.Methods A retrospective comparison of SR of an entire video with TOP100 image review was conducted by two experienced physician experts (ER) and one senior trainee (ST), each with >500 lifetime CE reads. The experts who reported the original CE, blindly reviewed the same CE TOP100 images in a short video format. The ST also independently read the TOP100 images, which was compared to the ER results. Agreement was assessed for significant findings, primary diagnosis and subsequent suggested management. Secondary outcomes included reading time and interobserver variability between SR and TOP100, and trainee versus expert TOP100 reads.Results A total of 50 CE videos were reviewed by each expert with TOP100. Findings were recorded and patient management was suggested based on TOP100 findings and CE indication. Each ER reviewed 25 videos of suspected SB bleeding and 25 videos of known or suspected inflammatory bowel disease. Both ERs had a substantial intraobserver agreement (κ 0.61-0.80) between SR and TOP100 for ulcers and active bleeding. One ER had only a fair (κ 0.37, p<0.01) intraobserver agreement for angiodysplasias while the other ER had a moderate agreement (κ 0.68, p<0.01) for the same category. Agreement on patient management based on TOP100 readings was similar for the two ERs (75%, 78%). The interobserver agreement between the ST and ER when using TOP100 was substantial for ulcers (κ 0.90, p<0.01), angiodysplasias (κ 0.61, p<0.01) and overall diagnostic yield (κ 0.69, p<0.01). Agreement was moderate for active bleeding (κ 0.54,p<0.01) and erosions (κ 0.58, p<0.01). The mean reading time per CE video with TOP100 was 1m48s vs 40m for SR.Conclusion The TOP 100 demonstrated moderate intraobserver agreement with standard reading for final diagnosis and management among ERs. A moderate to substantial agreement between ST and expert TOP100 reads suggests that the TOP100 could be an adjunct to trainees in reporting CE. Further studies with a larger pool of differing experience of trainees would help substantiate our findings.References Freitas M, Arieira C, Carvalho PB, et al. Simplify to improve in capsule endoscopy - TOP 100 is a swift and reliable evaluation tool for the small bowel inflammatory activity in Crohn’s disease. Scand J Gastroenterol. 2020;55:408–413.Giordano A, Escapa M, Urpi-Ferreruela M, et al. Diagnostic accuracy of artificial intelligence-aided capsule endoscopy (TOP100) in overt small bowel bleeding. Surg Endosc. 2023;37:7658–7666.