Study of capsule endoscopy delivery at scale through enhanced artificial intelligence-enabled analysis (the CESCAIL...

Area of Science:

• Gastroenterology
• Medical Imaging
• Artificial Intelligence

Background:

• Lower gastrointestinal (GI) diagnostics face significant capacity constraints, exacerbated by the COVID-19 pandemic, leading to diagnostic backlogs.
• Advancements in deep neural networks (DNNs) and AI present an opportunity to automate and enhance the efficiency of capsule video analysis.
• AI applications in small bowel capsule endoscopy demonstrate potential for similar improvements in colon capsule analysis.

Purpose of the Study:

• To determine the feasibility, accuracy, and productivity of AI-enabled analysis tools (AiSPEED) for polyp detection in colon capsule endoscopy (CCE).
• To compare the performance of AI-enabled analysis against the conventional care pathway involving clinician analysis (the gold standard).

Main Methods:

• A multi-centre diagnostic accuracy study recruiting 674 participants (retrospectively and prospectively).
• Colon capsule videos are analyzed via two pathways: AI-enabled analysis and conventional clinician analysis.
• Comparison of reports for accuracy (sensitivity, specificity) and reading time (in the prospective cohort).

Main Results:

• The study is currently recruiting participants and collecting data across multiple UK centres.
• Validation of the AI tool (AiSPEED) for polyp detection is ongoing.
• Observational data on real-world performance and pathway execution will be gathered.

Conclusions:

• The standard diagnostic accuracy study poses no additional risk to patients, as standard care pathways remain unaffected.
• AI-enabled tools have the potential to significantly improve the efficiency and accuracy of lower GI diagnostic services.
• The CESCAIL study will provide crucial data on the real-world utility of AI in colon capsule endoscopy.