Agentic artificial intelligence in inflammatory bowel disease: toward autonomous and adaptive care

Area of Science:

• Gastroenterology and Computational Medicine

Background:

• Inflammatory bowel disease (IBD) management is complex, requiring continuous monitoring and treatment adjustments.
• Current IBD care relies on fragmented, intermittently collected data, leading to reactive decision-making.
• Existing artificial intelligence (AI) applications in IBD are often task-specific and fail to capture disease's longitudinal nature.

Purpose of the Study:

• To introduce agentic AI as a novel approach for continuous and adaptive care in IBD.
• To highlight the potential of agentic AI in integrating diverse data streams for an evolving disease state representation.
• To explore how agentic AI can facilitate early detection of changes and support proactive interventions in IBD.

Main Methods:

• Utilizing agentic AI to integrate heterogeneous data sources (clinical, biochemical, imaging, molecular).
• Developing a closed-loop system linking monitoring, interpretation, and action for IBD management.
• Focusing on capturing the longitudinal nature of IBD through continuous data analysis.

Main Results:

• Agentic AI enables a unified representation of IBD disease activity by integrating diverse data streams.
• The proposed system supports early detection of disease flares and facilitates proactive therapeutic adjustments.
• Demonstrates a shift from reactive to proactive clinical decision-making in IBD management.

Conclusions:

• Agentic AI holds significant potential for transforming IBD care through continuous, adaptive, and proactive management.
• Addressing challenges in data quality, interpretability, and clinical integration is crucial for successful implementation.
• This approach promises to enhance patient outcomes by providing a more holistic and dynamic view of IBD.