AI-Driven View Guidance System in Intra-Cardiac Echocardiography Imaging.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Cardiology

Background:

• Intra-cardiac echocardiography (ICE) is vital for electrophysiology (EP) and structural heart disease (SHD) interventions.
• Effective ICE catheter manipulation demands significant expertise, leading to variable outcomes for less experienced operators.

Purpose of the Study:

• To develop an AI-driven guidance system for ICE catheter navigation.
• To assist users in achieving desired ICE views without specialized knowledge through closed-loop, human-in-the-loop feedback.

Main Methods:

• A novel AI method models relative position and orientation vectors between arbitrary and defined ICE views.
• The system provides real-time guidance for ICE catheter manipulation in a closed-loop configuration.
• Performance evaluated using simulation with real clinical data and semi-simulation with human-in-the-loop testing.

Main Results:

• The AI guidance system achieved an 89% success rate across 6,532 simulated test cases.
• Human-in-the-loop testing confirmed the feasibility of continuous, discrete guidance.
• The system demonstrated effective integration into clinical workflows.

Conclusions:

• The proposed AI system significantly enhances the accuracy and efficiency of ICE imaging procedures.
• This technology has the potential to democratize expertise in ICE-guided interventions.
• Continuous closed-loop guidance offers a promising approach for complex cardiac interventions.