Real-time 3D US-CT fusion-based semi-automatic puncture robot system: clinical evaluation

Area of Science:

• Medical robotics
• Image-guided therapy
• Surgical navigation

Background:

• Percutaneous radiofrequency ablation (PRFA) faces challenges in safe and accurate puncture due to image limitations and organ displacement.
• Existing systems struggle with precision, impacting treatment efficacy and patient safety.

Purpose of the Study:

• To evaluate a novel semi-automatic robotic puncture system integrating real-time ultrasound (US) and computed tomography (CT) images.
• To assess the system's clinical usefulness and accuracy in guiding PRFA procedures through a pilot study.

Main Methods:

• Development of an improved U-net model for image analysis, validated with a 0.87 Dice coefficient.
• Integration of the AI model into a robotic system for semi-automatic puncture guidance.
• Conducting a pilot clinical experiment with five participants using the integrated US-CT system.

Main Results:

• The 3D US-CT fusion process evaluated centroid distances with an average of 1.97 mm (range: 0.38–4.81 mm).
• All measured centroid distances were within the 5.00 mm ablation margin required for PRFA.

Conclusions:

• The robotic system shows potential for accurate puncture navigation in PRFA, meeting critical safety margins.
• The AI model demonstrated potential generalization performance, adaptable to variations in US images and patient anatomy.