Real-time respiratory motion compensated roadmaps for hepatic arterial interventions

Area of Science:

• Medical Imaging
• Interventional Radiology
• Biomedical Engineering

Background:

• Hepatic arterial interventions rely on static roadmap overlays for catheter navigation.
• Respiratory motion causes discrepancies between the roadmap and actual anatomy, complicating procedures.

Purpose of the Study:

• To develop and evaluate a dynamic respiratory motion compensated device guidance system.
• To improve accuracy and real-time performance in hepatic arterial interventions.

Main Methods:

• A respiratory motion model of the hepatic vasculature was created from X-ray sequences.
• Deep learning segmented guidewires and catheters from live fluoroscopic images.
• Real-time compensated displays (vessel or device) were evaluated in a porcine model.

Main Results:

• The motion compensated system achieved a median error of 1.08 mm.
• Guidewire tip accuracy in the correct vessel branch was significantly higher (86% vs. 69%) with compensated displays.
• Average processing time was 17 ms per frame.

Conclusions:

• The developed system enhances the accuracy of device positioning relative to hepatic vasculature.
• The integrated display modes simplify physician workflow by combining vessel and device information.
• Processing times are clinically feasible for real-time application.