Visualization of active devices and automatic slice repositioning ("SnapTo") for MRI-guided interventions

Area of Science:

• Medical Imaging
• Interventional Radiology
• Magnetic Resonance Imaging (MRI)

Background:

• Accurate visualization of interventional devices is critical for MRI-guided procedures.
• Current visualization methods may lack the speed and robustness required for real-time interventions.

Purpose of the Study:

• To introduce an improved visualization method for active interventional devices in MRI.
• To develop a fast and robust algorithm for reconstructing device trajectories.

Main Methods:

• Developed "CurveFind," an iterative prediction-correction algorithm using orthogonal projection images.
• Implemented features like "SnapTo" for automatic slice repositioning and 3D curve plotting.
• Integrated the system into a real-time MRI environment for in vitro and in vivo testing.

Main Results:

• CurveFind reconstructs 3D device trajectories from projection images in under a second.
• The algorithm demonstrates robustness to signal inhomogeneities by varying step size and search direction.
• Successful in vitro and in vivo (pig model) experiments were conducted with various devices.
• Achieved an average device curve reconstruction error of approximately 2 mm, even for complex geometries.

Conclusions:

• The CurveFind method significantly enhances the visualization of interventional devices during MRI.
• This real-time system improves the safety and effectiveness of MRI-guided interventions.
• The algorithm's speed, robustness, and accuracy make it suitable for clinical application.