A real-time cardiac surface tracking system using Subspace Clustering.

Area of Science:

• Medical Imaging
• Cardiac Electrophysiology
• Computational Geometry

Background:

• Catheter-based radiofrequency ablation for atrial fibrillation demands high-accuracy, real-time 3D cardiac surface tracking.
• Existing systems lack the sub-millimeter precision required for this complex procedure.

Purpose of the Study:

• To propose and validate a novel system for real-time, high-accuracy 3D tracking of cardiac surfaces.
• To address the limitations of current technologies in cardiac surface visualization during ablation procedures.

Main Methods:

• Utilized Subspace Clustering algorithm on pre-operative MRI data for cardiac surface deformation analysis during a training phase.
• Employed low-density outer cardiac surface samples in the tracking phase to identify active deformation subspaces.
• Reconstructed complete inner and outer cardiac surfaces in real-time using a least squares formulation.

Main Results:

• The proposed system demonstrates the capability for real-time 3D cardiac surface tracking with high accuracy.
• Successful application of the system to a real patient dataset validates its clinical potential.
• Achieved sub-millimeter accuracy in tracking cardiac surfaces, meeting critical procedural requirements.

Conclusions:

• The developed system offers a significant advancement for catheter-based atrial fibrillation ablation by providing accurate real-time 3D cardiac surface tracking.
• This technology has the potential to improve the safety and efficacy of cardiac ablation procedures.
• The proposed method overcomes current technological limitations, paving the way for enhanced cardiac interventions.