Toward multiple catheters detection in fluoroscopic image guided interventions

Area of Science:

• Medical Imaging
• Computational Biology
• Cardiovascular Interventions

Background:

• Catheter navigation in cardiac electrophysiology (EP) procedures is crucial for interventions like ablation.
• Catheter deformation during procedures due to patient and physician factors complicates real-time tracking.
• Low fluoroscopic image quality and instrument clutter present significant challenges for automated catheter tracking.

Purpose of the Study:

• To develop a robust and efficient method for automatic detection and tracking of cardiac catheter sheaths and tips.
• To improve the speed and accuracy of catheter tracking in fluoroscopic images for EP applications.
• To address the challenges posed by low image quality and complex clinical environments in cardiac interventions.

Main Methods:

• A computationally efficient framework was developed to trace the catheter sheath and detect multiple catheter tips simultaneously.
• The approach utilizes knowledge of the clinical setup to constrain the search space, enhancing tracking speed and accuracy.
• A modified fast marching weighted distance computation was employed to calculate geodesic properties, followed by a cascade classifier for tip detection.

Main Results:

• The proposed technique achieved multiple catheter tracking at a rate of 10 images per second.
• Validation on 1107 fluoroscopic images across four clinics demonstrated a very low false positive rate of 1.06%.
• The method proved effective in challenging clinical environments with low-quality images and multiple instruments.

Conclusions:

• The developed method offers a robust and efficient solution for automatic catheter sheath and tip detection and tracking in fluoroscopic images.
• This advancement has the potential to significantly benefit electrophysiology (EP) clinical applications by enabling fast and accurate catheter guidance.
• The technique's high speed and accuracy, validated on extensive clinical data, suggest its clinical viability for real-time interventional guidance.