Date of Award
Spring 2019
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Mathematics, Statistics and Computer Science
Program
Computational Sciences
First Advisor
Merrill, Stephen J.
Second Advisor
Sra, Jasbir
Third Advisor
Rowe, Daniel
Abstract
As an important reference for the physician during catheter ablation, the electrode-tissue contact force (CF), is one of the key points for the success of the catheter ablation. With the guide of CF sensing, the ablation procedure can be safer and more effective. Techniques and apparatus have been refined since catheter ablation was invented to treat cardiac arrhythmia. In the review part, different techniques for evaluating the electrode-tissue CF are discussed, including both direct and indirect measurement. Sensor-based direct measurement is broadly applied but restricted by the high cost. Surrogate markers of catheter-tissue contact such as impedance, electrogram (EGM) quality, catheter tip temperature and so on, are taken as reference evaluating CF as well, but each of them has their own drawbacks. In this dissertation, our approach estimating the CF is based on the moving pattern of the catheter tip in the heart chamber. The factors determining the catheter tip motion, include the cardiac and respiratory cycles, blood flow, and so on. If the position of the catheter tip can be recorded, then the motion of the catheter tip can be tracked and analyzed. Based on our collected data, the moving pattern of the catheter tip is different when the electrode-tissue CF level varies. Features extracted from catheter tip motion are significant for CF evaluation. There are different features selected to describe the moving pattern of the catheter tip, which are identified to best represent the movement by checking the corresponding CF as reference. In summary, if the feature has a strong correlation with the CF, then it can be taken as a good feature. Using the features as input, the CF evaluating mechanism is based on a multi-class classification decision tree to make an optimum and comprehensive estimation.