Simultaneous mapping of endocardium and epicardium from multielectrode intrachamber and intravenous catheters: a computer simulation-based validation

Abstract

A multielectrode basket-shaped contact catheter (MBC) provides simultaneous recordings of unipolar or bipolar electrograms from within the heart chambers. Another catheter-based mapping approach uses the multielectrode intravenous catheters (MIVCs), which are widely used to diagnose and treat supraventricular arrhythmias. It is also known that mapping techniques are usually limited to one surface at a time. Therefore, an approach that can be used for simultaneous mapping of left and right endocardial surfaces and epicardial surface will be beneficial to characterize and discriminate the endocardial and epicardial sources of the arrhythmias more accurately. In this study, we used statistical estimation method to map the endocardial and epicardial surfaces simultaneously based on combined usage of the MBC and MIVC. The statistical estimation method is based on high-resolution training data set to hypothesize the relationship between catheter measurements and inaccessible sites. To test this approach, we created a high-resolution map database consisting of computer simulation results of Aliev-Panfilov model of cardiac electrical activity on 3-dimensional Auckland canine heart geometry. The simulation database included 2590 maps each paced from a unique endocardial or epicardial site. Fifty or five percent of the database was used as the training data set and the remaining as test data set in the statistical estimation procedure. We selected 64 sites on the left and 64 on the right endocardial surfaces of the model heart geometry and used them as the surrogate MBC measurement sites. Ninety-one sites on the epicardium corresponding to the major coronary veins served as the surrogate MIVC leads. Finally, we tested the success of the method to determine the source of the arrhythmias using the correlation coefficient between the original and estimated activation maps and linear distance between their earliest activated sites. The performance of this approach was promising, such as when MBC on the left endocardium and MIVC were used together, the average linear distance was ∼2.4 mm and mean correlation coefficient was 0.995. It was possible to locate 95% of epicardial arrhythmia cases correctly on the epicardium. Ninety-nine percent of left endocardially originating arrhythmias were correctly located on the left endocardium. The results of this study showed that this approach is feasible and requires further effort.