Utilization of spatial interpolation and statistical estimation on catheter based cardiac mapping

Because the catheter based cardiac mapping does not require open-chest surgery, recently it has become more and more popular and brought novelty towards arrhythmia diagnosis and therapy (ablation). However, due to long procedure times and complexity in the heart anatomy, it is not possible to acquire high spatial resolution measurements. In this study by using Aliev-Panfilov?s mathematical model we created a computer simulation database and extracted activation time data from endocardial and epicardial surfaces. From limited number of selected points on cardiac surfaces we reconstructed high resolution activation data. In the first part of our study we used Newton?s linear, Hardy?s, Laplacian and spline spatial interpolation methods. Consequently, several error criteria between the original data and reconstructed data were computed, such as correlation coefficient (CC), root mean squares error (RMSE), relative error (RE) and localization error (LE). The average results obtained by using basket catheter were; CC: 0,992 ± 0,003, RMSE: 1,793 ± 0,330, RE: 0,044 ± 0,008 and LE: 3,998 ± 1,853. Secondly, instead of interpolation approaches we investigated the usage of statistical estimation method in order to reconstruct the activation time distribution on both endocardial and epicardial surfaces simultaneously. The worst results we obtained were CC: 0,995 ± 0,004, RMSE: 2,568 ± 0,727, RE: 0,033 ± 0,010, and LE: 2,386 ± 2,101. In addition, we tested the success of the method to determine the origin of the arrhythmia. As a result 100% of left endocardially and 98.8% of right endocardially originating arrhythmias were correctly located on their region.KEY WORDS: Cardiac mapping, spatial interpolation, statistical estimation, simulation of cardiac electrical activity, Aliev-Panfilov model.