Multiparametric optimization of the localization of the source of ectopic heart cycles using a simple dipole: A case study

Investigators: Svehlikova, J. – Zelinka, J. – Tysler M. – Tino, P.

In an informal collaboration with P. Tín from the University of Birmingham, we investigated the possibility of using two parameters simultaneously for non-invasive inverse localization of the stimulation site in a patient with stimulated heart activity using surface ECG maps and chest geometry and structure obtained from CT as input data. In the traditional solution, the only criterion function is the relative error (RRE) between the input potential map and the map generated by the inversely calculated dipole, the position of which represents the position of the source of cardiac activity. In this study, we used all the measured signals of cardiac cycles and their correlation to determine the second criterion parameter – the so-called. reliability score (RS), which evaluated the stability of the orientation of inversely determined dipole moments. Both parameters were calculated at each time point of the examined signal for all predefined dipole positions and the Pareto front was calculated from the given parameters. The resulting optimization was made from the graph of Pareto fronts of all examined time points by determining the so-called knees of the resulting Pareto front. Using this optimization method, the localization error of cardiac activity was improved from 16 mm (for RRE) to 11 mm (for RRE + RS) for the homogeneous chest model and from 13 to 9 mm for the non-homogeneous chest model.

 

 

 

 

 

 

 

Fig.: a) All measured ECG signals used in the calculations, b) Pareto queues (red dots) of the results for one time point, c) the resulting Pareto queues from all monitored time points.

Related projects:

• VEGA 2/0125/19 “Measurement and modeling of the electric field of the heart for non-invasive identification and interpretation of structural changes in the ventricular myocardium leading to ventricular arrhythmias.”
• APVV-14-0875 “Non-invasive localization of ectopic ventricular arrhythmias using ECG mapping and its use for causal treatment.”

Publication:

• ŠVEHLÍKOVÁ, Jana** – ZELINKA, Ján – TYŠLER, Milan – TIŇO, P. Multiobjective optimization approach to localization of ectopic beats by single dipole: Case study. In Computing in Cardiology, 2019, vol. 46, 4 p. (2018: 0.202 – SJR). ISSN 2325-8861. Typ: ADMB