My research activities as a part of PersonalizeAF team mainly focus on detection and localization of regions in atria exhibiting some specific types of electrical activities or conduction patterns as we call them. We aim to develop computational tools that can robustly pinpoint these regions. This localization can help us have a better understanding of atrial fibrillation (AF) and also guide physicians to additional ablation targets in cases where common ablation strategies do not work. I would like to start my series of blogs which will concentrate of local AF driver detection techniques ablation strategies.
In 1998, Haïssaguerre and his team first reported that the pulmonary vein ectopy might be the trigger of AF. In 45 patients they evaluated, 94% of all ectopic sites were located in the vicinity of pulmonary veins (see Fig. 1). Then the electrical isolation of this region, a therapy called as Pulmonary Vein Isolation (PVI), could eliminate AF. The team has performed radio frequency ablation on 38 of these patients and by the end of 7 months follow-up, 62% had no AF episodes without any drug therapy. This was a promising step and the start of the “era of non-pharmacologic AF therapies” .
However, success rates of PVI in the longer period were very low. Besides, ablation in patients with persistent AF has been observed to be significantly less effective with a recurrence rate of %60. Such results have canalized clinicians into more aggressive ablation options in addition to PVI. To this day, there is still no reproducible statistical evidence about the efficiency of such alternative ablation approaches. We are failing to spot these sites and this is probably due to the lack of understanding of AF dynamics. So, can we make better choices?
Complex Fractionated Atrial Electrograms (CFAE) are one of the popular choices. CFAE is defined as an atrial electrogram (AEG) comprising two or more deflections with continuous baseline activity or with a cycle length (CL) smaller than 120 ms (which can be approximated by dominant frequency). CFAE is thought to be representing phenomena that facilitate AF, such as slow zones of slow conduction, pivot points or wave breaks . As clearly seen in Fig. 2, CFAEs are morphologically complex signals compared to their normal counterparts. Based on these morphological differences, detection of such complexity and distinguishing CFAE from normal AEG, poses a relatively simple signal processing problem (!). We possess some time series complexity-based techniques as well as spectral techniques for this purpose. Furthermore, artificial intelligence-based approaches are emerging for this task. In the next stages of this blog series, I will summarize these!
But, let’s cut it here for now. I will be back, soon!
Input your search keywords and press Enter.