Automated quantification of atrial fibrillation complexity by probabilistic electrogram analysis and fibrillation wave reconstruction

Abstract

The analysis of high-density activation maps of atrial fibrillation (AF) provides fundamental insights into the fibrillation wave propagation patterns and thus the mechanisms of AF. Current annotation of local activations in unipolar atrial electrograms and the construction of fibrillation waves require labor-intensive manual editing. To enhance the possibilities for spatiotemporal analysis of AF, we developed a rapid and fully automated procedure to accurately identify local, intrinsic atrial deflections and construct fibrillation waves based on these deflections. In this study, the automated procedure was validated using manually annotated electrograms and wave maps. We show that the novel procedure accurately detects intrinsic deflections (sensitivity=87%, positive predictive value=89%) and that reconstructed wave maps correlate well with manually edited wave maps in terms of number of waves (r=0.96), intra-wave conduction velocity (r=0.97), AF cycle length (r=0.97), and wave size (r=0.96) (p<0.01 in all cases). The automated procedure is therefore an adequate substitute for manual annotation.