Epiphenomenal Re-Entry and Spurious Focal Activation Detection by Atrial Fibrillation Mapping Algorithms

The purpose of this study was to validate the ability of mapping algorithms to detect rotational activations (RoA) and focal activations (FoA) during fibrillatory conduction (FC) and atrial fibrillation (AF) and understand their mechanistic relevance.

Mapping algorithms have been proposed to detect RoA and FoA to guide AF ablation.

Rapid left atrial pacing created FC-fibrillatory electrograms-with and without AF induction in dogs (n = 17). Activation maps were constructed using Topera (Abbott, St. Paul, Minnesota) or CARTOFINDER (Biosense Webster, Irvine, California) algorithms. Mapping strategies included: panoramic noncontact mapping with a basket catheter (CARTOFINDER n = 6, Topera n = 5); and sequential contact mapping using 8-spline OctaRay catheter (Biosense Webster) (n = 6). Offline frequency and spectral analysis were also performed. Algorithm-detected RoA was manually verified.

The right atrium (RA) consistently exhibited fibrillatory signals during FC. FC with and without AF had similar left-to-right frequency gradients. Basket maps were either uninterpretable (847 of 990 Topera, 132 of 148 Cartofinder) or had unverifiable RoA. OctaRay contact mapping showed 4% RoA (n = 30 of 679) and 63% FoA (n = 429 of 679). Verified RoA clustered at consistent sites, was more common in the RA than left atrium (odds ratio: 3.5), and colocalized with sites of frequency breakdown in the crista terminalis and RA appendage. During pacing, spurious FoA sites were identified around the atria, but not at the actual pacing sites. RoA and FoA site distribution was similar during pacing with and without induction, and during induced AF.

Mapping algorithms were unable to detect pacing sites as true drivers of FC, and detected epiphenomenal RoA and FoA sites unrelated to AF induction or maintenance. Algorithm-detected RoA and FoA did not identify true AF drivers.