Abstract 644: Dominant Frequency Analysis of Atrial Fibrillation: Comparison of Contact and Noncontact Mapping
Introduction: The ability to acquire a dominant frequency (DF) map during atrial fibrillation (AF) instantaneously using noncontact mapping has significant advantages over the current sequential contact mapping approach; however, the relationship between DFs determined from contact bipolar compared to noncontact unipolar recordings is unknown. We sought to determine the difference between DFs determined using contact bipolar (Bi), contact unipolar (Uni), noncontact Uni, and noncontact virtual Bi recordings.
Methods: Sustained AF was induced in five canines with 12 weeks of atrial tachy-pacing at 400 ppm. A noncontact multielectrode array was positioned in the left atrium (LA). Two simultaneous contact signals (Uni and Bi) and three noncontact signals (Uni, Virtual Bi, and Bi Laplacian) of 7 seconds duration were recorded from multiple LA sites and classified into one of five regions (posterior, septal, pulmonary veins, roof, or appendage). All Uni signals underwent QRS subtraction. Fourier analysis was performed and the DFs of contact Bi signals were compared to each of the other recording modalities. To exclude signals with broad spectra, signals with a regularity index (RI) <0.2 were excluded.
Results: Recordings were obtained from 389 LA locations in 5 canines. Overall, there was no significant difference in the mean DF between contact Bi and contact Uni QRS signals (351 points, 11.8±1.9Hz vs. 11.8±1.8Hz; p=NS). Similarly, contact Bi and noncontact Uni DFs had no significant differences (342 points, 11.82±1.83Hz vs. 11.84±1.71Hz; p=NS). Comparing contact Bi and noncontact virtual Bi, there was a significant difference in the measured DF (211 points, 11.6±1.8Hz vs. 11.4±1.7Hz; p=0.0012), due mainly to DF differences in PV regions (11.8±2.4Hz vs. 11.2±2.1Hz; p<0.01). When comparing contact Bi and the noncontact Bi Laplacian, there was also a significant difference in DF (223 points, 11.39±1.75Hz vs. 11.12±1.62Hz; p<0.01).
Conclusions: We found that
estimation of DF using noncontact mapping is feasible
QRS subtracted noncontact unipolar signals perform better than virtual bipolar signals at estimating contact bipolar DFs. This has important implications for developing algorithms for noncontact frequency mapping of AF.