Cardiostim 2012 Termination of Atrial Fibrillation by Catheter Ablation Can Be Successfully Predicted from Baseline ECG Buttu A. 1 , Van Zaen J. 1 , Viso A. 1 , Forclaz A. 2 , Pascale P. 2 , Narayan S. 3 , Vesin J. 1 , Pruvot E. 2 1 Applied Signal Processing Group, Swiss Federal Institute of Technology EPFL, Lausann 2 Department of Cardiology, University Hospital Center Vaudois CHUV, Lausanne – Switze 3 University of California, San Diego - USA
Transcript
Slide 1
Cardiostim 2012 Termination of Atrial Fibrillation by Catheter
Ablation Can Be Successfully Predicted from Baseline ECG Buttu A.
1, Van Zaen J. 1, Viso A. 1, Forclaz A. 2, Pascale P. 2, Narayan S.
3, Vesin J. 1, Pruvot E. 2 1 Applied Signal Processing Group, Swiss
Federal Institute of Technology EPFL, Lausanne Switzerland 2
Department of Cardiology, University Hospital Center Vaudois CHUV,
Lausanne Switzerland 3 University of California, San Diego -
USA
Slide 2
Introduction The success rate of stepwise radiofrequency
ablation (step- CA) for patients (pts) with long-standing
persistent atrial fibrillation (LS-pAF) appears limited. Multiple
parameters have been used to predict the outcome of step-CA (AF
cycle length AFCL, AF duration..). Limited success. Aim of our
study: To develop innovative indices from baseline ECG recordings
(i.e. before ablation) that can predict the termination of AF
during step-CA.
Slide 3
Methods 17 consecutive male patients included. Clinical
characteristics: Study population Age (y) 60 5 AF duration (y) 7 5
Sustained AF (month) 21 13 BMI (kg/m 2 ) 30 6 LVEF (%) 48 11 LA
volume (ml) 173 37
Slide 4
Methods Electrophysiological study: Effective anticoagulation
therapy for > 1 month. Antiarrhythmic drugs (except amiodarone
and beta- blockers) were discontinued 5 half-lives before the
procedure. General anesthesia. Catheter for mapping and ablation:
3.5 mm cooled-tip Navistar (Webster). Chest lead V 6 was placed in
the back (V 6b ), within the cardiac silhouette.
Slide 5
Methods Ablation protocol: Procedural end point: Termination of
AF into sinus rhythm (SR) or atrial tachycardia (AT). Non
terminated AF were cardioverted electrically.
Slide 6
Methods Signal processing: adaptive harmonic frequency tracking
Power spectrum density Dominant frequency? Time-frequency
representation Dominant frequency First harmonic How to extract the
frequency content? Adaptive harmonic frequency tracking
Slide 7
Methods
Slide 8
Methods Organization Measurements Quantifies the cyclicity of
the oscillations Two organization measurements: 1.Adaptive
organization index (AOI): ratio between the power of the extracted
components and the total power of the signal. Mean 0.7 =
Slide 9
Methods Organization Measurements 2.Variance of the phase
difference (PD): variance of the slope of the phase difference.
Variance = 6.5 10 -6 Quantifies the regularity of the
oscillations
Slide 10
Methods AOI and PD were compared to classical indices: AFCL
computed from the inverse of the dominant frequency (classical
method) of chest leads V 1 and V 6b after QRST cancellation 2.
Organization index (OI) 1 : ratio of the power in a 1-Hz band
centered on the dominant peak to the total power in the spectrum
(FFT). All the considered measures were computed from 10- sec ECG
recordings at baseline, after QRST cancellation 2. 1 Everett T. H.
et al. IEEE J BME 2001; 48; 969-78 2 Lemay M. et al. IEE Trans
Biomed Eng 2007; 54; 542-6 OI = 0.24
Slide 11
Results Study population Left terminated (LT = 11, 65%) Right
terminated (RT = 2, 12%) Not terminated (NT = 4, 23%) Age (y) 60 5
59 657 163 5 AF duration (y) 7 5 8 67 04 2 Sustained AF (month) 21
13 17 811 239 11 BMI (kg/m 2 ) 30 6 32 628 427 7 LVEF (%) 48 11 44
1058 454 13 LA volume (ml) 173 37 179 26178 87164 33 Cumulative
ablation time (min) 55 14 49 1163 2165 15
Slide 12
Results
Slide 13
Conclusions Adaptive algorithms based on the instantaneous
tracking of the dominant frequency (and its harmonics) improve the
assessment of organization during AF. Our findings are suggestive
of a higher baseline bi-atrial organization in LT patients.
Innovative adaptive indices appear as promising parameters to
predict patients that can be left terminated. Validation on a
larger population is needed.
Slide 14
Thank you
Slide 15
Ventricular activity cancellation Improves the reliability of
ECG analysis during AF. Single-beat method 2 : QRS complexes and
T-waves are treated separetly: QRS cancellation is an interpolation
of atrial activity with weighted sinusoids. T-wave cancellation
based on a dominant T-wave approach. Advantages: Independent of the
length the ECG recording. No discontinuities and QRS residues in
the resulting signals. 2 Lemay M. et al. IEE Trans Biomed Eng 2007;
54; 542-6
Slide 16
Ventricular activity cancellation Example (signal duration
10-sec, sampling frequency: 1kHz) : Chest lead V 1 Sample Frequency
(Hz)
Slide 17
Ventricular activity cancellation Example (signal duration
10-sec, sampling frequency: 1kHz): Dorsal lead V 6b Sample
Frequency (Hz)
Slide 18
Study presented at HRS 2012: Contribution of Left and Right
Atrial Appendage Activities to ECG Fibrillation Waves. Methods:
Clinical characteristics: prior to ablation, catheters (CAT) were
introduced in 10 consecutive patients (605 y, continuous AF
duration 2214 m): 1.Quadripolar CAT in the RAA. 2.Decapolar CAT in
the coronary sinus (CS). 3.Dudecapolar CAT in the LAA. 10-sec
epochs for a total duration of 270 sec were used
Slide 19
Signal processing: EGMs: automatic computation of AFCL from
LAA, RAA and CS recordings. ECG: after QRST cancellation,
computation of AFCL on all chest leads (V 1 to V 6b ). Correlation:
using Pearsons correlation coefficient, the correlation between
AFCL was computed for each combination of chest leads and
EGMs.
Slide 20
Results RAA AFCL was best correlated with chest lead V 1 (R =
0.96) and progressively dropped until V 5 (R = 0.26).
Interestingly, LAA AFCL showed the opposite pattern with the
highest correlation in V 6b (R = 0.95) and the lowest one in V 2 (R
= 0.26). V1 reflects the activity from the RAA and the dorsal lead
V6b reflects the LAA activity
Slide 21
Clinical results Sites of AF termination for LT and RT Left
terminated (LT = 11, 65%) Right terminated (RT = 2, 12%) Left
Atrium Roof 2 LAA 1 Coronary sinus 2 Mitral isthmus 6 Right Atrium
Cavotricuspid isthmus 1 RAA 1