Accepted Manuscript A QRS Axis-Based Algorithm to Identify the Origin of Scar-Related Ventricular Tachycardia in an American Heart Association 17-Segment Model David Andreu, MSc, PhD, Juan Fernández-Armenta, MD, PhD, Juan Acosta, MD, Diego Penela, MD, PhD, Beatriz Jáuregui, MD, David Soto-Iglesias, MSc, PhD, Vladimir Syrovnev, MD, Elena Arbelo, MD, PhD, José María Tolosana, MD, PhD, Antonio Berruezo, MD, PhD PII:
S1547-5271(18)30577-0
DOI:
10.1016/j.hrthm.2018.06.013
Reference:
HRTHM 7623
To appear in:
Heart Rhythm
Received Date: 22 March 2018
Please cite this article as: Andreu D, Fernández-Armenta J, Acosta J, Penela D, Jáuregui B, SotoIglesias D, Syrovnev V, Arbelo E, Tolosana JM, Berruezo A, A QRS Axis-Based Algorithm to Identify the Origin of Scar-Related Ventricular Tachycardia in an American Heart Association 17-Segment Model, Heart Rhythm (2018), doi: 10.1016/j.hrthm.2018.06.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A QRS Axis-Based Algorithm to Identify the Origin of Scar-
2
Related Ventricular Tachycardia in an American Heart
3
Association 17-Segment Model
4
David Andreu, MSc, PhD1*; Juan Fernández-Armenta, MD, PhD2*; Juan Acosta, MD3;
5
Diego Penela, MD, PhD1,4; Beatriz Jáuregui, MD1,4; David Soto-Iglesias, MSc, PhD1,4,
6
Vladimir Syrovnev, MD1,4; Elena Arbelo, MD, PhD1,4,5; José María Tolosana, MD,
7
PhD1,4,5; Antonio Berruezo, MD, PhD1,4,5
RI PT
1
8
Institut Clínic Cardiovascular, Hospital Clínic, Barcelona. Spain 2
10 3
11
13
5
Hospital Universitario Puerta del Mar, Cádiz. Spain
Hospital Universitario Virgen del Rocío, Sevilla. Spain
Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS)
M AN U
4
12
SC
1
9
Centro de Investigación Biomédica en Red Cardiovascular (CIBERCV), Instituto de
14
Salud Carlos III (ISCIII), Madrid, Spain
15 16
* Both authors contributed equally to this work.
18
TE D
17
Short Title: Axis-based algorithm for scar-related VTs.
19
Author for correspondence:
21
Antonio Berruezo, MD, PhD
22
Arrhythmia Section, Cardiovascular Institute, Hospital Clínic
23
University of Barcelona
24
C/ Villarroel, 170
25
08036 – Barcelona
26
Phone: (+34) 932275551
27
Fax: (+34) 934513045
28
Email:
[email protected]
AC C
EP
20
29 30
Word count: 5000
1
ACCEPTED MANUSCRIPT ABSTRACT
32
Background: Previously proposed algorithms to predict the ventricular tachycardia
33
(VT) exit site have been based on diverse left ventricle (LV) models, but none of them
34
identify the precise region of origin on electroanatomical mapping (EAM). Moreover,
35
no ECG algorithm has been tested to predict the region of origin of scar-related VTs in
36
non-ischemic patients.
37
Objective: To validate a simple electrocardiographic algorithm to identify the segment
38
of origin (SgO) VT relative to the 17-segment American Heart Association (AHA)
39
model in patients with structural heart disease (SHD).
40
Methods: The study included 108 consecutive patients with documented VT and SHD
41
(71% coronary artery disease). A novel frontal-plane axis-based ECG algorithm
42
(highest positive or negative QRS voltage) together with the polarity in V3-V4 was used
43
to predict the SgO of the VT. The actual of the VT was obtained from the analysis of
44
the EAM during the procedure. Conventional VT mapping techniques were used to
45
identify the VT exit.
46
Results: In total, 149 12-lead ECGs of successfully ablated VT were analyzed. ECG-
47
suggested SgO matched with the actual SgO in 122 (82%) of the 149 VTs. In 21
48
(77.8%) of the 27 mismatching ECG-suggested SgO, the actual SgO was adjacent to the
49
segment suggested by the ECG. There were no differences in accuracy of the algorithm
50
depending on the SgO or the type of SHD.
51
Conclusion: This novel QRS axis-based algorithm accurately identifies the SgO of VT
52
in the 17-segment AHA model in patients with SHD.
EP
TE D
M AN U
SC
RI PT
31
53 54 KEYWORDS
56
Electrocardiogram; ventricular tachycardia; electroanatomical mapping; myocardial
57
infarction; cardiomyopathy; catheter ablation
AC C
55
2
ACCEPTED MANUSCRIPT INTRODUCTION
59
The ECG characteristics of the QRS during ventricular tachycardia (VT) can be used to
60
identify the beginning of the normal myocardium activation, which is contiguous to the
61
exit site from the slow conduction path in the scarred tissue, in patients with structural
62
heart disease. This information can be rapidly obtained from the 12-lead surface ECG
63
and guide VT mapping and ablation. Several ECG-based algorithms have been
64
previously proposed. The first algorithm developed by Miller et al.1 was restricted to
65
anterior or inferior myocardial infarction. Kuchar et al.2 developed a new algorithm in
66
post-myocardial infarction patients, using pacemapping as a surrogate of VT exit. The
67
algorithm was tested in a second group of patients and accurately (80-90%)
68
differentiated between anterior/lateral/septal/inferior regions but the precise location
69
was determined only in 39%.1 The precise VT-circuit exit site identification by using
70
contact electroanatomical mapping, which provides better anatomical accuracy than
71
conventional fluoroscopic mapping, has not been related to the ECG features of the
72
scar-related VTs.2,3 In addition, no ECG algorithm has been tested to predict site of
73
origin of scar-related VTs in non-ischemic patients.
74
Previously proposed algorithms to predict the VT site of origin have been based on
75
diverse left ventricle (LV) models.1,2,4 The LV 17-segment American Heart Association
76
(AHA) model is the standard model used in cardiac imaging.5 Since cardiac imaging is
77
increasingly used in the electrophysiology laboratory to guide interventional
78
procedures, it appears appropriate to use this model as the reference in VT ablation.6,7
79
The aim of this prospective study was to test the accuracy of a simplified ECG
80
algorithm to locate the segment of origin (SgO) using the 17-segment model, the
81
information from the electroanatomical map, and the exit site of the VT circuit in
82
patients with structural heart disease (SHD).
SC
M AN U
TE D
EP
AC C
83
RI PT
58
3
ACCEPTED MANUSCRIPT METHODS
85
Patient Sample
86
Consecutive patients with a documented VT in a 12-lead surface ECG and SHD
87
referred for VT ablation were prospectively included in this study. Patients were
88
considered for VT ablation if they met one of the following criteria: incessant VT,
89
repetitive episodes of sustained monomorphic VT, appropriate ICD therapy or
90
symptomatic frequent premature ventricular complex (PVC) despite the use of
91
antiarrhythmic drugs. Patients with an idiopathic VT or arrhythmogenic right
92
ventricular dysplasia were excluded from the study. The study was approved by the
93
Institutional Committee on Human Research at the authors' institutions.
SC
RI PT
84
94
Segment of Origin Definition Using the Axis-Based Algorithm
96
The novel algorithm uses the QRS axis in the frontal plane to locate the VT origin in the
97
LV short axis (inferior vs anterior, septal vs lateral) and the polarity in V3-V4 for its
98
location in the LV longitudinal plane (basal, medial or apical). Figure 1 shows the axis-
99
based algorithm. A 17-segment model of the LV is superimposed on a representation of
100
the QRS axis and the limb leads (segment number 17 is not shown because it is
101
unreachable from the endocardium).
102
Two steps are needed to define the SgO of a VT using this algorithm, as follows:
103
1) Identify the limbs lead with the highest voltage magnitude (positive or negative). If
104
this magnitude is I, II, or III, the adjacent leads must be considered, as the major axis is
105
facing the boundary between two groups of segments. The adjacent limb lead with
106
higher magnitude will determine the group of segments where the VT may originate. In
107
cases where there was not a clear lead with R or QS pattern, the Q, R or S wave with the
108
higher amplitude was taken into the account. In Figure 2, the highest QRS magnitude
109
(negative) is in lead III. The QRS magnitude in aVL (positive) and aVF (negative)
110
should be compared. As the absolute amplitude of QRS in aVF is higher than in aVL,
111
the potential SgO are segments 4, 10 and 15.
112
2) Identify the positivity or negativity of the precordial leads V3 and V4; concordance
113
indicates a basal or apical origin, respectively. Other combinations indicate a medial
114
origin. Figure 2 shows an example of a VT in which the algorithm correctly identified
115
the SgO in basal segment 4.
116
Two electrophysiologists followed the two-step algorithm to analyze all recorded 12-
117
lead ECGs of induced VTs of patients included in the study. They were blinded to the
AC C
EP
TE D
M AN U
95
4
ACCEPTED MANUSCRIPT 118
SgO of the VT. In case of discordance, a third electrophysiologist analyzed the ECG.
119
An ECG with polymorphic beats or unclear axis was classified by the observers as
120
undefined. If any one of them classified the ECG as undefined, the SgO of the VT
121
according to the ECG information was also classified as undefined.
122 Electrophysiology Study
124
The electrophysiology study and ablation procedure were performed using intravenous
125
conscious sedation or general anesthesia, except in patients in whom the index
126
arrhythmia was a frequent PVC. In those cases, activation maps were performed
127
without intravenous sedation and a bolus of fentanyl was administered intravenously
128
before radiofrequency ablation. A navigation system (CARTO3 system, Biosense
129
Webster, Diamond Bar, California, or Ensite NavX system, St. Jude Medical,
130
Minneapolis, Minnesota, USA) was used to guide the VT ablation. The stimulation
131
protocol consisted of programmed ventricular stimulation from the right ventricle (RV)
132
apex at three drive-cycle lengths with up to three extrastimuli and incremental burst
133
pacing at a cycle length up to 200 ms. If the clinical VT was not inducible, intravenous
134
isoproterenol was used.
135
A 3.5-mm electrode irrigated-tip catheter (Thermocool Navistar®, Biosense Webster)
136
was introduced through transseptal or retrograde aortic access for LV endocardial
137
mapping. A complete activation or substrate LV map was obtained in all cases. Briefly,
138
a percutaneous epicardial access was performed for epicardial mapping and ablation
139
when endocardial VT ablation was unsuccessful, when the endocardial mapping did not
140
identify a VT substrate, when an exclusively epicardial scar was observed in pre-
141
procedural contrast enhanced-cardiac magnetic resonance, or when a post-myocardial
142
infarction patient exhibited a transmural scar in conventional preprocedural imaging.
SC
M AN U
TE D
EP
AC C
143
RI PT
123
144
Identification of Segment of Origin during the Procedure
145
The actual SgO of each VT was identified in the EAM according to the following
146
criteria: 1) Presence of presystolic local electrograms not earlier than 50 ms before the
147
beginning of the QRS and termination of the VT during RF ablation or slow conducting
148
channel exit site confirmed through entrainment maneuvers together with VT
149
termination during RF ablation; 2) Achieving a 12/12 QRS morphology concordance
150
during pacing with the 12-lead ECG of the VT from a site with no more than 50 ms
151
delay between the stimulus artifact and the beginning of the QRS. A single operator 5
ACCEPTED MANUSCRIPT 152
blinded to the ECG readings retrospectively analyzed the EAM, and magnetic resonance
153
or scan images when available, to classify the SgO of the VTs.
154 Scar Area Measurements
156
In a subgroup of patients who underwent complete LV substrate mapping, the scar area
157
was obtained using the area measurement tool from the CARTO3 software. The
158
standard voltage thresholds were used to differentiate between healthy tissue and scar.8
RI PT
155
159 Statistical Analysis
161
Analysis was performed using SPSS 17.0 software (SPSS Inc., Chicago, Illinois).
162
Continuous variables are presented as mean±SD, unless otherwise specified. Accuracy
163
of algorithm predictions depending on actual SgO location and cardiomyopathy and
164
actual scar size was assessed using Chi-square and U-Mann-Whitney test, respectively.
165
A P value
