Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Automatic Detection of Slow Conducting Channels during Substrate Ablation of Scar-Related Ventricular Arrhythmias

Automatic Detection of Slow Conducting Channels during Substrate Ablation of Scar-Related... Hindawi Journal of Interventional Cardiology Volume 2020, Article ID 4386841, 13 pages https://doi.org/10.1155/2020/4386841 Research Article Automatic Detection of Slow Conducting Channels during Substrate Ablation of Scar-Related Ventricular Arrhythmias 1,2,3 4 4 5 Alejandro Alcaine , Beatriz Ja´uregui, David Soto-Iglesias, Juan Acosta, 6 7 8 9 Diego Penela, Juan Ferna´ndez-Armenta, Markus Linhart, David Andreu, 10,11,12 1,2 3 1,2 Lluı´s Mont, Pablo Laguna , Oscar Camara, Juan Pablo Martı´nez , and Antonio Berruezo CIBER en Bioingenier´ıa, Biomateriales y Nanomedicina (CIBER-BBN), Zaragoza, Spain BSICoS Group, Arago´n Institute of Engineering Research (I3A), IIS Arago´n, Universidad de Zaragoza, Zaragoza, Spain BCN MedTech Unit, PhySense Group, Department of Information and Communication Technologies, Universitat Pompeu Fabra, Barcelona, Spain Teknon Medical Center, Barcelona, Spain Hospital Universitario Virgen del Roc´ıo, Sevilla, Spain Ospedale Guglielmo da Saliceto, Piacenza, Italy Hospital Puerta del Mar, Ca´diz, Spain Arrhythmia Section, Cardiology, Hospital Universitari Doctor Josep Trueta, Girona, Spain Boston Scientific, Madrid, Spain Hospital Cl´ınic, Universitat de Barcelona, Barcelona, Spain Institut d’Investigacions Biom`ediques August Pi i Sunyer (IDIBAPS), Barcelona, Spain CIBER en Enfermedades Cardiovasculares (CIBER-CV), Barcelona, Spain Correspondence should be addressed to Antonio Berruezo; antonio.berruezo@quironsalud.es Received 12 November 2019; Revised 7 March 2020; Accepted 19 March 2020; Published 29 May 2020 Academic Editor: Jochen Wo¨hrle Copyright © 2020 Alejandro Alcaine et al. +is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Voltage mapping allows identifying the arrhythmogenic substrate during scar-related ventricular arrhythmia (VA) ablation procedures. Slow conducting channels (SCCs), defined by the presence of electrogram (EGM) signals with delayed components (EGM-DC), are responsible for sustaining VAs and constitute potential ablation targets. However, voltage mapping, as it is currently performed, is time-consuming, requiring a manual analysis of all EGMs to detect SCCs, and its accuracy is limited by electric far-field. We sought to evaluate an algorithm that automatically identifies EGM-DC, classifies mapping points, and creates new voltage maps, named “Slow Conducting Channel Maps” (SCC-Maps). Methods. Retrospective analysis of elec- troanatomic maps (EAM) from 20 patients (10 ischemic, 10 with arrhythmogenic right ventricular dysplasia/cardiomyopathy) was performed. EAM voltage maps were acquired during sinus rhythm and used for ablation. Preprocedural contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging was available for the ischemic population. +ree mapping modalities were analysed: (i) EAM voltage maps using standard (EAM standard) or manual (EAM screening) thresholds for defining core and border zones; (ii) SCC-Maps derived from the use of the novel SCC-Mapping algorithm that automatically identify EGM-DCs measuring the voltage of the local component; and (iii) Ce-CMR maps (when available). +e ability of each mapping modality in identifying SCCs and their agreement was evaluated. Results. SCC-Maps and EAM screening identified a greater number of SCC entrances than EAM standard (3.45± 1.61 and 2.95± 2.31, resp., vs. 1.05± 1.10; p< 0.01). SCC-Maps and EAM screening highly correlate with Ce-CMR maps in the ischemic population when compared to EAM standard (Lin’s correlation = 0.628 and 0.679, resp., vs. 0.212, p< 0.01). Conclusion. +e SCC-Mapping algorithm allows an operator-independent analysis of EGM signals showing better identification of the arrhythmogenic substrate characteristics when compared to standard voltage EAM. 2 Journal of Interventional Cardiology Table 1: Baseline characteristics of the study population. 1. Introduction Entire Voltage mapping allows the characterization of myocar- Ischemic ARVD/C population p-value (n � 10) (n � 10) dial scar, being a useful tool for ablation of scar-related (n � 20) ventricular arrhythmias (VA) [1–3]. Small bundles of Age (years) 57± 15 69± 8 45± 9 <0.001 viable cardiac myocytes within the scar create slow Sex (male) 15 (75%) 9 (90%) 6 (60%) 0.303 conducting channels (SCCs) that are responsible of the Hypertension 7 (35%) 6 (60%) 1 (10%) 0.057 formation of reentrant circuits promoting VA [3–9]. (n) Computer software for electroanatomical mapping Dyslipidemia 8 (40%) 7 (70%) 1 (10%) 0.020 (EAM) allows the quantification of local electrogram (n) LVEF (%) 44± 16 35± 20 49± 14 0.193 (EGM) voltages as the peak-to-peak differences of each EAM points (n) 532± 212 438± 208 626± 180 0.076 bipolar EGM signal [10]. However, far-field activity from surrounding healthy tissue can result in underestimation Values are given as mean± standard deviation or n (%). p-value refers to the comparison between ischemic and ARVD/C populations. ARVD/C: of the scar area and may lead to a worse definition of EGM arrhythmogenic right ventricular dysplasia/cardiomyopathy; LVEF: left signals with delayed components (EGM-DC), thus ventricular ejection fraction; and EAM: electroanatomical mapping. masking the presence of SCCs. +e “scar dechanneling” technique has been introduced 2.2. Mapping and Ablation Procedure. Electroanatomical as a substrate ablation strategy for scar-related VAs, either for ischemic or nonischemic cardiomyopathy [8, 9, 11]. maps (EAM) were obtained with the CARTO3 navigation system (Biosense Webster, Inc., Diamond Bar, CA, USA) Briefly, this technique is based on bipolar voltage mapping of the scar during sinus rhythm (SR), analysis of EGMs to using a 3.5 mm irrigated-tip +ermoCool SmartTouch ® ® identify SCCs, and ablation of all the identified SCC en- catheter (Biosense Webster, Inc., Diamond Bar, CA, USA) trances. Outcomes of the “scar dechanneling” technique for mapping and ablation. Bipolar electrograms were filtered depend on the correct identification and elimination of all from 30 to 250 Hz. +e 12-lead surface electrocardiogram present SCCs [9]. +is can be a time-consuming and skill- (ECG) and EGM signals from the mapping catheter were demanding task, being subject to significant interoperator displayed and stored for prospective analysis. Endocardial EAM maps were acquired for all patients belonging to the variability. We hypothesize that an automatic system able to identify EGM-DC within the substrate could simplify and ischemic subpopulation and in seven of the patients from the ARVD/C subpopulation, the rest of EAM maps of the standardize VA ablation procedures. In this study, we present and evaluate the performance ARVD/C subpopulation were obtained from the epicardium. of a novel algorithm for automatic EGM analysis so called “Slow Conducting Channel Mapping” algorithm, or Ablation was performed under conscious sedation or “SCC-Mapping.” +is algorithm dichotomizes normal general anaesthesia when epicardial access was required or from abnormal bipolar EGMs, automatically identifying anticipated. Bipolar voltage maps were obtained during SR the presence of EGM-DC within the substrate. By mea- and scar was identified using standard voltage thresholds defining scar core zone (CZ) (<0.5 mV), border zone (BZ) suring the bipolar voltage belonging to the local-field component, the SCC-Mapping algorithm may obtain (<1.5 mV), and healthy tissue (≥1.5 mV). +e “scar dechanneling” ablation technique was used for identification more accurate bipolar voltage maps. +us, a better scar characterization could help to guide scar-related VA and ablation of SCC entrances, thus isolating the VA isthmuses [9]. Identification of SCC was performed man- ablation procedures. ually by the EAM navigation system operator. After ablation of all the SCC entrances, a remap procedure was performed 2. Methods to detect any residual SCC and ablate them if needed. When finished, a programmed stimulation protocol was per- 2.1. Patient Sample. Twenty patients (fifteen males) with VA formed, remapping again the substrate in case any sustained who underwent catheter-based radiofrequency ablation VAs were found inducible, until noninducibility was were included in the study. Ten patients (nine males) had achieved. Ablation was performed in temperature-con- ischemic cardiomyopathy. Ten (six males) fulfilled Task trolled mode with 45 C temperature and 50 W power limit at Force criteria for arrhythmogenic right ventricular dyspla- 26 mL/min irrigation rate (40 W and 17 mL/min at sia/cardiomyopathy (ARVD/C). Ischemic patients were epicardium). selected from our database of VA substrate ablation as consecutive patients having preprocedural contrast-en- hanced cardiac magnetic resonance (Ce-CMR) imaging 2.3. Ce-CMR Acquisition and Processing. A preprocedural study. Basal characteristics of the population are summa- late gadolinium Ce-CMR was acquired in all ischemic cases rized in Table 1. +e study complied with the Declaration of and used to localize the arrhythmogenic substrate [12]. +e Helsinki, and the local ethics committee approved the study preprocedural Ce-CMR studies were obtained using a 3T protocol. All participants included in the study provided scanner (MAGNETOM Trio , Siemens Healthcare, ® ® informed written consent. Erlangen, Germany). Contrast-enhanced images were Journal of Interventional Cardiology 3 acquired 10 minutes after bolus injection of 0.2 mmol/kg distance between the main deflection of far- and local-field Gadobutrol (Gadovist , Bayer Hispania, Barcelona, Spain) component (i.e., the first and the second component) di- chotomized between d-EGMs and f-EGMs (see Figure 1(a)). using a commercially available, free-breathing, ECG-gated, navigator-gated, 3D inversion-recovery, gradient-echo +e cut-off threshold was set to 25 ms as a trade-off for good technique. identification of f-EGM and d-EGM signals. f-EGMs were Ce-CMR images were analysed as previously described ablation targets according to the “scar dechanneling” [12]. Briefly, a full left ventricular (LV) volume was technique, as they constitute the typical pattern at SCC reconstructed in the axial orientation, and the resulting entrances [9]. images were processed with the commercially available +e outcome of the algorithm is the identification label ADAS-3D software (Galgo Medical, Barcelona, Spain). on the type of EGM assigned to each mapping point. +ese Ten concentric surface layers (from 10% to 90%) were labels were then colour-coded and integrated into the 3D created automatically from endo- to epicardium of the LV EAM with the following criteria: small white spheres for wall thickness. A 3D shell was obtained for each layer. normal EGM mapping points; blue big spheres for d-EGM mapping points, and black big spheres for f-EGMs mapping Pixel signal intensity (PSI) maps based on Ce-CMR im- ages were projected to each shell, following a trilinear points. interpolation algorithm, and colour-coded. To identify the scar areas, a PSI-based algorithm was applied to char- 2.5. Construction of SCC-Maps. +e SCC-Mapping algo- acterize the hyperenhanced area as CZ, BZ or healthy rithm identifies the existence of potential EGM-DC, tissue using 40%± 5% and 60%± 5% of the maximum allowing to measure the bipolar amplitude of far- and intensity as thresholds [12]. +e BZ channels (i.e., SCCs) local-field components individually. +e projection of were defined by the ADAS-3D software as continuous these voltages on a 3D SCC-Map was performed using an 3D corridors (across all the Ce-CMR layers) of BZ (with additional decision tree (Figure 1(b)) with two branches: the specified PSI threshold) surrounded by scar core/ one for single-component EGMs and another for EGM- mitral annulus [12]. DC (either d-EGMs or f-EGMs). +e single-component EGM branch considered any mapping point <1.5 mV as far-field remote signals measured within the dense scar 2.4. @e “Slow Conducting Channel Mapping” (“SCC-Map- coming from the surrounding healthy myocardium, ping”) Algorithm. +e SCC-Mapping algorithm is based on automatically setting its bipolar voltage to zero. +e an EGM detector and delineator algorithm previously de- double-component EGM branch projects on the 3D SCC- veloped by our team [13]. +is detector automatically Map the bipolar voltage of the local-field component only identifies and delineates the onset and end landmarks of the if it is included within a window of interest. +is window bipolar EGM signal using the QRS complex of the 12-lead of interest was defined between the 5th and 95th per- surface ECG as the reference searching window, a method centiles of the onsets and ends of all identified local-field that has been already validated for activation mapping of components, respectively. If the local-field component of focal VA [13, 14]. +e entire processing algorithm was an EGM-DC mapping point did not meet this criterion, implemented in MATLAB (MATLAB R2016a, Math- then the same criterion as for single-component EGM Works, Inc., Natick, MA, USA). +e results were obtained mapping points was applied. Additionally, the algorithm offline; therefore, ablation outcomes were independent of includes a spatial coherence protection. +is protection the presented results. checks, for close EGM-DC mapping points (dis- Starting from an initial delineation of the mapping point tance< 6 mm), if the local-field components are similar in EGM signal using our EGM detector/delineator algorithm activation time and shape. When these criteria are met, [13], the SCC-Mapping algorithm uses a decision tree il- only the highest bipolar value is represented in the 3D lustrated in Figure 1(a). +is decision tree is based on two SCC-Map. main characteristics of the bipolar EGM signal: the delin- eated length and the bipolar voltage. For short-duration EGMs (<65 ms, based on [15]), 2.6. SCC Detection Agreement Evaluation. In this study, normal mapping points are distinguished from those can- three mapping modalities were considered: (i) EAM voltage didates to be an EGM-DC by the measured bipolar voltage. maps, (ii) SCC-Maps derived from the SCC-Mapping Al- +erefore, those mapping points showing a bipolar voltage gorithm; and (iii) Ce-CMR PSI maps (when available). An ≥3.5 mV were considered normal EGMs [15], whereas the expert operator visually evaluated the ability of each map- rest were candidates for being classified as EGM-DC or ping modality to identify SCC entrances from both the 3D remained as normal EGMs. Long-duration EGMs (>65 ms) coloured map and the acquired EGM signals. were always considered potential candidates for EGM-DC, regardless of their voltage value. In order to label an EGM-DC candidate with true 2.6.1. Identification of SCC Entrances Derived from the 3D delayed (d-EGM) or fused (f-EGM) components, the al- Coloured Map. For the EAM voltage maps, SCC entrance gorithm searched for the existence of a second EGM identification was performed using (1) the standard component based on the EGM detector/delineator algo- thresholds that define the presence of scar CZ and BZ rithm [13]. If a second EGM component was found, the time tissue (named “EAM standard”); and (2) using a manual 4 Journal of Interventional Cardiology Delineated EGM length ≤65 ms Yes No No Bipolar voltage ≥3.5 mV? Searching process: Yes Is there a second component? Yes No Distance 1st-to-2nd component ≤25 ms? Yes No Delayed Normal Fused EGM EGM EGM (a) Fused Normal Delayed EGM EGM EGM 2nd component within the Bipolar voltage <1.5 mV? window of interest? No Yes No Yes V = V V = V V = 0 SCC-MAP Far-Field SCC-MAP Local-Field SCC-MAP (b) Figure 1: +e “Slow Conducting Channel Mapping Algorithm.” (a) Decision tree for electrogram (EGM) signals with delayed components (EGM-DC) searching protocol and (b) algorithm for the reconstruction of “Slow Conducting Channel Maps” (SCC-Maps) on patient’s 3D anatomical map. VSCC-Map: bipolar voltage projected on patient’s 3D anatomical map. voltage screening process that dynamically modifies the identified (labelled) mapping points close to the BZ area. standard thresholds for BZ and CZ definition in order to +erefore, this evaluation could only be done for EAM voltage maps and SCC-Maps. enhance the presence of SCCs (named “EAM screening”) [3, 16]. For the SCC-Maps, this process was done directly from the coloured map. It should be noted that, due to the more precise bipolar voltage measurement method of 2.7. Statistical Analysis. Continuous data are shown in SCC-Maps, the CZ tissue threshold was set to ≤0.1 mV, mean± standard deviation, unless otherwise indicated. yielding a higher range of bipolar voltage measurements Categorical data are shown as percentages. Comparison (shown and discussed in the following sections). +ere- between different populations was given by the Wilcox- fore, the voltage screening process was not necessary for on–Mann–Whitney test or by the Fisher exact test when SCC-Maps. For the Ce-CMR PSI maps, the threshold appropriate. For evaluation of the agreement in the SCC definitions for tissue heterogeneity identification (i.e., BZ entrances identification among the different mapping mo- tissue that conforms SCCs) were those explained in the dalities, Wilcoxon–Mann–Whitney test, Lin’s concordance corresponding section. correlation factor “ρ” [17], and Bland–Altman plot analysis [18] were used. A p-value of ≤0.05 was considered as a cut- 2.6.2. Identification of SCC Entrances Derived from the off value for statistical significance. Statistics were obtained Analysed EGM Signals. +is process was done by manual using the MATLAB statistics toolbox (MATLAB R2016a, inspection of the presence or absence of f-EGMs in the MathWorks, Inc., Natick, MA, USA). Journal of Interventional Cardiology 5 in Table 2. However, no significant differences were found 3. Results between EAM screening and SCC-Maps versus Ce-CMR 3.1. Population Characteristics. Twenty patients were in- PSI maps (p � 0.202 and p � 1.0, resp.). Nevertheless, cluded in the study. 75% were male, and mean age was Bland–Altman plot analysis reveals a tendency of these 57± 15 years. Mean LV ejection fraction was 44± 16%, with three mapping modalities towards an underestimation of no significant differences between ischemic patients and the number of SCC entrances compared with Ce-CMR PSI those with ARVD/C. Table 1 summarizes the baseline maps (Pearson’s R � 0.63, p � 0.049; R � 0.85, p � 0.002; characteristics of the study population. and R � 0.61, p � 0.05 for the comparisons between SCC detection on SCC-Maps, EAM standard, and EAM screening against Ce-CMR PSI maps, resp., Figure 2(c)). 3.2. SCC Detection Agreement between EAM and SCC- Moreover, Supplementary Table 1 confirms the high Mapping. Table 2 and Figure 2(a) show the agreement in agreement of EAM screening and SCC-maps with Ce-CMR the number of SCC entrances identified from the 3D PSI maps compared with EAM standard (ρ � 0.679 and coloured maps between all the studied mapping modal- ρ � 0.628, resp., vs. ρ � 0.212, p< 0.01). Figure 4 shows two ities. EAM voltage maps with standard thresholds (“EAM examples where SCC-Maps had higher agreement with Ce- standard”) presented a significant lower number of SCC CMR PSI maps in identifying SCC, as compared to EAM entrances than EAM voltage maps with manual voltage standard maps. screening (“EAM screening”) (p< 0.01, 0.04, and 0.03 for the entire population, ischemic, and ARVD/C, resp.). Additionally, EAM standard maps also had less SCC 3.4. Effect of Selective Bipolar Voltage Measurement. +e entrances than SCC-Maps (p< 0.01, < 0.01, and 0.02 for SCC-Mapping algorithm was capable of detecting EGM- the entire population, ischemic, and ARVD/C, resp.). DC providing the bipolar voltage of the local component, However, there were no significant differences in the thus obtaining accurate voltage maps (SCC-Maps). +is number of identified SCC entrances between EAM more selective approach enlarged the voltage range dis- screening maps and SCC-Maps (p � 0.29, 0.10, and 0.87 played and thus improved the degree of detail of the for the entire population, ischemic, and ARVD/C, resp.). arrhythmogenic substrate when compared with standard Lin’s concordance correlation factor analysis shown in EAM voltage maps. Figure 5(b) illustrates the loss of scar Supplementary Table 1 supports these findings, showing a details in EAM voltage maps when compared to SCC- higher concordance between EAM screening and SCC- Maps, which display a higher range of voltage measure- Maps (ρ � 0.665, 0.528, and 0.877 for the entire pop- ments. Additionally, as depicted in Figure 4, SCC-Maps ulation, ischemic, and ARVD/C, resp.) than EAM stan- matched better with information obtained from Ce-CMR dard maps. PSI maps in the ischemic population than EAM standard +e Bland–Altman analysis shown in Figure 2(a) illus- maps. trates the agreement in SCC entrance identification among the different studied mapping modalities and populations. 4. Discussion +ere was a low bias in the number of SCC entrances identified between SCC-Maps and EAM screening maps, 4.1. Reviewing Current Bipolar Voltage Mapping. EAM with a small trend towards a SCC subidentification of SCC- systems are useful tools to map scar-related VAs, since Maps compared with EAM screening maps (Pearson’s they allow calculating the peak-to-peak local EGM signal R � 0.48, p � 0.033), which is in concordance with the amplitude and representing this value, colour-coded, on findings of Table 2. the cardiac anatomy, thus helping to identify and char- Table 3 and Figure 2(b) list the agreement in the number acterize the scar [3, 5, 6, 8–10, 19, 20]. However, myo- of SCC entrances identified from the f-EGM points between cardial scars are often surrounded by a considerable mapping modalities. No significant differences were found amount of healthy tissue, which may lead to local EGMs between EAM standard maps and SCC-Maps, confirmed by being masked by the presence of far-field signals. the Bland–Altman analysis shown in Figure 2(b) and the +erefore, regular voltage mapping with standard high correlation shown in Supplementary Table 2 (ρ � 0.918, thresholds may underestimate of the scar size and lose 0.871, and 0.936 for the entire population, ischemic, and significant scar details. +is increases the likelihood of ARVD/C, resp.). missing SCCs within the substrate and the need of ex- Figure 3 illustrates two examples of the electrical tensive operator analysis. propagation sequences along SCCs identified from the au- Examples of this phenomenon are illustrated in Figure 6. tomatically labelled mapping points on SCC-Maps com- In normal EGMs, the voltage map reflects the peak-to-peak pared to the SCC manual identification performed on EAM voltage (a); however, when healthy myocardium EGM (i.e., maps where the identification of these SCCs requires ex- far-field) has a higher amplitude than the late potential (i.e., tensive operator analysis. local-field), the voltage map reflects the far-field, high- amplitude component voltage (b). Moreover, when the far- 3.3. SCC Detection Agreement with Ce-CMR. +e number of and local-field components show comparable amplitudes, SCC entrances identified in EAM standard was signifi- voltage map may reflect the peak-to-peak amplitude of either cantly lower compared with Ce-CMR PSI maps, as shown the local or the far-field component (c), or a mix of both (d). 6 Journal of Interventional Cardiology Table 2: Analysis of colour-coded maps. Number of SCC entrances identified per patient and agreement between mapping modalities. ∗ † ‡ EAM standard EAM screening SCC-Map Ce-CMR PSI maps p-value p-value p-value Entire population (n � 20) 1.05± 1.10 2.95± 2.31 3.45± 1.61 N/A <0.01 <0.01 0.29 Ischemic (n � 10) 0.60± 1.00 2.20± 1.75 3.60± 1.43 3.70± 2.45 0.04 <0.01 0.10 ARVD/C (n � 10) 1.50± 1.08 3.70± 2.63 3.30± 1.83 N/A 0.03 0.02 0.87 ∗ † Number of SCC entrances per patient are given as mean± standard deviation. Differences between EAM standard and EAM screening. Differences between EAM standard and SCC-Maps. Differences between EAM screening and SCC-Maps. ARVD/C: arrhythmogenic right ventricular dysplasia/cardiomyopathy; Ce-CMR: contrast-enhanced cardiac magnetic resonance; EAM: electroanatomical mapping; N/A: not applicable; PSI: pixel signal intensity; and SCC: slow conducting channel. EAM standard vs. SCC-maps EAM screening vs. SCC-maps EAM standard vs. EAM screening 5 5 5 0 0 0 –5 –5 –5 0 2468 0 2468 0 2468 (EAM standard + SCC-maps)/2 (EAM screening + SCC-maps)/2 (EAM standard + EAM screening)/2 Ischemic ARVD/C (a) EAM standard f-EGMs vs. SCC-maps f-EGMs –5 0 510 15 (EAM standard + SCC-maps)/2 Ischemic ARVD/C (b) Figure 2: Continued. EAM standard-SCC-maps EAM screening-SCC-maps EAM standard-SCC-maps EAM standard-EAM screening Journal of Interventional Cardiology 7 Ce-CMR PSI maps vs. SCC-maps Ce-CMR PSI maps vs. EAM standard Ce-CMR PSI maps vs. EAM screening 5 5 5 0 0 0 –5 –5 –5 0 246802 2468 0 468 (Ce-CMR PSI maps + (Ce-CMR PSI maps + (Ce-CMR PSI maps + SCC-maps)/2 EAM standard)/2 EAM screening)/2 Ischemic (c) Figure 2: Bland–Altman plots for assessing the agreement in the identification of slow conducting channel (SCC) entrances (a) from the colour-coded 3D maps between the different mapping modalities: electroanatomical mapping (EAM) system maps with standard voltage thresholds (EAM standard), EAM maps with voltage screening (EAM screening), and “Slow Conducting Channel Maps” (SCC-Maps). (b) From the analysis of the presence of fused electrograms (f-EGM) components between EAM standard maps and SCC-Maps and (c) from the colour-coded 3D map between the different mapping modalities and the pixel signal intensity (PSI) maps derived from contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging in the ischemic population. Red solid line indicates mean and red dashed lines indicate mean± 2 standard deviations of the difference in the number of identified SCC entrances. ARVDC: arrhythmogenic right ventricular dysplasia/cardiomyopathy. Table 3: Analysis of EGM-DC and identification of f-EGMs. Number of SCC entrances identified per patient and agreement between mapping modalities. EAM maps SCC-Map p-value Entire population (n � 20) 6.10± 2. 81 5.35± 2.70 0.430 Ischemic (n � 10) 5.50± 2.17 4.70± 2.11 0.422 ARVD/C (n � 10) 6.70± 3.34 6.00± 3.16 0.790 Number of SCC entrances per patient are given as mean± standard deviation. ARVD/C: arrhythmogenic right ventricular dysplasia/cardiomyopathy; EAM: electroanatomical mapping; EGM-DC: electrograms with delayed components; f-EGM: fused electrograms; and SCC: slow conducting channel. A1 A2 Channel 1 Channel 2 V2 V2 1 1 2 2 3 3 4 9 7 12 0.5 0.1 1.5 1.5 EAM voltage map (mV) SCC-map (mV) (a) Figure 3: Continued. Ce-CMR PSI maps-SCC-maps Ce-CMR PSI maps-EAM standard Ce-CMR PSI maps-EAM screening 8 Journal of Interventional Cardiology B1 B2 Channel 1 Channel 2 V3 V3 1 1 4 3 0.5 0.1 10 1.5 1.5 EAM voltage map (mV) SCC-map (mV) (b) Figure 3: Examples of slow conducting channel (SCC) identification from the automatic mapping point labelling on “Slow Conducting Channel Maps” (SCC-Maps). (a) Endocardial electroanatomical map (EAM) of an ischemic patient showing two SCCs identified on SCC- Map. (b) Epicardial EAM from an arrhythmogenic right ventricular dysplasia/cardiomyopathy patient showing two SCCs identified on SCC-Map. A1 A2 A3 AV AV AV 0.5 1.5 0.1 1.5 60% 40% EAM voltage map (mV) SCC-map (mV) PSI map 10% wall layer (a) Figure 4: Continued. Journal of Interventional Cardiology 9 B1 B2 B3 0.5 1.5 0.1 1.5 60% 40% EAM voltage map (mV) SCC-map (mV) PSI map 10% wall layer (b) Figure 4: Agreement between electroanatomical mapping (EAM) voltage maps and “Slow Conducting Channel Maps” (SCC-Maps) against pixel signal intensity (PSI) maps derived from contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging. A1 and B1 show the electroanatomical mapping (EAM) voltage maps obtained with the EAM system from two different patients. A2 and B2 show the cor- responding SCC-Map and A3 and B3 show the acquired Ce-CMR PSI map. AV: aortic valve. 0.1 1.5 0.5 1.5 0.1 1.5 EAM voltage map (mV) EAM voltage map SCC-map (mV) modified th. (mV) (a) (b) (c) Figure 5: Endocardial substrate map from patient with myocardial infarction. (c) illustrates the richest scar details shown by the “Slow Conducting Channel Map” (SCC-Map) compared with electroanatomical mapping (EAM) voltage maps using the standard voltage thresholds (a) and using modified voltage thresholds (b). 10 Journal of Interventional Cardiology V6 V6 Normal EGM Local-field masked by healthy myocardium far-field 4 1 d − 2 d − 2 V = 0.53 mV V = 1.99 mV Local EAM V = 6.19 mV –1 EAM –2 –4 –2 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Time (ms) Time (ms) (a) (b) V6 V6 Local-field masked by comparable far-field Voltage measurement from both components 0.2 0.4 0.1 0.2 V = 0.49 mV d − 2 EAM V = 0.41 mV Local d − 2 V = 0.24 mV Local –0.1 –0.2 V = 0.33 mV EAM –0.2 –0.4 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Time (ms) Time (ms) (c) (d) Figure 6: Examples of (a) normal electrogram (EGM) bipolar voltage measurement by the electroanatomical mapping (EAM) system and (b–d) different bipolar EGM-DC signals with incorrect bipolar voltage measurement by the EAM system: (b) local-field component masked by high-amplitude far-field component. (c-d) Comparable amplitude of far-field and local-field components. +ese examples show the need of a more selective approach 4.3. SCC Detection and Agreement between Mapping to measure bipolar voltage for substrate mapping of scar- Modalities. SCC-Maps were highly correlated with EAM related VAs. standard maps when these were obtained after a manual voltage screening process (EAM screening), whereas raw EAM standard maps (EAM standard) correlated worse and identified a significant lower number of SCC en- 4.2. Main Findings. +e present study evaluates a novel trances. +e algorithm’s accuracy is illustrated by the fact automatic EGM signal analysis algorithm aiming to im- that SCC-Maps can display a detailed scar without the prove the accuracy of current voltage mapping obtained need of manual voltage screening, with good agreement with EAM systems. +is algorithm allows obtaining voltage with EAM after manual labelling of the mapping points maps with higher voltage range, hence depicting more corresponding to SCC entrances (i.e., those showing detailed scar characteristics, which may be useful to f-EGM signals). identify VA isthmuses during ablation procedures. +e Figure 3 shows ischemic and ARVD/C patient ex- main findings of the study are as follows: (1) the proposed amples where two different SCCs can be found. Both SCC-Mapping algorithm automatically identified SCC examples illustrate the superiority of SCC-Maps over entrances at the same level as manual EAM voltage EAM standard maps, allowing the identification of those screening; (2) the SCC-Mapping algorithm provided SCC- signal corridors by direct inspection of the colour map. Maps that match Ce-CMR PSI maps better than current +e EAM standard map of Figure 3(a) (panel A1) depicts a EAM voltage maps; and (3) the SCC-Mapping algorithm dense scar in the area where a second SCC can be found improves the definition of the scar CZ and BZ areas by following the activation sequence of the d-EGMs. +is allowing a higher voltage range. SCC can be easily identified by the colour map and Amplitude (mV) Amplitude (mV) Amplitude (mV) Amplitude (mV) Journal of Interventional Cardiology 11 devices in the ARVD/C population. On the other hand, automatic labelling in Figure 3(a) (panel A2). Similarly, the EAM standard map of Figure 3(b) (panel B1) does not the number of available patients with detailed EAM and quality Ce-CMR data has reduced lately, as detailed EAM allow to identify those SCCs that can be seen in the SCC- Map of Figure 3(b) (panel B2). Moreover, SCC-Maps and acquisition is a time-consuming and highly operator- EAM-screening correlate better than EAM-standard maps dependent task, while Ce-CMR-guided catheter substrate with Ce-CMR PSI maps. +ese examples illustrate the ablation has gained more interest [26]. Despite this trend, need of manual analysis of the EGM signals by the system the endpoints for determining the ablation targets are still operator in order to identify all the possible SCCs present based on EAM findings where the presented algorithm can in the substrate (i.e., using techniques like manual voltage play an important and complementary role. Also, this screening and individual EGM labelling). +is procedure algorithm paves the way for better integration of Ce-CMR can be guided and shortened by the proposed automatic and EAM data to improve scar-related VT ablation procedures. SCC-Mapping algorithm. Moreover, automatic and ob- jective identification becomes mandatory when using the Additionally, the algorithm was designed and tested using data from substrate-based VA ablation during SR. increasingly popular multielectrode mapping (MEM) catheters where tenths of simultaneous signals per beat Hence, no data from VA mapping was used or analysed with can be acquired. this algorithm and therefore, other possible VA isthmuses were not explored with this algorithm [27]. In this work, EAMs were acquired using a standard 4.4. Voltage @resholds for Scar Definition. +e proposed 3.5 mm irrigated-tip mapping catheter, which has a SCC-mapping algorithm provides a more precise quanti- longer interelectrode distance as compared with high- fication of the local-field voltage. +is aspect allows density MEM catheters that can better discriminate lo- changing the threshold definition for CZ tissue without cal- from far-field components [21]. However, manual losing scar information, thus improving the ability to annotation of multiple simultaneous signals obtained detect of SCC (Figure 4). +is effect is comparable to the with MEM is a nonaffordable task, for which an auto- one obtained with current MEM catheters [21], but using matic approach (like the SCC-Mapping algorithm) be- a regular electrode-size catheter in conjunction with an comes necessary. automatic algorithm to distinguish the far- and local-field components of the measured EGMs. In contrast, MEM 5. Conclusions catheters still need extensive operator analysis in order to identify and/or enhance the presence of SCCs. +is fact +e proposed automatic analysis of EGM signals using the could be mitigated if high-density mapping is combined “Slow Conducting Channel Mapping Algorithm” improves with an automatic algorithm as the presented in this the accuracy of bipolar voltage measurements within the scar work. area, achieving a more detailed tissue characterization and Figure 5 illustrates the loss of scar definition when the being an operator-independent tool for accurate identifi- modified voltage threshold for BZ and CZ tissue is used on cation of SCCs. +is last feature encourages the use of the EAM voltage maps. +e higher voltage range displayed by algorithm together with EAM navigation systems as a re- SCC-Maps also facilitates the voltage screening process producible approach for guiding VA ablation procedures in for SCC identification. However, although SCC-Maps and daily practice. EAM screening maps provide similar insights, the former were obtained without any manual intervention, thus Data Availability being operator independent. Additionally, as shown in Figures 3–5, it can be observed that SCC-Maps provided a +e electroanatomical mapping and image data used to better defined scar delineation than current EAM voltage support the findings of this study are restricted by the mapping. ´ Hospital Clınic Local Ethical Committee in order to protect +e proposed algorithm was evaluated when using the patient privacy. Data are available from Dr. Llu´ıs Mont, “scar dechanneling” ablation technique. However, the fact PhD, Arrhythmias Unit, Hospital Clinic, Carrer de Villar- that a higher voltage range can be described with SCC- roel, 170, 08036 Barcelona, Spain, for researchers who meet Maps allows a more detailed scar characterization, which the criteria for access to confidential data. suggests that the algorithm can also be useful for other ablation approaches [5, 22–25]. Eventually, the SCC- Conflicts of Interest Mapping algorithm could effectively improve the guid- Dr. A. Berruezo and Dr. L. Mont are stockholders in Galgo ance of pacing/entrainment manoeuvres for VA isthmus Medical SL. David Soto-Iglesias is an employee of Biosense identification based on the tagged data and SCC-Map Webster, Inc. +e authors declare that there are no conflicts information. of interest. 4.5. Study Limitations. +e main limitation of this study Acknowledgments was the relatively small sample size. Comparison of EAM maps against Ce-CMR PSI maps was only possible in the +is study was supported by personal grants to A. Alcaine ischemic population due to the presence of implantable (Refs. BES-2011-046644 and EEBB-I-15-09466); by 12 Journal of Interventional Cardiology pp. 301–309, John Wiley & Sons, Hoboken, NJ, USA, 3rd projects PID2019-104881RB-I00 from Ministerio de edition, 2009. Ciencia e Innovacio´n (Spain) and DPI2016-75458-R from [8] A. Berruezo, J. Fernandez-Armenta, ´ L. Mont et al., “Com- Ministerio de Economıa y Competitividad (Spain); and by bined endocardial and epicardial catheter ablation in Gobierno de Arago´n (Grupo Referencia BSICoS ref.: arrhythmogenic right ventricular dysplasia incorporating scar T39_20R) cofounded by FEDER 2014–2020. +is work was dechanneling technique,” Circulation: Arrhythmia and Elec- also supported in part by the project PI14/00759, integrated trophysiology, vol. 5, no. 1, pp. 111–121, 2012. in the Plan Nacional de I+D+i and cofounded by the [9] A. Berruezo, J. Fernandez-Armenta, ´ D. Andreu et al., “Scar Instituto de Salud Carlos III (ISCIII)-Subdireccio´n General dechanneling: new method for scar-related left ventricular de Evaluacio´n and European Regional Development Fund tachycardia substrate ablation,” Circulation: Arrhythmia and (European Union). +e computation was performed by the Electrophysiology, vol. 8, no. 2, pp. 326–336, 2015. ICTS NANBIOSIS, more specifically by the High Perfor- [10] Z. F. Issa, J. M. Miller, and D. P. Zipes, Clinical Arrhythmology mance Computing Unit of the CIBER in Bioengineering, and Electrophysiology: A Companion to Braunwald’s Heart Disease, Saunders, Philadelphia, PA, USA, 2nd edition, 2012. Biomaterials and Nanomedicine (CIBER-BBN) at the [11] J. Fernandez-Armenta, ´ D. Andreu, D. Penela et al., “Sinus University of Zaragoza. +e CIBER-BBN is an initiative of rhythm detection of conducting channels and ventricular Instituto de Salud Carlos III. tachycardia isthmus in arrhythmogenic right ventricular cardiomyopathy,” Heart Rhythm, vol. 11, no. 5, pp. 747–754, Supplementary Materials ´ ´ [12] D. Andreu, J. T. Ortiz-Perez, J. Fernandez-Armenta et al., “3D Supplementary Table 1: Lin’s concordance correlation factor delayed-enhanced magnetic resonance sequences improve on the number of SCC entrances identified between map- conducting channel delineation prior to ventricular tachy- ping modalities (rows against columns) from the analysis of cardia ablation,” EP Europace, vol. 17, no. 6, pp. 938–945, the colour-coded maps. Supplementary Table 2: Lin’s con- cordance correlation factor on the number of SCC entrances [13] A. Alcaine, D. Soto-Iglesias, M. Calvo et al., “A wavelet-based identified between EAM standard maps and SCC-Maps electrogram onset delineator for automatic ventricular acti- vation mapping,” IEEE Transactions on Biomedical Engi- from the analysis of EGM-DC and identification of f-EGMs. neering, vol. 61, no. 12, pp. 2830–2839, 2014. (Supplementary Materials) [14] A. Alcaine, D. Soto-Iglesias, J. Acosta et al., “Automatic ac- tivation mapping and origin identification of idiopathic References outflow tract ventricular arrhythmias,” Journal of Electro- cardiology, vol. 51, no. 2, pp. 239–246, 2018. [1] D. J. Callans, J.-F. Ren, J. Michele, F. E. Marchlinski, and [15] D. M. Cassidy, J. A. Vassallo, F. E. Marchlinski, A. E. Buxton, S. M. Dillon, “Electroanatomic left ventricular mapping in the W. J. Untereker, and M. E. Josephson, “Endocardial mapping porcine model of healed anterior myocardial infarction. in humans in sinus rhythm with normal left ventricles: ac- Correlation with intracardiac echocardiography and patho- tivation patterns and characteristics of electrograms,” Cir- logical analysis,” Circulation, vol. 100, no. 16, pp. 1744–1750, culation, vol. 70, no. 1, pp. 37–42, 1984. [16] S. E. Mountantonakis, R. E. Park, D. S. Frankel et al., “Re- [2] V. Y. Reddy, D. Wrobleski, C. Houghtaling, M. E. Josephson, lationship between voltage map “channels” and the location of and J. N. Ruskin, “Combined epicardial and endocardial critical isthmus sites in patients with post-infarction car- electroanatomic mapping in a porcine model of healed diomyopathy and ventricular tachycardia,” Journal of the myocardial infarction,” Circulation, vol. 107, no. 25, American College of Cardiology, vol. 61, no. 20, pp. 2088–2095, pp. 3236–3242, 2003. [3] A. Arenal, S. del Castillo, E. Gonzalez-Torrecilla et al., [17] L. Lin, A. S. Hedayat, B. Sinha, and M. Yang, “Statistical “Tachycardia-related channel in the scar tissue in patients methods in assessing agreement: models, issues, and tools,” with sustained monomorphic ventricular tachycardias: in- Journal of the American Statistical Association, vol. 97, fluence of the voltage scar definition,” Circulation, vol. 110, no. 457, pp. 257–270, 2002. no. 17, pp. 2568–2574, 2004. [18] D. G. Altman and J. M. Bland, “Measurement in medicine: the [4] J. M. de Bakker, F. J. van Capelle, M. J. Janse et al., “Reentry as analysis of method comparison studies,” @e Statistician, a cause of ventricular tachycardia in patients with chronic vol. 32, no. 3, pp. 307–317, 1983. ischemic heart disease: electrophysiologic and anatomic [19] F. E. Marchlinski, D. J. Callans, C. D. Gottlieb, and E. Zado, correlation,” Circulation, vol. 77, no. 3, pp. 589–606, 1988. “Linear ablation lesions for control of unmappable ventricular [5] A. Arenal, E. Glez-Torrecilla, M. Ortiz et al., “Ablation of tachycardia in patients with ischemic and nonischemic car- electrograms with an isolated, delayed component as treat- diomyopathy,” Circulation, vol. 101, no. 11, pp. 1288–1296, ment of unmappable monomorphic ventricular tachycardias in patients with structural heart disease,” Journal of the American College of Cardiology, vol. 41, no. 1, pp. 81–92, 2003. [20] J. Fernandez-Armenta, A. Berruezo, D. Andreu et al., “+ree- dimensional architecture of scar and conducting channels [6] E. M. Cronin, S. M. Al-Khatib, E. Anter et al., “2019 HRS/ EHRA/APHRS/LAHRS expert consensus statement on based on high resolution ce-CMR: insights for ventricular tachycardia ablation,” Circulation: Arrhythmia and Electro- catheter ablation of ventricular arrhythmias,” Europace, vol. 8, pp. 1–147, 2019. physiology, vol. 6, no. 3, pp. 528–537, 2013. [21] J. Acosta, D. Penela, D. Andreu et al., “Multielectrode vs. [7] J. V. Wylie, T. W. Smith, and M. E. Josephson, “Substrate mapping for ablation of ventricular tachycardia in coronary point-by-point mapping for ventricular tachycardia substrate ablation: a randomized study,” EP Europace, vol. 20, no. 3, artery disease,” in Cardiac Mapping, M. Shenasa, G. Hindricks, M. Borggrefe, and G. Breithardt, Eds., pp. 512–519, 2018. Journal of Interventional Cardiology 13 [22] L. Di Biase, P. Santangeli, D. J. Burkhardt et al., “Endo-epi- cardial homogenization of the scar versus limited substrate ablation for the treatment of electrical storms in patients with ischemic cardiomyopathy,” Journal of the American College of Cardiology, vol. 60, no. 2, pp. 132–141, 2012. [23] W. S. Tzou, D. S. Frankel, T. Hegeman et al., “Core isolation of critical arrhythmia elements for treatment of multiple scar- based ventricular tachycardias,” Circulation: Arrhythmia and Electrophysiology, vol. 8, no. 2, pp. 353–361, 2015. [24] F. Bogun, E. Good, S. Reich et al., “Isolated potentials during sinus rhythm and pace-mapping within scars as guides for ablation of post-infarction ventricular tachycardia,” Journal of the American College of Cardiology, vol. 47, no. 10, pp. 2013–2019, 2006. [25] P. Jaıs, P. Maury, P. Khairy et al., “Elimination of local ab- normal ventricular activities : a new end point for substrate modification in patients with scar-related ventricular tachy- cardia,” Circulation, vol. 125, no. 18, pp. 2184–2196, 2012. [26] C. Blomstrom-Lundqvist, ¨ A. Auricchio, J. Brugada et al., “+e use of imaging for electrophysiological and devices proce- dures: a report from the first European heart rhythm asso- ciation policy conference, jointly organized with the European association of cardiovascular imaging (EACVI), the council of cardiovascular imaging and the European society of cardiac radiology,” Europace, vol. 15, no. 7, pp. 927–936, 2013. [27] R. Martin, P. Maury, C. Bisceglia et al., “Characteristics of scar-related ventricular tachycardia circuits using ultra-high- density mapping,” Circulation: Arrhythmia and Electrophys- iology, vol. 11, no. 10, Article ID e006569, 2018. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Interventional Cardiology Hindawi Publishing Corporation

Loading next page...
 
/lp/hindawi-publishing-corporation/automatic-detection-of-slow-conducting-channels-during-substrate-zDDmdpSgul

References (35)

Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2020 Alejandro Alcaine et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
1540-8183
eISSN
0896-4327
DOI
10.1155/2020/4386841
Publisher site
See Article on Publisher Site

Abstract

Hindawi Journal of Interventional Cardiology Volume 2020, Article ID 4386841, 13 pages https://doi.org/10.1155/2020/4386841 Research Article Automatic Detection of Slow Conducting Channels during Substrate Ablation of Scar-Related Ventricular Arrhythmias 1,2,3 4 4 5 Alejandro Alcaine , Beatriz Ja´uregui, David Soto-Iglesias, Juan Acosta, 6 7 8 9 Diego Penela, Juan Ferna´ndez-Armenta, Markus Linhart, David Andreu, 10,11,12 1,2 3 1,2 Lluı´s Mont, Pablo Laguna , Oscar Camara, Juan Pablo Martı´nez , and Antonio Berruezo CIBER en Bioingenier´ıa, Biomateriales y Nanomedicina (CIBER-BBN), Zaragoza, Spain BSICoS Group, Arago´n Institute of Engineering Research (I3A), IIS Arago´n, Universidad de Zaragoza, Zaragoza, Spain BCN MedTech Unit, PhySense Group, Department of Information and Communication Technologies, Universitat Pompeu Fabra, Barcelona, Spain Teknon Medical Center, Barcelona, Spain Hospital Universitario Virgen del Roc´ıo, Sevilla, Spain Ospedale Guglielmo da Saliceto, Piacenza, Italy Hospital Puerta del Mar, Ca´diz, Spain Arrhythmia Section, Cardiology, Hospital Universitari Doctor Josep Trueta, Girona, Spain Boston Scientific, Madrid, Spain Hospital Cl´ınic, Universitat de Barcelona, Barcelona, Spain Institut d’Investigacions Biom`ediques August Pi i Sunyer (IDIBAPS), Barcelona, Spain CIBER en Enfermedades Cardiovasculares (CIBER-CV), Barcelona, Spain Correspondence should be addressed to Antonio Berruezo; antonio.berruezo@quironsalud.es Received 12 November 2019; Revised 7 March 2020; Accepted 19 March 2020; Published 29 May 2020 Academic Editor: Jochen Wo¨hrle Copyright © 2020 Alejandro Alcaine et al. +is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Voltage mapping allows identifying the arrhythmogenic substrate during scar-related ventricular arrhythmia (VA) ablation procedures. Slow conducting channels (SCCs), defined by the presence of electrogram (EGM) signals with delayed components (EGM-DC), are responsible for sustaining VAs and constitute potential ablation targets. However, voltage mapping, as it is currently performed, is time-consuming, requiring a manual analysis of all EGMs to detect SCCs, and its accuracy is limited by electric far-field. We sought to evaluate an algorithm that automatically identifies EGM-DC, classifies mapping points, and creates new voltage maps, named “Slow Conducting Channel Maps” (SCC-Maps). Methods. Retrospective analysis of elec- troanatomic maps (EAM) from 20 patients (10 ischemic, 10 with arrhythmogenic right ventricular dysplasia/cardiomyopathy) was performed. EAM voltage maps were acquired during sinus rhythm and used for ablation. Preprocedural contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging was available for the ischemic population. +ree mapping modalities were analysed: (i) EAM voltage maps using standard (EAM standard) or manual (EAM screening) thresholds for defining core and border zones; (ii) SCC-Maps derived from the use of the novel SCC-Mapping algorithm that automatically identify EGM-DCs measuring the voltage of the local component; and (iii) Ce-CMR maps (when available). +e ability of each mapping modality in identifying SCCs and their agreement was evaluated. Results. SCC-Maps and EAM screening identified a greater number of SCC entrances than EAM standard (3.45± 1.61 and 2.95± 2.31, resp., vs. 1.05± 1.10; p< 0.01). SCC-Maps and EAM screening highly correlate with Ce-CMR maps in the ischemic population when compared to EAM standard (Lin’s correlation = 0.628 and 0.679, resp., vs. 0.212, p< 0.01). Conclusion. +e SCC-Mapping algorithm allows an operator-independent analysis of EGM signals showing better identification of the arrhythmogenic substrate characteristics when compared to standard voltage EAM. 2 Journal of Interventional Cardiology Table 1: Baseline characteristics of the study population. 1. Introduction Entire Voltage mapping allows the characterization of myocar- Ischemic ARVD/C population p-value (n � 10) (n � 10) dial scar, being a useful tool for ablation of scar-related (n � 20) ventricular arrhythmias (VA) [1–3]. Small bundles of Age (years) 57± 15 69± 8 45± 9 <0.001 viable cardiac myocytes within the scar create slow Sex (male) 15 (75%) 9 (90%) 6 (60%) 0.303 conducting channels (SCCs) that are responsible of the Hypertension 7 (35%) 6 (60%) 1 (10%) 0.057 formation of reentrant circuits promoting VA [3–9]. (n) Computer software for electroanatomical mapping Dyslipidemia 8 (40%) 7 (70%) 1 (10%) 0.020 (EAM) allows the quantification of local electrogram (n) LVEF (%) 44± 16 35± 20 49± 14 0.193 (EGM) voltages as the peak-to-peak differences of each EAM points (n) 532± 212 438± 208 626± 180 0.076 bipolar EGM signal [10]. However, far-field activity from surrounding healthy tissue can result in underestimation Values are given as mean± standard deviation or n (%). p-value refers to the comparison between ischemic and ARVD/C populations. ARVD/C: of the scar area and may lead to a worse definition of EGM arrhythmogenic right ventricular dysplasia/cardiomyopathy; LVEF: left signals with delayed components (EGM-DC), thus ventricular ejection fraction; and EAM: electroanatomical mapping. masking the presence of SCCs. +e “scar dechanneling” technique has been introduced 2.2. Mapping and Ablation Procedure. Electroanatomical as a substrate ablation strategy for scar-related VAs, either for ischemic or nonischemic cardiomyopathy [8, 9, 11]. maps (EAM) were obtained with the CARTO3 navigation system (Biosense Webster, Inc., Diamond Bar, CA, USA) Briefly, this technique is based on bipolar voltage mapping of the scar during sinus rhythm (SR), analysis of EGMs to using a 3.5 mm irrigated-tip +ermoCool SmartTouch ® ® identify SCCs, and ablation of all the identified SCC en- catheter (Biosense Webster, Inc., Diamond Bar, CA, USA) trances. Outcomes of the “scar dechanneling” technique for mapping and ablation. Bipolar electrograms were filtered depend on the correct identification and elimination of all from 30 to 250 Hz. +e 12-lead surface electrocardiogram present SCCs [9]. +is can be a time-consuming and skill- (ECG) and EGM signals from the mapping catheter were demanding task, being subject to significant interoperator displayed and stored for prospective analysis. Endocardial EAM maps were acquired for all patients belonging to the variability. We hypothesize that an automatic system able to identify EGM-DC within the substrate could simplify and ischemic subpopulation and in seven of the patients from the ARVD/C subpopulation, the rest of EAM maps of the standardize VA ablation procedures. In this study, we present and evaluate the performance ARVD/C subpopulation were obtained from the epicardium. of a novel algorithm for automatic EGM analysis so called “Slow Conducting Channel Mapping” algorithm, or Ablation was performed under conscious sedation or “SCC-Mapping.” +is algorithm dichotomizes normal general anaesthesia when epicardial access was required or from abnormal bipolar EGMs, automatically identifying anticipated. Bipolar voltage maps were obtained during SR the presence of EGM-DC within the substrate. By mea- and scar was identified using standard voltage thresholds defining scar core zone (CZ) (<0.5 mV), border zone (BZ) suring the bipolar voltage belonging to the local-field component, the SCC-Mapping algorithm may obtain (<1.5 mV), and healthy tissue (≥1.5 mV). +e “scar dechanneling” ablation technique was used for identification more accurate bipolar voltage maps. +us, a better scar characterization could help to guide scar-related VA and ablation of SCC entrances, thus isolating the VA isthmuses [9]. Identification of SCC was performed man- ablation procedures. ually by the EAM navigation system operator. After ablation of all the SCC entrances, a remap procedure was performed 2. Methods to detect any residual SCC and ablate them if needed. When finished, a programmed stimulation protocol was per- 2.1. Patient Sample. Twenty patients (fifteen males) with VA formed, remapping again the substrate in case any sustained who underwent catheter-based radiofrequency ablation VAs were found inducible, until noninducibility was were included in the study. Ten patients (nine males) had achieved. Ablation was performed in temperature-con- ischemic cardiomyopathy. Ten (six males) fulfilled Task trolled mode with 45 C temperature and 50 W power limit at Force criteria for arrhythmogenic right ventricular dyspla- 26 mL/min irrigation rate (40 W and 17 mL/min at sia/cardiomyopathy (ARVD/C). Ischemic patients were epicardium). selected from our database of VA substrate ablation as consecutive patients having preprocedural contrast-en- hanced cardiac magnetic resonance (Ce-CMR) imaging 2.3. Ce-CMR Acquisition and Processing. A preprocedural study. Basal characteristics of the population are summa- late gadolinium Ce-CMR was acquired in all ischemic cases rized in Table 1. +e study complied with the Declaration of and used to localize the arrhythmogenic substrate [12]. +e Helsinki, and the local ethics committee approved the study preprocedural Ce-CMR studies were obtained using a 3T protocol. All participants included in the study provided scanner (MAGNETOM Trio , Siemens Healthcare, ® ® informed written consent. Erlangen, Germany). Contrast-enhanced images were Journal of Interventional Cardiology 3 acquired 10 minutes after bolus injection of 0.2 mmol/kg distance between the main deflection of far- and local-field Gadobutrol (Gadovist , Bayer Hispania, Barcelona, Spain) component (i.e., the first and the second component) di- chotomized between d-EGMs and f-EGMs (see Figure 1(a)). using a commercially available, free-breathing, ECG-gated, navigator-gated, 3D inversion-recovery, gradient-echo +e cut-off threshold was set to 25 ms as a trade-off for good technique. identification of f-EGM and d-EGM signals. f-EGMs were Ce-CMR images were analysed as previously described ablation targets according to the “scar dechanneling” [12]. Briefly, a full left ventricular (LV) volume was technique, as they constitute the typical pattern at SCC reconstructed in the axial orientation, and the resulting entrances [9]. images were processed with the commercially available +e outcome of the algorithm is the identification label ADAS-3D software (Galgo Medical, Barcelona, Spain). on the type of EGM assigned to each mapping point. +ese Ten concentric surface layers (from 10% to 90%) were labels were then colour-coded and integrated into the 3D created automatically from endo- to epicardium of the LV EAM with the following criteria: small white spheres for wall thickness. A 3D shell was obtained for each layer. normal EGM mapping points; blue big spheres for d-EGM mapping points, and black big spheres for f-EGMs mapping Pixel signal intensity (PSI) maps based on Ce-CMR im- ages were projected to each shell, following a trilinear points. interpolation algorithm, and colour-coded. To identify the scar areas, a PSI-based algorithm was applied to char- 2.5. Construction of SCC-Maps. +e SCC-Mapping algo- acterize the hyperenhanced area as CZ, BZ or healthy rithm identifies the existence of potential EGM-DC, tissue using 40%± 5% and 60%± 5% of the maximum allowing to measure the bipolar amplitude of far- and intensity as thresholds [12]. +e BZ channels (i.e., SCCs) local-field components individually. +e projection of were defined by the ADAS-3D software as continuous these voltages on a 3D SCC-Map was performed using an 3D corridors (across all the Ce-CMR layers) of BZ (with additional decision tree (Figure 1(b)) with two branches: the specified PSI threshold) surrounded by scar core/ one for single-component EGMs and another for EGM- mitral annulus [12]. DC (either d-EGMs or f-EGMs). +e single-component EGM branch considered any mapping point <1.5 mV as far-field remote signals measured within the dense scar 2.4. @e “Slow Conducting Channel Mapping” (“SCC-Map- coming from the surrounding healthy myocardium, ping”) Algorithm. +e SCC-Mapping algorithm is based on automatically setting its bipolar voltage to zero. +e an EGM detector and delineator algorithm previously de- double-component EGM branch projects on the 3D SCC- veloped by our team [13]. +is detector automatically Map the bipolar voltage of the local-field component only identifies and delineates the onset and end landmarks of the if it is included within a window of interest. +is window bipolar EGM signal using the QRS complex of the 12-lead of interest was defined between the 5th and 95th per- surface ECG as the reference searching window, a method centiles of the onsets and ends of all identified local-field that has been already validated for activation mapping of components, respectively. If the local-field component of focal VA [13, 14]. +e entire processing algorithm was an EGM-DC mapping point did not meet this criterion, implemented in MATLAB (MATLAB R2016a, Math- then the same criterion as for single-component EGM Works, Inc., Natick, MA, USA). +e results were obtained mapping points was applied. Additionally, the algorithm offline; therefore, ablation outcomes were independent of includes a spatial coherence protection. +is protection the presented results. checks, for close EGM-DC mapping points (dis- Starting from an initial delineation of the mapping point tance< 6 mm), if the local-field components are similar in EGM signal using our EGM detector/delineator algorithm activation time and shape. When these criteria are met, [13], the SCC-Mapping algorithm uses a decision tree il- only the highest bipolar value is represented in the 3D lustrated in Figure 1(a). +is decision tree is based on two SCC-Map. main characteristics of the bipolar EGM signal: the delin- eated length and the bipolar voltage. For short-duration EGMs (<65 ms, based on [15]), 2.6. SCC Detection Agreement Evaluation. In this study, normal mapping points are distinguished from those can- three mapping modalities were considered: (i) EAM voltage didates to be an EGM-DC by the measured bipolar voltage. maps, (ii) SCC-Maps derived from the SCC-Mapping Al- +erefore, those mapping points showing a bipolar voltage gorithm; and (iii) Ce-CMR PSI maps (when available). An ≥3.5 mV were considered normal EGMs [15], whereas the expert operator visually evaluated the ability of each map- rest were candidates for being classified as EGM-DC or ping modality to identify SCC entrances from both the 3D remained as normal EGMs. Long-duration EGMs (>65 ms) coloured map and the acquired EGM signals. were always considered potential candidates for EGM-DC, regardless of their voltage value. In order to label an EGM-DC candidate with true 2.6.1. Identification of SCC Entrances Derived from the 3D delayed (d-EGM) or fused (f-EGM) components, the al- Coloured Map. For the EAM voltage maps, SCC entrance gorithm searched for the existence of a second EGM identification was performed using (1) the standard component based on the EGM detector/delineator algo- thresholds that define the presence of scar CZ and BZ rithm [13]. If a second EGM component was found, the time tissue (named “EAM standard”); and (2) using a manual 4 Journal of Interventional Cardiology Delineated EGM length ≤65 ms Yes No No Bipolar voltage ≥3.5 mV? Searching process: Yes Is there a second component? Yes No Distance 1st-to-2nd component ≤25 ms? Yes No Delayed Normal Fused EGM EGM EGM (a) Fused Normal Delayed EGM EGM EGM 2nd component within the Bipolar voltage <1.5 mV? window of interest? No Yes No Yes V = V V = V V = 0 SCC-MAP Far-Field SCC-MAP Local-Field SCC-MAP (b) Figure 1: +e “Slow Conducting Channel Mapping Algorithm.” (a) Decision tree for electrogram (EGM) signals with delayed components (EGM-DC) searching protocol and (b) algorithm for the reconstruction of “Slow Conducting Channel Maps” (SCC-Maps) on patient’s 3D anatomical map. VSCC-Map: bipolar voltage projected on patient’s 3D anatomical map. voltage screening process that dynamically modifies the identified (labelled) mapping points close to the BZ area. standard thresholds for BZ and CZ definition in order to +erefore, this evaluation could only be done for EAM voltage maps and SCC-Maps. enhance the presence of SCCs (named “EAM screening”) [3, 16]. For the SCC-Maps, this process was done directly from the coloured map. It should be noted that, due to the more precise bipolar voltage measurement method of 2.7. Statistical Analysis. Continuous data are shown in SCC-Maps, the CZ tissue threshold was set to ≤0.1 mV, mean± standard deviation, unless otherwise indicated. yielding a higher range of bipolar voltage measurements Categorical data are shown as percentages. Comparison (shown and discussed in the following sections). +ere- between different populations was given by the Wilcox- fore, the voltage screening process was not necessary for on–Mann–Whitney test or by the Fisher exact test when SCC-Maps. For the Ce-CMR PSI maps, the threshold appropriate. For evaluation of the agreement in the SCC definitions for tissue heterogeneity identification (i.e., BZ entrances identification among the different mapping mo- tissue that conforms SCCs) were those explained in the dalities, Wilcoxon–Mann–Whitney test, Lin’s concordance corresponding section. correlation factor “ρ” [17], and Bland–Altman plot analysis [18] were used. A p-value of ≤0.05 was considered as a cut- 2.6.2. Identification of SCC Entrances Derived from the off value for statistical significance. Statistics were obtained Analysed EGM Signals. +is process was done by manual using the MATLAB statistics toolbox (MATLAB R2016a, inspection of the presence or absence of f-EGMs in the MathWorks, Inc., Natick, MA, USA). Journal of Interventional Cardiology 5 in Table 2. However, no significant differences were found 3. Results between EAM screening and SCC-Maps versus Ce-CMR 3.1. Population Characteristics. Twenty patients were in- PSI maps (p � 0.202 and p � 1.0, resp.). Nevertheless, cluded in the study. 75% were male, and mean age was Bland–Altman plot analysis reveals a tendency of these 57± 15 years. Mean LV ejection fraction was 44± 16%, with three mapping modalities towards an underestimation of no significant differences between ischemic patients and the number of SCC entrances compared with Ce-CMR PSI those with ARVD/C. Table 1 summarizes the baseline maps (Pearson’s R � 0.63, p � 0.049; R � 0.85, p � 0.002; characteristics of the study population. and R � 0.61, p � 0.05 for the comparisons between SCC detection on SCC-Maps, EAM standard, and EAM screening against Ce-CMR PSI maps, resp., Figure 2(c)). 3.2. SCC Detection Agreement between EAM and SCC- Moreover, Supplementary Table 1 confirms the high Mapping. Table 2 and Figure 2(a) show the agreement in agreement of EAM screening and SCC-maps with Ce-CMR the number of SCC entrances identified from the 3D PSI maps compared with EAM standard (ρ � 0.679 and coloured maps between all the studied mapping modal- ρ � 0.628, resp., vs. ρ � 0.212, p< 0.01). Figure 4 shows two ities. EAM voltage maps with standard thresholds (“EAM examples where SCC-Maps had higher agreement with Ce- standard”) presented a significant lower number of SCC CMR PSI maps in identifying SCC, as compared to EAM entrances than EAM voltage maps with manual voltage standard maps. screening (“EAM screening”) (p< 0.01, 0.04, and 0.03 for the entire population, ischemic, and ARVD/C, resp.). Additionally, EAM standard maps also had less SCC 3.4. Effect of Selective Bipolar Voltage Measurement. +e entrances than SCC-Maps (p< 0.01, < 0.01, and 0.02 for SCC-Mapping algorithm was capable of detecting EGM- the entire population, ischemic, and ARVD/C, resp.). DC providing the bipolar voltage of the local component, However, there were no significant differences in the thus obtaining accurate voltage maps (SCC-Maps). +is number of identified SCC entrances between EAM more selective approach enlarged the voltage range dis- screening maps and SCC-Maps (p � 0.29, 0.10, and 0.87 played and thus improved the degree of detail of the for the entire population, ischemic, and ARVD/C, resp.). arrhythmogenic substrate when compared with standard Lin’s concordance correlation factor analysis shown in EAM voltage maps. Figure 5(b) illustrates the loss of scar Supplementary Table 1 supports these findings, showing a details in EAM voltage maps when compared to SCC- higher concordance between EAM screening and SCC- Maps, which display a higher range of voltage measure- Maps (ρ � 0.665, 0.528, and 0.877 for the entire pop- ments. Additionally, as depicted in Figure 4, SCC-Maps ulation, ischemic, and ARVD/C, resp.) than EAM stan- matched better with information obtained from Ce-CMR dard maps. PSI maps in the ischemic population than EAM standard +e Bland–Altman analysis shown in Figure 2(a) illus- maps. trates the agreement in SCC entrance identification among the different studied mapping modalities and populations. 4. Discussion +ere was a low bias in the number of SCC entrances identified between SCC-Maps and EAM screening maps, 4.1. Reviewing Current Bipolar Voltage Mapping. EAM with a small trend towards a SCC subidentification of SCC- systems are useful tools to map scar-related VAs, since Maps compared with EAM screening maps (Pearson’s they allow calculating the peak-to-peak local EGM signal R � 0.48, p � 0.033), which is in concordance with the amplitude and representing this value, colour-coded, on findings of Table 2. the cardiac anatomy, thus helping to identify and char- Table 3 and Figure 2(b) list the agreement in the number acterize the scar [3, 5, 6, 8–10, 19, 20]. However, myo- of SCC entrances identified from the f-EGM points between cardial scars are often surrounded by a considerable mapping modalities. No significant differences were found amount of healthy tissue, which may lead to local EGMs between EAM standard maps and SCC-Maps, confirmed by being masked by the presence of far-field signals. the Bland–Altman analysis shown in Figure 2(b) and the +erefore, regular voltage mapping with standard high correlation shown in Supplementary Table 2 (ρ � 0.918, thresholds may underestimate of the scar size and lose 0.871, and 0.936 for the entire population, ischemic, and significant scar details. +is increases the likelihood of ARVD/C, resp.). missing SCCs within the substrate and the need of ex- Figure 3 illustrates two examples of the electrical tensive operator analysis. propagation sequences along SCCs identified from the au- Examples of this phenomenon are illustrated in Figure 6. tomatically labelled mapping points on SCC-Maps com- In normal EGMs, the voltage map reflects the peak-to-peak pared to the SCC manual identification performed on EAM voltage (a); however, when healthy myocardium EGM (i.e., maps where the identification of these SCCs requires ex- far-field) has a higher amplitude than the late potential (i.e., tensive operator analysis. local-field), the voltage map reflects the far-field, high- amplitude component voltage (b). Moreover, when the far- 3.3. SCC Detection Agreement with Ce-CMR. +e number of and local-field components show comparable amplitudes, SCC entrances identified in EAM standard was signifi- voltage map may reflect the peak-to-peak amplitude of either cantly lower compared with Ce-CMR PSI maps, as shown the local or the far-field component (c), or a mix of both (d). 6 Journal of Interventional Cardiology Table 2: Analysis of colour-coded maps. Number of SCC entrances identified per patient and agreement between mapping modalities. ∗ † ‡ EAM standard EAM screening SCC-Map Ce-CMR PSI maps p-value p-value p-value Entire population (n � 20) 1.05± 1.10 2.95± 2.31 3.45± 1.61 N/A <0.01 <0.01 0.29 Ischemic (n � 10) 0.60± 1.00 2.20± 1.75 3.60± 1.43 3.70± 2.45 0.04 <0.01 0.10 ARVD/C (n � 10) 1.50± 1.08 3.70± 2.63 3.30± 1.83 N/A 0.03 0.02 0.87 ∗ † Number of SCC entrances per patient are given as mean± standard deviation. Differences between EAM standard and EAM screening. Differences between EAM standard and SCC-Maps. Differences between EAM screening and SCC-Maps. ARVD/C: arrhythmogenic right ventricular dysplasia/cardiomyopathy; Ce-CMR: contrast-enhanced cardiac magnetic resonance; EAM: electroanatomical mapping; N/A: not applicable; PSI: pixel signal intensity; and SCC: slow conducting channel. EAM standard vs. SCC-maps EAM screening vs. SCC-maps EAM standard vs. EAM screening 5 5 5 0 0 0 –5 –5 –5 0 2468 0 2468 0 2468 (EAM standard + SCC-maps)/2 (EAM screening + SCC-maps)/2 (EAM standard + EAM screening)/2 Ischemic ARVD/C (a) EAM standard f-EGMs vs. SCC-maps f-EGMs –5 0 510 15 (EAM standard + SCC-maps)/2 Ischemic ARVD/C (b) Figure 2: Continued. EAM standard-SCC-maps EAM screening-SCC-maps EAM standard-SCC-maps EAM standard-EAM screening Journal of Interventional Cardiology 7 Ce-CMR PSI maps vs. SCC-maps Ce-CMR PSI maps vs. EAM standard Ce-CMR PSI maps vs. EAM screening 5 5 5 0 0 0 –5 –5 –5 0 246802 2468 0 468 (Ce-CMR PSI maps + (Ce-CMR PSI maps + (Ce-CMR PSI maps + SCC-maps)/2 EAM standard)/2 EAM screening)/2 Ischemic (c) Figure 2: Bland–Altman plots for assessing the agreement in the identification of slow conducting channel (SCC) entrances (a) from the colour-coded 3D maps between the different mapping modalities: electroanatomical mapping (EAM) system maps with standard voltage thresholds (EAM standard), EAM maps with voltage screening (EAM screening), and “Slow Conducting Channel Maps” (SCC-Maps). (b) From the analysis of the presence of fused electrograms (f-EGM) components between EAM standard maps and SCC-Maps and (c) from the colour-coded 3D map between the different mapping modalities and the pixel signal intensity (PSI) maps derived from contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging in the ischemic population. Red solid line indicates mean and red dashed lines indicate mean± 2 standard deviations of the difference in the number of identified SCC entrances. ARVDC: arrhythmogenic right ventricular dysplasia/cardiomyopathy. Table 3: Analysis of EGM-DC and identification of f-EGMs. Number of SCC entrances identified per patient and agreement between mapping modalities. EAM maps SCC-Map p-value Entire population (n � 20) 6.10± 2. 81 5.35± 2.70 0.430 Ischemic (n � 10) 5.50± 2.17 4.70± 2.11 0.422 ARVD/C (n � 10) 6.70± 3.34 6.00± 3.16 0.790 Number of SCC entrances per patient are given as mean± standard deviation. ARVD/C: arrhythmogenic right ventricular dysplasia/cardiomyopathy; EAM: electroanatomical mapping; EGM-DC: electrograms with delayed components; f-EGM: fused electrograms; and SCC: slow conducting channel. A1 A2 Channel 1 Channel 2 V2 V2 1 1 2 2 3 3 4 9 7 12 0.5 0.1 1.5 1.5 EAM voltage map (mV) SCC-map (mV) (a) Figure 3: Continued. Ce-CMR PSI maps-SCC-maps Ce-CMR PSI maps-EAM standard Ce-CMR PSI maps-EAM screening 8 Journal of Interventional Cardiology B1 B2 Channel 1 Channel 2 V3 V3 1 1 4 3 0.5 0.1 10 1.5 1.5 EAM voltage map (mV) SCC-map (mV) (b) Figure 3: Examples of slow conducting channel (SCC) identification from the automatic mapping point labelling on “Slow Conducting Channel Maps” (SCC-Maps). (a) Endocardial electroanatomical map (EAM) of an ischemic patient showing two SCCs identified on SCC- Map. (b) Epicardial EAM from an arrhythmogenic right ventricular dysplasia/cardiomyopathy patient showing two SCCs identified on SCC-Map. A1 A2 A3 AV AV AV 0.5 1.5 0.1 1.5 60% 40% EAM voltage map (mV) SCC-map (mV) PSI map 10% wall layer (a) Figure 4: Continued. Journal of Interventional Cardiology 9 B1 B2 B3 0.5 1.5 0.1 1.5 60% 40% EAM voltage map (mV) SCC-map (mV) PSI map 10% wall layer (b) Figure 4: Agreement between electroanatomical mapping (EAM) voltage maps and “Slow Conducting Channel Maps” (SCC-Maps) against pixel signal intensity (PSI) maps derived from contrast-enhanced cardiac magnetic resonance (Ce-CMR) imaging. A1 and B1 show the electroanatomical mapping (EAM) voltage maps obtained with the EAM system from two different patients. A2 and B2 show the cor- responding SCC-Map and A3 and B3 show the acquired Ce-CMR PSI map. AV: aortic valve. 0.1 1.5 0.5 1.5 0.1 1.5 EAM voltage map (mV) EAM voltage map SCC-map (mV) modified th. (mV) (a) (b) (c) Figure 5: Endocardial substrate map from patient with myocardial infarction. (c) illustrates the richest scar details shown by the “Slow Conducting Channel Map” (SCC-Map) compared with electroanatomical mapping (EAM) voltage maps using the standard voltage thresholds (a) and using modified voltage thresholds (b). 10 Journal of Interventional Cardiology V6 V6 Normal EGM Local-field masked by healthy myocardium far-field 4 1 d − 2 d − 2 V = 0.53 mV V = 1.99 mV Local EAM V = 6.19 mV –1 EAM –2 –4 –2 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Time (ms) Time (ms) (a) (b) V6 V6 Local-field masked by comparable far-field Voltage measurement from both components 0.2 0.4 0.1 0.2 V = 0.49 mV d − 2 EAM V = 0.41 mV Local d − 2 V = 0.24 mV Local –0.1 –0.2 V = 0.33 mV EAM –0.2 –0.4 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 Time (ms) Time (ms) (c) (d) Figure 6: Examples of (a) normal electrogram (EGM) bipolar voltage measurement by the electroanatomical mapping (EAM) system and (b–d) different bipolar EGM-DC signals with incorrect bipolar voltage measurement by the EAM system: (b) local-field component masked by high-amplitude far-field component. (c-d) Comparable amplitude of far-field and local-field components. +ese examples show the need of a more selective approach 4.3. SCC Detection and Agreement between Mapping to measure bipolar voltage for substrate mapping of scar- Modalities. SCC-Maps were highly correlated with EAM related VAs. standard maps when these were obtained after a manual voltage screening process (EAM screening), whereas raw EAM standard maps (EAM standard) correlated worse and identified a significant lower number of SCC en- 4.2. Main Findings. +e present study evaluates a novel trances. +e algorithm’s accuracy is illustrated by the fact automatic EGM signal analysis algorithm aiming to im- that SCC-Maps can display a detailed scar without the prove the accuracy of current voltage mapping obtained need of manual voltage screening, with good agreement with EAM systems. +is algorithm allows obtaining voltage with EAM after manual labelling of the mapping points maps with higher voltage range, hence depicting more corresponding to SCC entrances (i.e., those showing detailed scar characteristics, which may be useful to f-EGM signals). identify VA isthmuses during ablation procedures. +e Figure 3 shows ischemic and ARVD/C patient ex- main findings of the study are as follows: (1) the proposed amples where two different SCCs can be found. Both SCC-Mapping algorithm automatically identified SCC examples illustrate the superiority of SCC-Maps over entrances at the same level as manual EAM voltage EAM standard maps, allowing the identification of those screening; (2) the SCC-Mapping algorithm provided SCC- signal corridors by direct inspection of the colour map. Maps that match Ce-CMR PSI maps better than current +e EAM standard map of Figure 3(a) (panel A1) depicts a EAM voltage maps; and (3) the SCC-Mapping algorithm dense scar in the area where a second SCC can be found improves the definition of the scar CZ and BZ areas by following the activation sequence of the d-EGMs. +is allowing a higher voltage range. SCC can be easily identified by the colour map and Amplitude (mV) Amplitude (mV) Amplitude (mV) Amplitude (mV) Journal of Interventional Cardiology 11 devices in the ARVD/C population. On the other hand, automatic labelling in Figure 3(a) (panel A2). Similarly, the EAM standard map of Figure 3(b) (panel B1) does not the number of available patients with detailed EAM and quality Ce-CMR data has reduced lately, as detailed EAM allow to identify those SCCs that can be seen in the SCC- Map of Figure 3(b) (panel B2). Moreover, SCC-Maps and acquisition is a time-consuming and highly operator- EAM-screening correlate better than EAM-standard maps dependent task, while Ce-CMR-guided catheter substrate with Ce-CMR PSI maps. +ese examples illustrate the ablation has gained more interest [26]. Despite this trend, need of manual analysis of the EGM signals by the system the endpoints for determining the ablation targets are still operator in order to identify all the possible SCCs present based on EAM findings where the presented algorithm can in the substrate (i.e., using techniques like manual voltage play an important and complementary role. Also, this screening and individual EGM labelling). +is procedure algorithm paves the way for better integration of Ce-CMR can be guided and shortened by the proposed automatic and EAM data to improve scar-related VT ablation procedures. SCC-Mapping algorithm. Moreover, automatic and ob- jective identification becomes mandatory when using the Additionally, the algorithm was designed and tested using data from substrate-based VA ablation during SR. increasingly popular multielectrode mapping (MEM) catheters where tenths of simultaneous signals per beat Hence, no data from VA mapping was used or analysed with can be acquired. this algorithm and therefore, other possible VA isthmuses were not explored with this algorithm [27]. In this work, EAMs were acquired using a standard 4.4. Voltage @resholds for Scar Definition. +e proposed 3.5 mm irrigated-tip mapping catheter, which has a SCC-mapping algorithm provides a more precise quanti- longer interelectrode distance as compared with high- fication of the local-field voltage. +is aspect allows density MEM catheters that can better discriminate lo- changing the threshold definition for CZ tissue without cal- from far-field components [21]. However, manual losing scar information, thus improving the ability to annotation of multiple simultaneous signals obtained detect of SCC (Figure 4). +is effect is comparable to the with MEM is a nonaffordable task, for which an auto- one obtained with current MEM catheters [21], but using matic approach (like the SCC-Mapping algorithm) be- a regular electrode-size catheter in conjunction with an comes necessary. automatic algorithm to distinguish the far- and local-field components of the measured EGMs. In contrast, MEM 5. Conclusions catheters still need extensive operator analysis in order to identify and/or enhance the presence of SCCs. +is fact +e proposed automatic analysis of EGM signals using the could be mitigated if high-density mapping is combined “Slow Conducting Channel Mapping Algorithm” improves with an automatic algorithm as the presented in this the accuracy of bipolar voltage measurements within the scar work. area, achieving a more detailed tissue characterization and Figure 5 illustrates the loss of scar definition when the being an operator-independent tool for accurate identifi- modified voltage threshold for BZ and CZ tissue is used on cation of SCCs. +is last feature encourages the use of the EAM voltage maps. +e higher voltage range displayed by algorithm together with EAM navigation systems as a re- SCC-Maps also facilitates the voltage screening process producible approach for guiding VA ablation procedures in for SCC identification. However, although SCC-Maps and daily practice. EAM screening maps provide similar insights, the former were obtained without any manual intervention, thus Data Availability being operator independent. Additionally, as shown in Figures 3–5, it can be observed that SCC-Maps provided a +e electroanatomical mapping and image data used to better defined scar delineation than current EAM voltage support the findings of this study are restricted by the mapping. ´ Hospital Clınic Local Ethical Committee in order to protect +e proposed algorithm was evaluated when using the patient privacy. Data are available from Dr. Llu´ıs Mont, “scar dechanneling” ablation technique. However, the fact PhD, Arrhythmias Unit, Hospital Clinic, Carrer de Villar- that a higher voltage range can be described with SCC- roel, 170, 08036 Barcelona, Spain, for researchers who meet Maps allows a more detailed scar characterization, which the criteria for access to confidential data. suggests that the algorithm can also be useful for other ablation approaches [5, 22–25]. Eventually, the SCC- Conflicts of Interest Mapping algorithm could effectively improve the guid- Dr. A. Berruezo and Dr. L. Mont are stockholders in Galgo ance of pacing/entrainment manoeuvres for VA isthmus Medical SL. David Soto-Iglesias is an employee of Biosense identification based on the tagged data and SCC-Map Webster, Inc. +e authors declare that there are no conflicts information. of interest. 4.5. Study Limitations. +e main limitation of this study Acknowledgments was the relatively small sample size. Comparison of EAM maps against Ce-CMR PSI maps was only possible in the +is study was supported by personal grants to A. Alcaine ischemic population due to the presence of implantable (Refs. BES-2011-046644 and EEBB-I-15-09466); by 12 Journal of Interventional Cardiology pp. 301–309, John Wiley & Sons, Hoboken, NJ, USA, 3rd projects PID2019-104881RB-I00 from Ministerio de edition, 2009. Ciencia e Innovacio´n (Spain) and DPI2016-75458-R from [8] A. Berruezo, J. Fernandez-Armenta, ´ L. Mont et al., “Com- Ministerio de Economıa y Competitividad (Spain); and by bined endocardial and epicardial catheter ablation in Gobierno de Arago´n (Grupo Referencia BSICoS ref.: arrhythmogenic right ventricular dysplasia incorporating scar T39_20R) cofounded by FEDER 2014–2020. +is work was dechanneling technique,” Circulation: Arrhythmia and Elec- also supported in part by the project PI14/00759, integrated trophysiology, vol. 5, no. 1, pp. 111–121, 2012. in the Plan Nacional de I+D+i and cofounded by the [9] A. Berruezo, J. Fernandez-Armenta, ´ D. Andreu et al., “Scar Instituto de Salud Carlos III (ISCIII)-Subdireccio´n General dechanneling: new method for scar-related left ventricular de Evaluacio´n and European Regional Development Fund tachycardia substrate ablation,” Circulation: Arrhythmia and (European Union). +e computation was performed by the Electrophysiology, vol. 8, no. 2, pp. 326–336, 2015. ICTS NANBIOSIS, more specifically by the High Perfor- [10] Z. F. Issa, J. M. Miller, and D. P. Zipes, Clinical Arrhythmology mance Computing Unit of the CIBER in Bioengineering, and Electrophysiology: A Companion to Braunwald’s Heart Disease, Saunders, Philadelphia, PA, USA, 2nd edition, 2012. Biomaterials and Nanomedicine (CIBER-BBN) at the [11] J. Fernandez-Armenta, ´ D. Andreu, D. Penela et al., “Sinus University of Zaragoza. +e CIBER-BBN is an initiative of rhythm detection of conducting channels and ventricular Instituto de Salud Carlos III. tachycardia isthmus in arrhythmogenic right ventricular cardiomyopathy,” Heart Rhythm, vol. 11, no. 5, pp. 747–754, Supplementary Materials ´ ´ [12] D. Andreu, J. T. Ortiz-Perez, J. Fernandez-Armenta et al., “3D Supplementary Table 1: Lin’s concordance correlation factor delayed-enhanced magnetic resonance sequences improve on the number of SCC entrances identified between map- conducting channel delineation prior to ventricular tachy- ping modalities (rows against columns) from the analysis of cardia ablation,” EP Europace, vol. 17, no. 6, pp. 938–945, the colour-coded maps. Supplementary Table 2: Lin’s con- cordance correlation factor on the number of SCC entrances [13] A. Alcaine, D. Soto-Iglesias, M. Calvo et al., “A wavelet-based identified between EAM standard maps and SCC-Maps electrogram onset delineator for automatic ventricular acti- vation mapping,” IEEE Transactions on Biomedical Engi- from the analysis of EGM-DC and identification of f-EGMs. neering, vol. 61, no. 12, pp. 2830–2839, 2014. (Supplementary Materials) [14] A. Alcaine, D. Soto-Iglesias, J. Acosta et al., “Automatic ac- tivation mapping and origin identification of idiopathic References outflow tract ventricular arrhythmias,” Journal of Electro- cardiology, vol. 51, no. 2, pp. 239–246, 2018. [1] D. J. Callans, J.-F. Ren, J. Michele, F. E. Marchlinski, and [15] D. M. Cassidy, J. A. Vassallo, F. E. Marchlinski, A. E. Buxton, S. M. Dillon, “Electroanatomic left ventricular mapping in the W. J. Untereker, and M. E. Josephson, “Endocardial mapping porcine model of healed anterior myocardial infarction. in humans in sinus rhythm with normal left ventricles: ac- Correlation with intracardiac echocardiography and patho- tivation patterns and characteristics of electrograms,” Cir- logical analysis,” Circulation, vol. 100, no. 16, pp. 1744–1750, culation, vol. 70, no. 1, pp. 37–42, 1984. [16] S. E. Mountantonakis, R. E. Park, D. S. Frankel et al., “Re- [2] V. Y. Reddy, D. Wrobleski, C. Houghtaling, M. E. Josephson, lationship between voltage map “channels” and the location of and J. N. Ruskin, “Combined epicardial and endocardial critical isthmus sites in patients with post-infarction car- electroanatomic mapping in a porcine model of healed diomyopathy and ventricular tachycardia,” Journal of the myocardial infarction,” Circulation, vol. 107, no. 25, American College of Cardiology, vol. 61, no. 20, pp. 2088–2095, pp. 3236–3242, 2003. [3] A. Arenal, S. del Castillo, E. Gonzalez-Torrecilla et al., [17] L. Lin, A. S. Hedayat, B. Sinha, and M. Yang, “Statistical “Tachycardia-related channel in the scar tissue in patients methods in assessing agreement: models, issues, and tools,” with sustained monomorphic ventricular tachycardias: in- Journal of the American Statistical Association, vol. 97, fluence of the voltage scar definition,” Circulation, vol. 110, no. 457, pp. 257–270, 2002. no. 17, pp. 2568–2574, 2004. [18] D. G. Altman and J. M. Bland, “Measurement in medicine: the [4] J. M. de Bakker, F. J. van Capelle, M. J. Janse et al., “Reentry as analysis of method comparison studies,” @e Statistician, a cause of ventricular tachycardia in patients with chronic vol. 32, no. 3, pp. 307–317, 1983. ischemic heart disease: electrophysiologic and anatomic [19] F. E. Marchlinski, D. J. Callans, C. D. Gottlieb, and E. Zado, correlation,” Circulation, vol. 77, no. 3, pp. 589–606, 1988. “Linear ablation lesions for control of unmappable ventricular [5] A. Arenal, E. Glez-Torrecilla, M. Ortiz et al., “Ablation of tachycardia in patients with ischemic and nonischemic car- electrograms with an isolated, delayed component as treat- diomyopathy,” Circulation, vol. 101, no. 11, pp. 1288–1296, ment of unmappable monomorphic ventricular tachycardias in patients with structural heart disease,” Journal of the American College of Cardiology, vol. 41, no. 1, pp. 81–92, 2003. [20] J. Fernandez-Armenta, A. Berruezo, D. Andreu et al., “+ree- dimensional architecture of scar and conducting channels [6] E. M. Cronin, S. M. Al-Khatib, E. Anter et al., “2019 HRS/ EHRA/APHRS/LAHRS expert consensus statement on based on high resolution ce-CMR: insights for ventricular tachycardia ablation,” Circulation: Arrhythmia and Electro- catheter ablation of ventricular arrhythmias,” Europace, vol. 8, pp. 1–147, 2019. physiology, vol. 6, no. 3, pp. 528–537, 2013. [21] J. Acosta, D. Penela, D. Andreu et al., “Multielectrode vs. [7] J. V. Wylie, T. W. Smith, and M. E. Josephson, “Substrate mapping for ablation of ventricular tachycardia in coronary point-by-point mapping for ventricular tachycardia substrate ablation: a randomized study,” EP Europace, vol. 20, no. 3, artery disease,” in Cardiac Mapping, M. Shenasa, G. Hindricks, M. Borggrefe, and G. Breithardt, Eds., pp. 512–519, 2018. Journal of Interventional Cardiology 13 [22] L. Di Biase, P. Santangeli, D. J. Burkhardt et al., “Endo-epi- cardial homogenization of the scar versus limited substrate ablation for the treatment of electrical storms in patients with ischemic cardiomyopathy,” Journal of the American College of Cardiology, vol. 60, no. 2, pp. 132–141, 2012. [23] W. S. Tzou, D. S. Frankel, T. Hegeman et al., “Core isolation of critical arrhythmia elements for treatment of multiple scar- based ventricular tachycardias,” Circulation: Arrhythmia and Electrophysiology, vol. 8, no. 2, pp. 353–361, 2015. [24] F. Bogun, E. Good, S. Reich et al., “Isolated potentials during sinus rhythm and pace-mapping within scars as guides for ablation of post-infarction ventricular tachycardia,” Journal of the American College of Cardiology, vol. 47, no. 10, pp. 2013–2019, 2006. [25] P. Jaıs, P. Maury, P. Khairy et al., “Elimination of local ab- normal ventricular activities : a new end point for substrate modification in patients with scar-related ventricular tachy- cardia,” Circulation, vol. 125, no. 18, pp. 2184–2196, 2012. [26] C. Blomstrom-Lundqvist, ¨ A. Auricchio, J. Brugada et al., “+e use of imaging for electrophysiological and devices proce- dures: a report from the first European heart rhythm asso- ciation policy conference, jointly organized with the European association of cardiovascular imaging (EACVI), the council of cardiovascular imaging and the European society of cardiac radiology,” Europace, vol. 15, no. 7, pp. 927–936, 2013. [27] R. Martin, P. Maury, C. Bisceglia et al., “Characteristics of scar-related ventricular tachycardia circuits using ultra-high- density mapping,” Circulation: Arrhythmia and Electrophys- iology, vol. 11, no. 10, Article ID e006569, 2018.

Journal

Journal of Interventional CardiologyHindawi Publishing Corporation

Published: May 29, 2020

There are no references for this article.