Application of Ripple Mapping with an Electroanatomic Mapping System for Diagnosis of Atrial Tachycardias

Transcription

1 1361 Application of Ripple Mapping with an Electroanatomic Mapping System for Diagnosis of Atrial Tachycardias SHAHNAZ JAMIL-COPLEY, M.R.C.P., NICK LINTON, M.R.C.P., MICHAEL KOA-WING, Ph.D., M.R.C.P., PIPIN KOJODJOJO, Ph.D., M.R.C.P., PHANG BOON LIM, Ph.D., M.A., M.R.C.P., LOUISA MALCOLME-LAWES, B.Sc., M.R.C.P., ZACHARY WHINNETT, Ph.D., M.R.C.P., IAN WRIGHT, B.Sc., WYN DAVIES, M.D., F.R.C.P., NICHOLAS PETERS, M.D., F.R.C.P., F.H.R.S., DARREL P. FRANCIS, M.D., F.R.C.P., and PRAPA KANAGARATNAM, Ph.D., M.R.C.P. From the St. Marys Hospital, Imperial College Healthcare NHS Trust, London, UK Ripple Mapping: A Novel 3D EGM Display. Background: Three-dimensional (3D) mapping is often used to guide ablation in atrial tachycardia (AT), but maps can be susceptible to annotation and interpolation errors. Ripple Mapping (RM) is a technique that displays electrogram time voltage data simultaneously as dynamic bars on the surface shell to overcome these limitations. Objectives: We hypothesized that RM would be superior to established 3D activation mapping. Methods: CARTO-XP TM maps of ATs were collected without any manual annotation and studied on a CARTO-based offline RM system. Paired unannotated CARTO-XP and Ripple Maps were presented to experienced CARTO users with limited RM training. These assessors were allowed to annotate the CARTO-XP maps, but were blinded to conventional EP data. Results: CARTO-XP maps of AT (10 patients) were studied in RM format and the diagnosis was confirmed by entrainment in all cases and with termination of tachycardia in 9/10 cases. Blinded assessors (n = 11) reached the correct diagnosis using RM in 35/44 (80%) compared to 22/44 (50%) using CARTO-XP (P = 0.029). The time to the correct diagnosis was also shorter with RM (136 seconds vs. 212 seconds; P = 0.022). The causes of diagnostic errors using RM (insufficient point density, particularly in low-voltage areas, and the operator not assessing all available views) were overcome with an improved MatLab version showing both scar and dynamic bars on the same shell. Conclusion: RM does not need any manual annotation of local activation time and enables rapid diagnosis of AT with higher diagnostic accuracy than conventional 3D activation mapping. (J Cardiovasc Electrophysiol, Vol. 24, pp , December 2013) activation mapping, atrial tachycardia, catheter ablation Introduction Three-dimensional (3D) activation mapping systems have been developed to aid the localization of the source of focal tachycardias and the critical isthmuses of reentrant circuits. 1-6 Dr. Shahnaz Jamil-Copley is funded by a grant (PG/10/37/28347) from the British Heart Foundation. This work was supported by the NIHR Biomedical Research Centre, and the ElectroCardioMaths Programme of the Imperial BHF Centre of Research Excellence. Imperial College owns the Intellectual Property rights of Ripple Mapping on behalf of Dr. Kanagaratnam, Dr. Francis, and Dr. Linton. Dr. Kanagaratnam, Dr. Francis, Dr. Linton, and Dr. Jamil-Copley have received consulting fees from Biosense Webster with regard to this software. Dr. Davies reports serving as a consultant to and is a stockholder in Rhythmia Medical. Other authors: No disclosures. Address for correspondence: Dr. Prapa Kanagaratnam, Department of Cardiology, Mary Stanford Wing, St. Mary s Hospital, Imperial College Healthcare NHS Trust, London W2 1NY, UK. Fax: ; p.kanagaratnam@imperial.ac.uk Manuscript received 27 February 2013; Revised manuscript received 17 June 2013; Accepted for publication 12 July doi: /jce Typically these systems display electrical data from mapping catheters using local activation time (LAT) to create an activation map or isochronal map. Alternatively, the local voltage can be displayed as isopotential maps. These techniques display the parameter being measured using a color-scale on a 3D representation of the cardiac chamber of interest. The conversion of the raw electrical data to clean activation maps requires an expert operator to validate the LAT assigned by the software algorithm. Errors can be introduced by incorrect assignment of the LAT, inappropriate selection of the window-of-interest (WOI), inappropriate voltage thresholds or interpolation within unmapped regions. We previously investigated techniques for overcoming these limitations and presented Ripple Mapping (RM) as a potential theoretical solution. 7 RM displays each electrogram at its corresponding 3D coordinate on the surface of the cardiac chamber of interest as a dynamic surface bar that changes in length and color according to the electrogram voltage time relationship that is time-gated to a selected fiduciary reference electrogram. Both positive and negative electrogram deflections are shown protruding outward from the surface. If multiple points are collected in a small area, then each bar height changes according to the local voltage in a sequential fashion creating a ripple effect as the movement traverses from one bar to the next. This allows the operator to appreciate the

2 1362 Journal of Cardiovascular Electrophysiology Vol. 24, No. 12, December 2013 Figure 1. Ripple bar height versus electrogram voltage: The top panels (A F) show increasing Ripple bar height in correlation to the increasing magnitude of the corresponding electrogram voltage in the annotation window (lower window). Panel A shows the time-cursor (white bar) at an isoelectric point ofthe chosen electrogram (lower panel) with no visible corresponding ripple bar (upper panel). As the time-cursor moves along the electrogram, the bar always moves outward from the cardiac shell irrespective of whether it is a negative or positive bipolar electrogram polarity. Panel C represents deflection of the electrogram below the isolelectric line corresponding to a negative bipolar value ( 0.86 mv); however, the bar is clearly protruding outward from the shell. The color can be correlated to the reference color bar seen to the left of the panel with low voltage represented by red and high voltage represented by purple. dominant propagation pattern without any software processing or annotation. Furthermore, the system does not interpolate within unmapped regions. All the components of the electrogram are preserved and represented because it does not use a single LAT. Therefore, a sequence of small potential changes in a fractionated electrogram can be temporally linked to its adjacent neighbors. We demonstrated the proof-of-concept using locally developed MatLab software, but to be used in clinical studies it required incorporation with a 3D navigation system. 7 The CARTO-XP TM system (Biosense Webster, Inc., Diamond Bar, CA, USA) is widely used in clinical practice and the RM concept was introduced onto this platform as an offline version for validation purposes. We tested the hypothesis that RM accurately represents activation sequences during simple and complex atrial arrhythmias without the need to specify a WOI or LAT assignment, thereby improving diagnostic accuracy compared to conventional 3D activation mapping. Methods RM Module on CARTO-XP A RM module was developed on the CARTO-XP TM system (Biosense-Webster, Haifa, Israel) and was added to the current options for visualization of electrogram data as activation, voltage, or propagation maps. Studies were collected on a standard CARTO-XP TM system with a suitable WOI and standard filter thresholds. The operator was asked to create detailed maps to reduce interpolation errors. The LAT was designated by the proprietary automatic algorithm and was not manually adjusted. The studies were then uploaded onto a stand-alone computer with the RM module. The collected data were not modified or filtered further prior to its display as a Ripple Map. We first validated the representation of electrogram data as Ripple bars and then the diagnostic capability of RM was prospectively assessed in 10 patients undergoing clinically indicated ablation procedures for incessant atrial tachycardia (AT). These offline paired Ripple Maps and CARTO-XP maps were then presented to independent assessors for direct comparison of the diagnostic accuracy of conventional CARTO mapping versus RM. Software Validation RM displays the electrogram corresponding to each CARTO-XP point as a dynamic bar at its 3D location, the height and color of which correspond to the magnitude of the electrogram bipolar voltage at that time point in the annotation window. Both positive and negative deflections of the bipolar electrogram are represented by an outward movement of the bar from the cardiac shell (Fig. 1). The software algorithm was validated by visual assessment (for example, ripple bar not visible at isoelectric point of the electrogram; ripple bar lengthened when magnitude of the electrogram voltage increased in value whether positive or negative; ripple bar shortened when magnitude of the electrogram voltage reduced whether positive or negative) of the concordance of the ripple bar height with the electrogram voltage (n = 100). For each electrogram (n = 100) a visual examination was made to confirm that the color corresponded with the absolute electrogram voltage. In 20 electrograms the voltage in the time window after each change in color was documented and plotted against the color spectrum. RM is critically dependent on an accurate temporal relationship between each dynamic surface bar and the fiduciary time point. The CARTO-XP activation maps were reviewed to identify areas of uniform planar activation and these were compared with the activation sequence of the dynamic surface bars on the Ripple Maps to check for consistency. This was also confirmed by the conventional method of assessing the direction of activation by comparing the timing of the map electrograms to the fiduciary electrogram (usually the coronary sinus) in the annotation window.

3 Jamil-Copley et al. Ripple Mapping: A Novel 3D EGM Display 1363 TABLE 1 Patient Demographics Previous ablation Clinical AT Age for clinical cycle length Tachycardia CARTO Patient (years) Sex Background AT (ms) mechanism points (n) Circuit locations 1* 73 F Aortic and mitral valve No 278 Macro-reentry 90 LA: Roof dependent replacement and surgical AF ablation 2* 55 M Percutaneous RFA for AF Yes 219 Macro-reentry 136 LA: Perimitral (PVI) M Mitral valve repair 2 No 251 Localized Reentry 130 RA: CTI Surgical Maze RFA of CTI 4* 70 M Mitral valve repair Yes 318 Macro-reentry 215 LA: Roof dependent Surgical AF ablation 5* 50 M Percutaneous RFA for No 288 Localized reentry 197 LA: Base of LAA persistant AF M CABG Mitral valve repair No 275 Macro-reentry 150 RA: CTI 7 33 F Percutaneous Yes 428 Focal 149 RA: Crista Terminalis cryoablation for AT 8 64 F Percutaneous No 423 Localized reentry 157 RA: Septa 1 HRA cryoablation for AF 9 80 M CABG No 253 Macro-reentry 193 RA: CTI 10* 61 M Percutaneous RFA for persistant AF 2and CTI No 207 Macro-reentry 268 LA: Roof dependent AF = atrial fibrillation; CABG = coronary artery bypass graft; CL = cycle length; CTI= cavo tricuspid isthmus; HRA = high right atrium; LA = left atrium; LAA = left atrial appendage; PVI = pulmonary vein isolation; RA = Right atrium; RFA = radiofrequency ablation. *Cases selected for independent assessor analysis. Prospective Clinical Evaluation The clinical operator was blinded to the CARTO-XP map data and allowed to review the Ripple Maps to make a diagnosis. A designated CARTO operator then proceeded to annotate the CARTO-XP map and entrainment maneuvers were performed to make a final diagnosis, which was confirmed by successful ablation. Comparative Evaluation of RM versus Conventional Activation Mapping We could not identify any published literature that has investigated the diagnostic accuracy of 3D activation mapping without concurrent conventional electrophysiological data. RM was compared with CARTO-XP for diagnostic accuracy. Up to 5 maps of the original 10 cases with successful tachycardia termination during ablation were selected for analysis by 11 independent expert CARTO operators. Seven assessors were from external high volume EP centers routinely using CARTO-XP for diagnosing complex atrial arrhythmias. Four assessors were from our center but had no exposure to the live clinical cases or the CARTO-XP-based RM system. Cases selected for the blinded assessment included both reentry (with minimal to complex underlying scar and therefore associated circuit) and focal/localized reentry mechanisms with varying density of maps to allow examination of the ease of use of Ripple Maps in varying situations (Table 1 asterisk cases). Assessors were given minimal RM training (<10 minutes) following which the CARTO-XP and Ripple Maps were presented to them in a randomly sequenced unpaired manner. They were blinded to the patients clinical and EP data but were told which chamber had been mapped. The assessors were allowed to alter any aspect of the CARTO-XP map as would be the case during a clinical procedure. On the Ripple Maps, they could select the views, select the pairs of views and alter the propagation map playback speed. The time taken to reach a diagnosis and the accuracy of this diagnosis was assessed for both CARTO-XP and Ripple Maps. Studies selected for this analysis had a stable cycle length with the correct diagnosis confirmed by conventional entrainment maneuvers. During the comparative evaluation by independent assessors sources of interpretation errors were documented. Statistical Analysis Categorical variables are expressed as percentages. Continuous data are expressed as a mean ± standard deviation. Statistical analysis was performed using IBM SPSS Statistics software. Continuous variables were analyzed using the Wilcoxon signed ranks test and categorical variables analyzed using the McNemar test. A value of P < 0.05 was considered significant. Local research ethics committee approval was granted for the study. Software Validation Results The surface ripple bars (n = 100) changed height in direct proportion to the corresponding electrogram voltage. Scatter plots of 20 electrograms in 5 different patients confirmed a correlation between the voltage at a given time point along the electrogram and the color of the corresponding surface ripple bar. The activation sequence was validated utilizing 5 randomly distributed ripple bars and their corresponding electrograms across a selected atrial wall in 5 patients. The Ripple Map dynamic bar activation sequences in all cases concurred

4 1364 Journal of Cardiovascular Electrophysiology Vol. 24, No. 12, December 2013 Figure 2. CARTO-XP and ripple activation map of a left atrial tachycardia: The unannotated CARTO-XP LAT map (top image 2A) was difficult to interpret and a tachycardia circuit could not be identified. However, the unprocessed data displayed as a Ripple Map (image 2B) revealed a roof-dependent tachycardia (arrows on the maps display the direction of ripple bar activation) with an isthmus near the RSPV and the superior view displayed activation spreading from right to left. The first entrainment maneuver performed at the roof near the LSPV revealed a PPI-TCL of 20 milliseconds suggesting its involvement in the circuit; however, due to activation seen on the Ripple Map (roof/anterior wall activation from right to left) entrainment was also performed at the roof nearthe RSPV and the floor near the RIPV, which revealed a PPI-TCL of 0 and 12 milliseconds, respectively. Despite insufficient point collection on the anteroseptal wall of the left atrium, the Ripple Map was able to display the overall activation pattern. In this example, the CARTO-XP activation map was non-diagnostic, and retrospective annotation performed after the diagnosis was made using Ripple Mapping, demonstrated the roof-dependent macro-reentrant circuit (Fig. 2C). LSPV = left superior pulmonary vein; RIPV = right inferior pulmonary vein; RSPV = right superior pulmonary vein. with the activation direction from multielectrode catheters and the corresponding CARTO-XP activation maps. Prospective Clinical Evaluation The same diagnosis was reached using RM and entrainment maneuvers in all 10 cases and confirmed by successful ablation in 9 cases (Table 1). In 1 patient the diagnosis of roof-dependent left AT was evident on the Ripple Map and confirmed with entrainment maneuvers; however, ablation did not terminate the tachycardia as roofline block could not be achieved by ablation. Figure 2 illustrates a patient with 3 failed ablation attempts at another center for incessant AT following surgical MV repair and AF ablation. The auto-annotated CARTO- XP map was difficult to interpret; however, the same data displayed as a Ripple Map demonstrated a roof-dependent macro-reentrant tachycardia with a narrow isthmus near the right upper pulmonary vein with passive activation of the remaining roof. Ablation at this site successfully terminated tachycardia and the patient has remained free of recurrent arrhythmias. Corresponding CARTO-XP activation maps confirmed the same diagnosis but only after detailed annotation and postprocessing.

5 Jamil-Copley et al. Ripple Mapping: A Novel 3D EGM Display 1365 Mean time to a diagnosis (s) [n = 42] Correct diagnosis [n = 44] Mean time to the correct diagnosis* (s) [n = 17] TABLE 2 Diagnostic Data CARTO XP Ripple Map p value 281 (±281) 152 (±115.6) (50%) 35 (80%) (±137.7) (±96.7) Mean time to a diagnosis, diagnostic yield and the mean time to correct diagnosis all significantly favor Ripple Mapping vs. CARTO-XP TM. *Analysis of data where the correct diagnosis was reached for the same patient on both CARTO-XP and Ripple Mapping. Comparative Evaluation of RM versus Conventional 3D Activation Mapping Expert CARTO operators (n = 11) examined 22 maps (each as CARTO-XP and Ripple Maps), providing 44 paired timing data. A diagnosis was reached on both CARTO-XP and Ripple Maps in 42 cases giving paired data for time to diagnosis. A diagnosis was not reached or an incorrect diagnosis was made in 9 Ripple Maps and 22 CARTO-XP maps. The accurate diagnostic yield was significantly higher with the Ripple Maps (80% vs 50%; P = 0.029) (Table 2). The time to diagnosis included any time used to annotate the maps and was found to be significantly longer for CARTO- XP than for the Ripple Maps (281 seconds vs 152 seconds; P = 0.002). A correct diagnosis was reached by both modalities in 17 cases and the time taken to reach a correct diagnosis was also significantly longer for CARTO-XP (212 seconds vs 136 seconds; P = 0.022) (Table 2). Discussion We have demonstrated that RM accurately displays electrograms as dynamic bars on a 3D cardiac shell and that the ripple activation sequence concurs with the corresponding CARTO-XP activation maps. Analyzing ATs with this offline CARTO-XP system incorporating RM, we were able to make a diagnosis in all cases followed by confirmation with entrainment. In a comparative study, expert CARTO operators were faster and more accurate using the Ripple Maps when cases were reviewed using only the 3D mapping data. This clinical validation study confirms the capability of RM for displaying AT and that it may be better than conventional activation maps for diagnostic interpretation. We analyzed the cause of failure to achieve 100% diagnostic yield. The following potential causes of error were observed. The Ripple effect was not apparent in regions of insufficient point collection, making it difficult to visualize the complete circuit. It should be noted that although CARTO- XP activation maps interpolate into such unmapped regions they assume that all activation between the collected points is uniform throughout and consistent with the direction suggested by the collected points, thereby introducing a potential source of error. Density of maps created by operators depends predominantly upon the complexity of the ablation history and the extent of regions of slow conduction. Interpretation of data within low-amplitude scar regions was difficult as there appeared to be a tendency for operators to avoid high density point collection in these areas. The demarcation of such scarred area was not shown on the 3D geometry so the location of potential isthmuses was difficult to interpret. Secondly, alteration in the Ripple bar color to reflect changing voltage frequencies along an electrogram appeared not to be helpful to the assessors as this was easily appreciated by the change in bar height alone. Finally, the presence of bystander and blind alley activation caused diagnostic difficulties when the assessor did not review the entire geometry from all views available. The use of specific paired views (Postero-anterior and Left Lateral; Left Anterior Oblique and Postero-anterior; Superior and Antero-posterior; Inferior and Right Anterior Oblique) followed by operator choice appeared to improve diagnostic accuracy. In order to overcome these issues, we developed a custom MatLab Ripple Mapping software that allowed the display of both the Ripple Map and a bipolar voltage map on a single geometry. These maps displayed low-amplitude scar regions in a different color on the surface geometry. Furthermore, this system was capable of importing CARTO3 TM maps with more detailed geometry. We studied a further 5 cases of AT with an aim to collect equidistant points throughout the chamber per map (Table 3). This improved the quality of the Ripple Maps and again all cases were successfully diagnosed with RM alone. We observed that the Ripple Maps showed complete macrorentrant circuits when >250 points were collected and 150 points appeared sufficient for focal tachycardias with evenly spaced points <1 cm apart. Figure 3 and supporting videos show an example of the Ripple Map created using autoannotated TABLE 3 CARTO 3 Patient Demographics and Tachycardia Characteristics Age AT CL AT CARTO Patient (years) Sex Background (ms) mechanism points (n) Circuit location 1 33 F Cryoablation of Crista 440 Focal 189 RA: Crista Terminalis Terminalis EAT 2 54 F PVI and Perimitral AT 360 Macro-reentry 284 LA: Peri-mitral Ablation 3 72 F PVI 350 Focal 256 LA: Left anteroseptal wall near previous WACA 4 56 F CTI and PVI 220 Macro-reentry 355 LA: Dual-loop re-entry: roof dependant and peri-mitral 5 67 M PVI and Re-do PVI 264 Macro-reentry 366 LA: Peri-mitral AT = atrial tachycardia; CL = cycle length; EAT = ectopic atrial tachycardia; LA = left atrium; PVI = pulmonary vein isolation; RA = right atrium; RFA = radiofrequency ablation.

6 1366 Journal of Cardiovascular Electrophysiology Vol. 24, No. 12, December 2013 Figure 3. Dual loop reentry atrial tachycardia: A left atrial CARTO-3 activation map was created with no operator annotation during point collection. This was analyzed on the MatLab Ripple Mapping software to guide ablation. (A) Twelve-lead surface ECG of the clinical atrial tachycardia (cycle length 220 milliseconds; recording speed 25 mm s 1 ). (B) LAO view of the left atrium using custom MatLab Ripple Mapping software. Counter-clockwise ripple bar propagation is seen around the mitral valve annulus (see supporting video) in keeping with a peri-mitral atrial tachycardia. (C) LL view of the left atrium using custom MatLab Ripple Mapping software. Two simultaneous wave-fronts are evident. Ripple bar propagation is seen caudo-cranially on the posterior wall and cranio-caudally on the anterior wall confirming a roof-dependent macrorentry circuit. Due to low-voltage electrograms on the roof and posterior roof the ripple bars in this region are short and difficult to appreciate in the images (see supporting video). Simultaneously, a second ripple bar wave-front is seen traversing the peri-mitral region from the posterior LA floor. (D) PA view of the left atrium using custom MatLab Ripple Mapping software. An island (isthmus) of healthy (purple) tissue* is visible between 2 heavily scarred (cyan) regions on the posterior left atrium near the roof. Ripple bar propagation simultaneously travels from the floor caudo-cranially toward the roof and toward the anterior mitral valve region.

7 Jamil-Copley et al. Ripple Mapping: A Novel 3D EGM Display 1367 Figure 3. (Continued) (E) Two surface ECG leads (avf, V6) and intracardiac bipolar (filtered) Hz) electrograms from the coronary sinus (CS) and ablation (Map) catheter are displayed (100 mm s 1 ). The results of entrainment maneuvers are shown. The first image shows entrainment from the distal CS bipole with a PPI-TCL of 14 milliseconds. The middle image shows entrainment from the isthmus on the posterior roof* with a PPI-TCL of zero. The final image shows a PPI-TCL of 15 milliseconds during entrainment from the anterior mitral valve region below the LAA. (F) Ablation performed to transect the isthmus on the posterior roof* caused an increase in the tachycardia cycle length by 12 milliseconds confirming successful ablation of the roof-dependent circuit but persistence of the perimitral bystander. A subsequent mitral isthmus line terminated the perimitral tachycardia to sinus rhythm. (G) AP image of the left atrium with individual electrograms selected (creating a virtual catheter) and marked with a colored bar at that location (this is not a ripple map but an additional feature available on the MATLAB RM software to study the local electrograms). These electrograms are shown against a fiducial time point with the electrogram presented in the same color as the bar representing its location. Late electrogram activation is seen on the mid-anterior MV region secondary to the perimitral loop with earlier electrograms on the anterior LA wall from the roof-dependent wave-front. This confirms that the roof-dependent wavefront is the first to activate the shared isthmus at the floor of the LA and the perimitral wavefront is the secondary circuit being entrained by the former. CS =coronary sinus; CTI = cavotricuspid isthmus; LAA = left atrial appendage; LAO =left anterior oblique; LL = left lateral; MV = mitral valve; PA = poster-anterior; PPI-TCL = postpacing interval minus tachycardia cycle length; TCL = tachycardia cycle length.

8 1368 Journal of Cardiovascular Electrophysiology Vol. 24, No. 12, December 2013 Figure 4. Left atrial focal tachycardia. Panels A I demonstrate the centrifugal Ripple bar propagation during a focal atrial tachycardia. This map was created using CARTO-MEM and 1,433 points collected in the left atrium. CARTO3 data illustrating a double loop reentry circuit. Figure 4 and the supporting video ( focal AT ) display the centrifugal activation pattern of the ripple bars in this left atrial focal tachycardia created using CARTO-MEM. The lower diagnostic yield of CARTO-XP in the absence of the surface ECG, the mapping catheter and entrainment data can be explained by the following. Isochronal activation mapping using the CARTO TM system requires the operator to select a WOI using the cycle length of the tachycardia relative to a fiduciary electrogram from a reference catheter. 7 The activation map is constructed using automatically assigned LAT that can be manually adjusted. However, a small number of incorrectly assigned points due to fractionated electrograms and far-field components can dramatically alter the appearance of an activation map. Isochronal patches are colored according to the timing of the central point and the interpolation into unmapped regions is susceptible to errors but can be avoided by increasing the number of points collected, which in turn increases analysis time. Activation maps can be converted to propagation maps, which display moving activation wavefronts that are more intuitive but result in an even greater false sense of resolution of the map. 8,9-15 There are no other published studies subjecting any of the currently established 3D mapping systems to this type of analysis. Identification of the mechanism and site for ablation in ATs can be achieved using entrainment maneuvers utilizing the postpacing interval (PPI) to rapidly determine if the pacing site is part of a reentrant circuit. 9,11,16 A short PPI minus tachycardia cycle length (PPI-TCL) is likely to be near the circuit, but a long PPI-TCL can be deceptive if it is near a line of block alongside the circuit. Entrainment maneuvers also run the risk of accelerating or terminating the tachycardia. 11,12 A site identified as being part of the circuit from entrainment maneuvers is not necessarily the optimal location to deliver lesions. For example the left atrial roof is frequently a part of macroreentrant circuits, but long linear lesions have a significant failure rate for both acute and chronic bidirectional block making them proarrhythmic. Furthermore, entrainment mapping can be more challenging when the isthmus is in an unpredictable location Conversely, activation mapping may take longer to establish the complete circuit but will demonstrate the isthmus of slow conduction and therefore potentially the most effective site for ablation. This was apparent in the cases shown in Figures 2 and 3. RM should enable a single expert operator to collect points using a foot pedal and then diagnose the tachycardia circuit without the need for annotation. Entrainment can be used to confirm the chosen ablation site and even if this causes acceleration or termination, a diagnosis has already been reached to guide therapeutic ablation. Limitations This was a retrospective study because the system is currently an offline tool and therefore direct comparison with current 3D mapping systems needs to be made cautiously. In this validation study, the offline tool was used in a small

9 Jamil-Copley et al. Ripple Mapping: A Novel 3D EGM Display 1369 number of patients and needs an online system to be tested prospectively in a larger cohort. As experienced with most other currently established mapping systems, variations in tachycardia cycle length during mapping are likely to cause errors. Far-field signals may cause misinterpretation but can be easily visually isolated from the rest of the activation sequence. Conclusions RM is a novel method for displaying 3D cardiac activation patterns. We have confirmed its feasibility for use with the CARTO TM navigation system and its accuracy at displaying the activation sequence of ATs. This preliminary study also suggests a higher diagnostic yield using RM in experienced CARTO operators. Acknowledgments: We are grateful to the team at Biosense-Webster, Haifa, Israel, for developing the prototype Ripple Mapping module. The authors thank the operators who participated as blinded assessors from St. Mary s Hospital, Imperial College Healthcare NHS Trust, London, UK; Heart Hospital, University College London, UK; Providence Healthcare, Vancouver, Canada and Brigham and Women s Hospital, Boston, MA, USA. References 1. Stevenson WG, Delacretaz E, Friedman PL, Ellison KE: Identification and ablation of macroreentrant ventricular tachycardia with the CARTO electroanatomic mapping system. Pacing Clin Electrophysiol 1998;21: Marchlinski F, Callans D, Gottlieb C, Rodriguez E, Coyne R, Kleinman D: Magnetic electroanatomical mapping for ablation of focal atrial tachycardias. Pacing Clin. Electrophysiol 1998;21: Earley MJ, Showkathali R, Alzetani M, Kistler P, Gupta D, Abrams D, Horrocks J, Harris S, Schilling R: Radiofrequency ablation of arrhythmias guided by non-fluoroscopic catheter location: a prospective randomized trial. Eur heart J 2006;27: Oral H, Pappone C, Chugh A, Good E, Bogun F, Pelosi F, Bates E, Lehmann M, Vicedomini G, Augello G, Agricola E, Sala S, Santinelli V, Morady F: Circumferential pulmonary-vein ablation for chronic atrial fibrillation. N Engl J Med 2006;354: Schilling R, Peters N, Davies D: Feasibility of a noncontact catheter for endocardial mapping of human ventricular tachycardias. Circulation 1999;99: De Ponti R, Verlato R, Bertaglia E, Del Greco M, Fusco A, Bottoni N, Drago F, Sciarra L, Ometto R, Mantovan R, Salerno-Uriarte JA: Treatment of macro-reenatrant atrial tachycardia based on electroanatomic mapping: Identification and ablation of the mid-diastolic isthmus. Europace 2007;9: Linton NW, Koa-Wing M, Francis DP, Kojodjojo P, Lim PB, Salukhe TV, Whinnett Z, Davies DW, Peters NS, O Neill MD, Kanagaratnam P: Cardiac ripple mapping: A novel three-dimensional visualization method for use with electroanatomic mapping of cardiac arrhythmias. Heart Rhythm 2009;6: Irtel TA, Delacrétaz E: Intra-atrial reentrant tachycardia with ambiguous data from activation mapping: What to do next? Heart Rhythm 2005;2: Stevenson WG, Sager PT, Friedman PL: Entrainment techniques for mapping atrial and ventricular tachycardias. J Cardiovasc Electrophysiol 1995;6: Miyazaki H, Stevenson WG, Stephenson K, Soejima K, Epstein LM: Entrainment mapping for rapid distinction of left and right atrial tachycardias. Heart Rhythm 2006;3: Waldo AL, Plumb VJ, Arciniesgas JG, MacLean WA, Cooper TB, Priest MF, James TN: Transient entrainment and interruption of atrioventricular bypass pathway type of paroxysmal atrial tachycardia: A model for under-standing and identifying reentrant arrhythmias. Circulation 1983;67: Jaïs P, Shah DC, Haïssaguerre M, Hocini M, Tian Peng J, Takashashi A, Garrigue S, Le Métayer P, Clémenty J: Clinical investigation and reports: Mapping and ablation of left atrial flutters. Circulation 2000;101: Shah DC, Jais P, Haissaguerre M, Chouairi S, Takahashi A, Hocini M, Garrigue S, Clementy J: Three-dimensional mapping of the common atrial flutter in the right atrium. Circulation 1997;96: Shpun S, Gepstein L, Hayam G, Ben-Haim SA: Guidance of radiofrequency endocardial ablation with realtime three-dimensional magnetic navigation system. Circulation 1997;96: Nakagawa H, Jackman W: Use of a three-dimensional, non-fluoroscopic mapping system for catheter ablation of typical atrial flutter. Pacing Clin Electrophysiol 1998;21: Saoudi N, Cosio F, Waldo A, Chen SA, Iesaka Y, Lesh M, Saksena S, Salerno J, Schoels W: Classification of atrial flutter and regular atrial tachycardia according to electrophysiologic mechanism and anatomic bases: A statement from a joint expert group from the Working Group of Arrhythmias of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. J Cardiovasc Electrophysiol 2001;12: Shah D, Jäis P, Takahashi A, Hocini M, Peng JT, Clementy J, Haissaguerre M: Dual-loop intra-atrial reentry in humans. Circulation 2000;101: Stevenson IH, Kistler PM, Spence SJ, Vohra JK, Sparks PB, Morton JB, Kalman JM: Scar-related right atrial macroreentrant tachycardia in patients without prior atrial surgery: Electroanatomic characterization and ablation outcome. Heart Rhythm 2005;2: Nakagawa H, Jackman WM: Use of a 3-dimensional electroanatomical mapping system for catheter ablation of macroreentrant right atrial tachycardia following atriotomy. J Electrocardiol 1999;32: Jais P, Matsuo S, Knecht S, Weerasooriya R, Hocini M, Sacher F, Wright M, Nault I, Lellouche N, Klein G, Clémenty J, Haïssaguerre M: A deductive mapping strategy for atrial tachycardia following atrial fibrillation ablation: Importance of localized reentry. J Cardiovasc Electrophysiol 2009;20: Triedman JK, Alexander M, Berul CI, Bevilacqua LM, Walsh EP. Electroanatomic mapping of entrained and exit zones in patients with repaired congenital heart disease and intra-atrial reentrant tachycardia. Circulation 2001;103: Supporting Information Additional Supporting information may be found in the online version of this article at the publisher s website: Video S1: Left lateral view of the left atrium. Video S2: Left anterior oblique view of the left atrium. Video S3: Postero-anterior view of the left atrium. Video S4: Focal Atrial Tachycardia.