Author + information
- Received February 3, 2015
- Revision received April 7, 2015
- Accepted April 16, 2015
- Published online August 1, 2015.
- Jacob S. Koruth, MD∗,
- Christopher Schneider, MEng†,
- Boaz Avitall, MD, PhD‡,
- Leonardo Ribeiro, MD∗,
- Srinivas Dukkipati, MD∗,
- Gregory P. Walcott, MD§,
- Patrick Phillips, PhD†,
- Hugh T. McElderry, MD§ and
- Vivek Y. Reddy, MD∗∗ ()
- ∗Helmsley Electrophysiology Center, Mount Sinai Medical Center, New York, New York
- †VytronUS, Inc., Sunnyvale, California
- ‡University of Illinois at Chicago, Chicago, Illinois
- §University of Alabama at Birmingham, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Vivek Y. Reddy, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1030, New York, New York 10029.
Objectives The purpose of this study was to assess the feasibility of pulmonary vein (PV) isolation using low-intensity collimated ultrasound.
Background Contemporary approaches to PV isolation are limited by the technical complexity of mapping and ablation. We describe a novel approach to left atrial anatomic rendering and PV isolation that aims to overcome some of these limitations by using low-intensity collimated ultrasound (LICU) system, which allows for near real-time geometry creation and automated ablation in a porcine model.
Methods Twenty swine were anesthetized, and the LICU ablation catheter was placed in the left atrium via percutaneous transseptal access. Ultrasound M-mode–based anatomies of the inferior PVs were successfully created, and ablation was performed under automatic robotic control along a user-defined lesion path. One animal was excluded because of device failure.
Results All target PVs in the 19 remaining animals were isolated acutely, requiring a mean of 1.6 applications. Ten animals were sacrificed acutely, and the remaining 9 survived for 35 ± 11 days. Of these 9, 1 animal was excluded from analysis because the index lasso position could not be reliably recreated. PVs in 5 of 8 animals remained isolated at sacrifice. Of the 77 total histological sections, 62 lesions (80.5%) were noted to be transmural. Lesions were homogeneous and characterized by coagulative necrosis and fibrous tissue. The mean myocardial thickness was 2.66 ± 1.80 mm, and the mean lesion depth was 4.28 ± 1.97 mm. No extra cardiac or collateral lesions were noted.
Conclusions This study demonstrates the safety and efficacy of a novel noncontact ultrasound mapping and ablation system to produce continuous transmural lesions that can isolate PVs in a porcine model.
Currently available approaches to pulmonary vein (PV) isolation can be broadly separated into 2 categories: 1) point-by-point ablation approaches that aim to create circular perivenous lesions that are, in aggregate, electrically isolating (1); and 2) single-application PV isolation approaches using either balloons or multipolar catheters (2–4). The success of the former approach depends on factors such as catheter tip design, contact force sensing capability, accurate left atrial (LA) and PV mapping, and, importantly, operator expertise in catheter manipulation (5,6). On the other hand, the single-application ablation approaches were designed to require somewhat less operator dependency to attain adequate expertise. However, these approaches are largely limited to isolation at the level of the PV ostia and with minimal effect on the PV antral regions and posterior LA (7). In light of these various limitations, it would be desirable to have a single integrated anatomic mapping and automated ablation system capable of creating uninterrupted transmural lesions along any user-defined path in the LA without these limitations.
To that end, we describe a novel approach to PV isolation that is capable of creating 1) near-real-time, noncontact 3-dimensional ultrasound anatomic renderings of the LA and PVs; followed by 2) robotically controlled perivenous ultrasound ablation. This technology aims to create continuous and transmural lesions along user-defined trajectories within the LA. In this preclinical study, we aim to assess the feasibility of this technology to achieve acute and chronic PV isolation followed by physiologic and histological assessment.
The experimental protocol was approved by the respective institutional animal care and use committees at Mt. Sinai Hospital in New York, New York; University of Alabama at Birmingham, Alabama; and Sutter Institute for Medical Research in Sacramento, California. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Low-intensity collimated ultrasound system
The low-intensity collimated ultrasound ablation system (LICU) (VytronUS, Inc., Sunnyvale, California) comprises the following major components: 1) a deflectable trans-septal sheath; 2) a robotic tip open-irrigated catheter with an additional manually deflectable section; and 3) an electronic control console (Figure 1A). The deflectable sheath is 13-F and capable of 135° of deflection. The distal tip of the catheter contains the proprietary LICU ultrasound transducer that is driven at approximately 10 MHz. The design of the transducer allows for creation of a highly collimated ultrasound beam. Thus the acoustic intensity (energy density) is effectively constant throughout the length of the beam from the face of the transducer to the maximum therapeutic range of 17 mm. The choice of frequency is such that blood is largely “transparent” to the beam with negligible ultrasound attenuation in the blood pool. The maximum therapeutic and imaging ranges of the ultrasound beam are 17 mm and 40 mm, respectively (Figure 1B).
The system produces endocardial geometry by robotically scanning the LA and PVs with the ultrasound-equipped catheter tip. A proprietary edge detection algorithm identifies the endocardial surface in individual M-mode lines; this range is then combined with tip position coordinate data to construct 2- and 3-dimensional maps of the chamber (Figures 2A and 2B). The anatomic map is superimposed with a color-coded range map that displays those sections that are within therapeutic ablation range of the transducer (purple, blue, green) and those that are beyond therapeutic range (yellow, orange, red). On this 2-dimensional mapped anatomy, the user then describes free-form lesion trajectories that are automatically ablated under robotic control without the need for continuous operator manipulation. The catheter is positioned in such a manner that the robotic movement of the catheter along the trajectory is uninterrupted. The interaction with the LA walls can be detected fluoroscopically or by interruptions during the smooth continuous creation of the anatomy. As described previously, the color-coded range map allows identification of regions of interest that are within the therapeutic range.
Because the maps and lesion trajectories are all self-referencing and nearly real time, they are minimally sensitive to external spatial error introduced by cardiac motion or fluid intake, among others. An approximately constant ultrasound beam speed at the endocardial surface is achieved by first scanning the target tissue along the desired trajectory to map the endocardial topography, and then using these data to robotically control the catheter tip’s velocity. The highly directional ultrasound beam generates thermal injury by absorption and dissipation of ultrasound energy as sound waves propagate through the target tissue. Because of the negligible dissipation of ultrasound energy in blood, lesion formation is relatively insensitive to the distance from the catheter tip to the target tissue.
The endocardial surface is cooled by both circulating blood in the LA and the saline infused through the catheter tip during ablation. The flow rate during mapping is 2 ml/min and during ablation is 15 ml/min. The LICU beam produces an acoustic radiation pressure that induces a jet of fluid in the direction of the beam. These mechanisms aim to prevent the formation of char or thrombus.
After an overnight fast, 20 swine (mean weight, 74 kg) were pre-medicated, intubated, and mechanically ventilated. General anesthesia was maintained with isoflurane (1% to 3%) for the duration of the procedure. Percutaneous venous access was obtained, and transseptal puncture was performed in typical fashion after systemic heparin administration to gain access to the LA. The deflectable sheath was advanced into the LA over a 0.035-inch guidewire and positioned at the ostium of the inferior common PV (ICPV). Contrast was then injected to obtain a pulmonary venogram (Figures 3A and 3B). A circular mapping catheter was placed in the target PV to document electrograms at baseline and after ablation to assess electrical isolation (Figure 4). Creation of an electroanatomic map using a magnetic or impedance-based system for additional guidance was left to the operator’s discretion (this was performed in 10/20 swine). The ablation catheter was then advanced into the LA through the deflectable sheath and positioned in the LA chamber facing the PV. The deflection of the catheter’s robotic tip was registered with the anatomic axes of the LA to guide catheter manipulation and map optimization. Ultrasound-based geometry was then created and optimized (by manipulating catheter position) until all structures of interest were visualized within the catheter’s field of view and therapeutic range (Figure 2). Once an acceptable beam trajectory around the ostium of the ICPV was designated by the operator, the saline flow rate was increased, and the lesion trajectory was traversed under robotic control with the transducer in therapeutic mode, which produced a maximum acoustic intensity of 2.5 W/mm2 with a beam transit speed of 0.25 mm/s at the endocardial surface. These energy settings were based on data obtained during in vitro testing. If isolation was not achieved after completion of the encircling lesion, additional applications were performed.
The acute animals (10 of 20) were humanely sacrificed at the end of the procedure according to institutional protocols. The chronic animals (10 of 20) survived and underwent repeat catheterization at approximately 4 weeks to assess for isolation and PV stenosis. They were then euthanized, and pathological examination was performed.
The chest was opened and all structures (pericardium, lungs, aorta, and so on) were carefully inspected for signs of injury. The heart and lungs were then removed en bloc and photographed. In a selection of animals, the heart was immersion-stained with triphenyl tetrazolium chloride to enhance contrast between ablated tissue and adjacent viable tissue. The LA was then opened, photographed, and examined for catheter manipulation and ablation-related injury, after which the heart was immersion-fixed in 10% neutral buffered formalin.
In the chronic animals, the target veins were serially sectioned along the axis of the PV. Labeled cassettes were submitted for paraffin embedding, microtoming, staining, and histological analysis. Slides were stained with hematoxylin and eosin and Movat’s pentachrome stains. Lesion depth, lesion area, and myocardial thickness were quantified using light microscopy morphometry.
All continuous variables are presented as mean ± SD.
ICPVs in all 20 swine were accessed, and LA-PV electrical potentials were present in all cases. Ultrasound-based 2- and 3-dimensional anatomies of the inferior LA and ICPVs were successfully created in all animals.
Nineteen of these swine were then successfully ablated. One animal was excluded from analysis because of a technical failure of the LICU system, resulting in failure of energy delivery for ablation.
All targeted inferior common PVs were successfully isolated. Procedure and fluor-oscopy times specific to mapping and ablation were not recorded because the experiments frequently incorporated multiple interruptions in the work flow related to technical aspects of this first-generation system. A total of 31 ablations were performed in 19 target veins (mean 1.6 ablation trajectories per vein) to achieve acute isolation. If PV isolation was not achieved after the initial encircling ablation, additional ablation was performed. Gaps were identified based on circular mapping catheter electrograms and fluoroscopy. Locations of gaps were noted both anteriorly and posteriorly in the PVs. The mean length of the perivenous circumferential lesion set was 94.6 ± 35.5 mm, and the mean duration of total ablation per vein (including additional ablation if needed) was 936 ± 378 s. The duration of the initial perivenous lesions set alone was 1,068 ± 258 s. An example of electrical PV isolation is depicted in Figure 4. Ten swine (10 of 20) were then sacrificed.
The remaining 9 of 20 swine survived a mean of 35 ± 11 days after initial PV isolation. One additional animal was excluded from this group because of difficulty reproducing the index procedure’s lasso catheter position and thus fairly assessing isolation of the vein. Of the remaining 8 swine, the target veins in 5 (62.5%) remained durably isolated at follow-up.
No instances of PV stenosis, cardiac perforation, or collateral damage to adjacent structures (lung, great vessels, and so on) were noted in either of the cohorts. Adverse events are shown in Table 1. Animals in both cohorts suffered ventricular tachycardia or fibrillation during ablation of the ostium of the ICPV (in swine, the ostium of the ICPV is located near the mitral annulus, where inadvertent ablation of the ventricular myocardium can result in ventricular arrhythmias).
Gross examinations of both acute and chronic swine revealed antral circumferential PV lesions with no evidence of endocardial thrombus or char formation. The lesions were visually continuous and well demarcated from normal atrial tissue (Figure 5).
Histological characteristics of the lesions in the chronic swine experiment are described in Table 2. Although only 8 veins qualified for assessment of chronic PV isolation, histology was assessed on all veins where ablation was successfully performed (i.e., 9 PVs). A mean of 8.6 sections per vein were obtained. On a per-vein basis, lesion transmurality was 80.7 ± 16.3%. When assessed per section (n = 77), 62 of 77 sections (80.5%) demonstrated transmural lesions. The lesions consisted of coagulative necrosis that was partially replaced by fibrous connective tissue as demonstrated in Figure 6. The mean myocardial thickness for all sections was 2.66 ± 1.80 mm, and the mean lesion depth was 4.28 ± 1.97 mm. The coagulative necrosis often extended into the adjacent epicardial adipose tissue. Additionally, the lesions demonstrated a mild inflammatory response. The interface between the lesion and the uninjured myocardium was sharply demarcated. Some sections demonstrated necrotic myocardium and/or small foci of dystrophic calcification. There were no areas of crater formation within the lesion to suggest steam pop occurrence.
This porcine study demonstrates the feasibility of a novel approach to PV isolation using LICU. We demonstrate in this study the successful creation of: 1) near-real-time anatomic renderings of the LA-PV antrum; followed by: 2) automatic robotic ablation to achieve electrical PV isolation.
Feasibility of PV mapping and isolation
In this study, we successfully created ultrasound-based 3-dimensional geometries of the ICPV-LA junction in all swine (20 of 20). Following this, we were able to achieve an acute PV isolation rate of 100% in all 19 swine.
Ablation lesions were noted to be well demarcated, contiguous, homogeneous, and without endocardial charring; in addition, they were noted to be along the pre-defined perivenous trajectory in the swine that were examined acutely. This ability to achieve acute PV isolation provides proof of concept for this novel approach to PV isolation.
Durable PV isolation was observed in 62.5% (5 of 8 swine) with this approach. On histology, 19% of all sections were noted to be nontransmural, which likely accounts for the PV reconnections that were noted in the remaining 3 of 8 swine at the follow-up examination. Design changes in this first-generation system that occurred during the course of the study—as well as unintended variations in the velocity of the robotically driven tip resulting from limited deflection control—may have both contributed to the lack of consistent lesion transmurality. In addition, it is possible that gaps of unablated tissue may not have been detected during sectioning. It is anticipated that the use of different dosing strategies and future system refinements will improve transmurality and chronic PV isolation rates. Nonetheless, these data provide proof of concept that electrical PV isolation can be achieved with a reasonable rate of chronic durability with this novel LICU system.
Comparison to current PV isolation technologies
The system tested in this study has several potential advantages over available technologies. Specifically, these are:
1. Near-real-time mapping. A limitation of catheter ablation using electroanatomic mapping is the occurrence of anatomic inaccuracies in the map onto which ablation lesions are then registered. These inaccuracies are known to result from cardiac motion, respiratory motion, volume-related chamber enlargement, and technology-specific map drifts/shifts (8–10); importantly, these inaccuracies can affect the contiguity of point-by-point lesion sets. The ability to create near-real-time maps using LICU allows for possibly more accurate rendering of the targeted anatomy, which may improve lesion contiguity.
2. Robotically controlled ablation process. This capability optimizes the ability of the system to create lesions without the need for continuous operator involvement and skill, factors that influence outcomes with nearly all other available PV isolation technologies.
3. Contact-independent ablation technology. This capability allows the system to ablate without needing to maintain contact or a specified force with the target tissue. Both of these variables have been shown to be responsible for inadequate lesion formation with traditional radiofrequency ablation (11).
Although not tested in this study, this approach is also capable of a significant degree of flexibility in the lesion set deployed for atrial fibrillation ablation. For example, although extraostial PV isolation may be the desired ablation approach for paroxysmal atrial fibrillation, patients with persistent atrial fibrillation may benefit from other lesion sets, such as “posterior LA box” isolation or other linear lesions such as the roof line or mitral isthmus line (12,13). Creation of these lesion sets using single-shot ablation technologies is typically not possible, whereas point-by-point ablation is technically demanding and laborious and suffers from a high incidence of chronic reconduction across the lesion sets (14).
However, it is important to mention that although the LICU system has the potential to overcome these traditional shortcomings of contemporary PV isolation technologies, there is a considerable amount of validation work that will be necessary before these putative advantages can be realized.
Adverse events and Study limitations
Ventricular tachycardia or fibrillation occurred in 7 of 20 swine, likely from unintended ablation of the left ventricular myocardium that lies in close proximity to the anterior border of the inferior common PV in swine. These ventricular arrhythmias are due to the species-specific propensity for ventricular arrhythmias in this model in our collective experience. Indeed, ventricular fibrillation almost invariably occurs during radiofrequency ablation in the porcine ventricular myocardium, whereas it is uncommon in human ventricular ablation. This, along with anatomic separation of human PVs from the ventricle, makes it unlikely that ventricular tachycardia or fibrillation will be a clinically relevant adverse event. Two swine also developed ST-segment elevation: one that was transient and resolved, whereas the other resulted in ventricular fibrillation with subsequent death. These were presumed to be due to air embolism related to catheter insertion/withdrawal. These events clearly emphasize the importance of careful management of large-French sheaths to avoid inadvertent introduction of air into left-sided circulation.
The use of ultrasound-based PV ablation in the past, specifically high-intensity focused ultrasound (HIFU), caused significant issues with esophageal injury resulting in its withdrawal from clinical use (15). This raises the question of similar collateral damage with LICU. There are significant differences between HIFU and LICU that allow for an improved safety profile. 1) The LICU beam operates at lower acoustic intensity than HIFU. At the lower acoustic intensity cavitation (mechanical tissue trauma resulting from acoustic pressure fluctuations) does not occur. In addition, at this lower intensity, lesion formation occurs in a more gradual fashion compared with HIFU. 2) The collimated nature of LICU further allows for gradual lesion creation. These characteristics therefore allow for monitoring of collateral damage (phrenic nerve pacing, esophageal temperature monitoring) during lesion creation. The previously described HIFU system, on the other hand, sonicated all points within the focus of the HIFU beam simultaneously and approximately 10 times more quickly than the LICU beam does. This prevented cessation of ablation in response to early signs of collateral damage.
Although collateral damage was not seen in this study, further studies are needed to fully understand the margin of safety that exists beyond which collateral damage can occur. Human studies are planned, and measures to avoid collateral damage such as phrenic nerve pacing, esophageal temperature monitoring, and esophageal deviation will be used, in addition to the wall thickness–based dose titration mentioned previously.
This study demonstrates the feasibility of a novel, noncontact, semirobotic, near-real-time anatomic rendering and ablation system using LICU to isolate pulmonary veins in a porcine model. Further system optimization is necessary to improve the transmurality and durable electrical isolation rates.
COMPETENCY IN MEDICAL KNOWLEDGE: Contemporary approaches to mapping and ablation for PV isolation have significant limitations. LICU provides an opportunity to improve upon these limitations by combining mapping with automated ablation.
TRANSLATIONAL OUTLOOK: This study demonstrated the safety and efficacy of this noncontact system in achieving PV isolation in a porcine model. Additional clinical studies will be needed to validate the safety and incremental value of this approach.
This study was supported by a grant from VytronUS, Inc. Dr. Koruth has received grant support from VytronUS. Drs. Reddy, McElderberry, and Avitall have received consulting income and stock options from VytronUS. Drs. Phillips and Schneider are employees of VytronUS. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- high-intensity focused ultrasound
- inferior common pulmonary vein
- left atrial
- low-intensity collimated ultrasound ablation system
- pulmonary vein
- Received February 3, 2015.
- Revision received April 7, 2015.
- Accepted April 16, 2015.
- American College of Cardiology Foundation
- Van Belle Y.,
- Janse P.,
- Rivero-Ayerza M.J.,
- et al.
- Reddy V.Y.,
- Neuzil P.,
- Themistoclakis S.,
- et al.
- Scaglione M.,
- Caponi D.,
- Anselmino M.,
- et al.
- Malchano Z.J.,
- Neuzil P.,
- Cury R.C.,
- et al.
- Kiuchi K.,
- Kircher S.,
- Watanabe N.,
- et al.
- Sawhney N.,
- Anand K.,
- Robertson C.E.,
- Wurdeman T.,
- Anousheh R.,
- Feld G.K.
- Neven K.,
- Schmidt B.,
- Metzner A.,
- et al.