Author + information
- Received April 18, 2016
- Revision received June 8, 2016
- Accepted July 14, 2016
- Published online September 7, 2016.
- Henry Chubb, MBBS, MAa,
- James L. Harrison, MB BCh, MA, PhDb,
- Steffen Weiss, MScb,
- Sascha Krueger, PhDb,
- Peter Koken, Dipl Ingb,
- Lars Ø. Bloch, MDc,
- Won Yong Kim, MDc,
- Gregg S. Stenzel, BSEE, MBAd,
- Steven R. Wedan, MSEEd,
- Jennifer L. Weisz, BSEE, MSTMd,
- Jaswinder Gill, MB BChir, MDa,e,
- Tobias Schaeffter, PhDa,
- Mark D. O’Neill, MB BCh, DPhila,e and
- Reza S. Razavi, MB BS, MDa,e,∗ ()
- aDivision of Imaging Sciences and Biomedical Engineering, King’s College London, London, United Kingdom
- bResearch Laboratories, Philips Technologie GmbH, Hamburg, Germany
- cDepartment of Cardiology, Aarhus University Hospital, Skejby, Denmark
- dImricor Medical Systems, Inc., Burnsville, Minnesota
- eDepartment of Cardiology, St. Thomas’ Hospital, London, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Reza S. Razavi, Division of Imaging Sciences and Biomedical Engineering and Cardiovascular Medicine, 4th Floor, Lambeth Wing, St. Thomas’ Hospital, Westminster Bridge Road, London SE1 7EH, United Kingdom.
Objectives This study sought to develop an actively tracked cardiac magnetic resonance-guided electrophysiology (CMR-EP) system and perform first-in-human clinical ablation procedures.
Background CMR-EP offers high-resolution anatomy, arrhythmia substrate, and ablation lesion visualization in the absence of ionizing radiation. Implementation of active tracking, where catheter position is continuously transmitted in a manner analogous to electroanatomic mapping (EAM), is crucial for CMR-EP to take the step from theoretical technology to practical clinical tool.
Methods The setup integrated a clinical 1.5-T scanner, an EP recording and ablation system, and a real-time image guidance platform with components undergoing ex vivo validation. The full system was assessed using a pre-clinical study (5 pigs), including mapping and ablation with histological validation. For the clinical study, 10 human subjects with typical atrial flutter (age 62 ± 15 years) underwent MR-guided cavotricuspid isthmus (CTI) ablation.
Results The components of the CMR-EP system were safe (magnetically induced torque, radiofrequency heating) and effective in the CMR environment (location precision). Targeted radiofrequency ablation was performed in all animals and 9 (90%) humans. Seven patients had CTI ablation completed using CMR guidance alone; 2 patients required completion under fluoroscopy, with 2 late flutter recurrences. Acute and chronic CMR imaging demonstrated efficacious lesion formation, verified with histology in animals. Anatomic shape of the CTI was an independent predictor of procedural success.
Conclusions CMR-EP using active catheter tracking is safe and feasible. The CMR-EP setup provides an effective workflow and has the potential to change the way in which ablation procedures may be performed.
Catheter ablation has become the cornerstone of curative treatment for tachyarrhythmia. However, despite many technological advances, there remains a significant recurrence rate, particularly for more complex arrhythmias, including ventricular tachycardia and atrial fibrillation (AF). There is therefore an imperative for a paradigm shift, and cardiac magnetic resonance (CMR) imaging guidance of ablation procedures provides a potential solution by enhanced imaging of the anatomy and arrhythmia substrate, and real-time visualization of the intracardiac catheter and ablation lesion formation.
Over the last decade, centers worldwide have sought to establish the core technologies and techniques for CMR-guided electrophysiological (CMR-EP) procedures. Building upon the principles of interventional CMR procedures (1), platforms have been developed that are capable of performing diagnostic and interventional EP procedures in a CMR environment in animals (2–10) and humans (3,11–14). Although relatively simple in concept, CMR-EP presents a range of complex technical challenges. Most obviously, the operating environment is within a strong magnetic field, and therefore the use of magnetic material is severely curtailed. Other important challenges include radiofrequency (RF) safety, gradient field safety, and minimization of mutual electromagnetic interference between CMR and EP components. All this must be achieved while maintaining the multiple functions of the ablation catheter (maneuverability, position tracking, measurement of low-amplitude intracardiac potentials, pacing, temperature sensing, RF ablation energy delivery). Equally importantly, CMR imaging using various suitable contrasts and CMR device tracking must be developed and combined into a clinically viable workflow that replaces conventional fluoroscopy and exploits the unique capabilities of CMR for imaging of cardiac anatomy and arrhythmia substrate.
Because of these challenges, overall progress in the field of CMR-EP has been relatively slow. It has been possible to demonstrate use of ablation catheters within CMR scanners (2,3,10) for performance of ablation in animals (5,6,8,9,15) and to perform CMR-guided ablation of atrial flutter in humans (7,13,14). However, in order to make the leap from research to viable clinical tool, the realization of RF-safe active tracking in a clinical-grade catheter is paramount. Passive tracking relies upon CMR visualization of the ablation catheter and is therefore slow and prone to localization errors. Active tracking enables automation of the tracking of the catheter, freeing the operator to work in real-time throughout the cardiac target field.
The achievement of active tracking opens up all the strengths of fast electroanatomic mapping (EAM), including local activation time (LAT) mapping and voltage mapping, which may then be combined synergistically with real-time imaging of cardiac anatomy, arrhythmia substrate, and surrounding structures. This study documents the development of the first CMR-guided electrophysiology platform using active catheter tracking for ablation of arrhythmia from development to pre-clinical validation to clinical translation.
The technical development of the underlying technologies for the CMR-EP catheter and guidance platform was performed between 2004 and 2010. This study details the pre-clinical optimization, testing, and system validation (2010 to 2013) and clinical studies with medium-term follow-up (2014 to 2015). Approximately 9 months after the start of the clinical study, the system that was developed and validated was made available to a second group, who started clinical studies in 2015 and recently published acute results of their first 6 cases (14). However, this paper describes the technical developments, their integration into a clinical system, and the pre-clinical validation and acute and medium-term results in patients.
The working CMR-EP setup combines a standard clinical 1.5-T MR scanner (Achieva, Philips Healthcare, Best, the Netherlands), a clinical patient monitoring system suitable for CMR use (Expression, Invivo, Gainesville, Florida), a standard clinical RF generator (IBI 1500, St. Jude Medical, St. Paul, Minnesota) with a new EP recording system specifically designed for CMR use (Horizon, Imricor Medical Systems, Burnsville, Minnesota), new dedicated CMR electrophysiology and ablation catheters (vision ablation catheter; Imricor) (Online Figure 3) and a new CMR-EP guidance platform for mapping and ablation of cardiac arrhythmias (interventional MRI Suite [iSuite]; Philips) (Figure 1). Technical specifications are detailed in the Online Methods.
The active tracking capability was provided by 2 miniature solenoid CMR receiver coils at the tip of the ablation catheter 2 mm and 11 mm proximal to the ring electrode. A dedicated tracking sequence (modified fast-field echo; in-plane resolution 0.83 mm, 10-Hz tracking rate) enabled the localization of the catheter position and orientation. The derived catheter position was overlaid on a pre-acquired balanced steady-state free precession 3D whole heart (bSSFP-3DWH) dataset.
Catheter tip position was also used for automatic derivation of slice position for the multiplanar reformatting of the bSSFP-3DWH image data (Figure 2A, Online Video 1) and for calculation of real-time imaging planes of the catheter-tissue interface (Figures 2B to 2D, Online Video 2). Intracardiac electrograms (IEGM) were transmitted by wire within the catheter shaft with a novel RF winding structure that passes small IEGM potentials while maintaining high impedance around the Larmor frequency at 1.5-T (64-MHz). Bipolar potentials were presented on the Horizon interface (Figure 3), and LAT relative to reference was used to generate activation maps (Figure 2A).
Ex Vivo Technical Development
Ex vivo testing and validation was performed at the Division of Imaging Sciences and Biomedical Engineering (King’s College London, United Kingdom), Philips, and Imricor.
Magnetically induced torque on the ablation catheter was assessed using a torsional spring apparatus and force by measuring magnetically induced deflection, both at 1.5 T. RF heating was assessed in gelled saline (conductivity 0.47 S/m) using a high specific absorption rate (SAR) sequence (balanced turbo field echo, whole body: 4 W/kg). Active catheter tracking performance was assessed in a phantom, and a high-resolution turbo spin echo (TSE) single-slice sequence was used as the gold standard for true catheter position evaluation. Three orthogonal imaging slices were obtained in order to pinpoint the true catheter tip and were compared with the catheter position as assessed by active tracking (600 measurements) (Online Methods).
The complete CMR-EP ablation system was tested prospectively in 5 Danish Landrace pigs (Aarhus University Hospital, Skejby, Denmark; approximately 40 kg). Studies complied with institutional and national guidelines for the care and use of animals.
Ten patients with typical atrial flutter were enrolled for ablation of isthmus-dependent atrial flutter under CMR guidance. Inclusion criteria were 18 to 80 years of age and undergoing first-time clinically indicated ablation therapy of documented paroxysmal or persistent counter clockwise (typical) right atrial flutter. Procedures were performed through uninterrupted warfarin or interrupted rivaroxaban therapy. Exclusion criteria included any contraindication to CMR imaging or gadolinium contrast, previous ablation, previous cardiac surgery, and any intracardiac mass including thrombus or myxoma. Patient characteristics are summarized in Table 1.
The clinical study was performed in the interventional MR suite at St. Thomas’ Hospital (London, United Kingdom), approved by the UK Health Research Authority (NRES Committee East of England, reference 14/EE/0031, UK Clinical Trials Gateway Study ID 16258), and informed consent was obtained from all patients.
Pre-clinical system evaluation
The procedure was performed using the CMR scanner without fluoroscopy. Animals were pre-sedated using azaperone (4 mg/kg) and midazolam (0.5 mg/kg). General anesthesia was induced using intravenous ketamine (5 mg/kg) and midazolam (0.5 mg/kg), and maintained with propofol (3 mg/kg/h) and fentanyl (15 μg/kg/h). Animals were intubated and mechanically ventilated. Two 9-F sheaths were placed percutaneously in the right femoral vein, followed by a bolus intravenous injection of 100 IU/kg heparin.
Pre-clinical: pre-ablation imaging
A bSSFP-3DWH dataset was acquired without contrast in order to provide a “road map” for the procedure. An acquisition window in ventricular diastole was selected, with a maximum window of 180 ms (sagittal orientation 2 × 2 × 2-mm resolution; reconstructed: 1.3 × 1.3 × 1.3 mm; T2 preparation; pencil respiratory navigation; 5-channel phased array coil). The RA was manually segmented from the bSSFP-3DWH by using freely available 3-dimensional (3D) medical image segmentation software (itk-SNAP version 2.2.0) and the shell imported into iSuite to act as a roadmap for mapping and ablation.
As a baseline for post-ablation imaging, T2-weighted (T2W) images were also acquired prior to ablation (multislice turbo spin echo [TSE], double-inversion recovery pre-pulse, SPIR fat suppression, echo time [TE]: 45 ms; repetition time [TR] twice the cardiac cycle length: 1.5 × 1.5 mm [reconstructed: 1.0 × 1.0 mm]; slice thickness: 3 mm).
Pre-clinical: electroanatomic mapping and ablation
Active catheter tracking was used to place investigational catheters in the coronary sinus (CS) and right atrium (RA) using the segmented roadmap. Activation data were acquired during CS pacing. For each sampling point, the time delay (LAT) from the pacing artifact to the local RA electrogram was measured on the EP recording system and automatically transmitted to the image guidance platform to produce a color-coded activation map on the RA shell.
Point-by-point RF ablation (35 W; 48°C; 60 s per lesion; 17 ml/min irrigation) was then performed from the superior vena cava (SVC) to the inferior vena cava (IVC) along the posterior wall of the RA and the location of each ablation point recorded on the RA shell. A linear set of RA ablation lesions rather than cavotricuspid isthmus (CTI) line was chosen due to the propensity of swine to develop ventricular fibrillation in response to right ventricular ablation. The foot pedal was used to allow rapid switching from active catheter tracking (for catheter navigation) to real-time imaging (to confirm the tracked catheter position before each RF delivery).
Following completion of the intercaval ablation lesion and post-ablation CMR, the activation map was repeated according to the same protocol.
Pre-clinical: post-ablation imaging and histology
T2W imaging was repeated immediately post-ablation with identical parameters. Three-dimensional late gadolinium enhancement (LGE) imaging was then performed approximately 20 min after administration of 0.2 ml/kg gadobutrol (Gadavist, Bayer Healthcare Pharmaceuticals, Berlin, Germany). Respiratory-navigated, electrocardiography (ECG)-triggered inversion recovery turbo field echo acquisition: 1.3 × 1.3 × 4 mm; reconstructed to 0.6 × 0.6 × 2 mm (TE: 3.0 ms; TR: 6.2 ms; flip angle 25°).
Following procedure completion, euthanasia was performed using an intravenous bolus of phenobarbital (80 mg/kg); the porcine hearts were explanted, and the RA was opened and photographed. The hearts were then fixed in formaldehyde. The ablation line and surrounding tissue were excised en bloc and cut into 4-mm sections perpendicularly to the ablation line. Each cross-section was photographed and then dehydrated, embedded in paraffin, sectioned (3-μm sections), and stained with hematoxylin and eosin (H&E) for microscopy examination.
For the clinical study, procedures were performed under general anesthesia. Beyond the 5-Gauss (500-μT) line, two 10-French long venous sheaths were placed percutaneously in the right femoral vein, and the patient was then moved to the CMR scanner bore.
Clinical study: Pre-ablation Imaging
bSSFP-3DWH imaging was performed without contrast as described in the preceding text (32-channel phased-array coil). Patients experiencing atrial flutter (7 of 10 subjects) at the start of procedure underwent DC cardioversion (50 to 100 J) before entry to the scanner in order to facilitate cardiac gating. An automated segmentation technique, using a shape-constrained deformable model, was used to derive the RA contour (SmartHeart, Philips). CS and IVC were delineated manually (itk-SNAP) and added to the cardiac model to assist the procedure.
The length and morphology of the CTI were measured on bSSFP-3DWH by using the technique detailed by Kirchhof et al. (16). IVC/CTI angle was measured as the angle between 2 tangential lines placed parallel to the CTI floor and the adjacent wall of the IVC. Measurements were performed twice by a single observer (H.C.) (1 month between measurements) and once by a second observer (J.H.) to assess reproducibility.
Clinical study: Electroanatomic Mapping and Ablation
Activation mapping was performed as described above prior to ablation. A point-by-point RF ablation (35 to 45 W; 60 s per lesion; 17 ml/min irrigation) was performed along a pre-planned CTI line during coronary sinus pacing. CTI conduction block was confirmed by a superior-to-inferior activation pattern at the lateral wall on CS pacing, with LAT mapping repeated following completion of ablation. Bidirectional block was confirmed by differential pacing from 2 sites lateral to the CTI line.
The protocol allowed a maximum of 2 h (from first ablation lesion) to obtain bidirectional CTI block. For patients in whom this was not achieved, the subject was moved to a conventional fluoroscopy suite, and ablation was completed using a conventional nonirrigated ablation catheter (8-mm tip, large curve Blazer II, Boston Scientific, Natick, Massachusetts) under fluoroscopic guidance.
Clinical study: Post-ablation Imaging
All clinical subjects underwent acute T2W imaging immediately after ablation (parameters as above). Three-dimensional LGE imaging was performed acutely post-ablation for all but 2 human subjects who underwent CMR-guided ablation; the time required for procedure completion precluded late acute imaging in these 2 subjects. LGE acquisition parameters were identical to those for the pre-clinical study but performed in a slice orientation parallel to the CTI. Multiple time-separated LGE acquisitions were performed, and the dataset commenced at 20 min post-contrast was used for RF lesion analysis. T2W enhancement was quantified using a threshold of 3.3 standard deviations (SDs) above the adjacent right ventricular myocardial signal, and ablation volume was assessed within the pseudo-3D dataset using Seg3D (University of Utah, Salt Lake City, Utah).
Chronic lesion imaging was performed 3 months post ablation. An ECG and respiratory-gated MR angiogram was performed using a previously described technique (17). This high-contrast sequence was used to derive a mask of the RA for interrogation of the 3D LGE sequence imaging the CTI, acquired along the axis of the CTI ablation line (acquisition parameters as for acute LGE post-ablation imaging, begun 20 min post-contrast injection). CMR-derived scar was interrogated using a maximum intensity projection technique (2 mm inside and outside the RA shell) and thresholded on the 3D scar mesh to a signal intensity of 3.3 SD above the blood pool mean (18).
The ablated area of the RA was assessed defining the superior margin of the RA floor as the most inferior axial slice lying above the CS os and inferior margin as the most superior axial slice below the RA floor. Measurements were performed using Paraview (Kitware, New York, New York). Lesion-specific scar was also assessed using a point-specific technique. The catheter tip site during RF delivery was recorded relative to the bSSFP-3DWH sequence, and then referenced to the chronic imaging using an affine registration technique (19). The registration technique was used to define the matrix for the transformation of the bSSFP-3DWH during ablation to the 3D MR angiogram sequence at chronic imaging, and the same matrix was applied to the RF delivery sites. The derived RF lesion sites were then projected to the closest surface on the 3D scar mesh, and the projection distance and associated scar were recorded. The scar value was recorded as the maximum signal intensity within a 3-mm radius of the projected lesion site.
Continuous variables are expressed as mean ± SD. Statistical analysis was performed using SPSS version 2 statistics (IBM, Armonk, New York).
Ex Vivo Technical Validation
The maximum magnetically induced torque and force on the investigational catheter at 1.5-T were <2.2 mNm and <7.7 Nm, respectively, well below regulatory thresholds (351 mNm and 79.5 Nm, respectively). The maximum temperature rise on any portion of the investigational catheter due to RF-induced heating was observed at the tracking coil locations at a catheter insertion depth of 45 cm and was measured at 2.1°C above background heating.
Average tip displacement, the discrepancy between actively tracked and gold-standard TSE-derived positions, along the axis of the catheter was 0.90 ± 0.58 mm. The angular deviation of catheter orientation from its true direction was 8.5° ± 3.6°.
Active catheter tracking was achieved in all animals, and the 2 investigational CMR catheters were positioned in the CS and RA without the requirement for fluoroscopy. Bipolar IEGMs were recorded with minimal CMR interference, and a LAT map was created during CS pacing in all animals. Using active tracking, activation maps during CS pacing were created in all animals (mean number of points 34 ± 5; mean acquisition time 20 ± 8 min) before ablation. Irrigated RF ablation was successfully performed from the SVC to IVC in all 5 animals. Activation maps following ablation demonstrated a change in the pattern of activation of the RA, with activation detour secondary to the linear ablation lesion.
All 5 animals survived until the end of the procedure, and there was no evidence of cardiac perforation, pericardial effusion, or cardiac tamponade.
Active catheter tracking, appropriate catheter manipulation, and activation mapping were achieved in all 10 subjects. SmartHeart was capable of creating a 3D roadmap automatically in all subjects, and LAT map (28 ± 7 points) was created in 24 ± 11 min. Cine CMR imaging was used to confirm catheter position, and there was no discrepancy detected in projected and imaged positions (Figures 2B to 2D), including during RF delivery. Bipolar EGMs of sufficient quality for local activation annotation were recorded (Figure 3).
CMR-guided RF ablation was performed in 9 of the 10 subjects. Ablation was not possible for 1 subject due to a persistent impedance error detected by the non-investigational RF generator, later attributed to an error in generator setup. The procedure was therefore performed in a conventional EP laboratory. For the remaining 9 subjects, 7 (78%) had an acutely successful procedure under CMR guidance alone, with a post-procedural transisthmus conduction time ranging from 124 to 180 ms (mean 160 ms). Two subjects (22%) required completion of the CTI line under fluoroscopic guidance, using a conventional nonirrigated ablation catheter (post–CMR-guided ablation transisthmus conduction times 73 and 95 ms, respectively). Radiation exposure in patients who underwent CMR-guided ablation alone was zero and was mean 90 (range: 64 to 106) cGy/cm2 for those 3 patients who required conventional fluoroscopically guided ablation.
Total procedure time was 314 ± 54 min, including the additional time for conventional fluoroscopically guided ablation (286 ± 29 min excluding fluoroscopy). Time from first to last ablation lesion under CMR guidance was 78 ± 40 min, with an average of 24 ± 9 lesions required under CMR guidance (total ablation time: 18.3 ± 9.1 min) (Figure 4). There were no significant safety concerns, and all patients were discharged within 24 h of the procedure.
Final ablation outcome is summarized in Figure 5. Two subjects had a late recurrence of atrial flutter. One subject returned as asymptomatic at 3 months, and the second subject presented as symptomatic to the emergency department. Both patients underwent a second ablation procedure under conventional EAM guidance (CARTO 3, Biosense Webster/Johnson & Johnson, New Brunswick, New Jersey) with evidence of low-voltage IEGMs in the region of prior ablation (Online Figure 1). In all other patients, there were no further arrhythmias detected on 24-h tape or 12-lead ECG at 3 months follow-up.
Intra-Procedural Ablation Assessment
Pre-ablation, no appreciable T2W enhancement was seen in any of the animals. Post-ablation, there was increased T2W enhancement and atrial wall thickness between the SVC and IVC in all animals.
Post-ablation T2W imaging was performed at a mean of 27 ± 17 min after the final ablation lesion. Mean volume of CTI T2W enhancement was 6.4 ± 4.0 ml, and there was no significant correlation with total ablation time (R2 = 0.009). Gaps in T2W enhancement along the ablation line could not be discerned in any patient, and enhancement position closely overlaid ablation lesion sites (Figure 6). Gadolinium-enhanced imaging was performed in 7 of 9 human subjects who underwent CMR-guided ablation. Early no-reflow in the CTI region was seen all 7 cases, with subsequent LGE (Figure 6). Again, there were no discernible gaps in the ablation line.
Final Lesion Assessment
The linear ablation lesion was inspected macroscopically, both immediately after heart explantation and following fixation in formaldehyde. There was close correspondence between the iSuite 3D RA shell, showing the intended location of ablation, and the sites directly visualized (Figures 7A and 7B). A macroscopic cross-section through the ablation line (showing the typical features of RF injury, a central zone of pallor and a surrounding hemorrhagic border zone) and microscopic sections (stained with H&E) are shown in Figures 7D and 7E. Microscopic examination demonstrated histological findings consistent with RF ablation.
Ablation scar at the CTI was found in all patients at 3 months on LGE imaging (Figure 8). There was no relationship between total ablation time and total scarred area (R2 = 0.003) or proportion of the floor of the right atrium occupied by scar (R2 = 0.02). For ablations performed using the investigational catheter versus conventional nonirrigated catheter (alone or in combination), there was no detected difference in surface area of scar (11.3 ± 11.7 cm2 vs. 13.8 ± 6.4 cm2, respectively).
For lesion-specific scar assessment, patients who underwent fluoroscopically guided ablation (either alone or in combination) were excluded (n = 3), leaving a total of 151 ablation lesions assessed. A total of 108 lesions (72%) were associated with scar on the 3D mesh, and projection distance to mesh was mean 6.2 ± 4.1 mm. CMR-assessed scar was significantly greater when the RF energy was delivered <10 mm away from the right atrial mesh (signal intensity: 5.25 SD above blood pool mean, versus 1.82 SD at >10 mm; p < 0.0001) (Figure 8). In predicting effective ablation, the area under the receiver-operating curve for distance from the mesh was 0.66.
Determinants of Successful Ablation
Table 2 details the relationship of CTI anatomy and acute ablation imaging parameters to the outcome of an acutely successful ablation procedure under CMR-guided ablation alone. The subject who underwent entirely fluoroscopically guided ablation (see the preceding text) was excluded from analysis. The IVC/CTI angle showed minimal overlap between successful/unsuccessful groups, but the small sample sizes preclude meaningful statistical analysis. Intraobserver and interobserver intraclass correlation coefficients for IVC/CTI angle measurement were 0.90 (95% confidence interval [CI]: 0.53 to 0.97) and 0.89 (95% CI: 0.54 to 0.98), respectively.
These studies have demonstrated the development, feasibility, and safety of an actively tracked CMR-EP set up capable of robust performance in the demanding CMR environment in both animals and humans. Many of the key attributes of conventional EAM systems, such as operator ease of use and intuitive data representation, have also been established. The clinical study has identified limitations in efficacy and highlighted discrete areas for further development, but demonstrates significant progress of CMR-EP towards clinical utility.
CMR-EP System Evaluation
Using this CMR-EP set up, active catheter tracking for ablation was achieved for the first time in humans (11). This is in contrast to passive tracking techniques that rely on detection of magnetic susceptibility artifacts or signal voids to locate the catheter tip, necessitating constant communication between the electrophysiologist and a skilled manipulator of the imaging planes. With passive tracking, complex movements of a curved catheter almost inevitably cause the catheter to leave the imaging plane, and in human studies skilled operators have struggled considerably to perform many relatively routine aspects of EP ablation procedures, such as selective intubation of the coronary sinus or completion of CTI block (13). All these factors affect procedural success rate, safety, and time.
With active catheter tracking, accurate and fast EAM was enabled with generation of activation time maps by the image guidance platform. In both humans and animals, up to 40 mapping points were recorded in less than 20 min, whereas passive tracking requires 2 to 5 min for each CMR-acquired mapping point (20). Furthermore, automatic tip alignment algorithms enabled rapid slice determination for optimal imaging of the ablation catheter within the soft tissue environment (21), a unique capability that may be invaluable for more complex ablation procedures and for imaging of real-time lesion formation during ablation. Imaging could be performed during energy delivery, and the impact of the 500-kHz RF energy source on CMR-imaging, based on the 64-MHz proton precession frequency at 1.5-T, was minimal.
Clinical Study: Procedural outcome
The overall time for the human procedures was long, averaging over 5 h for a procedure that rarely exceeds 1 h under conventional fluoroscopic guidance. It should be noted that imaging and mapping protocols were performed that would not typically be required for a simple ablation, but even accounting for these phases, the duration remains substantially longer than that for conventional methods.
Furthermore, the overall medium-term success rate for flutter ablation was low (72% for CMR-guided ablation alone; 56% overall) compared with 85% to 92% for conventional flutter ablation techniques. In all subjects who underwent CMR-guided ablation, there was significant scar at 3 months at the site of ablation. However, in 4 patients, there was either a need for further lesions under fluoroscopy or recurrence at medium-term follow-up of 3 months. The cause of failure in this subset of patients required detailed assessment and is likely to be related to 1 of 3 factors: failure to recognize the correct target for ablation, failure to reach the correct location, or failure to form effective ablation lesions.
Ablation target recognition
Diagnosis of the arrhythmia mechanism and selection of the CTI as the ablation target were based upon the 12-lead ECG P-wave morphology during atrial flutter in patients with normal ECG. The 2 recurrences were of confirmed typical atrial flutter, successfully treated with completion of a CTI line. The absence of activation mapping or entrainment to confirm the diagnosis is unlikely to have had an impact upon procedural outcome.
At the end of the CMR-guided ablation, the assessment of the adequacy of the ablation line was performed both electronically and through CMR imaging. The acquired IEGMs were acceptable and could demonstrate double potentials and local timings (Figure 3), and hence, electrical assessment of bidirectional block was easily performed. On imaging, however, no gaps in the ablation lines were identified, and therefore, locations for targeting for further CMR-guided ablation could not be delineated.
Catheter reach to ablation target
The reach of the current CMR-compatible catheter, which was designed as an all-purpose EP catheter, may have been insufficient in patients with an acute IVC/CTI angle (Table 2, Online Figure 2), and there was substantial difficulty in catheter manipulation to the IVC end of the CTI during procedures. The eustachian ridge was the site of highest voltage (least scar) for the 2 patients who returned for repeated ablation procedure (Online Figure 1).
Ablation lesion formation
The pre-clinical study clearly demonstrated efficacious lesion formation (Figure 7). This was confirmed in late assessment of ablation lesions in the clinical study, and both the electrical (CARTO) and the CMR assessments demonstrated effective chronic scar formation. Lesion-by-lesion analysis for the clinical study found that 72% of lesions were associated with scar. Comparison with conventional ablation is difficult as CMR-guided ablation is uniquely suited for lesion formation assessment, with virtual elimination of registration issues; the registration of EAM-guided ablation to imaging is subject to imperfect comparisons of variable landmarks. However, a 72% success rate is comparable to conventional ablation studies (22) where up to 25% of an ablation line may not be associated with chronic scar. Furthermore, for lesions where the catheter was far from the RA mesh, lesion formation was less likely.
Future Developments and Clinical Implications
CMR-guided ablation system
CMR-EP provides a platform for anatomically guided ablation procedures, informed by cardiac structure and arrhythmia substrate amenable to imaging, such as scar, with the tantalizing possibility of visualization of short-term lesion formation during the procedure. The implications for ablation procedures for AF (22,23) and ventricular tachycardia (24) are substantial, and the most complex ablation procedures are likely to reap the greatest benefit from the additional capabilities of CMR imaging. However, further complementary developments are also required.
A number of commercial and academic institutions are working on technologies, responding to challenges such as communication in the noisy CMR environment, CMR-compatible 12-lead ECG, CMR-compatible defibrillation, rapid and automated CMR-scar segmentation, and trans septal puncture equipment. These developments are at various stages of commercial development (25,26), and their realization will bring CMR-EP ever closer.
The complex structural and physiological responses to ablation, including interstitial edema, hyperemia, tissue coagulation, and microvascular obstruction can be visualized using CMR imaging. Short-term lesion imaging techniques may rely on native intrinsic contrast (T1 and/or T2 weighted) (2,27–30) or contrast agent enhancement (LGE) (2,9,28), but short-term imaging of lesions within the thin atrial myocardium remains particularly challenging (31,32). Purely T2W imaging is a blunt tool. No gaps in ablation line were identified in the short term with T2-weighted or LGE imaging, and this was despite a 56% clinical success rate.
The finding of low sensitivity of short-term atrial lesion imaging for gaps is in keeping with that of studies by other groups, and further developments in short-term post ablation MR imaging techniques are necessary (9,18,31,33). Identifying gaps within the ventricle is likely to be more easily achievable in the first instance. Lesion imaging in the future is likely to rely on the intrinsic T1–time shortening that occurs swiftly with effective lesion formation (28), and nonenhanced visualization techniques are particularly attractive for repeated imaging.
These studies were designed to take the new CMR-EP platform from development to pre-clinical to first-in-human feasibility trials. As such, there are limitations that should be acknowledged. Ablation for atrial flutter is primarily an anatomically driven procedure: this clinical ablation does not constitute a comprehensive assessment of the capability of surface ECG and IEGM fidelity to guide more complex EP procedures. IEGM fidelity will continue to improve towards that achievable for conventional EAM, but will require further engineering and time to match it.
In addition, the arrhythmia was not mapped, following the decision to cardiovert to sinus rhythm at the start of the procedure on account of the impact of arrhythmia on acquisition of ECG-gated CMR imaging. The primary concern was for the timely acquisition of the 3D whole heart, which was adversely affected by the clinical arrhythmia. However, active tracking technology is independent of heart rhythm and arrhythmia mapping should not be limited in further studies. Technical solutions for the swift acquisition of a high-resolution 3D road map in arrhythmia will require further development.
Finally, ablation procedures in both animals and humans were weighted toward achievement of effective and targeted RF lesions, and imaging protocols were relatively focused on more conventional imaging sequences that demonstrated feasibility rather than long-term durability of short-term lesions.
Real-time CMR imaging guidance of EP procedures is feasible, and active tracking technologies enable an approach and workflow that closely mimics conventional EAM. CMR-compatible catheters can be used to create effective ablation lesions, including for the treatment of atrial flutter in humans. The CMR-EP system is currently slower and less effective than conventional ablation, and this appears to be related primarily to the reach of the investigational catheter. However, CMR-EP provides contemporaneous, high-fidelity imaging of cardiac anatomy, fibrotic arrhythmia substrate, and ablation lesions. Further innovation of these new tools may lead to fundamental changes in the way in which both simple and complex ablation procedures are performed.
COMPETENCY IN MEDICAL KNOWLEDGE: Catheter ablation is a first-line treatment option for many heart rhythm disturbances. However, the success rate for ablation of more complex arrhythmia substrates—AF and ventricular tachycardia—is modest, and there is a need for the development of new ablation and guidance technologies.
TRANSLATIONAL OUTLOOK: The performance of ablation procedures in the CMR scanner poses many technical and engineering demands, but it has now been proven to be feasible. Technological solutions have enabled ablation procedures to be performed in a manner similar to that using conventional electroanatomic mapping systems. Adoption in standard clinical electrophysiology practice is now substantially closer. Further refinement of the technology is necessary to improve catheter maneuverability and intracardiac signals, and to develop a CMR-compatible 12-lead ECG and defibrillator. With these advances, there is the potential to revolutionize complex ablations, such as those for ventricular tachycardia, and to improve simple ablations such as those for atrial flutter.
The authors thank members of the team who were involved in performing the MR-guided procedures, including the team assisting with animal studies in Aarhus, Anne K. Grøndal and Steen F. Pedersen; anesthesiologists Shyamala Moganasundram and Daniel Taylor; radiographers Stephen Sinclair and Tracy Moon; and physiologists Julian Bostock and Shadada Begum.
For an expanded Methods section as well as supplemental figures and videos, please see the online version of this article.
This work was supported by Centre of Excellence in Medical Engineering, funded by Wellcome Trust, and Engineering and Physical Sciences Research Council grant WT 088641/Z/09/Z; by the Department of Health through a National Institute for Health Research comprehensive Biomedical Research Centre award to Guy's and St. Thomas' National Health Service Foundation Trust in partnership with King's College London and King’s College Hospital NHS Foundation Trust; and by British Heart Foundation Clinical Research Training Fellowship award FS/10/65/28404 (Dr. Harrison). Mr. Weiss, Dr. Krueger, and Mr. Koken are employees of Philips GmbH Innovative Technologies. Mr. Stenzel, Mr. Wedan, and Ms. Weisz are employees of Imricor and hold Imricor stock. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- balanced steady-state free precession 3D whole heart
- cardiac magnetic resonance
- cardiac magnetic resonance-guided electrophysiology
- coronary sinus
- cavotricuspid isthmus
- digital amplifier stimulator
- electroanatomic mapping
- intracardiac electrogram
- inferior vena cava
- local activation time
- late gadolinium enhancement
- right atrium
- superior vena cava
- echo time
- repetition time
- turbo spin echo
- Received April 18, 2016.
- Revision received June 8, 2016.
- Accepted July 14, 2016.
- American College of Cardiology Foundation
- Lardo A.C.,
- McVeigh E.R.,
- Jumrussirikul P.,
- et al.
- Nazarian S.,
- Kolandaivelu A.,
- Zviman M.M.,
- et al.
- Dukkipati S.R.,
- Mallozzi R.,
- Schmidt E.J.,
- et al.
- Schmidt E.J.,
- Mallozzi R.P.,
- Thiagalingam A.,
- et al.
- Hoffmann B.A.,
- Koops A.,
- Rostock T.,
- et al.
- Nordbeck P.,
- Beer M.,
- Köstler H.,
- et al.
- Ganesan A.N.,
- Selvanayagam J.B.,
- Mahajan R.,
- et al.
- Ranjan R.,
- Kholmovski E.G.,
- Blauer J.,
- et al.
- Oduneye S.O.,
- Pop M.,
- Shurrab M.,
- et al.
- Chubb H.,
- Harrison J.,
- Williams S.E.,
- et al.
- Sommer P.,
- Grothoff M.,
- Eitel C.,
- et al.
- Hilbert S.,
- Sommer P.,
- Gutberlet M.,
- et al.
- Nordbeck P.,
- Hiller K.-H.,
- Fidler F.,
- et al.
- Kirchhof P.,
- Özgün M.,
- Zellerhoff S.,
- et al.
- Groarke J.D.,
- Waller A.H.,
- Vita T.S.,
- et al.
- Harrison J.L.,
- Jensen H.K.,
- Peel S.A.,
- et al.
- Nordbeck P.,
- Bauer W.R.,
- Fidler F.,
- et al.
- Voigt T.,
- Koken P.,
- Harrison J.L.,
- Weiss S.,
- Krueger S.,
- Schaeffter T.
- Bisbal F.,
- Guiu E.,
- Cabanas-Grandío P.,
- et al.
- Nazarian S.,
- Bluemke D.A.,
- Lardo A.C.,
- et al.
- Rogers T.,
- Lederman R.J.
- Bhagirath P.,
- van der Graaf M.,
- Karim R.,
- et al.
- Arujuna A.,
- Karim R.,
- Caulfield D.,
- et al.
- Celik H.,
- Ramanan V.,
- Barry J.,
- et al.
- Williams S.E.,
- Harrison J.,
- Chubb H.,
- et al.
- Harrison J.L.,
- Whitaker J.,
- Chubb H.,
- et al.
- Harrison J.L.,
- Sohns C.,
- Linton N.W.,
- et al.