Background
Fused real-time three-dimensional transesophageal echocardiography and fluoroscopy has been used in adult patients during percutaneous mitral valve and aortic valve procedures. The use of fused echocardiographic/x-ray fluoroscopic imaging (FEX) in pediatric patients undergoing congenital heart disease catheterization has not been evaluated for feasibility and safety. The aims of this study were to assess the feasibility and safety of FEX for interventional guidance and to perform a comparison of atrial septal defect (ASD) device closure using this technology with traditional guidance methods.
Methods
Prospective evaluation of FEX in congenital cardiac interventions was conducted. A subset of patients with ASD closures were compared with patients with historical ASD closures with and without FEX. The interventionalist and echocardiographer rated the anatomic quality of the fusion imaging as (1) excellent, (2) good, or (3) poor. In addition, the utility of FEX procedural guidance was graded as (1) superior, (2) no added benefit, or (3) inferior to that of standard guidance by fluoroscopy and transesophageal echocardiography.
Results
FEX was successfully used in 26 procedures on 25 patients with congenital heart disease from January 2013 to February 2015. The median age was 9 years (range, 3–26 years), and the median weight was 29 kg (range, 16–77 kg). Twenty-six procedures were performed, including ASD closure, Fontan fenestration closure, and transcatheter valve placement in the tricuspid valve position. There was reduced fluoroscopy time and radiation dose in patients with ASDs who underwent imaging using this new technology ( P < .001 and P < .03, respectively). There were no statistically significant differences in procedural times between the two groups. Anatomic definition was rated as excellent in 20 of 26 procedures, with the remaining six rated was good. Twenty-one of 26 procedures were graded as superior (81%), and five of 26 (19%) were graded as providing no added benefit. There were no complications in any of the procedures.
Conclusions
In this early experience, FEX is feasible and safe in patients undergoing congenital heart disease catheterization and provides useful guidance in the majority of interventional procedures. There were relative reductions in fluoroscopy time and radiation dose with the use of FEX for ASD closure.
With the increased number and complexity of catheter-based interventions in congenital heart disease (CHD), there is increased collaboration between interventional and imaging cardiologists. While interventionalists have mastery of fluoroscopic images, imaging cardiologists have mastery of two-dimensional (2D) and real-time three-dimensional (3D) transesophageal echocardiographic (TEE) imaging, which displays cardiac soft tissue anatomy not seen in detail with fluoroscopic imaging. The spatial relationship of the two imaging modalities is different and uncorrelated. Left on an x-ray image might be right on an echocardiographic image. Up on an echocardiographic image might be down on an x-ray image. Each time interventional cardiologists switch their attention from one imaging modality to the other, they need to mentally reorient the images. The increasing complexity of catheter-based therapy, including device closure and percutaneous valve placement, comes with new demands from clinical and imaging perspectives. The collaborating physicians must interpret and communicate the images for each other, which at times can lead to misunderstanding of the anatomy during the procedure. Thus, a recently developed technology allows live RT 2D and 3D TEE images to be spatially registered and fused with live fluoroscopy (EchoNavigator; Philips Healthcare, Best, The Netherlands). This system therefore enables interventional cardiologists to visualize the harmonized images, while interventionalists can directly manipulate the 3D TEE images for a clearer understanding of the anatomy. These attributes should provide for more facile catheter navigation and device manipulation.
The EchoNavigator fused echocardiographic/x-ray fluoroscopic imaging (FEX) system has been used in adult patients for percutaneous mitral clip placement, atrial septal defect (ASD) device closures, transseptal puncture, left atrial appendage closure, and transcatheter aortic valve implantation. The aims of this study were to assess the feasibility, utility, and safety of using FEX for interventional guidance during CHD cardiac catheterization and to compare ASD device closure using the FEX system with traditional guidance methods of using only fluoroscopy and TEE imaging.
Methods
We prospectively enrolled subjects with CHD undergoing cardiac catheterization at Children’s Hospital Colorado between January 2013 and February 2015 who were suitable for the use of FEX. Per inclusion criteria, we enrolled subjects with CHD undergoing intracardiac interventions, such as ASD closure or dilation, Fontan fenestration closure or creation, ventricular septal defect closure, pulmonary artery or stent dilation, and transcatheter tricuspid valve-in-valve (TViV) placement, in whom we could safely pass the X7-2t 3D TEE probe (iE33; Philips Medical Systems, Andover, MA) without causing injury. Exclusion criteria were body size too small for 3D TEE probe placement and esophageal abnormalities.
The 10 subjects with ASDs who underwent ASD closure with use of FEX were compared with 20 historical ASD control subjects. Two historical control subjects were matched to one study subject. They were matched by ASD device size ± 2 mm, weight ± 10 kg, and age ± 5 years. The primary end point was fluoroscopy time. The secondary end points were procedural time and radiation dose. All procedures were performed under general anesthesia with mechanical ventilation. Signed consent was obtained for all study subjects, and assent was obtained as applicable. The study was conducted with approval from the institutional review board.
Echocardiography
During the interventional procedures, all study subjects underwent comprehensive 2D TEE and RT 3D TEE (X7-2tTEE probe) imaging using a commercially available echocardiographic system (iE33) to evaluate the respective congenital heart defects. Two-dimensional TEE imaging included standard midesophageal, transgastric, and deep transgastric views. RT 3D TEE studies included live 3D, 3D zoom mode, and full-volume wide-angle acquisition mode at different transducer positions.
Cardiac Catheterization and Intervention
Standard cardiac catheterization approaches and techniques were performed in all subjects. These approaches and techniques were based on the specific anatomy and/or intervention being performed and were not altered or deviated from on the basis of FEX.
FEX System Technology
The FEX system has smart technology that automatically finds and tracks the position of the 3D TEE probe ( Figure 1 ) in the fluoroscopic image. The system is able to automatically match and align a computer-based model of the probe head with the fluoroscopic image of the probe. By doing so, the system geometrically registers both imaging modalities. Each time fluoroscopy is activated; the system instantaneously and continuously re-registers the two modalities on the basis of the 3D TEE probe position in the fluoroscopic image. Proper registration is indicated when the system colors the head of the TEE probe green ( Figure 1 C). After registration, the spatial orientation of the 3D ultrasound field-of-view cone (the ultrasound volume) is displayed graphically in FEX’s fluoroscopic image ( Figure 1 D). This ultrasound cone overlaid on the fluoroscopic image allows a more rapid and clear understanding of the spatial relationship between the echocardiographic and fluoroscopic images by the interventionalist, which makes the interpretation and understanding of anatomic structures imaged by transesophageal echocardiography more straightforward.
In addition to tracking TEE probe movements, the FEX system will track C-arm gantry movements and produce RT images from the 3D TEE volume data set that correspond to that specific fluoroscopic projection. The system allows echocardiographic images in the same anatomic alignment as fluoroscopy from the corresponding C-arm position and will track changes in the C-arm position. This feature is called “follow C-arm” ( Figure 1 C). Changes in angulation, rotation, or position of the TEE probe are immediately registered and updated on fluoroscopic images. When fluoroscopy is idle, the system will continue to show images from the last position of active fluoroscopy. After an extended period of inactive fluoroscopy, the system will display the head of the probe in white, indicating that the system is unsure of exact registration. During fluoroscopy, the head of the probe will display in red if the systems are not registered.
An additional feature of the FEX system is the “marker feature.” Using 3D and “X-plane” orthogonal TEE images in registered FEX, specific soft tissue anatomy can be labeled, and, more important, markers can be placed ( Figure 2 , Video 1 ; available at www.onlinejase.com ). These labels and markers are then visible in all views of the 3D TEE volume data set, and they are visible on the fluoroscopic image in spatially correlated, anatomically correct locations. The markers serve as targets for more precise catheter and device manipulation.
Last, X-plane and 3D echocardiographic images can be overlaid on fluoroscopic images using the FEX system to better understand the soft tissue anatomy on fluoroscopic images ( Videos 2 and 3 ; available at www.onlinejase.com ). For the last five subjects enrolled, live fluoroscopic overlay from X-plane and 3D images was used to guide catheter manipulations and device delivery ( Videos 2 and 3 ; available at www.onlinejase.com ).
FEX Display
A feature of the FEX system that is novel to both transesophageal echocardiography and fluoroscopy is the ability to display multiple views of the live 3D TEE volume set and the live fluoroscopic view simultaneously ( Figure 1 ). These views include (1) the echo view: the3D TEE image displayed on the echocardiographic machine determined by the echocardiographer; (2) x-ray view: the fluoroscopic image seen on fluoroscopy; (3) the C-arm view: the 3D TEE image corresponding to the RT fluoroscopic projection of the C-arm gantry; and (4) the free view: a 3D TEE image of virtually any view at the discretion of the interventionalist. RT 3D TEE images of the 3D volume data set in the free and C-arm views can be manipulated by the interventionalist, such as cropping to remove portions of the anatomy that may obscure optimal visualization of specific anatomy of interest ( Figure 3 , Video 4 ; available at www.onlinejase.com ).
Grading of FEX Images
The interventional cardiologist and the echocardiographer graded the quality of the fusion images of each procedure and the ability to visualize spatial anatomic relationships; anatomic definition grades were poor, good, and excellent. The images were also graded for their overall ability to facilitate guidance of catheterization; procedure guidance grades were inferior, no added benefit, and superior compared with guidance by standard fluoroscopy and 2D or 3D TEE images alone. These images were reviewed by both the interventionalist and the echocardiographer, with a single consensus grade given to both the anatomic definition and guidance. The imaging and interventional operators prospectively judged collectively the advantages and disadvantages of the FEX system for each procedure.
Statistical Analysis
For descriptive statistics, categorical variables were summarized with counts and percentages and continuous variables were summarized with means and SDs. Chi-square tests were used for categorical variables when expected cell counts were greater than or equal to five; otherwise Fisher exact tests were used. T tests for unequal variances were used for continuous variables. Primary outcomes were analyzed using generalized estimating equations regression models to account for the 1:2 matched case-control study design. Jackknife variance estimates were used because of the small number of clusters. Data analysis was conducted using R version 3.2.3 (R Foundation for Statistical Computing, Vienna, Austria). Generalized estimating equations models were fit using the R package geepack.
Results
Feasibility of the FEX System
Thirty-one subjects were enrolled from January 2013 to February 2015. Six subjects were considered screen failures (one subject was too small to safely accept the 3D TEE probe, ASDs were deemed not amenable to device closure in three subjects, and TEE imaging was found not to be needed for the procedure in two subjects). Therefore, 25 subjects with 26 procedures were included, and FEX was feasible in all of these procedures. Table 1 shows the demographics of these subjects and their respective CHD. There were no complications in any of the procedures performed with the use of FEX.
| Variable | Value |
|---|---|
| Procedures | 26 |
| Male/female | 10:15 |
| Age (y) | 9 (3–26) |
| Weight (kg) | 29.25 (16–77) |
| ASDs | 10 |
| Tricuspid atresia s/p Fontan | 3 |
| Pulmonary atresia with IVS s/p Fontan | 2 |
| HLHS s/p Fontan | 2 |
| Ebstein’s anomaly | 2 |
| Tricuspid stenosis | 1 |
| Cor triatriatum dexter | 1 |
| Pulmonary hypertension | 1 |
| L-TGA s/p Senning and Rastelli | 1 |
| Aortic valve stenosis | 1 |
| TOF with absent RPA | 1 |
Primary and Secondary End Points
Twenty historical ASD control subjects were matched to 10 subjects with ASD who underwent FEX ( Table 2 ). The generalized estimating equations models showed significant differences between study groups for both fluoroscopy time and radiation dose. Fluoroscopy time averaged 6.01 min less for patients with ASD using FEX ( P < .001), and radiation dose averaged 5,198.3 mGy · cm 2 less in subjects with ASD undergoing FEX compared with control subjects ( Table 3 , Figure 4 ). There was no statistical difference between groups regarding age, gender, weight, and device used. There were no statistical differences in procedural time between the two groups ( Table 3 ). Although there were no statistical differences in procedural time, procedural time declined over time as the interventionalist progressed through the learning curve ( Figure 5 ). Procedural time had a missing value for one control subject, and radiation dose had missing values for 11 control subjects. The significant difference in radiation dose was sensitive to model specification. There was no significant difference if one form of variance estimation was used (sandwich or robust estimators), but there was significance with the other (jackknife estimators). Jackknife estimators are recommended when the number of clusters is <30. Given that we had only 10 clusters of three subjects each (two control subjects, one patient), the jackknife estimators were appropriate and recommended. Additionally, this result was consistent with the result of the fluoroscopy time model. Radiation dose also had 11 missing values, all in control patients. Chi-square tests and t tests were conducted using the control patients for all other variables by whether radiation dose was missing. None of these tests were significant, suggesting that the data were missing at random. Given that the jackknife estimators are recommended, that the results were consistent with another highly correlated outcome (fluoroscopy time correlation = 0.763), and that the missing data were missing at random, the radiation dose results are likely reliable and a true significant result.
| Overall | Case | Control | P | ||||
|---|---|---|---|---|---|---|---|
| n or mean | % or SD | n or mean | % or SD | n or mean | % or SD | ||
| Age (y) ( n = 30) | 7.67 | 3.55 | 8.80 | 3.05 | 7.10 | 3.71 | .1952 |
| Gender ( n = 30) | |||||||
| Female | 18 | 60.0 | 6 | 60.0 | 12 | 60.0 | 1.0000 |
| Male | 12 | 40.0 | 4 | 40.0 | 8 | 40.0 | |
| Weight ( n = 30) | 27.48 | 10.43 | 30.23 | 9.50 | 26.11 | 10.83 | .2975 |
| Device type ( n = 30) | |||||||
| ASO | 11 | 36.7 | 3 | 30.0 | 8 | 40.0 | .0147 |
| GSO | 6 | 20.0 | 5 | 50.0 | 1 | 5.0 | |
| Helex | 13 | 43.3 | 2 | 20.0 | 11 | 55.0 | |
| Device size (mm) ( n = 30) | 22.33 | 5.60 | 22.30 | 5.76 | 22.35 | 5.67 | .9823 |
| Mean | 95% CI | P | ||
|---|---|---|---|---|
| Lower | Upper | |||
| Fluoroscopy time (min) | .0005 | |||
| Control | 18.62 | 15.69 | 21.55 | |
| Case | 12.61 | 10.98 | 14.24 | |
| Procedure time (min) | .1145 | |||
| Control | 94.52 | 85.09 | 103.94 | |
| Case | 107.30 | 92.49 | 122.11 | |
| Radiation dose (mGy · cm 2 ) | .0291 | |||
| Control | 12,114.73 | 7,358.91 | 16,870.56 | |
| Case | 6,916.40 | 4,920.07 | 8,912.73 | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
