Two- and Three-Dimensional Transesophageal Echocardiography in Patient Selection and Assessment of Atrial Septal Defect Closure by the New DAS–Angel Wings Device
Initial Clinical Experience
Background Transcatheter closure of atrial septal defects (ASDs) has been feasible and successful. Two-dimensional echocardiography (2DE) was applied to patients before selection and during device deployment. Three-dimensional echocardiography (3DE) can provide unique anatomic perspectives that might aid in improving device closure of ASDs.
Methods and Results Twenty-two consecutive patients were enrolled in an initial protocol for ASD device closure by the new DAS–Angel Wings occluder device. On the basis of transesophageal (TEE) 2DE and 3DE, 13 patients were considered eligible for device closure (9 secundum ASDs and 4 with patent foramen ovale associated with a cerebral vascular accident). Maximal ASD diameter and surrounding rim tissues were compared by TEE 2DE and 3DE and with balloon sizing measurements at catheterization. ASD size measured by TEE 2DE and 3DE correlated well (y=1.0x+0.049, r=.95), with good limits of agreement. However, balloon-stretched diameter measurements were systematically larger than echocardiographic measurements. Rim tissue measurements correlated well; however, TEE 3DE could demonstrate the entire shape and perimeter of the defect. Two-dimensional imaging provided reliable information during device deployment and for closure of small ASDs. However, 3DE was superior for imaging the device, especially when abnormally placed.
Conclusions Three-dimensional imaging provides unique images and projections that were essential for understanding the spatial relationship of the device to the atrial septum. Three-dimensional echocardiography significantly enhanced our understanding of two-dimensional images and provided an imaging conceptualization that should aid in future development of device closures.
The utility of 2DE for patient selection and TEE 2DE for transcatheter device closure has been demonstrated.1 However, the characteristics of the defect and device need to be mentally conceptualized, from multiple orthogonal images, into a three-dimensional construct to best comprehend their shape, position, and spatial relationships with other anatomic structures.
Three-dimensional echocardiography provides unique en face views of ASDs.2 3 Such imaging of ASDs and devices might aid in patient selection and in assessment of device position after deployment. Enhanced imaging of the DAS–Angel Wings device, consisting of two square disks with a conjoined ring in the center,4 would require visualization of all four corners of each disk on the right and left atrial sides. Such imaging capabilities might be available with 3DE.
We report our experience in the application of 2DE and 3DE for (1) patient selection and (2) early assessment of ASD closure by the new Angel Wings device.
Twenty-two patients referred for catheter device closure of a secundum ASD, or a patent foramen ovale (PFO) with previous documented stroke, were enrolled. According to protocol, patients with a moderate secundum ASD, or a clinically suspected PFO and a positive contrast transthoracic echocardiogram, underwent TEE 2DE and 3DE studies to evaluate the morphological characteristics of the defects. Patients with a secundum ASD with a maximal diameter of 20 mm with right ventricular volume overload and at least 5 mm of tissue surrounding the defect or patients with a PFO and a positive contrast study were considered eligible for device closure. The study was performed under an FDA Investigational Device Exemption approved by the hospital institutional review board.
While the patient was under propofol anesthesia, TEE 2DE studies were performed with a multiplane probe (Hewlett Packard). We classified and measured the tissue surrounding the defect as follows (Fig 1A⇓): the superior anterior rim (SA) was the distance from the anterior border of the defect to the outer aortic wall closest to the defect; the inferior anterior (IA) rim was from the inferior border of the defect to the mid tricuspid valve annulus; the superior posterior rim (SP) was between the superior border of the defect to the midpoint of the inlet orifice diameter of the superior vena cava; and the inferior posterior rim (IP) was from the inferior border of the defect to the midpoint of the inlet orifice diameter of the inferior vena cava.
For 3DE reconstructions, the video output of the echocardiographic system was interfaced to a 3DE system (TomTec). With ECG and respiratory gating, rotational scanning was used, acquiring and storing two-dimensional images every 3 degrees. The reference position was a basal transverse plane to image the long axis of the right atrium. Digital data were stored in a conical volume, and specific cut planes were used for 3DE reconstructions to image en face the ASD and rims surrounding the defect from both the right and the left atrium,4 (Fig 1⇑, B through E). Measurement of ASD maximal and minimal diameters and surrounding rims were made on 3DE images obtained from right-sided cutting planes. The size and shape of the defect appeared to change during the cardiac cycle; therefore diameter and rim measurements were made when the defect appeared maximal in size. The ratio between the maximal and minimal diameters was obtained to approximate ASD shape.
Under general anesthesia, 13 patients underwent transcatheter closure. The catheterization procedure for deployment of the DAS–Angel Wings device (Microvena Corp) is described elsewhere.4 TEE 2DE was used to guide ASD stretched diameter measurements to ensure that the balloon was perpendicular and not oblique across the atrial septum. After device placement, TEE 2DE imaged each square disk of the device and its relationship to the atrial septum. In all patients a 3DE was performed immediately after device deployment. The corners of each disk were classified according to the anatomic position of the rims (Fig 2A⇓).
Two-dimensional and three-dimensional echocardiographic ASD maximal diameter measurements were done by blinded observers and also compared with balloon occlusive stretched diameter measurements. Data are expressed as mean±SD. Comparisons between measurements were done by paired t test for continuous variables, linear regression analysis, and calculation of the mean difference (Table⇓). Interobserver and intraobserver variability for measurement of three-dimensionally derived maximal diameters were determined in all studies. Comparison measurements were made on separate 3DE reconstructions.
Optimal TEE 2DE was performed safely in all patients, both before and during cardiac catheterization. Eighteen patients had a secundum ASD, and in 4 patients a PFO was confirmed. Three-dimensional reconstruction allowed optimal imaging of the secundum ASD in all 18 patients. En face views demonstrated the size, shape, and dynamic changes of the defect during the cardiac cycle (Fig 1⇑, B through D). Three-dimensional echocardiographic reconstruction of the atrial septum was possible in all 4 patients with a PFO. However, 3DE did not more clearly delineate the defect because of its small size and overlying septum primum tissue and therefore was not included in comparative measurements. Thirteen patients subsequently underwent device closure, 9 had a secundum ASD, and 4 had a PFO.
A very close correlation, with good limits of agreement, was found between ASD maximal diameter measurements between TEE 2DE and 3DE (Table⇑). Interobserver and intraobserver variability for 3DE maximal diameter measurements was 0.5% and 2.3%, respectively. The ratio of the maximal to minimal ASD diameter by 3DE was 1.4, approximating an oval defect shape. The mean maximal ASD stretched diameter (17.0±4.4 mm) was consistently larger than the maximal TEE 3DE diameter (11.6±3.1 mm, P=.01), the latter measured at the time of catheterization in the same patients (n=9). In these 9 patients, the maximal 3DE ASD diameters measured at the precatheterization TEE study correlated well with those measured at the time of catheterization (r=.94).
A close correlation was found between TEE 2DE and 3DE measurements for the SA, SP, and IA rims (Table⇑). A minimum of 5 mm was measured in all rims by 2DE and 3DE. However, the entire length of the IP rim was difficult to visualize by 2DE, since this necessitated a basal orientation with retroflexion of the probe, resulting in tangential imaging slices. Inadequate two-dimensional image acquisition resulted in suboptimal 3DE reconstructions; hence comparative measurements were not made for this rim. Transesophageal two-dimensional echocardiographic measurements of the SP rim were difficult because the entrance of the superior vena cava to the right atrium is curvilinear. To standardize this measurement in all patients, the midpoint of the diameter of the inlet orifice was chosen to depict the SVC entrance. From TEE 3DE we learned to appreciate that the SA rim and the outer aortic wall appeared to blend imperceptibly from en face projections. Inadvertent incorporation of the aortic wall as the atrial septum would result in overestimation of this rim. We also appreciated from both 2DE and 3DE that after placement, in certain patients the aortic wall appears to support the anterior surfaces of this device. The long-term effects of the disks against the walls of the aorta will require close observation.
On the basis of TEE 2DE and 3DE, all 4 patients with a PFO and 6 of 9 patients with a secundum ASD were diagnosed as having an optimal device placement. Two-dimensional echocardiography was able to simultaneously visualize portions of the device appropriately placed on both sides of the atrial septum. In these 10 patients, 3DE en face views depicted simultaneously all 4 corners and corresponding edges of each square disk of the device appropriately placed on the right and left atrial sides (Fig 2B⇑ and 2C⇑).
Three patients with secundum ASDs were diagnosed as having abnormal device placement. In the first patient, TEE 2DE examination revealed the left and right atrial disks deviated inferior into the right atrium (Fig 3A⇓). Specifically, on 3DE the right atrial en face view showed that the disks were not folded and that the SA, SP, and IP corners were closely juxtaposed to the atrial septum (Fig 3B⇓). The abnormal position of the inferior edge was clearly delineated from an oblique view (Fig 3⇓, C and D). The IP corners were in optimal position; however, midway the inferior edges of both right and left disks deviated into the right atrium. This three-dimensional imaging helped to better comprehend the two-dimensional images, showing more specifically where the deviation occurred and that the IP corners of both disks were well placed. From such imaging we conjecture that this resulted from selecting a device that was too small for the particular size of ASD. Because of a significant residual shunt, the patient underwent uncomplicated surgical closure of the ASD, in which the abnormal location of the IA corners of the right and left atrial disks, as depicted by 3DE, were confirmed.
In the second patient, TEE 2DE revealed the superior and inferior corners of both right and left atrial disks deviated away from the atrial septum (Fig 4A⇓). Color flow Doppler confirmed a small residual shunt. Transesophageal 3DE showed that all four corners of the left atrial disk were abnormally displaced, with the SP, IP, and IA corners folded inward toward each other (Fig 4⇓, B and C). The SA corner was so significantly folded downward that the left atrial device appeared triangular in shape. This specific geometric configuration could not be appreciated by two-dimensional imaging. From the 3DE we learned that such two-dimensional images may imply improper unfolding of the device from the delivery system.
In the third patient, TEE 2DE revealed that the posterior portions of the right-sided disk deviated away from the interatrial septum. A small residual shunt was confirmed by Doppler. Three-dimensional echocardiography demonstrated that the right-sided SP and IP corners folded toward each other, with the left disk well juxtaposed to the atrial septal surface. Although two-dimensional imaging showed the deviation of the posterior right atrial disk away from the septum, three-dimensional imaging demonstrated the specific corners folded toward each other, implying that the device may have been too large to unfold properly in the posterior right atrium.
Transesophageal 2DE was valuable in the preselection of patients and during deployment of the new Das Angel Wings device. Three-dimensional echocardiography aided in patient selection and in final determination of device placement. Further experience will be gained as we apply this knowledge in the selection of patients, devices, and techniques for device placement. However, 3DE did create thicker imaging structures.3 Changes in opacity and threshold settings can alter diameter measurements.5 Despite these pitfalls, TEE 3DE measurements of ASD maximal diameter compared closely with TEE 2DE measurements. We realize that our study lacks a “gold standard” for ASD diameter measurement. However, the accuracy of 3DE for determination of the size, shape, and position of ASDs has been reported in an in vitro study.6
Although TEE 2DE imaging provided maximal diameter measurements, three-dimensional en face views could clearly delineate the defect shape and spatial orientation. These views could simultaneously demonstrate the maximal and minimal diameters, which in most patients indicated that the defects were oval. Further studies need to determine if defect shape, as depicted by 3DE, will affect results of device closure.
In our study, stretched diameter measurements were larger than 3DE measurements. We speculate that coexistent septum primum tissue, or thin pliable septum secundum rim tissue, could easily be distended, yielding larger balloon-stretched diameter measurements. Because of imaging artifact created by the reverberations from the distended balloon, the rims of the atrial defect could not be adequately visualized to verify this hypothesis. Nonetheless, a minimum of 5 mm of rim tissue and a balloon-stretched diameter measurement may be the only necessary anatomic information for Angel Wings device closure in small to moderate ASDs. However, future device placement will be applied to larger ASDs, unusual defect shapes, and positions. Three-dimensional imaging may not only be advantageous in patient selection but perhaps may be applied to custom production of devices.
Transesophageal 2DE appeared to be adequate for assessment of the device when successfully placed. However, only a linear aspect of each disk could be simultaneously visualized on the corresponding side of the atrial septum. Additionally, one cannot exactly determine where the two-dimensional ultrasound beam insonates the square disk, other than referencing the cite to other anatomic landmarks. Furthermore, two-dimensional imaging does not allow for visualization of the edges that merge to create a corner. Hence, multiple orthogonal images were necessary to appreciate, albeit indirectly, the placement of the corners and edges of each disk. True, simultaneous depiction of the edges that merge to create a corner require a volume-based data set. Direct en face three-dimensional views could simultaneously and directly image all four corners and corresponding edges of either disk from the right or left atrial septal surface, almost exactly as the device appears in spatial reality. Direct visualization of the entire edges of both squares of the device and their proximity to the atrial septum could be seen by additional unique cutting planes and projections.
Three-dimensional echocardiography was far superior for determining abnormal device placement. Two-dimensional imaging could only depict that portions of the disks were deviated away from the atrial septum; 3DE more directly visualized the specific abnormalities. Such imaging may provide important information as to the causality of abnormal device placement. We acknowledge that the anatomic findings were confirmed in only one patient; therefore, extensive conclusions comparing 2DE and 3DE and the potential causes for abnormal placement would not be appropriate from this study.
From this protocol, we have learned how to apply three-dimensional conceptualization and imaging to visualize and better comprehend the anatomy of the atrial septum and to apply that understanding to device closure. However, perhaps the most important aspect of this technology was that 3DE significantly enhanced our understanding of two-dimensional imaging. From a conceptualization of the en face views, we initiated a terminology for the ASD and rims that we now apply to TEE 2DE imaging. This has provided for a universal language and understanding for ASD device closure. Three-dimensional conceptualization should be an important process in future initiatives for device closures.
Selected Abbreviations and Acronyms
|ASD||=||atrial septal defect|
- Received June 16, 1997.
- Revision received July 3, 1997.
- Accepted July 15, 1997.
- Copyright © 1997 by American Heart Association
Das GS, Voss G, Jarvis G, Wyche K, Gunther R, Wilson RF. Experimental atrial septal defect closure with a new, transcatheter, self-centering device. Circulation.. 1993;88:1754-1764.
Sugeng L, Cao QL, Delabays A, Magni G, Marx G, Ludomirsky A, Vogel A, Pandian N. Evaluation of the accuracy of new quantitative image processing methods in measuring the size of ventricular septal defects directly on three-dimensional echocardiograms and factors influencing its reliability. J Am Coll Cardiol.. 1995;25:185A. Abstract.