Comparison of Imaging Techniques to Assess Appendage Anatomy and Measurements for Left Atrial Appendage Closure Device Selection

Abstract: Background. The adequate device size selection for left atrial appendage closure is crucial to ensuring adequate implantation and for avoiding the need for multiple attempts that increase the risk of complications. Our aim was to evaluate the information obtained using different imaging techniques to select the size of the closure device in a clinical environment. Methods. Thirty-seven patients who consecutively underwent implantation of Amplatzer cardiac plug (ACP) devices were studied. All patients were examined using computed tomography (CT) prior to intervention. Measurements were compared to those obtained using intraoperative transesophageal echocardiography (IOTEE) and angiography. Size was determined by the longest axis of the appendage ostium. The influence of all techniques on the correct selection of final size was assessed. Results. The measurements taken using the three techniques agreed in only 21.6% of the cases, leading to accurate selection of device size. Two techniques coincided as follows: IOTEE-CT in 45.9%, angiography-CT in 35.13%, and angiography-IOTEE in 24.3%. Measurements using CT were definitive for ACP selection in 75.7% of cases, angiography in 48.6%, and echocardiography in  51.4%. Device size was undermeasured with angiography in 35.1% of cases, and with IOTEE in 24.3%; CT overmeasured 21.6% of cases. The combination of angiography-CT was the most accurate for selection of device size. Conclusion. CT most often predicts the appropriate device size. If it fails, it usually overestimates the size. Agreement of measurements with all three techniques is the most accurate situation; when two agree, the most accurate combination is angiography and CT.

J INVASIVE CARDIOL 2014;26(9):462-467

Key words: Amplatzer occluder device, atrial appendage, computer-generated 3D imaging,

transesophageal echocardiography, angiography, magnetic resonance imaging, computed tomography

_______________________________________

Atrial fibrillation (AF) of non-valvular origin occurs in about 1%-2% of patients over 60 years old and in over 10% of patients over 80 years old.^1,2 Most of these patients need oral anticoagulants in order to avoid thromboembolic complications of their AF. However, despite recent advances in anticoagulation therapies, a considerable number of patients experience major or minor bleeding (around 3% and 15%, respectively). Moreover, these treatments are also contraindicated in other patients presenting with a history of serious hemorrhage (intracranial, digestive, etc).^3,4 Left atrial appendage (LAA) closure is a good treatment option for these groups of patients, because thrombi are formed in this structure in over 90% of cases. The prospective, randomized Protect AF study showed that appendage occlusion with the Watchman device (Boston Scientific) was, at least, non-inferior to oral anticoagulation with warfarin, and even superior to warfarin in reducing intracranial hemorrhages.⁵ 

Anatomical knowledge of LAA is particularly important for interventional cardiologists facing this new LAA closure technique. In addition, the selection of appropriate device size for implantation is a major factor for successful intervention. Size selection is not a simple issue, partially due to the high variability in the anatomical shape of the LAAs and also due to the variable information that different imaging techniques provide for ostium and appendage measurement. Although the available imaging techniques are of great value, the data obtained using different methods may not match, which may not facilitate the task of choosing the correct size. 

The aim of this article is to present our single-center experience with the Amplatzer cardiac plug (ACP) (St Jude Medical) and to study the usefulness of the three measurement techniques: intraoperative transesophageal echocardiography (IOTEE), angiography, and three-dimensional (3D) imaging (computed tomography [CT]), comparing their predictive value and reliability.   

Methods

LAA closure with ACP devices in 37 consecutive patients in our service was planned using three different imaging techniques. IOTEE and angiography were performed in all cases to obtain the necessary parameters for the selection of the size of the occluding device; additionally, images were obtained using a three-dimensional technique (CT) a few days before the intervention. The protocol of this study was approved by the Clinical Research Ethics Committee at the hospital where it was performed, and all procedures have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. All participating patients gave their informed consent. 

Cardiac CT scans were performed with a 64-slice scanner (Light Speed VCT; GE Healthcare, Inc). CT scans were not electrocardiogram (ECG)-synchronized, and the images were of good quality for anatomical analysis of the LAA with low radiation doses (range of effective doses in our patients 4-6 mSv for mean body mass index of 27 kg/m²). The cardiac scans were digitally sent to a network server, and then imported into an electrophysiology workstation with CARTO software, where they were processed with the CartoMerge Image Integration Module (Biosense Webster, Inc). A 3D reconstruction of the heart was performed, followed by segmentation of different structures to remove those not necessary, leaving only the 3D image of the left atrium (LA) with the LAA and pulmonary veins (this is unrelated to performing a CARTO-3D anatomical map of the LA). This system has been validated in previously published studies.^6,7 The ostium was measured on the long and short axes using the internal image that was considered to be the landing zone. To avoid the diagonalization of the measurement, it was corrected very strictly with the external image, so that the line was completely perpendicular to the cut in the selected plane (Figure 1). The longest axis was chosen for device selection in the analyses. 

Time required for CT acquisition, image post-processing, and obtention of measurements was 55-65 min. Transesophageal echocardiography (TEE) tests (iE33 echocardiography system; Philips Healthcare) were performed 48 hours before surgery to rule out the presence of thrombi in the LAA, but the final TEE measurements in two planes were those obtained the day of intervention with the patient anesthetized. A first scan of 40°-135° with the probe in midesophageal position was completed to obtain both ostium diameters: minor (usually between 45°-70°) and major (at approximately 135°). The measurement is usually performed in the ostium above the point corresponding to the circumflex artery, and again about 5-10 mm into the appendage, because some LAAs become wider or narrower at this level, depending on their shape and axis orientation. The 3D ultrasound measurement was available in only the last 22 patients, and provided control of the sealing of the margins. This system also permitted verification that the device was not affecting other anatomical structures, such as the mitral valve annulus or the left pulmonary vein, and to exclude complications such as pericardial effusion. Selective angiography of the LAA was performed, with a volume of contrast similar to that of a left coronary artery usually in a 25° right anterior oblique (RAO) view, caudal angulation of 20°, RAO 25°, and cranial angulation of 20°, although the reference is usually the first projection (Xcelera; Philips Medical Systems). The calibration reference was a radiopaque ball of 23 mm at the midaxillary line. 

TEE and angiography measurements were compared with those obtained using CT, with special attention to the major axis, which is usually the superoinferior (Figures 1-3). The ACP device size was selected by means of the range tables provided by the manufacturer, and based on the imaging results, the measurements were selected using the following criteria: (1) matching measurements of all three techniques (CT, angiography, and TEE); (2) if all three did not match, matching measurements of two techniques. This was not carried out in an exact mathematical manner; instead, there was a tendency toward the greater measurement if the differences were small, since according to the selection table of device sizes, the jump from one to another involves differences of tenths of a millimeter (Table 1). For practical purposes, it is important to tend toward the selection of a larger device when the measurements with the different techniques are around the size of the upper limit of the previous size, since as can be seen in Figure 3, no part of the ostium may be left uncovered. Moreover, if the device covers the larger axis adequately, there is not usually any transversal overstretching of the smaller axis; instead, the mesh of the device is compressed forward. The following examples illustrate device choice and subsequent criteria for evaluating which technique gave the most reliable prediction: (1) Angiography shows a diameter of 18.5 mm, IOTEE indicates 18.6 mm. These diameters suggest size 20 or size 22, respectively (Table 1). The diameter as measured by CT is 20 mm. Two techniques indicate device size 22, so we use this one. (2) However, if the diameter were 18.4 mm by IOTEE (size 20), the angiography was just at the upper limit for size 20 (18.5 mm) and CT indicated 20 mm (size 22), we would still choose size 22 for the practical reasons stated above. In the first example, two techniques indicate size 22: if this implantation is successful, CT and IOTEE measurements are considered correct. In the second scenario, we would consider that only CT predicted the appropriate size. 

Angiographic measurements were performed at the time of the procedure by the actual device implanter after deciding the exact implantation point, by means of a measurement system accredited by Philips (Philips Electronics NV). The same operator performed all procedures. There were no problems with any transeptal puncture, which was quick in all cases, and this had therefore no impact on other variables.   

Definition of success and satisfactory outcome of the implantation. Implantation was considered successful and satisfactory when the following criteria were fulfilled: (1) device placement with a suitable distance between the lobe and the outer disc, with the lobe in the neck area and the disc covering the LAA ostium; (2) image showing slight compression of the lobe with a pneumatic effect, but avoiding a “raspberry effect” due to excessive compression; (3) absence of flow between the LA and the appendage with no peridevice leaks or <3 mm, as measured with Doppler color flow; (4) proof of firm anchorage with persistent position in LAA by means of a sustained tug; and (5) absence of cardiac complications. The size of the finally implanted occluder was considered the correct size (variable “device”), because there was no significant residual leakage in all those cases with successful implantation, which would have indicated too small a device, nor were there any significant complications in neighboring structures, which would have indicated a too large occluder. 

Statistical analysis. All quantitative data are expressed as mean ± standard deviation. Student’s t-test was used for continuous data. The relation between any two methods was determined using Pearson’s correlation. Reliability was assessed using the intraclass correlation coefficient (ICC). The Bland-Altman method was used to measure the limits of agreement between two methods. P-value <.05 was considered for statistical significance. Statistical reliability was assessed using the intraclass correlation coefficient.   

Results 

Table 1 shows the selection of ACP device sizes according to the range of the LAA ostium diameter. The baseline characteristics of the 37 patients and the measurements obtained using the different procedures are presented in Table 2. The mean age was 74.7 ± 7.6 years and 51% were female. Atrial fibrillation was permanent in 29 patients and 8 presented with paroxysmal AF. The mean values of CHADS₂, CHA₂DS₂-VASc and HAS-BLED⁸ are also presented in Table 2. The measurements obtained with all three techniques and the appropriate device size in all 37 cases of our series are shown in Table 3. The distances measured by intraprocedure IOTEE were very similar to those of the angiographic measurements (not significant) and were somewhat shorter than the superoinferior axis by CT (P<.02), with the anteroposterior axis the shortest of all four distances (Table 2). 

 We analyzed the data on the overlap of the measurements obtained with the different techniques and the ability to predict the appropriate device size with each of them. A size different from the originally selected size had to be used in 5 cases (13.5%; a smaller size was used in 4 cases, and a bigger size was used in 1 case). As shown in Table 4, the diameter measurement of the superoinferior axis by CT helped appropriate size selection in  75.7% of cases, angiography in  48.6%, IOTEE in 51.4%, and the anteroposterior diameter by CT in 30.6%. However, this analysis is based on strictly numeric values so that the results can be rationalized on a solely medical or statistical basis, which does not precisely reflect the reality of the choice, which is somewhat intuitive, as explained in the Methods section. 

The figures obtained by two techniques matched in 8/37 patients (21.6%) with perfect device selection. Agreement was as follows: angiography and CT in 35.1%, IOTEE and CT in 45.9%, and IOTEE and angiography in 24.3%.  

Although three imaging techniques were assessed (angiography, TEE, and CT/MRI), the reliability of four measurements was analyzed: the major diameter of the ostium, as obtained by angiography, the major axis measured by TEE, and both the superoinferior and anteroposterior axes of the ostium by CT. 

The anteroposterior diameter by CT showed the lowest percentage of agreement with the appropriate size, but in those cases in which the anteroposterior diameter was greater than or equal to the measure of the superoinferior axis (5 cases or 13.5%), 100% success with a double-checked diameter was also obtained.

The information provided by IOTEE was very similar to that obtained by angiography (means: 19.03 ± 3.37 mm vs 19.10 ± 3.67 mm; P=.80), and there was a high correlation between both techniques (Pearson correlation coefficient, 0.87; P<.001; intraclass correlation coefficient, 0.87; 95% confidence interval, 0.77-0.93). We observed that the superoinferior axis measured by CT tended to be larger than the measurements obtained from angiography (means: 19.98 ± 3.38 mm vs 19.10 ± 3.67 mm; P<.02). This tendency and the difference in the measurements between CT and angiography are reduced depending on the greater magnitude of the measurement (Figure 4, bottom). The correlation between angiography and CT (superoinferior axis) was high (Pearson correlation coefficient, 0.82; P<.001; intraclass correlation coefficient, 0.77; 95% confidence interval, 0.56-0.88) (Figure 4, top).

Multivariate analysis showed that the combination of techniques that provided more reliable prediction of the correct device size was ostium measurement of the right caudal oblique projection by angiography and measurement of the longest axis by CT (Figure 4).

Discussion  

Importance of the proper device selection and integration of different techniques. 

Role of third 3D imaging techniques (CT or MRI). As the interventional group’s experience increases, complications during implantation are reduced,^9-11 but this experience-based success could benefit from detailed knowledge of the reliability of imaging techniques and the limitations of measurements used. The irregularity of LAA anatomy causes discrepancies in measurements of ostium diameters obtained using the various techniques. With the usual two-dimensional TEE technique, at least two perpendicular planes have to be measured to obtain the long superoinferior and short anteroposterior diameters. However, the approach does not guarantee that these axes are measured at their maximum length, the implementation of 3D echocardiography has provided only a partial solution to this problem because the calculations are influenced by the 2D measurements, and angiography is dependent on the selected projection and the proper calibration of measurements with a reference. In our experience, the use of a radiopaque ball 25-30 mm in diameter is a more accurate reference than the catheter, which may enhance calibration errors. On the other hand, we have observed that measurements obtained in a caudal RAO projection with approximately 20°/20° angulation correlate well with the values of the major diameter by CT. An interesting and recently published study by Nietlispach et al shows good results in device size selection using only angiography. However, they present higher percentages of device embolization and greater need to add another device in order to close residual shunts than in our series.¹²  

These last years have seen excellent publications on imaging studies describing LAA, especially with the use of MRI and CT and their potential usefulness in occluder device implantation.^13-15 However, none of these imaging studies was followed by actual device implantation, and our series presents this additional value. These previous reports already comment that the ostium measurements for adequate selection of device size are at times difficult, as the border between LA and LAA may not be clearly defined. This situation leads to an arbitrary definition of the LAA ostium, especially in some anatomical types of appendage. In our experience, this is often the case in those with a greater limbus originating from a thin and marked lateral ridge.¹⁶ The measurements obtained by CT are slightly longer than those obtained by TEE and angiography.¹⁵ These studies also note that the area of the LAA ostium is larger in patients with permanent atrial fibrillation than in those with paroxysmal or persistent fibrillation, and that the greater the LA, the greater the LAA ostium.^13,17 As shown in Table 3, the measurement of the major diameter by CT was most often in accordance with the correct device size. In our series, when the results of the three techniques coincided or the lengths of the superoinferior and anteroposterior axis obtained by CT were the same (13 of 37 patients; 35.13%), the estimation was 100% correct. A size different from the originally selected size had to be used in 5 cases.¹⁸ The interim report of the European ACP Registry indicated that a size different from the one actually needed had been preselected in 17% of cases,¹⁰ and other studies report this situation in as many as 24% of procedures (thus requiring from 1 to 4 attempts).¹⁹ Although the actual size needed is larger than the one preselected in more than half of the cases, the new size is smaller in a not inconsiderable number of interventions, reflecting the complexity in selecting an appropriate diameter. 

These data do not diminish the great importance of TEE as the main measurement technique, which is increasingly becoming the routine reference with the incorporation of 3D reconstruction. However, the integration of the measurements obtained by angiography, 2D or 3D TEE, and MRI or CT may be useful in some circumstances. Although logistical reasons do not allow this in some centers, this procedure can be recommended in complex cases in which the first implantation attempt failed or in those featuring a high discrepancy between measurements. Non-ECG synchronized CT studies result in radiation dose exposure of approximately one-third the dose used for ECG-synchronized cardiac CT (4-6 mSv for a mean body mass index of 27 kg/m² compared to 12-15 mSv, respectively). This, together with the mature or older age of these patients, makes radiation dose of less importance than in other study and population types. 

The comparison of our results with the scant data available in the literature on the subject cannot prove a reduced percentage of required calculations and size changes during the interventions or fewer cases of embolization. We are aware that a randomized study would be appropriate, comparing two groups, one with measurements only by angiography and TEE and another also including 3D imaging, in order to determine whether the use of the latter results in lower percentage of reattempts and different device size selection. 

Acknowledgments. The authors would like to thank Dr Blanca Piedrafita of Medical Statistics Consulting (Valencia, Spain) for her editorial and writing assistance.  

References 

1. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) study. JAMA. 2001;285(18):2370-2375.
2. Cea-Calvo L, Redon J, Lozano JV, et al. Prevalence of atrial fibrillation in the Spanish population aged 60 years or more. The PREV-ICTUS study. Rev Esp Cardiol. 2007;60(6):616-624.
3. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361(12):1139-1151 (Epub 2009 Aug 30).
4. Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365(10):883-891 (Epub 2011 Aug 10).
5. Holmes DR, Reddy VY, Turi ZG, et al. Percutaneous closure of the left atrial appendage versus warfarin therapy for prevention of stroke in patients with atrial fibrillation: a randomised non-inferiority trial. Lancet. 2009;374(9689):534-542.
6. Sra J. Registration of three dimensional left atrial images with interventional systems. Heart. 2005;91(8):1098-1104.
7. Wang Y, Di Biase L, Horton RP, Nguyen T, Morhanty P, Natale A. Left atrial appendage studied by computed tomography to help planning for appendage closure device placement. J Cardiovasc Electrophysiol. 2010;21(9):973-982.
8. Lip GY. Implications of the CHA(2)DS(2)-VASc and HAS-BLED scores for thromboprophylaxis in atrial fibrillation. Am J Med. 2011;124(2):111-114 (Epub 2010 Sep 29).
9. Park JW, Bethencourt A, Sievert H, et al. Left atrial appendage closure with Amplatzer cardiac plug in atrial fibrillation: initial European experience. Catheter Cardiovasc Interv. 2011;77(5):700-706 (Epub 2011 Mar 8).
10. Park JW, Sievert H, Schillinger W, Lickfett L, Lopez-Minguez JR, Omran H. Interim data from AMPLATZER Cardiac Plug Registry. J Am Coll Cardiol. 2011;58:Abstract TCT-112.
11. Reddy VY, Holmes D, Doshi SK, Neuzil P, Kar S. Safety of percutaneous left atrial appendage closure: results from the Watchman Left Atrial Appendage System for Embolic Protection in Patients with AF (PROTECT AF) clinical trial and the Continued Access Registry. Circulation. 2011;123(4):417-424 (Epub 2011 Jan 17).
12. Nietisplach F, Gloekler S, Krause R, et al. Amplatzer left atrial appendage occlusion: single center 10-year experience. Catheter Cardiovasc Interv. 2013;82(2):233-289 (Epub 2013 Mar 22).
13. Budge LP, Shaffer KM, Moorman JR, Lake DE, Ferguson JD, Mangrum JM. Analysis of in vivo left atrial appendage morphology in patients with atrial fibrillation: a direct comparison of transesophageal echocardiography, planar cardiac CT, and segmented three-dimensional cardiac CT. J Interv Card Electrophysiol. 2008;23(2):87-93 Epub 2008 Aug 7).
14. Heist EK, Refaat M, Danik SB, Holmvang G, Ruskin JN, Mansour M. Analysis of the left atrial appendage by magnetic resonance angiography in patients with atrial fibrillation. Heart Rhythm. 2006;3(11):1313-1318 (Epub 2006 Aug 3).
15. Krishnaswamy A, Patel NS, Ozkan A, et al. Planning left atrial appendage occlusion using cardiac multidetector computed tomography. Int J Cardiol. 2012;158(2):313-317 (Epub 2012 May 21).
16. López Mínguez JR, González Fernández R, Fernández Vegas C, et al. Anatomical classification of left atrial appendages in specimens applicable to CT imaging techniques for implantation of Amplatzer Cardiac Plug. J Cardiovasc Electrophysiol. 2014 Apr 9 [Epub ahead of print].
17. Nucifora G, Faletra FF, Regoli F, et al. Evaluation of the left atrial appendage with real-time 3-dimensional transesophageal echocardiography: implications for catheter-based left atrial appendage closure. Circ Cardiovasc Imaging. 2011;4(5):514-523 (Epub 2011 Jul 7).
18. López-Mínguez JR, Eldoayen-Gragera J, Reyes González-Fernández R, et al. Immediate and one-year results in 35 consecutive patients after closure of left atrial appendage with ACP device. Rev Esp Cardiol (Engl Ed). 2012;66(2):90-97 (Epub 2012 Aug 29).
19. Sievert H, Bayard YL. Percutaneous closure of the left atrial appendage: a major step forward. JACC Cardiovasc Interv. 2009;2(7):601-602.

_________________________________________

From the ¹Interventional Cardiology Section, Cardiology Service, Infanta Cristina University Hospital, Badajoz, Spain; and ²Department of Anatomy and Cell Biology, Medicine Faculty, Extremadura University, Spain.  

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted September 8, 2013, provisional acceptance given October 9, 2013, final version accepted November 20, 2013.

Address for correspondence: Dr José Ramón López-Mínguez, Infanta Cristina University Hospital, Badajoz, Spain. Email: lopez-minguez@hotmail.com