| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Cardiovascular Center Bethanien (CCB), Im Pruefling 23, D-60389
Frankfurt/Main, Germany.
2 German Cancer Research Center (DKFZ), Heidelberg, Germany.
3 University of Oxford Centre for Clinical MR Research (OCMR), Oxford, United
Kingdom.
Received September 24, 2004;
accepted after revision January 4, 2005.
Address correspondence to O. K. Mohrs
(o.mohrs{at}gmx.de).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. The image quality was good for both 2D perfusion (grade 4 ± 1) and 3D turbo fast low-angle shot (FLASH) (grade 4 ± 1, n.s.). Compared with TEE, 2D perfusion, 3D turboFLASH, and the combination of both techniques yielded sensitivities of 47/35/44%, specificities of 50/67/67%, positive predictive values of 73/75/80%, and negative predictive values of 25/27/29%, respectively. The size of the thrombus was overestimated by 2D perfusion (66%) and by 3D turboFLASH (25%) and agreement for location and shape of thrombus was 50% and 75% for 2D perfusion and 75% and 50% for 3D turboFLASH, respectively. The TEE thrombus size was significantly larger in patients with true-positive diagnoses by 2D perfusion (148%) and by 3D turboFLASH (151%) when compared with patients with false-negative diagnoses (p < 0.05 for both). No such difference was found for image quality, time delay between TEE and MRI examination, and location and shape of thrombi. Contrast-enhanced MRI lacks diagnostic accuracy for the detection of thrombi in the left atrial appendage. Future technical improvements are essential to establish this technique as a noninvasive alternative to TEE.
Keywords: atrial mass • atrium • MRI • thrombus • transesophageal echocardiography
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Currently, transesophageal echocardiography (TEE) is the accepted clinical reference to detect thrombi in the left atrial appendage with high diagnostic accuracy (sensitivity, specificity, and positive and negative predictive values, 100%, 99%, and 86% and 100%, respectively) for large thrombi when compared with surgical confirmation [6, 7]. TEE, however, is semiinvasive, and the diagnosis and estimation of the size of a thrombus in the left atrial appendage remain challenging because of its complex anatomy, with its variable configuration of lobes and branches [8]. In contrast, cardiovascular MRI allows noninvasive evaluation of cardiac morphology and function without geometric assumptions [9]. Recent studies have shown the reliability of MRI to detect left ventricular thrombus [10, 11], and high diagnostic accuracy was suggested for the detection of thrombi in the left atrial appendage using unenhanced MRI [12]. However, unenhanced MRI overestimates thrombus size by an average of 23% and might be feasible only for thrombi of certain age—that is, thrombi exhibiting a certain degree of organization.
Contrast-enhanced MRI techniques might be advantageous in determining the presence and size of thrombi in the left atrial appendage, similar to the superiority of contrast-enhanced over unenhanced MRI for the detection of left ventricular thrombi, particularly of small thrombi. For MRI to be established as an alternative to TEE, important sources of systemic embolization, such as atrial thrombi and patent foramen ovale, would have to be detectable. As shown recently, MRI is capable of depicting patent foramen ovale [13], but its potential to detect atrial thrombi has not yet been investigated, to our knowledge. We aimed to assess the ability of contrast-enhanced MRI to depict thrombi in the left atrial appendage and to compare its diagnostic accuracy with that of TEE.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Study Protocol
Each patient underwent TEE and MRI afterward with a time delay of 1
± 2 days (mean ± SD). After oral administration of 0.02 mg of
lidocaine (Xylocaine Spray, AstraZeneca), TEE was performed with a 5-MHz
phased multiplanar probe (Vingmed System Five, GE Healthcare) providing an
in-plane resolution of 0.4 x 2.1 mm2. Images were obtained in
standard imaging planes, as recommended by the guidelines of the American
Society of Echocardiography
[14].
All MRI examinations were performed on a 1.5-T MR scanner (Magnetom Sonata Maestro Class, Siemens Medical Solutions). For signal detection, the combination of a six-channel body phased-array coil and a two-channel spine phased-array coil was used. Because of the variability in cycle lengths in patients with atrial fibrillation, we adapted our ECG triggering (Magnitude 3150, In-Vivo Research) according to the individual minimal cycle length over a short period of time. In patients with a minimal cycle length of 600 msec, every heart cycle was used for image reconstruction prospectively. In patients with a cycle length of less than 600 msec, we used a doubled R-R interval as the acquisition window. An ECG-gated segmented fast imaging with steady-state free precession (trueFISP) cine sequence (TR/TE, 2.7/1.2; temporal resolution, 34 msec; voxel size, 1.7 x 1.3 x 6.0 mm3) was applied for localization of angulated axial and sagittal planes and for evaluation of left atrial appendage anatomy.
During administration of a 10-mL bolus of gadolinium diethylene triamine pentaacetic acid ([Magnevist, Schering] mean dose, 0.06 ± 0.01 [SD] mmol/kg of body weight) followed by 20 mL of saline (rate, 5 mL/sec) into an antecubital vein, 40 consecutive images were acquired in the same two angulated planes using a contrast-enhanced 2D perfusion study (saturation-recovery trueFISP sequence); the parameters were a TE of 2.7 msec, an inversion time (TI) of 217 msec, a flip angle of 50°, temporal resolution of 832 msec, and voxel size of 1.8 x 1.4 x 6.5 mm3. Immediately after the administration of an additional 10 mL of gadolinium diethylene triamine pentaacetic acid (i.e., total mean dose of gadolinium diethylene triamine pentaacetic acid, 0.13 ± 0.02 mmol/kg of body weight), a 3D turbo fast low-angle shot (FLASH) sequence was acquired in the same planes using the following parameters: TE of 1.5 msec, typical minimum acquisition window of 451 msec, flip angle of 10°, matrix of 152 x 256, and voxel size of 1.9 x 1.4 x 4.0 mm3. TI was optimized (typical values, 200–350 msec).
Data Analysis
All TEE images were recorded on videotape and evaluated by consensus of two
experienced observers blinded to diagnosis and MRI data. Thrombus was
identified as an echo-dense mass and categorized according to its location,
shape, and size (Table 1). The
proximal aspect of a thrombus in the left atrial appendage was located in its
ostium, body, or apex. According to the classification of Abe et al.
[15], the shape of the
thrombus was divided into three types: mobile ball type, the thrombus is
shaped like a ball and is mobile with the heartbeat; fixed ball type, the
thrombus is shaped like a ball, but is not mobile with the heartbeat; and
mountain type, the thrombus is shaped like a mountain that has a broad base
and is not mobile with the heartbeat. The size of the thrombus was measured by
planimetry in the view with the largest area.
|
MR images were evaluated by consensus of two experienced observers blinded to the diagnosis and the TEE findings. The MR image quality was graded as 1, not assessable; 2, poor; 3, moderate; 4, good; or 5, excellent. Two-dimensional perfusion, 3D turboFLASH, and the combination of both sequences were analyzed at three separate time points with at least 7 days between sessions. Low-signal-intensity filling defects in the left atrial appendage were interpreted as thrombus, and MRI perfusion and 3D turboFLASH images were analyzed for thrombus location, shape, and size, as described earlier.
Statistical Analysis
All data are presented as mean ± SD. Sensitivity, specificity, and
positive and negative predictive values were calculated in the standard
manner. TEE and contrast-enhanced MRI size determination was tested for
association using Pearson's correlation coefficient. The paired Student's
t test was used to test for differences between TEE and
contrast-enhanced MRI thrombus size and differences between 2D perfusion and
3D turboFLASH sequences in image quality. The chi-square test (nominal data,
such as location and shape) and the Student's t test for unpaired
samples (continuous data, such as size of thrombus, image quality, and time
delay between TEE and MRI examinations) were applied to test for differences
between patients with true-positive and false-negative MRI results to identify
causes for false-negative results. A p value of less than 0.05 was
considered statistically significant. All computations were performed using
SPSS software (version 11.5, Statistical Package for the Social Sciences).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Detection of Thrombus in the Left Atrial Appendage
Using contrast-enhanced 2D perfusion imaging—Seventeen (74%)
of the 23 assessable patients showed a thrombus in the left atrial appendage
on TEE. MRI detected eight of the 17 thrombi and correctly excluded thrombus
in three of six patients. Two-dimensional perfusion imaging yielded a
sensitivity of 47%, a specificity of 50%, a positive predictive value of 73%,
and a negative predictive value of 25%.
Of the eight patients with a true-positive diagnosis (Figs. 1A, 1B, and 1C) by 2D perfusion imaging, the location of four thrombi (50%) and the shape of six thrombi (75%) were correctly classified. Two-dimensional perfusion imaging (247 ± 250 mm2) overestimated the size of thrombus by 66% compared with TEE measurements (149 ± 115 mm2, not significant). This resulted in a moderate correlation (r = 0.65, p = 0.08). The three false-positive results in 2D perfusion imaging measured small thrombi (28 ± 18 mm2). Small thrombus size measured by TEE was associated with false-negative thrombi diagnosis by 2D perfusion imaging (Figs. 2A, 2B, and 2C). The TEE-measured size of the thrombus was 148% higher in patients with true-positive (149 ± 115 mm2) compared with those with false-negative (60 ± 45 mm2, p < 0.05) results. There were no significant differences between patients with true-positive and false-negative results for location of thrombus, shape of thrombus, image quality, and time delay between TEE and MRI examinations (Figs. 3A, 3B, 3C, and Table 2).
|
|
|
|
|
|
|
|
|
|
Using contrast-enhanced 3D turboFLASH imaging—Seventeen (74%) of the 23 assessable patients showed a thrombus in the left atrial appendage on TEE. MRI detected six of 17 thrombi and correctly excluded thrombus in four of six patients. Three-dimensional turboFLASH imaging yielded a sensitivity of 35%, a specificity of 67%, a positive predictive value of 75%, and a negative predictive value of 27%.
Of the six patients with a true-positive diagnosis by 3D turboFLASH imaging, the location of four thrombi (75%) and the shape of three thrombi (50%) were correctly classified. Three-dimensional turboFLASH imaging (214 ± 220 mm2) overestimated the size of thrombus by 25% compared with TEE measurements (171 ± 119 mm2, not significant). This resulted in a strong correlation (r = 0.88, p = 0.02). The two false-positive results in 3D turboFLASH imaging measured small thrombi (49 ± 10 mm2). Small thrombus size measured by TEE was associated with false-negative thrombi diagnosis by 3D turboFLASH imaging. The TEE-measured size of the thrombus was 151% higher in patients with true-positive (171 ± 119 mm2) compared with those with false-negative results (68 ± 52 mm2, p < 0.05). There were no significant differences between patients with true-positive and false-negative results for location of thrombus, shape of thrombus, image quality, and time delay between TEE and MRI examinations (Table 2).
Using both contrast-enhanced techniques combined—Eighteen (75%) of the 24 assessable patients showed a thrombus in the left atrial appendage on TEE. MRI detected eight of 18 thrombi and correctly excluded thrombus in four of six patients. The combination of 2D perfusion imaging and 3D turboFLASH imaging yielded a sensitivity of 44%, a specificity of 67%, a positive predictive value of 80%, and a negative predictive value of 29%.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The image quality of both the 2D perfusion and 3D turboFLASH examinations was good, despite a high percentage of patients with atrial fibrillation, which can restrict image quality. However, thrombus in the left atrial appendage is common in patients with atrial fibrillation and one possible explanation for the preserved image quality in our study might be reduced movement of the left atrial appendage in such patients [16].
Our findings indicate that the size of thrombi in the left atrial appendage may be overestimated by both contrast-enhanced MRI techniques. A possible explanation for this phenomenon could be that both MRI techniques have a poorer spatial resolution than TEE—by a factor of 3. The 2D and 3D nature of contrast-enhanced imaging techniques could explain the difference of size estimation between those two MRI sequences. Other potential explanations for the discrepancies in size measurements and in diagnostic accuracy for the detection of thrombi include contrast agent dosage, effects of atrial fibrillation, imaging plane selection, and coverage of the left atrial appendage. Ohyama and coworkers [12] reported high diagnostic accuracy for the detection of thrombi in the left atrial appendage using unenhanced MRI performed the same day as TEE. These techniques similarly overestimated the size of thrombus by 23%. The blood suppression with inversion recovery sequences used by that group was dependent on inflowing blood, and this is a potential source of error for false-positive diagnoses and apparently oversized thrombi [17, 18]. However, despite being the clinical reference for the detection of atrial thrombi, TEE might underestimate the size of particularly large thrombi because of the complex anatomy of the left atrial appendage and the 2D nature of the data. Because the determination of thrombus size measured by TEE compared with surgical specimens has not been validated, we can only report the discrepancy between TEE and contrast-enhanced MRI findings; we cannot conclude whether TEE underestimates or contrast-enhanced MRI overestimates thrombus size or both.
Our results show reasonable agreement for the determination of the location and shape of left atrial appendage thrombi among 2D perfusion imaging, 3D turboFLASH imaging, and TEE. This information is of clinical importance because the shape of thrombi in the left atrial appendage has been shown to have prognostic significance [15]. Whether unenhanced MRI allows precise localization and description of the shape of thrombi remains unclear.
Our data suggest that the size of thrombus is the single most important parameter for false-negative diagnoses by 2D perfusion imaging and 3D turboFLASH imaging. Other parameters, such as image quality, time delay between TEE and MRI investigation, and the location and shape of thrombus, do not appear to be associated with false-negative diagnoses. These findings underline the importance of sufficient spatial resolution, correct plane selection, and image coverage to detect thrombi in the left atrial appendage. Three-dimensional turboFLASH imaging is based on delayed or absent contrast agent uptake in the thrombus. Therefore, older, organized, and thus perfused thrombi might be missed because of early uptake of contrast agent. Delayed imaging using the same technique might then show an enhancing thrombus due to a delayed contrast agent washout, a phenomenon described previously [11].
False-positive diagnoses of thrombi by contrast-enhanced MRI could be caused by artifacts; marked trabeculations; or the complex anatomy, such as atypical branches, of the left atrial appendage. Similarly, previous reports indicate that even TEE is prone to false-positive diagnoses due to artifacts [19] or atypical branches [8] in the left atrial appendage.
Our findings indicate the need for improvement of current contrast-enhanced MRI techniques. The limited spatial resolution may be overcome by the implementation of parallel imaging, by the use of higher field strengths, and by future developments of MR hardware.
Limitations
Our study was designed as a feasibility study to evaluate contrast-enhanced
MRI to detect thrombi in the left atrial appendage. We therefore studied a
highly selected population with evidence of such thrombi on TEE and compared
that population with a small control group. Larger studies in an unselected
population will be necessary once some of the identified problems have been
solved or alleviated.
We are unable to provide comparative data of unenhanced to contrast-enhanced MRI. Future studies with improved contrast-enhanced MRI should investigate the diagnostic accuracy of unenhanced and of contrast-enhanced MRI.
TEE is the clinical reference to detect thrombi in the left atrial appendage. However, this technique has been shown to cause false-positive and false-negative findings when compared with the gold standards of surgery or autopsy [8, 19]. Unfortunately, we cannot provide such data and, therefore, cannot exclude the possibility that some of the false-negative or false-positive findings by MRI might have been correct.
Despite a short time interval between TEE and MRI (1 ± 2 days) and no evidence of symptomatic embolization, we cannot exclude the possibility of real differences in thrombus, its presence and size, between those two study time points. Thrombi can change over a short time period—for example, due to asymptomatic embolization from the constantly moving atrium or due to the dynamic process of clot formation and remodeling, particularly during oral anticoagulation therapy. Because Ohyama and coworkers [12] used an even shorter time interval between TEE and unenhanced MRI, this might in part explain the better diagnostic accuracy reported.
In conclusion, we have shown that contrast-enhanced cardiovascular MRI currently lacks the diagnostic accuracy for the detection of thrombi in the left atrial appendage, mainly because of insufficient spatial resolution. Implementation of parallel imaging and advanced MR hardware may help to establish this technique as a noninvasive alternative to TEE in the future.
Acknowledgments
This study contains data from the doctoral thesis by Matthias Welsner.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  O. K. Mohrs, S. E. Petersen, D. Erkapic, A. Victor, T. Schlosser, B. Nowak, G. Kauffmann, T. Voigtlaender, and H.-U. Kauczor Dynamic Contrast-Enhanced MRI Before and After Transcatheter Occlusion of Patent Foramen Ovale Am. J. Roentgenol., March 1, 2007; 188(3): 844 - 849. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
