Two-Layered Susceptibility Vessel Sign on 3-Tesla T2*-Weighted Imaging Is a Predictive Biomarker of Stroke Subtype
Background and Purpose—A susceptibility vessel sign (SVS) on 1.5-tesla (T)-T2*-weighted images may predict cardioembolism. It has also been detected in patients with large artery atherosclerosis. In patients with major vessel occlusion, the SVS was comprised 2 layers on 3T-T2*-weighted images. We assessed the efficacy of 2-layered SVS on 3T-T2*-weighted imaging scans for predicting cardioembolism.
Methods—Our study included 132 patients who had ischemic stroke within the preceding 24 hours and presented with internal carotid artery or middle cerebral artery occlusion because of cardioembolism or large artery atherosclerosis. We compared 2-layered SVS and SVS on 3T-T2*-weighted imaging scans for their sensitivity, specificity, and diagnostic odds ratio for predicting cardioembolism.
Results—We enrolled 132 patients (72 men; mean age, 74.5 years); of these, 63 (47.7%) were presented with cardioembolism. Although the sensitivity of SVS and 2-layered SVS for cardioembolism and large artery atherosclerosis was not statistically different (74.6% and 58.0%, respectively), the sensitivity of 2-layered SVS was significantly higher in patients with cardioembolism (42.9%) than those with large artery atherosclerosis (2.9%; P<0.001). The specificity and diagnostic odds ratio for 2-layered SVS for cardioembolism were 97.1% and 25.1; for SVS they were 42.0% and 2.1, respectively.
Conclusion—The specificity of 2-layered SVS for cardioembolism was high. It may be useful for predicting cardioembolism and for the management of patients with acute ischemic stroke.
- diagnostic imaging
- internal carotid artery
- intracranial embolism
- magnetic resonance imaging
- middle cerebral artery
Cardioembolism and large artery atherosclerosis (LAA) are major factors in cerebral arterial occlusion. Identification of the ischemic stroke subtype is of utmost importance for the appropriate management of patients with acute stroke1 and 24-hour ECGs and echocardiograms are required.2 Although some MRI studies showed biomarkers of ischemic stroke, for example, the vessel status or penumbral enhancement,3,4 these markers did not help to identify the stroke subtype. However, MRI predictors of cardioembolism are useful because they are objective and yield information such as the infarct volume and the site(s) of arterial obstruction.
The composition of clots in obstructed arteries varies depending on whether the embolic source is cardioembolism or LAA.5 Although the susceptibility vessel sign (SVS) on 1.5T-T2*-weighted images (T2*-WI) was reported to be associated with cardioembolism, it was also detected in some patients with LAA.6 Some patients with acute cerebral arterial occlusion manifested 2-layered SVS on 3T-T2*-WI scans. We assessed the efficacy of the 2-layered SVS on 3T-T2*-WI scan for predicting cardioembolism and examined whether its value for predicting the stroke subtype is greater than of SVS.
This study was approved by the local ethics committee. Our inclusion criteria, MRI parameters, and statistical analysis methods are provided in the online-only Data Supplement. Only consecutive patients with internal carotid artery or middle cerebral artery occlusion because of cardioembolism or LAA were included in this study. We compared SVS and 2-layered SVS observed on 3T-T2*-WI scans for their sensitivity, specificity, and diagnostic odds ratio7 for cardioembolism. SVS was defined as a hypointense signal in a vessel cistern on T2*-WI scans that was larger than the contralateral arterial diameter, 2-layered SVS as an SVS that contained a low-intensity core surrounded by a signal of higher intensity (Figure 1).
We included 132 patients (72 men; mean age, 74.5±12.1 years). Of these, 63 (47.7%) were in the cardioembolism group. Their baseline characteristics differed significantly only with respect to atrial fibrillation at the time of admission (Table I in the online-only Data Supplement). The sensitivity of SVS was not statistically different in patients with cardioembolism (74.6%) or LAA (58.0%). However, the sensitivity of 2-layered SVS was significantly higher in the cardioembolism group than in the LAA group (42.9% versus 2.9%, P<0.001; Figure 2). In addition, the specificity and diagnostic odds ratio of 2-layered SVS were superior to SVS in patients with cardioembolism (specificity, 97.1% versus 42.0%; diagnostic odds ratio, 25.1 versus 2.1; Table).
The diagnostic indices are shown in Table II in the online-only Data Supplement. Of 45 patients without known atrial fibrillation at the time of admission in cardioembolism group, 9 (20%) manifested the 2-layered SVS and were diagnosed with cardioembolism. Representative cases are shown in Figure I in the online-only Data Supplement.
Although MRI can detect cerebral major vessel occlusion because of thrombus,8 it does not identify stroke subtypes. We looked for imaging biomarkers of the stroke subtype in patients with acute ischemic stroke. Although the SVS on 1.5T-T2*-WI was detected in 25.5% of patients with LAA,6 the 2-layered SVS exhibited a higher specificity for cardioembolism than SVS (97.1% versus 74.5%) suggesting that the 2-layered SVS is useful for predicting cardioembolism.
We performed T2*-WI with a longer echo time to emphasize susceptibility differences among tissues. To our knowledge, no abnormal findings on 3T-T2*-WI with a long echo time have been reported. Our imaging parameters are useful for detecting paramagnetic materials and contributed to our findings. The appearance of SVS may depend on the amount of clots and the paramagnetic material content within the clots. The amount of clots originating from cardioembolism is thought to be larger than of those because of LAA,9 and this may contribute to the higher sensitivity of the SVS for cardioembolism than LAA reported by Cho et al.6
We did not assess the histology of the endovascularly retrieved clots. Cardioembolism clots contain more fibrin than erythrocytes and LAA clots were associated with a high percentage of erythrocytes.10 Histologically, thrombi from patients with atrial fibrillation were heterogeneous and contained more fibrin than erythrocytes while LAA clots harbored a high proportion of erythrocytes.11 Deoxyhemoglobin and hemosiderin in the erythrocytes are paramagnetic materials that result in a signal reduction on T2*-WI scans.12 We suggest that 2-layered SVS reflects the heterogeneity of paramagnetic and diamagnetic materials within the clots.
Because the sensitivity of 2-layered SVS for predicting cardioembolism was not high (42.9%), we cannot exclude cardioembolism in our patients without such SVS. However, we demonstrated that the 2-layered SVS on MRI scans was highly specific and of positive predictive value and that this sign is useful for diagnosing cardioembolism even in patients without known atrial fibrillation. Kimura et al13 reported that the appearance of SVS on the middle cerebral artery after the administration of tissue-type plasminogen activator is predictive of no early recanalization. We did not assess the value of 2-layered SVS for therapeutic decision-making and studies to analyze its use for reperfusion therapy such as the administration of tissue-type plasminogen activator are underway in our laboratory. Finally, our study was conducted on a 3T scanner and as such instruments are not available at many centers, we are conducting additional studies using 1.5T scanners.
The 2-layered SVS on 3T-T2*-WI scans can predict cardioembolism. Additional studies are needed to assess the value of this sign for the selection of the appropriate reperfusion treatment.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.007227/-/DC1.
- Received September 4, 2014.
- Revision received November 3, 2014.
- Accepted November 6, 2014.
- © 2014 American Heart Association, Inc.
- Liao J,
- Khalid Z,
- Scallan C,
- Morillo C,
- O’Donnell M
- Kidwell CS
- Liebeskind DS,
- Sanossian N,
- Yong WH,
- Starkman S,
- Tsang MP,
- Moya AL,
- et al
- Cho KH,
- Kim JS,
- Kwon SU,
- Cho AH,
- Kang DW
- Kimura K,
- Sakamoto Y,
- Aoki J,
- Iguchi Y,
- Shibazaki K,
- Inoue T