Seung-Hoon Lee

Seung-Hoon Lee

A conversation with Seung-Hoon Lee, MD, PhD, Professor of Neurology, Seoul National University Hospital, about the role of the susceptibility vessel sign on SWI to predict stroke subtype and recanalization.

Interviewed by José G. Merino, MD, Associate Professor of Neurology, University of Maryland School of Medicine.

They will be discussing the paper, “Prediction of Stroke Subtype and Recanalization Using Susceptibility Vessel Sign on Susceptibility-Weighted Magnetic Resonance Imaging,” published in the June 2017 issue of Stroke.

Dr. Merino: Could you please briefly describe the study and summarize the key findings, putting them into context of what was known before you did the study?

Dr. Lee: I’m glad to talk about our research in this interview. Thrombi in the cerebral arteries appear hypointense on susceptibility-weighted MRI (SWMRI). We call them “the susceptibility vessel sign” (SVS). The methodological strength of this study is that SWI MRI is much more sensitive than GRE and thus can quantify the size of the SVS. In this study, we analyzed the relationship between the size of the SVS, the stroke mechanism, and whether successful recanalization occurred in patients receiving endovascular treatment. Cardiac emboli are large but fragile because they are rich in RBCs but have scant platelets. We hypothesized that because the SVS reflects the red blood cell component of the clot, patients with larger SVS are more likely to have a cardioembolic source and thus more likely to have successful recanalization. We found that as the SVS size increased, the probability of cardioembolic stroke was higher, but that SVS size did not show any positive or negative correlation with successful recanalization. This is probably due to the high recanalization rate with the stent-retrievers, irrespective of stroke etiology. No association between SVS size and recanalization can be partly explained by clot fragility in cardioembolic stroke.

Dr. Merino: Is the SVS size, as measured on SWMRI, larger than the actual clot size because of the susceptibility effect?

Dr. Lee: Of course, SVS has susceptibility artifacts, but comparison with actual clot size is case-dependent. This is because the size of the SVS depends not only on the size of the clot but also on its components. For example, in an RBC rich clot with high paramagnetic effect due to the iron in the hemoglobin, the SVS will appear larger than the clot. On the other hand, in a platelet-rich thrombus with low paramagnetic effect, the SVS may appear smaller than the actual clot.

Dr. Merino: What accounts for the finding that the increase in SVS diameter may be associated with cardioembolic stroke?

Dr. Lee: Cardioembolic clots may have more paramagnetic effects due to the large proportion of RBCs, so the SVS diameter may appear larger. Another reason is that the clot located in intracranial stenotic segment, which is more common in east-Asians, is relatively small; therefore, the size of SVS in this case may be small.

Dr. Merino: What pathological data supports the association of large SVS and RBC-rich cells?

Dr. Lee: Unfortunately, we did not have any pathological data regarding SVS. However, a previous study showed SVS in Gradient Echo imaging (GRE-SVS) was associated with RBC-rich thrombi (Liebeskind et al. Stroke 2011). One study also showed that GRE-SVS was associated with cardioembolism (Cho et al. Stroke 2005). Recent findings using stent-retriever showed that GRE-SVS was associated with both cardioembolism and RBC-rich clot (Kim et al, American Journal of Neuroradiology 2015). These are partly in line with our findings in this study. However, it is a limitation that all those studies used GRE imaging and only evaluated the presence of SVS, not the size of SVS. Therefore, further pathological data is needed to evaluate the association of SVS size and RBC-rich clot in SWMRI.

Dr. Merino: Would you have expected clot characteristics as visualized with SWMRI to predict recanalization? Why do you think that SVS diameter was not associated to recanalization?

Dr. Lee: Yes. We hypothesized that because the SVS reflects the RBC component of the clot, larger SVS would be found in patients with cardioembolic stroke, and they may be more likely to have successful recanalization. But SVS size did not show any positive or negative correlation with successful recanalization. We think this is because the recanalization rate after endovascular treatment in this study (78.6%) is quite high with the use of stent retrievers, regardless of the size of SVS.

Dr. Merino: Should the findings on pre-treatment SWMRI help guide the choice of etiologic work-up of stroke?

Dr. Lee: Yes. When SVS is considerably large or small in SWMRI, it may help to guide more accurate etiologic work-up.

Dr. Merino: Would clot characteristics differ if the parameters used to obtain the SWMRI differ?

Dr. Lee: The SVS may vary if the MRI parameters for obtaining the SWMRI are different. For example, in a study of cerebral microbleeds, smaller section thickness and higher magnetic field each yielded increased lesion contrast for cerebral microbleeds (Nandigam et al, Am J Neuroradiol, 2009). However, in our study, there was no significant difference in SVS presence or size according to the MR scanners.

Dr. Merino: Do all patients considered for endovascular embolectomy at your center have imaging with MRI before treatment?

Dr. Lee: Until 2016, we had used MRI based protocol, so we had taken MRI before endovascular treatment in almost all patients. However, multimodal CT based protocol has been mainly used since 2017.

Dr. Merino: As you know, most patients being considered for embolectomy in the U.S. are imaged with CT. Do you think that a dense MCA on CT may have the same predictive properties as the SVS?

Dr. Lee: A couple of studies showed that Hounsfield unit of the hyperdense sign in a noncontrast CT was associated with clot composition and post-endovascular treatment recanalization (Puig et al, American Journal of Neuroradiology, 2012; Moftakhar et al, Stroke, 2013). However, in those studies, ROIs were manually drawn and calibration was needed to obtain Hounsfield units of the clots. Since SVS in SWMRI can visualize clots more clearly, and its diameters and lengths could be easily measured, we think SVS in SWMRI is more useful than hyperdense sign in clinical settings.

Dr. Merino: Is there anything else you want to tell me that I have not asked you?

Dr. Lee: We want to give more emphasis to the methodology of our study. Conventionally, in studies using T2*-GRE, SVS was defined as presence only when the signals were larger than vessel diameter. However, since SWMRI can visualize SVS more sensitively and clearly, we could quantify SVS size even if it was smaller than vessel diameter. Our study suggests that we should consider not only presence of SVS, but also the size and shape of SVS together in the clinical practice. We think this is the way to make SWMRI more effective.

Dr. Merino: Thank you very much.