Research Dynamic changes of intramural hematoma in patients with acute spontaneous internal carotid artery dissection Mirjam R. Heldner1, Mila Nedelcheva1, Xin Yan1, Johannes Slotboom2, Etienne Mathier1, Justine Hulliger1, Rajeev K. Verma2, Matthias Sturzenegger1, Simon Jung1,2, Corrado Bernasconi1, Marcel Arnold1, Roland Wiest2†, and Urs Fischer1*† Background We prospectively investigated temporal and spatial evolution of intramural hematomas in patients with acute spontaneous internal carotid artery dissection using repeated magnetic resonance imaging over six-months. Aim The aim of the present study was to assess dynamic changes of intramural hematoma in patients with acute spontaneous internal carotid artery dissection at multiple follow-up time-points with T1w, PD/T2w, and magnetic resonance angiography. Methods We performed serial multiparametric magnetic resonance imaging in 10 patients with spontaneous internal carotid artery dissection on admission, at days 1, 3, 7–14 and at months 1·5, 3, and 6. We calculated the volume and extension of the hyperintense intramural hematoma using T1w and PD/T2w fat suppressed sequences and assessed the degree of stenosis due to the hematoma using magnetic resonance angiography. Results Mean interval from symptom onset to first magnetic resonance imaging was two-days (SD 2·7). Two patients presented with ischemic stroke, three with transient ischemic attacks, and five with pain and local symptoms only. Nine patients had a transient increase of the intramural hematoma volume, mainly up to day 10 after symptom onset. Fifty percent had a transient increase in the degree of the internal carotid artery stenosis on MRA, one resulting in a temporary occlusion. Lesions older than one-week were predominantly characterized by a shift from iso- to hyperintese signal on T2w images. At three-month follow-up, intramural hematoma was no longer detectable in 80% of patients and had completely resolved in all patients after six-months. Conclusions Spatial and temporal dynamics of intramural hematomas after spontaneous internal carotid artery dissection showed an early volume increase with concomitant progression of the internal carotid artery stenosis in 5 of 10 patients. Although spontaneous internal carotid artery dissection overall carries a good prognosis with spontaneous
Correspondence: Urs Fischer*, Department of Neurology, University Hospital Bern and University of Bern, Freiburgstrasse 4, CH-3010 Bern, Switzerland. E-mail: [email protected] 1 Department of Neurology, Inselspital, University Hospital Bern and University of Bern, Bern, Switzerland 2 University Institute of Diagnostic and Interventional Neuroradiology, Inselspital, University Hospital Bern and University of Bern, Bern, Switzerland Received: 22 May 2015; Accepted: 16 March 2015 †
These authors share senior authorship.
hematoma resorption in all our patients, early follow-up imaging may be considered, especially in case of new clinical symptoms. Key words: acute stroke therapy, arterial dissection, intramural hematoma, MR imaging, neurology, neuroradiology
Introduction Spontaneous cervical artery dissection (sCAD) accounts for about 2·5% of all strokes and 10–25% of ischemic strokes in young adults (1,2). Magnetic resonance imaging (MRI) has replaced conventional angiography for diagnosis of CAD and is considered as the gold standard (3). Intramural hematoma is visualized on T1w images with fat suppression technique as hyperintense, often crescent-shaped structure, adjacent to the vessel lumen and often spiraling along the length of the artery, with a very high detection sensitivity. Carotid ultrasonography and CT angiography are alternative methods to screen for CAD (4). Traditionally sCAD is considered to be a monophasic disorder with an initial intramural hematoma that decreases gradually over time. Therefore, most clinicians do not perform early follow-up imaging in patients with sCAD. However, there is increasing evidence that sCAD evokes a temporal arterial wall instability within the first three-months, with early recurrent dissections in up to 20% of patients (5). Early increase in intramural hematoma size elevates the risk of hemodynamic or embolic strokes due to stenoses or occlusions of brain supplying arteries. Therefore, early dynamic changes in intramural hematoma might influence diagnostic and therapeutic decisions. However, temporal evolution of intramural hematoma has not been investigated using multiparametric MRI from the acute stage up to resolution.
Aim The aim of the present study was to assess dynamic changes of intramural hematoma in patients with acute spontaneous internal carotid artery dissection (sICAD) at multiple follow-up timepoints with T1w, PD/T2w, and magnetic resonance angiography (MRA). We hypothesized that intramural hematoma in patients with sICAD is not a stable, but a dynamic condition which fluctuates within the first days after symptom onset.
Conflict of interest: None declared.
Methods
Funding: Swiss National Science Foundation (SPUM) and Swiss Heart Foundation.
Patients This study was based on the Bernese stroke database. From February 2006 to May 2008, 87 patients with sICAD were treated in
DOI: 10.1111/ijs.12553 © 2015 World Stroke Organization
Vol 10, August 2015, 887–892
887
Research the University Hospital Bern. Patients >18 years with sICAD confirmed by MRI were asked to participate. Exclusion criteria were: traumatic ICAD, defined as a consequence of a heavy trauma to the skull, neck, or throat; severe strokes [National Institutes of Health Stroke Scale (NIHSS) score of >20 on admission or >10 after 24 h]; contraindications for MR examinations; persisting occlusion of the internal carotid artery (ICA) during first days; time interval from symptom onset to admission >seven-days; and lack of MR capacity to perform multiple follow-up investigations or no informed consent. This study was approved by the local Ethical Committee. All patients or their relatives gave written informed consent. Details of the presenting event were recorded from the patient and their relatives, along with demographic data, risk factors, connective tissue disorders, history of trauma, local and cerebral ischemic symptoms and signs of sICAD, and time from first symptom onset. Clinical assessment was performed by neurologists using the NIHSS score (6). Some aspects of these patients have been published previously (7–9). Immediately after clinical evaluation, all patients underwent MRI and MRA to confirm sICAD. We performed follow-up images at days 1, 3, 7–14 and at months 1·5, 3, and 6 after admission using the same sequences. We categorized hemorrhages into hyperacute (3–7 days), late subacute (>7–14 days), and chronic (>14 days) (10). Therapy (antiplatelets or anticoagulants) was based on the treating physician’s decision. Follow-up information was obtained at outpatient visits or telephone interviews. Recurrent strokes and the modified Rankin Scale were recorded (11). Neuroimaging protocol MRI scans were performed both on 1·5T (Siemens Magnetom Avanto) or 3T (Siemens Magnetom Verio, Siemens Healthcare, Erlangen, Germany), according to the availability in the emergency setting. MRI protocols included an axial diffusion-weighted sequence, a time of flight angiogram, coronal T1w fat saturated turbo spin echo sequences, axial T1w fat saturated turbo spin echo sequences and double echo PD/T2w axial fat saturated spine echo sequences over the neck, and a 3-D fast low angle single shot (FLASH) sequence of the neck followed by a contrast enhanced first-pass gadolinium MRA. A detailed overview on the imaging sequence parameters is provided as supporting information. Diagnosis of sICAD was confirmed if a crescent-shaped high signal intensity within a vessel wall (intramural hematoma or double lumen) on T1w MR images with fat suppression technique was present (12–16). Measurement of intramural hematoma and assessment of the degree of stenosis Volumes of the intramural hematoma were assessed with the in-house developed software SCANalyze Version 5.1.r637 (J. S.) (17). Volume measurements were performed independently by research fellows (M. N., E. M., J. H.) and supervised by a senior neuroradiologist (R. W.), all blinded for clinical data and outcome. Intramural hematoma volume was measured using the following formula: {(cross-sectional areas of one slice demonstrating the intramural hematoma) × (slice thickness + interslice
888
Vol 10, August 2015, 887–892
M. R. Heldner et al. thickness)} × {n}, which is the number of slices demonstrating the intramural hematoma. The maximal axial extent, the maximal axial area, and the length of the intramural hematoma were determined using the same software. Moreover, the distance from the carotid bifurcation as well as the location of the distal end of the intramural hematoma were measured. Extracranial stenosis was measured with the North American Symptomatic Carotid Endarterectomy Trial technique (NASCET) by two research fellows (M. R. H., X. Y.), supervised by the senior neuroradiologist (R. W.), all blinded for clinical data (18). Classification of sICAD To investigate signal changes on axial T1w and PD/T2w fatsaturated MRI over time, two raters (R. W. and M. R. H.) categorized the signal intensities on the respective slices at the level of the maximum diameter of the vessel intramural hematoma on a semi-quantitative graduate-step intensity scale as follows: 3 = very hyperintense, 2 = moderately hyperintense, 1 = slightly hyperintense, 0 = isointense and −1 = hypointense (MR images of one patient are given in Fig. 4 as an example). Consensus reading based on shared interpretations was applied. Last but not least, the two raters (R. W. and M. R. H.) compared the rating of their consensus reading with the reports given by two independent neuroradiologists. Statistical analysis A descriptive statistical analysis of all evaluable assessments was performed. The interobserver agreement was assessed using intraclass correlation. For the analysis of volume and degree of stenosis, mean values of the raters were used. An analysis of the ability of the MRI signal to identify the age of the lesion was performed. A classifier was developed to distinguish hematomas up to the early subacute phase from older lesions (i.e. ≤ or >seven-days from the onset of symptoms) based on the composition of the MRI signal (T1, T2, PD) using logistic regression. Due to the absence of any relevant measurable lesion volume after day 45, only assessments up to that time-point were taken into consideration. The performance of the classifier was evaluated employing a receiver operating characteristics (ROC) analysis.
Results Ten patients with sICAD were included in this study (flowchart of in- and excluded patients in Fig. S1). Baseline characteristics are shown in Table 1. Mean interval from first symptom onset [pain or local symptoms only, transient ischemic attack (TIA), ischemic stroke] to first MRI was two-days (SD 2·7). Individual clinical patient data are presented in Table S2. One patient showed a new Horner’s syndrome at day eight (patient 2) and one a glossopharyngeal nerve palsy at day four (patient 4) after symptom onset. A new diffusion-weighted imaging (DWI) lesion during follow-up was observed in one patient at day nine (patient 8). All other patients had neither new DWI lesions, nor new clinical symptoms. Overall, intramural hematoma decreased in all patients during follow-up: after three-months, intramural hematoma was no longer visible in 8/10 patients, only small in the remaining two © 2015 World Stroke Organization
Research
M. R. Heldner et al.
Table 1 Baseline characteristics of patients with spontaneous dissections of the internal carotid artery Age (years) mean ± SD Female gender, n/% Fever at presentation, n/% Prior infection, n/% Respiratory Gastrointestinal Other Prior minor trauma, n/% Vascular risk factors, n/% Arterial hypertension Diabetes mellitus Hypercholesterolemia Current cigarette smoking Former cigarette smoking Known connective tissue disease Oral anticonception, n/% Migraine, n/% Without aura With aura
46 ± 10 3/30 2/20 5/50 3 0 2 3/30 1/10 0/0 5/50 2/20 3/30 0 1/10 1/10 2/20
Fig. 1 Evolution of the size of hematoma volume in individual patients with spontaneous internal carotid artery dissections.
patients, and at six-months it had resolved in all patients. An early increase in volume was detected in nine patients. Peak volume was reached at day 0 after symptom onset in one patient, at day 6 in one patient, at day 8 in three patients, at day 9 in one patient, at day 10 in three patients, and at month 1·5 in one patient. The evolution of the size of the hematoma volume in individual patients is shown in Fig. 1. Five patients showed a transitory increase in the degree of the ICA stenosis on MRA, reaching maximum values at day 3, day 4, day 8, day 9, and day 10 after symptom onset. One patient developed an ICA occlusion on day 10. Overall, the mean degree of stenosis was 70% (SD 25%) up to day 3, 75% (SD 10%) at >3–7 days, 70% (SD 30%) at >7–14 days after symptom onset, and 50% © 2015 World Stroke Organization
Fig. 2 Evolution of the degree of stenosis in individual patients with spontaneous internal carotid artery dissections.
(SD 15%) at months 1·5. The evolution of the degree of the stenosis in individual patients is shown in Fig. 2. The correlation of the hematoma volume and the degree of the stenosis is shown in Fig. S2. Interobserver agreement using intraclass correlation for measurement of the hematoma volume was 0·993 and for the degree of stenosis 0·862. The signal changes of intramural hematoma categorized by two consensus readers (R. W. and M. R. H.) and of one patient (patient 3) during follow-up are summarized in Figs 3 and 4. Lesions older than one-week were predominantly identified by a shift from iso- to hyperintense signal on T2w images. A detailed description of signal changes is provided in the supporting information. There was a disagreement in labeling signal changes in four images between the two consensus readers and the two independent neuroradiologists (in two cases: moderately vs. very hyperintense, in one case: slightly vs. moderately hyperintense, and in one case: isointense vs. slightly hyperintense). The logistic regression model of the MRI signal to identify lesions older than one-week led to a predictor with an area under the ROC of 0·85 and optimum sensitivity and specificity >80% (Fig. 3, bottom right). The coefficients (P values) of models were as follows: intercept 4·33 (P = 0·0045), T1 0·34 (P = 0·42), T2 −1·47 (P = 0·0024), PD −0·32 (P = 0·57). Addition of age and gender to the model did not improve the prediction substantially.
Discussion Traditionally, sCAD is considered as a monophasic disorder with an intramural hematoma that decreases over time. Our study shows that the intramural hematoma reveals early dynamic changes with an increase in volume in the majority of patients during the first two-weeks and with an increase in the degree of the ICA stenosis in half of patients. These findings might reflect the observations of an increased risk of recurrent TIA or stroke in the first weeks after sCAD (19). Vol 10, August 2015, 887–892
889
Research
M. R. Heldner et al.
Fig. 3 Alteration of signal intensity on T1 (top left), T2 (top right), and PD weighted images (bottom left). Receiver operating curve for sensitivity and specificity of the MRI signal to identify cervical artery dissections older than one-week (area under the curve: 0·85), bottom right.
We found a dynamic evolution of the intramural hematoma from the acute stage to the long term. At three-months, intramural hematoma was no longer detectable in 80% of patients, and at six-months in all patients. As most arterial abnormalities stabilize or resolve by three-months, and vessels that fail to reconstitute a normal lumen by six-months are highly unlikely to recover at later time-points (20), patients with acute intramural hematoma may require a closer clinical and imaging observation than those with persistent hematomas more than three-months. In our study, dynamic changes were mainly observed within the first days. After nine-days of symptom onset, no new symptoms nor new DWI lesions occurred. These findings may suggest that follow-up imaging to detect changes in hematoma volume and stenosis can be limited to the first few weeks after symptom onset, unless the patients complain of new signs or symptoms. Bachmann and coworkers described dynamic changes of intramural hematoma in patients with sICAD with a temporary increase in intramural hematoma (12). They performed follow-up MRI in 21 patients with 24 dissected arteries within two-weeks (range 7–30 days) after admission. They found an increase in stenosis in 2 (8%) arteries, no changes in 13 (54%) arteries, and a decrease in 5 (21%) arteries. Four (17%) initially
890
Vol 10, August 2015, 887–892
occluded arteries were recanalyzed at follow-up. In contrast to Bachmann and coworkers, we found an increase in the degree of ICA stenosis in 50% of patients. Given the early dynamic changes in carotid artery dissections, the increase in stenosis in Bachmann and coworkers is likely to be underestimated as they performed follow-up images mostly after two-weeks. In our study, all patients had multiple follow-up images and the majority of changes were observed within 10 days. In another study, Dittrich et al. performed MR follow-up studies in patients with sCAD after 16 days and seven-months (5). They did not focus on dynamic changes of intramural hematoma itself; however, they described recurrent, mostly asymptomatic dissections in 20% of patients within one-month and in up to 25% within sevenmonths. Both studies support our hypothesis and findings that sCAD should be considered as a dynamic disease, characterized by vessel instability during the first weeks rather than a monophasic disorder with an intramural hematoma which monotonically decreases over time. For clinical and forensic reasons, it is of major interest to classify the age of hematomas in patients with sICAD. Given the multiple follow-up images performed in the sub-/acute stage and the clinical information on time of symptom onset, our study has © 2015 World Stroke Organization
Research
M. R. Heldner et al.
Fig. 4 Signal changes of one patient (patient 3) during follow-up (first line: T1w, second line: T2w, third line: PD MR imaging).
the potential to test the ability to predict the age of the intramural hematoma based on different components of the MRI signal. The performance of a classifier to detect lesions older than one-week (i.e. at least in the late subacute phase) was good, especially for the signal change on the T2w images. Combined fat-saturated T1w and T2w imaging allows a further differentiation between different stages of hemorrhage according to the signal composition (21). Concerning signal intensity, mixed patterns of T1/T2w are frequently observed on initial and/or follow-up imaging, suggesting a dynamic evolution (12,22). Previous studies have investigated different stages of hemorrhage according to their appearance on T1 and T2w images, either in good (21) or moderate (23) in keeping with the histological classification of sCAD as proposed by Lusby and coworkers (24). A logistic regression model of the MRI signal indicated that lesions that are older than one-week can safely be predicted by changes on the T2w from isoto hyperintense signal. While intramural hematomas showed a wide variety of signal intensities on T1w sequences in the first exams, we were able to predict the subacute stage with a sensitivity and specificity of >0·80. Our study has limitations. First, as patients had a total of seven follow-up images with four of them performed within up to day 14 after symptom onset, we had to select patients willing to undergo repeated MRI. However, given our hypothesis, we focused on multiple follow-up images in a small cohort rather than on few images in a large cohort. Second, we had to exclude patients with severe strokes, as it was not possible to perform multiple follow-up images. Furthermore, patients with persisting © 2015 World Stroke Organization
occlusions of the ICA during first days were excluded. Third, although we incorporated a detailed imaging protocol, our conclusions must be considered as preliminary due to the small sample size. Our data do not justify firm conclusions regarding the need for serial imaging. Fourth, in patients with sICAD the exact time of dissection onset is always based on the first complaints of the patient, which does not necessarily correspond to the disruption of the vessel wall. Fifth, the majority of follow-up images were performed within the first 14 days after symptom onset and we therefore cannot rule out further changes of intramural hematoma shortly thereafter. Finally, our study is too small to assess the impact of heparin or aspirin on the intramural hematoma. Previous studies have shown a progression of intramural hematoma under heparin (25,26).
Conclusions No other study has investigated short and long-term changes of sICAD volumes and signal changes on both T1 and PD/T2 fs sequences with a baseline and three early follow-up images. We conclude that the intramural hematoma in patients with sICAD shows dynamic changes with a temporary increase in the volume in the vast majority of patients within the first days after symptom onset, and that a temporary increase in the degree of the ICA stenosis occurs in a relevant percentage of patients accompanied by T2 signal increases on fs sequences. After the first few weeks, MR follow-up may be of no additional value, especially after three-months, where the intramural hematoma has resolved in Vol 10, August 2015, 887–892
891
Research the majority of patients. Early follow-up imaging may be considered, especially in case of new clinical symptoms.
Author contributions M. R. Heldner and U. Fischer had full access to all the data and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design and supervision and acquisition of data: M. Sturzenegger; R. Wiest; U. Fischer. Analysis and interpretation of data: M. R. Heldner; M. Nedelcheva; X. Yan; J. Slotboom; E. Mathier; J. Hulliger; C. Bernasconi; R. Wiest; U. Fischer. Drafting of the manuscript: M. R. Heldner; C. Bernasconi; R. Wiest; U. Fischer. Critical revision of the manuscript for important intellectual content: M. R. Heldner; M. Nedelcheva; X. Yan; J. Slotboom; E. Mathier; J. Hulliger; R. K. Verma; M. Sturzenegger; S. Jung; C. Bernasconi; M. Arnold; R. Wiest; U. Fischer. Statistical analysis: M. R. Heldner; C. Bernasconi.
References 1 Lee VH, Brown RD Jr, Mandrekar JN, Mokri B. Incidence and outcome of cervical artery dissection: a population-based study. Neurology 2006; 67:1809–15. 2 Bigi S, Fischer U, Wehrli E et al. Differences in risk-factors, aetiology, and outcome between children and young adults with acute ischaemic stroke. Ann Neurol 2011; 70:245–54. 3 Brant-Zawadzki M. Routine MR imaging of the internal carotid artery siphon: angiographic correlation with cervical carotid lesions. AJR Am J Roentgenol 1990; 155:359–63. 4 Provenzale JM, Sarikaya B. Comparison of test performance characteristics of MRI, MR angiography, and CT angiography in the diagnosis of carotid and vertebral artery dissection: a review of the medical literature. AJR Am J Roentgenol 2009; 193:1167–74. 5 Dittrich R, Nassenstein I, Bachmann R et al. Polyarterial clustered recurrence of cervical artery dissection seems to be the rule. Neurology 2007; 69:180–6. 6 Lyden P, Raman R, Liu L, Emr M, Warren M, Marler J. National Institutes of Health Stroke Scale certification is reliable across multiple venues. Stroke 2009; 40:2507–11. 7 Baumgartner RW, Arnold M, Baumgartner I et al. Carotid dissection with and without ischemic events: local symptoms and cerebral artery findings. Neurology 2001; 57:827–32. 8 Arnold M, Kappeler L, Georgiadis D et al. Gender differences in spontaneous cervical artery dissection. Neurology 2006; 67:1050–2. 9 Arnold M, Pannier B, Chabriat H et al. Vascular risk factors and morphometric data in cervical artery dissection: a case-control study. J Neurol Neurosurg Psychiatry 2009; 80:232–4. 10 Gomori JM, Grossman RI, Yu-Ip C, Asakura T. NMR relaxation times of blood: dependence on field strength, oxidation state, and cell integrity. J Comput Assist Tomogr 1987; 11:684–90. 11 Van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988; 19:604–7. 12 Bachmann R, Nassenstein I, Kooijman H et al. High-resolution magnetic resonance imaging (MRI) at 3·0 Tesla in the short-term follow-up of patients with proven cervical artery dissection. Invest Radiol 2007; 42:460–6.
892
Vol 10, August 2015, 887–892
M. R. Heldner et al. 13 Schievink WI. Spontaneous dissection of the carotid and vertebral arteries. N Engl J Med 2001; 344:898–906. 14 Djouhri H, Guillon B, Brunereau L et al. MR angiography for the long-term follow-up of dissecting aneurysms of the extracranial internal carotid artery. AJR Am J Roentgenol 2000; 174:1137–40. 15 Kasner SE, Hankins LL, Bratina P, Morgenstern LB. Magnetic resonance angiography demonstrates vascular healing of carotid and vertebral artery dissections. Stroke 1997; 28:1993–7. 16 Kirsch E, Kaim A, Engelter S et al. MR angiography in internal carotid artery dissection: improvement of diagnosis by selective demonstration of the intramural haematoma. Neuroradiology 1998; 40:704–9. 17 Slotboom J, Schaer R, Ozdoba C et al. A novel method for analyzing DSCE-images with an application to tumor grading. Invest Radiol 2008; 43:843–53. 18 Staikov I, Arnold M, Mattle HP et al. Comparison of the ECST, CC, and NASCET grading methods and ultrasound for assessing carotid stenosis. European Carotid Surgery Trial. North American Symptomatic Carotid Endarterectomy Trial. J Neurol 2000; 247:681–6. 19 Biousse V, D’Anglejan-Chatillon J, Touboul PJ, Amarenco P, Bousser MG. Time course of symptoms in extracranial carotid artery dissections. A series of 80 patients. Stroke 1995; 26:235–9. 20 Schwartz NE, Vertinsky AT, Hirsch KG, Albers GW. Clinical and radiographic natural history of cervical artery dissections. J Stroke Cerebrovasc Dis 2009; 18:416–23. 21 Nedeltchev K, Bickel S, Arnold M et al. R2-recanalization of spontaneous carotid artery dissection. Stroke 2009; 40:499–504. 22 Chu B, Kampuschulte A, Ferguson MS et al. Hemorrhage in atherosclerotic carotid plaque: a high-resolution MRI study. Stroke 2004; 35:1079–84. 23 Habs M, Pfefferkorn T, Cyran CC et al. Age determination of vessel wall hematoma in spontaneous cervical artery dissection: a multisequence 3T cardiovascular magnetic resonance study. J Cardiovasc Magn Reson 2011; 28:13–76. 24 Lusby RJ, Ferrell LD, Ehrenfeld WK, Stoney RJ, Wylie EJ. Carotid plaque hemorrhage. Its role in production of cerebral ischemia. Arch Surg 1982; 117:1479–88. 25 Machet A, Fonseca AC, Oppenheim C et al. Does anticoagulation promote mural hematoma growth or delayed occlusion in spontaneous cervical artery dissections? Cerebrovasc Dis 2013; 35:175–81. 26 Dreier JP, Lurtzing F, Kappmeier M et al. Delayed occlusion after internal carotid artery dissection under heparin. Cerebrovasc Dis 2004; 18:296–303.
Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Fig S1. Flowchart of inclusion of patients. Fig S2. Correlation of the hematoma volume and the degree of the stenosis in patients with spontaneous internal carotid artery dissections. Table S1. Length of sICAD after admission. Table S2. Individual clinical patient data of 10 patients with spontaneous dissections of the internal carotid artery. Data S1. Imaging text.
© 2015 World Stroke Organization
Correspondence: Urs Fischer*, Department of Neurology, University Hospital Bern and University of Bern, Freiburgstrasse 4, CH-3010 Bern, Switzerland. E-mail: [email protected] 1 Department of Neurology, Inselspital, University Hospital Bern and University of Bern, Bern, Switzerland 2 University Institute of Diagnostic and Interventional Neuroradiology, Inselspital, University Hospital Bern and University of Bern, Bern, Switzerland Received: 22 May 2015; Accepted: 16 March 2015 †
These authors share senior authorship.
hematoma resorption in all our patients, early follow-up imaging may be considered, especially in case of new clinical symptoms. Key words: acute stroke therapy, arterial dissection, intramural hematoma, MR imaging, neurology, neuroradiology
Introduction Spontaneous cervical artery dissection (sCAD) accounts for about 2·5% of all strokes and 10–25% of ischemic strokes in young adults (1,2). Magnetic resonance imaging (MRI) has replaced conventional angiography for diagnosis of CAD and is considered as the gold standard (3). Intramural hematoma is visualized on T1w images with fat suppression technique as hyperintense, often crescent-shaped structure, adjacent to the vessel lumen and often spiraling along the length of the artery, with a very high detection sensitivity. Carotid ultrasonography and CT angiography are alternative methods to screen for CAD (4). Traditionally sCAD is considered to be a monophasic disorder with an initial intramural hematoma that decreases gradually over time. Therefore, most clinicians do not perform early follow-up imaging in patients with sCAD. However, there is increasing evidence that sCAD evokes a temporal arterial wall instability within the first three-months, with early recurrent dissections in up to 20% of patients (5). Early increase in intramural hematoma size elevates the risk of hemodynamic or embolic strokes due to stenoses or occlusions of brain supplying arteries. Therefore, early dynamic changes in intramural hematoma might influence diagnostic and therapeutic decisions. However, temporal evolution of intramural hematoma has not been investigated using multiparametric MRI from the acute stage up to resolution.
Aim The aim of the present study was to assess dynamic changes of intramural hematoma in patients with acute spontaneous internal carotid artery dissection (sICAD) at multiple follow-up timepoints with T1w, PD/T2w, and magnetic resonance angiography (MRA). We hypothesized that intramural hematoma in patients with sICAD is not a stable, but a dynamic condition which fluctuates within the first days after symptom onset.
Conflict of interest: None declared.
Methods
Funding: Swiss National Science Foundation (SPUM) and Swiss Heart Foundation.
Patients This study was based on the Bernese stroke database. From February 2006 to May 2008, 87 patients with sICAD were treated in
DOI: 10.1111/ijs.12553 © 2015 World Stroke Organization
Vol 10, August 2015, 887–892
887
Research the University Hospital Bern. Patients >18 years with sICAD confirmed by MRI were asked to participate. Exclusion criteria were: traumatic ICAD, defined as a consequence of a heavy trauma to the skull, neck, or throat; severe strokes [National Institutes of Health Stroke Scale (NIHSS) score of >20 on admission or >10 after 24 h]; contraindications for MR examinations; persisting occlusion of the internal carotid artery (ICA) during first days; time interval from symptom onset to admission >seven-days; and lack of MR capacity to perform multiple follow-up investigations or no informed consent. This study was approved by the local Ethical Committee. All patients or their relatives gave written informed consent. Details of the presenting event were recorded from the patient and their relatives, along with demographic data, risk factors, connective tissue disorders, history of trauma, local and cerebral ischemic symptoms and signs of sICAD, and time from first symptom onset. Clinical assessment was performed by neurologists using the NIHSS score (6). Some aspects of these patients have been published previously (7–9). Immediately after clinical evaluation, all patients underwent MRI and MRA to confirm sICAD. We performed follow-up images at days 1, 3, 7–14 and at months 1·5, 3, and 6 after admission using the same sequences. We categorized hemorrhages into hyperacute (3–7 days), late subacute (>7–14 days), and chronic (>14 days) (10). Therapy (antiplatelets or anticoagulants) was based on the treating physician’s decision. Follow-up information was obtained at outpatient visits or telephone interviews. Recurrent strokes and the modified Rankin Scale were recorded (11). Neuroimaging protocol MRI scans were performed both on 1·5T (Siemens Magnetom Avanto) or 3T (Siemens Magnetom Verio, Siemens Healthcare, Erlangen, Germany), according to the availability in the emergency setting. MRI protocols included an axial diffusion-weighted sequence, a time of flight angiogram, coronal T1w fat saturated turbo spin echo sequences, axial T1w fat saturated turbo spin echo sequences and double echo PD/T2w axial fat saturated spine echo sequences over the neck, and a 3-D fast low angle single shot (FLASH) sequence of the neck followed by a contrast enhanced first-pass gadolinium MRA. A detailed overview on the imaging sequence parameters is provided as supporting information. Diagnosis of sICAD was confirmed if a crescent-shaped high signal intensity within a vessel wall (intramural hematoma or double lumen) on T1w MR images with fat suppression technique was present (12–16). Measurement of intramural hematoma and assessment of the degree of stenosis Volumes of the intramural hematoma were assessed with the in-house developed software SCANalyze Version 5.1.r637 (J. S.) (17). Volume measurements were performed independently by research fellows (M. N., E. M., J. H.) and supervised by a senior neuroradiologist (R. W.), all blinded for clinical data and outcome. Intramural hematoma volume was measured using the following formula: {(cross-sectional areas of one slice demonstrating the intramural hematoma) × (slice thickness + interslice
888
Vol 10, August 2015, 887–892
M. R. Heldner et al. thickness)} × {n}, which is the number of slices demonstrating the intramural hematoma. The maximal axial extent, the maximal axial area, and the length of the intramural hematoma were determined using the same software. Moreover, the distance from the carotid bifurcation as well as the location of the distal end of the intramural hematoma were measured. Extracranial stenosis was measured with the North American Symptomatic Carotid Endarterectomy Trial technique (NASCET) by two research fellows (M. R. H., X. Y.), supervised by the senior neuroradiologist (R. W.), all blinded for clinical data (18). Classification of sICAD To investigate signal changes on axial T1w and PD/T2w fatsaturated MRI over time, two raters (R. W. and M. R. H.) categorized the signal intensities on the respective slices at the level of the maximum diameter of the vessel intramural hematoma on a semi-quantitative graduate-step intensity scale as follows: 3 = very hyperintense, 2 = moderately hyperintense, 1 = slightly hyperintense, 0 = isointense and −1 = hypointense (MR images of one patient are given in Fig. 4 as an example). Consensus reading based on shared interpretations was applied. Last but not least, the two raters (R. W. and M. R. H.) compared the rating of their consensus reading with the reports given by two independent neuroradiologists. Statistical analysis A descriptive statistical analysis of all evaluable assessments was performed. The interobserver agreement was assessed using intraclass correlation. For the analysis of volume and degree of stenosis, mean values of the raters were used. An analysis of the ability of the MRI signal to identify the age of the lesion was performed. A classifier was developed to distinguish hematomas up to the early subacute phase from older lesions (i.e. ≤ or >seven-days from the onset of symptoms) based on the composition of the MRI signal (T1, T2, PD) using logistic regression. Due to the absence of any relevant measurable lesion volume after day 45, only assessments up to that time-point were taken into consideration. The performance of the classifier was evaluated employing a receiver operating characteristics (ROC) analysis.
Results Ten patients with sICAD were included in this study (flowchart of in- and excluded patients in Fig. S1). Baseline characteristics are shown in Table 1. Mean interval from first symptom onset [pain or local symptoms only, transient ischemic attack (TIA), ischemic stroke] to first MRI was two-days (SD 2·7). Individual clinical patient data are presented in Table S2. One patient showed a new Horner’s syndrome at day eight (patient 2) and one a glossopharyngeal nerve palsy at day four (patient 4) after symptom onset. A new diffusion-weighted imaging (DWI) lesion during follow-up was observed in one patient at day nine (patient 8). All other patients had neither new DWI lesions, nor new clinical symptoms. Overall, intramural hematoma decreased in all patients during follow-up: after three-months, intramural hematoma was no longer visible in 8/10 patients, only small in the remaining two © 2015 World Stroke Organization
Research
M. R. Heldner et al.
Table 1 Baseline characteristics of patients with spontaneous dissections of the internal carotid artery Age (years) mean ± SD Female gender, n/% Fever at presentation, n/% Prior infection, n/% Respiratory Gastrointestinal Other Prior minor trauma, n/% Vascular risk factors, n/% Arterial hypertension Diabetes mellitus Hypercholesterolemia Current cigarette smoking Former cigarette smoking Known connective tissue disease Oral anticonception, n/% Migraine, n/% Without aura With aura
46 ± 10 3/30 2/20 5/50 3 0 2 3/30 1/10 0/0 5/50 2/20 3/30 0 1/10 1/10 2/20
Fig. 1 Evolution of the size of hematoma volume in individual patients with spontaneous internal carotid artery dissections.
patients, and at six-months it had resolved in all patients. An early increase in volume was detected in nine patients. Peak volume was reached at day 0 after symptom onset in one patient, at day 6 in one patient, at day 8 in three patients, at day 9 in one patient, at day 10 in three patients, and at month 1·5 in one patient. The evolution of the size of the hematoma volume in individual patients is shown in Fig. 1. Five patients showed a transitory increase in the degree of the ICA stenosis on MRA, reaching maximum values at day 3, day 4, day 8, day 9, and day 10 after symptom onset. One patient developed an ICA occlusion on day 10. Overall, the mean degree of stenosis was 70% (SD 25%) up to day 3, 75% (SD 10%) at >3–7 days, 70% (SD 30%) at >7–14 days after symptom onset, and 50% © 2015 World Stroke Organization
Fig. 2 Evolution of the degree of stenosis in individual patients with spontaneous internal carotid artery dissections.
(SD 15%) at months 1·5. The evolution of the degree of the stenosis in individual patients is shown in Fig. 2. The correlation of the hematoma volume and the degree of the stenosis is shown in Fig. S2. Interobserver agreement using intraclass correlation for measurement of the hematoma volume was 0·993 and for the degree of stenosis 0·862. The signal changes of intramural hematoma categorized by two consensus readers (R. W. and M. R. H.) and of one patient (patient 3) during follow-up are summarized in Figs 3 and 4. Lesions older than one-week were predominantly identified by a shift from iso- to hyperintense signal on T2w images. A detailed description of signal changes is provided in the supporting information. There was a disagreement in labeling signal changes in four images between the two consensus readers and the two independent neuroradiologists (in two cases: moderately vs. very hyperintense, in one case: slightly vs. moderately hyperintense, and in one case: isointense vs. slightly hyperintense). The logistic regression model of the MRI signal to identify lesions older than one-week led to a predictor with an area under the ROC of 0·85 and optimum sensitivity and specificity >80% (Fig. 3, bottom right). The coefficients (P values) of models were as follows: intercept 4·33 (P = 0·0045), T1 0·34 (P = 0·42), T2 −1·47 (P = 0·0024), PD −0·32 (P = 0·57). Addition of age and gender to the model did not improve the prediction substantially.
Discussion Traditionally, sCAD is considered as a monophasic disorder with an intramural hematoma that decreases over time. Our study shows that the intramural hematoma reveals early dynamic changes with an increase in volume in the majority of patients during the first two-weeks and with an increase in the degree of the ICA stenosis in half of patients. These findings might reflect the observations of an increased risk of recurrent TIA or stroke in the first weeks after sCAD (19). Vol 10, August 2015, 887–892
889
Research
M. R. Heldner et al.
Fig. 3 Alteration of signal intensity on T1 (top left), T2 (top right), and PD weighted images (bottom left). Receiver operating curve for sensitivity and specificity of the MRI signal to identify cervical artery dissections older than one-week (area under the curve: 0·85), bottom right.
We found a dynamic evolution of the intramural hematoma from the acute stage to the long term. At three-months, intramural hematoma was no longer detectable in 80% of patients, and at six-months in all patients. As most arterial abnormalities stabilize or resolve by three-months, and vessels that fail to reconstitute a normal lumen by six-months are highly unlikely to recover at later time-points (20), patients with acute intramural hematoma may require a closer clinical and imaging observation than those with persistent hematomas more than three-months. In our study, dynamic changes were mainly observed within the first days. After nine-days of symptom onset, no new symptoms nor new DWI lesions occurred. These findings may suggest that follow-up imaging to detect changes in hematoma volume and stenosis can be limited to the first few weeks after symptom onset, unless the patients complain of new signs or symptoms. Bachmann and coworkers described dynamic changes of intramural hematoma in patients with sICAD with a temporary increase in intramural hematoma (12). They performed follow-up MRI in 21 patients with 24 dissected arteries within two-weeks (range 7–30 days) after admission. They found an increase in stenosis in 2 (8%) arteries, no changes in 13 (54%) arteries, and a decrease in 5 (21%) arteries. Four (17%) initially
890
Vol 10, August 2015, 887–892
occluded arteries were recanalyzed at follow-up. In contrast to Bachmann and coworkers, we found an increase in the degree of ICA stenosis in 50% of patients. Given the early dynamic changes in carotid artery dissections, the increase in stenosis in Bachmann and coworkers is likely to be underestimated as they performed follow-up images mostly after two-weeks. In our study, all patients had multiple follow-up images and the majority of changes were observed within 10 days. In another study, Dittrich et al. performed MR follow-up studies in patients with sCAD after 16 days and seven-months (5). They did not focus on dynamic changes of intramural hematoma itself; however, they described recurrent, mostly asymptomatic dissections in 20% of patients within one-month and in up to 25% within sevenmonths. Both studies support our hypothesis and findings that sCAD should be considered as a dynamic disease, characterized by vessel instability during the first weeks rather than a monophasic disorder with an intramural hematoma which monotonically decreases over time. For clinical and forensic reasons, it is of major interest to classify the age of hematomas in patients with sICAD. Given the multiple follow-up images performed in the sub-/acute stage and the clinical information on time of symptom onset, our study has © 2015 World Stroke Organization
Research
M. R. Heldner et al.
Fig. 4 Signal changes of one patient (patient 3) during follow-up (first line: T1w, second line: T2w, third line: PD MR imaging).
the potential to test the ability to predict the age of the intramural hematoma based on different components of the MRI signal. The performance of a classifier to detect lesions older than one-week (i.e. at least in the late subacute phase) was good, especially for the signal change on the T2w images. Combined fat-saturated T1w and T2w imaging allows a further differentiation between different stages of hemorrhage according to the signal composition (21). Concerning signal intensity, mixed patterns of T1/T2w are frequently observed on initial and/or follow-up imaging, suggesting a dynamic evolution (12,22). Previous studies have investigated different stages of hemorrhage according to their appearance on T1 and T2w images, either in good (21) or moderate (23) in keeping with the histological classification of sCAD as proposed by Lusby and coworkers (24). A logistic regression model of the MRI signal indicated that lesions that are older than one-week can safely be predicted by changes on the T2w from isoto hyperintense signal. While intramural hematomas showed a wide variety of signal intensities on T1w sequences in the first exams, we were able to predict the subacute stage with a sensitivity and specificity of >0·80. Our study has limitations. First, as patients had a total of seven follow-up images with four of them performed within up to day 14 after symptom onset, we had to select patients willing to undergo repeated MRI. However, given our hypothesis, we focused on multiple follow-up images in a small cohort rather than on few images in a large cohort. Second, we had to exclude patients with severe strokes, as it was not possible to perform multiple follow-up images. Furthermore, patients with persisting © 2015 World Stroke Organization
occlusions of the ICA during first days were excluded. Third, although we incorporated a detailed imaging protocol, our conclusions must be considered as preliminary due to the small sample size. Our data do not justify firm conclusions regarding the need for serial imaging. Fourth, in patients with sICAD the exact time of dissection onset is always based on the first complaints of the patient, which does not necessarily correspond to the disruption of the vessel wall. Fifth, the majority of follow-up images were performed within the first 14 days after symptom onset and we therefore cannot rule out further changes of intramural hematoma shortly thereafter. Finally, our study is too small to assess the impact of heparin or aspirin on the intramural hematoma. Previous studies have shown a progression of intramural hematoma under heparin (25,26).
Conclusions No other study has investigated short and long-term changes of sICAD volumes and signal changes on both T1 and PD/T2 fs sequences with a baseline and three early follow-up images. We conclude that the intramural hematoma in patients with sICAD shows dynamic changes with a temporary increase in the volume in the vast majority of patients within the first days after symptom onset, and that a temporary increase in the degree of the ICA stenosis occurs in a relevant percentage of patients accompanied by T2 signal increases on fs sequences. After the first few weeks, MR follow-up may be of no additional value, especially after three-months, where the intramural hematoma has resolved in Vol 10, August 2015, 887–892
891
Research the majority of patients. Early follow-up imaging may be considered, especially in case of new clinical symptoms.
Author contributions M. R. Heldner and U. Fischer had full access to all the data and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design and supervision and acquisition of data: M. Sturzenegger; R. Wiest; U. Fischer. Analysis and interpretation of data: M. R. Heldner; M. Nedelcheva; X. Yan; J. Slotboom; E. Mathier; J. Hulliger; C. Bernasconi; R. Wiest; U. Fischer. Drafting of the manuscript: M. R. Heldner; C. Bernasconi; R. Wiest; U. Fischer. Critical revision of the manuscript for important intellectual content: M. R. Heldner; M. Nedelcheva; X. Yan; J. Slotboom; E. Mathier; J. Hulliger; R. K. Verma; M. Sturzenegger; S. Jung; C. Bernasconi; M. Arnold; R. Wiest; U. Fischer. Statistical analysis: M. R. Heldner; C. Bernasconi.
References 1 Lee VH, Brown RD Jr, Mandrekar JN, Mokri B. Incidence and outcome of cervical artery dissection: a population-based study. Neurology 2006; 67:1809–15. 2 Bigi S, Fischer U, Wehrli E et al. Differences in risk-factors, aetiology, and outcome between children and young adults with acute ischaemic stroke. Ann Neurol 2011; 70:245–54. 3 Brant-Zawadzki M. Routine MR imaging of the internal carotid artery siphon: angiographic correlation with cervical carotid lesions. AJR Am J Roentgenol 1990; 155:359–63. 4 Provenzale JM, Sarikaya B. Comparison of test performance characteristics of MRI, MR angiography, and CT angiography in the diagnosis of carotid and vertebral artery dissection: a review of the medical literature. AJR Am J Roentgenol 2009; 193:1167–74. 5 Dittrich R, Nassenstein I, Bachmann R et al. Polyarterial clustered recurrence of cervical artery dissection seems to be the rule. Neurology 2007; 69:180–6. 6 Lyden P, Raman R, Liu L, Emr M, Warren M, Marler J. National Institutes of Health Stroke Scale certification is reliable across multiple venues. Stroke 2009; 40:2507–11. 7 Baumgartner RW, Arnold M, Baumgartner I et al. Carotid dissection with and without ischemic events: local symptoms and cerebral artery findings. Neurology 2001; 57:827–32. 8 Arnold M, Kappeler L, Georgiadis D et al. Gender differences in spontaneous cervical artery dissection. Neurology 2006; 67:1050–2. 9 Arnold M, Pannier B, Chabriat H et al. Vascular risk factors and morphometric data in cervical artery dissection: a case-control study. J Neurol Neurosurg Psychiatry 2009; 80:232–4. 10 Gomori JM, Grossman RI, Yu-Ip C, Asakura T. NMR relaxation times of blood: dependence on field strength, oxidation state, and cell integrity. J Comput Assist Tomogr 1987; 11:684–90. 11 Van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988; 19:604–7. 12 Bachmann R, Nassenstein I, Kooijman H et al. High-resolution magnetic resonance imaging (MRI) at 3·0 Tesla in the short-term follow-up of patients with proven cervical artery dissection. Invest Radiol 2007; 42:460–6.
892
Vol 10, August 2015, 887–892
M. R. Heldner et al. 13 Schievink WI. Spontaneous dissection of the carotid and vertebral arteries. N Engl J Med 2001; 344:898–906. 14 Djouhri H, Guillon B, Brunereau L et al. MR angiography for the long-term follow-up of dissecting aneurysms of the extracranial internal carotid artery. AJR Am J Roentgenol 2000; 174:1137–40. 15 Kasner SE, Hankins LL, Bratina P, Morgenstern LB. Magnetic resonance angiography demonstrates vascular healing of carotid and vertebral artery dissections. Stroke 1997; 28:1993–7. 16 Kirsch E, Kaim A, Engelter S et al. MR angiography in internal carotid artery dissection: improvement of diagnosis by selective demonstration of the intramural haematoma. Neuroradiology 1998; 40:704–9. 17 Slotboom J, Schaer R, Ozdoba C et al. A novel method for analyzing DSCE-images with an application to tumor grading. Invest Radiol 2008; 43:843–53. 18 Staikov I, Arnold M, Mattle HP et al. Comparison of the ECST, CC, and NASCET grading methods and ultrasound for assessing carotid stenosis. European Carotid Surgery Trial. North American Symptomatic Carotid Endarterectomy Trial. J Neurol 2000; 247:681–6. 19 Biousse V, D’Anglejan-Chatillon J, Touboul PJ, Amarenco P, Bousser MG. Time course of symptoms in extracranial carotid artery dissections. A series of 80 patients. Stroke 1995; 26:235–9. 20 Schwartz NE, Vertinsky AT, Hirsch KG, Albers GW. Clinical and radiographic natural history of cervical artery dissections. J Stroke Cerebrovasc Dis 2009; 18:416–23. 21 Nedeltchev K, Bickel S, Arnold M et al. R2-recanalization of spontaneous carotid artery dissection. Stroke 2009; 40:499–504. 22 Chu B, Kampuschulte A, Ferguson MS et al. Hemorrhage in atherosclerotic carotid plaque: a high-resolution MRI study. Stroke 2004; 35:1079–84. 23 Habs M, Pfefferkorn T, Cyran CC et al. Age determination of vessel wall hematoma in spontaneous cervical artery dissection: a multisequence 3T cardiovascular magnetic resonance study. J Cardiovasc Magn Reson 2011; 28:13–76. 24 Lusby RJ, Ferrell LD, Ehrenfeld WK, Stoney RJ, Wylie EJ. Carotid plaque hemorrhage. Its role in production of cerebral ischemia. Arch Surg 1982; 117:1479–88. 25 Machet A, Fonseca AC, Oppenheim C et al. Does anticoagulation promote mural hematoma growth or delayed occlusion in spontaneous cervical artery dissections? Cerebrovasc Dis 2013; 35:175–81. 26 Dreier JP, Lurtzing F, Kappmeier M et al. Delayed occlusion after internal carotid artery dissection under heparin. Cerebrovasc Dis 2004; 18:296–303.
Supporting Information Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Fig S1. Flowchart of inclusion of patients. Fig S2. Correlation of the hematoma volume and the degree of the stenosis in patients with spontaneous internal carotid artery dissections. Table S1. Length of sICAD after admission. Table S2. Individual clinical patient data of 10 patients with spontaneous dissections of the internal carotid artery. Data S1. Imaging text.
© 2015 World Stroke Organization
