High-Resolution Double Inversion Recovery Black-Blood Imaging of Cervical Artery Dissection Using 3T MR Imaging

SUMMARY: We performed high-resolution DIR-BBI of the cervical arteries at 3T in 19 subjects with cervical dissection. It offered excellent visualization of both the lumen and arterial wall, allowing detection of the primary and secondary features of dissection. We suggest that this is a highly useful technique for diagnosis of cervical dissection, either routinely or in equivocal cases of suspected dissection. It also offers further insight into the pathogenesis of this disorder.

D IR-BBI has proved to be an excellent technique for structural assessment of the vascular system, nulling signal intensity from flowing blood and perivascular fat. When performed with a surface coil at 3T, it offers a high-resolution assessment of the cervical artery wall and lumen. The technique has been widely used for assessment of carotid atheromatous plaque, allowing plaque subcomponents and substructural changes to be identified. 1 We were interested in applying this technique to cervical artery dissection, a pathology that arises in the arterial wall, to see if DIR-BBI might either assist in the diagnosis and/or add to the known pathogenesis and imaging appearances of this condition. Dissection is increasingly recognized as an underlying cause of stroke, [2][3][4] which is of significant importance because it is treatable. Various imaging modalities have been used to diagnosis this condition, including digital subtraction angiography 5 CTA, 6 and MR imaging (MR angiography or T1-weighted fat-saturated imaging). 7 The diagnosis of dissection can be challenging for all modalities, often relying on secondary luminal findings. Only CTA and MR imaging can assess the primary site of pathology, the arterial wall, though MR imaging offers a much higher contrast resolution for detection of hemorrhage.

Methods and Technique
We performed DIR-BBI of the cervical arteries in subjects with arterial dissection. A high-resolution approach was used, with 3T MR imaging (Signa Excite; GE Healthcare, Milwaukee, Wisconsin) and a surface coil (phased-array 4-channel; Herman Flick, Flick Engineering Solution BV, the Netherlands). 2D time-of-flight angiography of the neck vessels was used to plan axial 2D BBI (T1-weighted and T2*-weighted).
We applied the technique to 19 subjects (12 men, 7 women; mean age, 47 Ϯ 16 years) with cervical artery dissection in at least 1 vessel, confirmed by CTA (64-section Brilliance CT scanner; Philips Healthcare, Best, the Netherlands) (0.625-mm sections; bolus tracking, 75 mL of contrast at 5 mL/s). Each of the 4 cervical arteries on CTA and BBI was independently reviewed for signs of dissection: 1) intimal flap or double lumen, 2) wall thickening Ϯ T1 hyperintensity of subacute hemorrhage (BBI only), and 3) stenosis, occlusion, or pseudoaneurysm. A conclusion on the presence or absence of dissection was reached for each vessel (CTA, 1 observer; MR imaging, 2 observers, by consensus), and any other potentially relevant observations were noted. Statistical analyses were performed by using the Statistical Package for the Social Sciences software (SPSS, Chicago, Illinois).

Results
Demographics, clinical findings, and brain imaging findings are summarized in Table 1. Vascular findings are summarized in Table 2. MR imaging was performed a median of 4 days after CTA (interquartile range, 1-5 days).
There was no significant difference between CTA and BBI in the number of subjects or individual vessels with dissection (subjects: CTA, n ϭ 17; BBI, n ϭ 19; vessels: CTA, n ϭ 20; BBI, n ϭ 24; Mc-Nemar test: P Ͼ 0.05). (On CTA review, 2 subjects were thought to show imaging features that could have resulted from atherosclerosis or dissection).
There was no difference in the ability of CTA and MR imaging to detect occlusion (CTA, n ϭ 8; MR imaging, n ϭ 7) or stenosis (CTA, n ϭ 13; MR imaging, n ϭ 15) (McNemar test, no significant difference). A pseudoaneurysm was seen in only 1 dissected CA, on both modalities (Fig 2 A, -B).

Discussion
This study confirms that BBI is a useful technique for detection of the pathologic findings of dissection, allowing a confident diagnosis to be made. We chose to perform this technique on subjects with known dissection, as confirmed by CTA, providing a high concentration of pathologic vessels on which to evaluate this sequence. While our study was not designed to compare the sensitivity and specificity of CTA and MR imaging in diagnosing dissection, 8 BBI did offer improved detection of intimal flaps and wall thickening, suggesting it might offer some improved sensitivity. This technique allows clear visualization of the vessel lumen: Flowing blood is saturated by inversion pulses, giving a true black-blood appearance, improved over spin-echo sequences and offering a high sensitivity to the detection of intraluminal pathology. It also offers excellent spatial and contrast resolution of the vessel wall, allowing better detection of wall thickening and hemorrhage. High T1 contrast resolution allows detection of subacute hemorrhage. We did notice some variation in the strength of T1 hyperintensity, as yet unexplained and not in our data, related to symptom duration. We included a T2*-weighted sequence to identify the low signal intensity of deoxyhemoglobin in acute dissection, though the numbers were too small to analyze.
The high-resolution DIR-BBI technique may allow us to examine the pathogenesis of dissection further. The traditional belief is that a tear occurs within the intima, allowing entry of arterial blood, creating a false plane of cleavage. This explains why patients with inherent structural collagen weakness 9,10 are more prone to the condition, but in total, this accounts for only 1%-5% of those affected; in our series only 1 patient (5%) had such connective tissue disease. Contrary to this theory is that of spontaneous rupture of the vasa vasorum. Following this initial event, dissection can extend either between the intima and media, resulting in stenosis or occlusion, or alternatively and less frequently between media and thinner adventitia, which may lead to pseudoaneurysm formation and occasionally rupture and subarachnoid hemorrhage if intracranial. 11 Spontaneous rupture of the vasa vasorum, as an underlying cause of dissection might explain why a flap is uncommonly seen at autopsy. However, we identified a flap in nearly one-third of subjects, suggesting that its presence is more common than previously thought and adding support for an intimal tear as a significant primary pathogenic process of arterial dissection.
With the high-resolution DIR-BBI technique, we found it possible to clearly differentiate intramural hematoma from  intraluminal thrombus; furthermore, we could sometimes differentiate intramedial from subintimal hemorrhage.
An observational finding commonly noted was T2 hyperintense periadventitial signal-intensity change. We postulate that during acute dissection, the artery is surrounded by a variable degree of inflammatory edema. A similar response has been demonstrated in spontaneous coronary artery dissection, 12,13 and our study confirms similar observation in cervical dissection, as recently reported. 14