- Email: [email protected]
Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
Journal of Neuroradiology (2016) xxx, xxx—xxx
Available online at
ScienceDirect www.sciencedirect.com
ORIGINAL ARTICLE
Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T G. Zhou a, M.H. Li a, C. Lu b, Y.L. Yin c, Y.Q. Zhu a, X.E. Wei a, H.T. Lu a, Q.Q. Zheng d, W.W. Gao e,∗ a
Department of diagnostic and interventional radiology, Shanghai Jiao Tong university affiliated Sixth People’s hospital, Shanghai, China b School of radiology, Taishan medical college, Taian, Shandong, China c Department of anesthesiology, The Military general hospital of Beijing PLA, Beijing, China d Department of hematology, Shanghai Jiao Tong university affiliated Sixth People’s hospital, Shanghai, China e Department of neurosurgery, Shanghai Jiao Tong university affiliated Sixth People’s hospital, MS, No. 600, Yi Shan road, 200233 Shanghai, China
KEYWORDS Spinal dural arteriovenous fistulas; Dynamic contrast-enhanced MR angiography; Digital subtraction angiography; Angioarchitecture; Accuracy
Summary Objective: This study was undertaken to evaluate the accuracy of dynamic contrast-enhanced magnetic resonance angiography (DCE-MRA) in the precise location and demonstration of fistulous points in spinal dural arteriovenous fistulas (SDAVFs). Methods: Fifteen patients (14 men, 1 woman; age range: 40—78 years; mean: 55.5 years) harboring SDAVF who underwent preoperative DCE-MRA and spinal digital subtraction angiography (DSA) between January 2012 and January 2015 were evaluated retrospectively. Two reviewers independently evaluated the level and side of the arteriovenous fistula and feeding artery on 3 T DCE-MRA and DSA images. The accuracy of DCE-MRA was assessed by comparing its findings with those from DSA and surgery in each case. Results: All 15 patients underwent DCE-MRA and DSA. DSA was unsuccessful in two patients due to technical difficulties. All cases were explored surgically, guided by the DCE-MRA. Surgery confirmed that 14 AVF sites were located in the thoracic spine, 5 in the lumbar spine, and 1 in the cervical spine. The origin of the fistulas and feeding arteries was accurately shown by DCE-MRA in 11 of the 15 patients. DCE-MRA also detected dilated perimedullary veins in all 15 patients. Overall, DCE-MRA facilitated DSA catheterization in 10 cases. In six patients, the artery of Adamkiewicz could be observed. In 15 out of 20 fistulas (75%), both readers agreed
Abbreviations: SDAVF, Spinal dural arteriovenous fistula; DCE-MRA, Dynamic contrast-enhanced magnetic resonance angiography; DSA, Digital subtraction angiography; FOV, Field of view. ∗ Corresponding author. Tel.: +86 021 64844183; fax: +86 021 64844183. E-mail address: [email protected] (W.W. Gao). http://dx.doi.org/10.1016/j.neurad.2016.10.002 0150-9861/© 2016 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
2
G. Zhou et al. on the location on DCE-MRA images, and the coefficient of the interobserver agreement was 0.67 (95% confidence interval [CI], 0.16—0.87). In 13 of 16 shunts (75%), the DCE-MRA consensus findings and DSA findings coincided. The intermodality agreement was 0.77 (95% CI: 0.35—0.92). Conclusions: Our DCE-MRA studies benefited from the use of a high-field 3 T MR imaging unit and reliably detected and localized the SDAVF and feeding arteries. As experience with this technique grows, it may be possible to replace DSA with DCE-MRA if surgery is the planned treatment. © 2016 Elsevier Masson SAS. All rights reserved.
Introduction Spinal dural arteriovenous fistulas (SDAVF) can render devastating neurological consequences and lead to considerable morbidity. SDAVF is the most common type of spinal vascular malformation in which an arteriovenous connection is typically located within the dura mater of a nerve root sleeve, underneath the vertebral body pedicle in the neural foramen between the radicular artery and the medullary vein [1,2]. Diagnosis and treatment prior to irreversible cord ischemia or infarction are extremely important, since the symptoms at that point can be reversed. Accurately locating the fistula is of crucial importance before neurosurgery or superselective embolization. However, despite preoperative MRI and DSA, the precise localization and anatomic delineation of the fistula site remains challenging due to the complex angioarchitecture and small vascular caliber. The aim of the following article is to compare the agreement between intra-arterial DSA and DCE-MRA at 3.0 T in localizing the shunting point of SDAVFs.
Methods The protocol of this study was approved by the Ethical Committee of our institute. Between January 2012 and June 2015, a total of 15 patients (14 men and 1 woman; age range, 40—78 years; mean, 55.5 years) at a single institution were confirmed to have SDAVFs. All patients presented with congestive myelopathy (the mean length of the illness course was 6 months) and underwent MR imaging (MRI) and DSA. Our inclusion criteria were a diagnosis of SDAVF on the basis of spinal DSA scans, which was verified by surgery after spinal 3 T DCE-MRA. The fast DCE-MRA examinations were performed on an Achieva 3.0 T MRI system (Philips Healthcare, Amsterdam, The Netherlands). A testbolus (1 mL of contrast) technique positioned through the aorta at T10 vertebral body level was employed to evaluate the cycle time before examination. The contrast agent was injected through the antecubital veins using a highpressure syringe at a rate of 3 mL/s and a total volume of 20 mL (gadodiamide, 0.5 mmol/mL) based on the calculated cycle time using a test bolus injection. The technique was then implemented using a 3D radiofrequency-spoiled fast gradient-echo volume acquisition using the following parameters: TR/TE, 3.1 ms/1.1 ms; flip angle, 20◦ ; voxel size, 1.0 × 0.9 × 0.9 mm; field of view (FOV), 357 × 440 mm; four slabs (180 slices); SENSE factor, 2. The FOV was positioned to cover the entire T2 hyperintense cord and serpentine
flow voids. The placement of the FOV may also base on clinical suspicion of the location of the fistula. The data acquisition lasted 2 minutes and was performed during the contrast agent injection. The acquired image data sets were then transferred to a workstation (IntelliVue Guardian Early Warning Score; Philips Healthcare), in which reconstruction using maximum intensity projection and volume-rendering DCE-MRA images was performed using a 3D specialized software package (Volume Inspection, Philips Medical Systems, Amsterdam, The Netherlands). Fistulous points in these subjects were determined via maximum-intensity projection(MIP), volume rendering technique(VRT), multiplanar reformations (MPR) and their source images. Two neuroradiologists (W.-W. Gao and H.-T. Lu, with 25 and 10 years of experience in neuroimaging, respectively) were blinded to the subjects to analyze the images independently. They provided the diagnosis including the lesion range, feeding arteries, and fistulas. DSA was performed on a monoplanar unit (Axion Artis) with a 1024 × 1024 matrix and 17-20-cm FOV using conventional methods. The interval between DCEMRA and DSA studies ranged from 0—8 days (mean, 3 days). Two authors (W.-W. Gao and G. Zhou) compared the DCEMRA results with the DSA and surgical findings. Interobserver agreement on the angiographic findings was determined by calculating the coefficient. The accuracy of DCE-MRA was evaluated by checking whether the level and side of the AVF that it suggested corresponded with the DSA and surgical findings. All analyses were performed using the STATA 12 statistical package (StataCorp, College Station, TX, USA).
Results DCE-MRA and surgery were performed successfully in 15 patients. Intra-arterial DSA was unsuccessful in two patients due to technical difficulties, mainly related to coexisting aortic atherosclerosis and tortuosity. These two patients were treated surgically only on the basis of the DCE-MRA results. Surgery confirmed that 14 AVF sites were located in the thoracic spine, 5 in the lumbar spine, and 1 in the cervical spine (Tables 1 and 2). The locations of the feeders and fistulas identified on the DCE-MRA and DSA images are shown in Table 1. DCE-MRA demonstrated a precise fistula angioarchitectural configuration in 11 (11/15) cases. In four cases, the small shunts inside the nidus could not be individualized. The draining radicular vein could be followed in all cases from the fistula site to the coronal venous plexus around the spinal cord. DCE-MRA was effective in demonstrating the feeding artery, fistula site, and dilated perimedullary veins in most cases. In 11 (75%) patients, the
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
Consensus MRA shunt Site
DSA shunt Site
Surgery shunt Site
Level of feeder
No. of feeder
Location of AKA
1 2 3
40, M 54, M 41, M
T2-10 T4-12 T8-L1
T9-10, rt T11-12, rt Unsuccessful
T7-11 T9-12 T12-L1
T10, rt T9, rt T4, rt T6, lt T12, lt T12, rt T9, lt T12, lt
L1, lt T9, rt Cannot find
63, F 42, M 66, M
1 1 2
T10, rt L2, lt T12, lt
7
69, M
T6-L1
Cannot find
59, 40, 46, 60, 53, 54,
T7-11 T8-L2 T9-12 T6-L2 T11-L4 T11-L3
L3, lt T9, lt T5, lt T11, lt T12, rt L2, rt T8, lt T8, rt
2
8 9 10 11 12 13
1 1 1 1 1 2
Cannot Cannot Cannot T9, lt Cannot Cannot
14 15
68, M 78, M
T9-10, rt T11-12, rt T4, rt T9, lt L1, lt T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt T7-8, lt L3-4, rt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5
1 1 3
4 5 6
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, lt L3-4, rt T12, lt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5
1 1
Cannot find Cannot find
M M M M M M
T5-L1 T9-12
T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt Unsuccessful L3-4, rt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5
V2, rt Lateral sacral artery, lt
find find find find find
M: male; F: female; lt: left; rt: right; T: thoracic spine; L: lumbar spine; C: cervical spine; V: vertebral artery; AKA: Adamkiewicz artery.
ARTICLE IN PRESS
Range of cord Signal abnormality
+Model
Age (y), Sex
NEURAD-601; No. of Pages 7
Case no.
Summary of the MRA and DSA findings from 15 SDAVF patients.
Localization of spinal dural arteriovenous fistulas
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
Table 1
3
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
4
G. Zhou et al. Table 2
Summary of the localization of SDAVF at MRA and DSA.
Case no.
1 2 3
4 5 6 7 8 9 10 11 12 13 14 15
MRA shunt Site
Reader 1
Reader 2
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T12, lt L4, lt T9, rt T10-L1, lt T11-T12, lt C2, rt L4-5, lt
T9-10, lt T10-11, lt T4-5, rt T6-7, lt T11-T12, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T9, rt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
Interobserver agreement
Consensus MRA shunt Site
DSA shunt Site
Intermodality Agreement
16 (80%) = 0.79 [0.69—0.89]
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T12, lt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
T9-10, rt T11-12, rt Unsuccessful
12 (75%) = 0.73 [0.61—0.86]
T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt Unsuccessful L3-4, lt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
M: male; F: female; lt: left; rt: right; T: thoracic spine; L: lumbar spine; C: cervical spine; V: vertebral artery; AKA: Adamkiewicz artery.
DCE-MRA-derived level and side of the AVF corresponded with the level and side of the AVF and its feeding artery as determined by the DSA and surgical findings. In four cases, the AVF was fed by multiple feeders derived from different segmental arteries. Six of the 15 patients showed the origin of Adamkiewicz artery (AKA) via DCE-MRA at the T9-L2 level (Table 1). In 15 of 20 fistulas (75%), both readers agreed on the location on DCE-MRA images, and the coefficient of the interobserver agreement was 0.67 (95% confidence interval [CI], 0.16-0.87). In 13 of 16 shunts (75%), the DCE-MRA consensus findings and DSA findings coincided. The intermodality agreement was 0.77 (95% CI, 0.35—0.92) (Table 2).
Discussion To our knowledge, there is still a dearth of updated information regarding the accuracy of DCE-MRA in demonstrating the presence of spinal dural arteriovenous fistulas. This study, therefore, provides a retrospective series to assess the reliability of DCE-MRA in the depiction of the vasculature of SDAVFs. The findings in our report indicate that DCE-MRA can play a key role in the investigation of these lesions. The level and side of the shunt, feeding artery, draining veins, and the AKA, are key issues to be determined via this method [3]. The precise location and anatomic delineation of the fistula site remains challenging for MR angiography, because the fistula does not have a high flow or a large feeder lumen size [4]. DSA is the gold standard for diagnostic imaging of SDAVF. However, it is an invasive
method. Selective catheter angiography failed in the two cases because of the aortic atherosclerosis and tortuosity, as patients within this group present with an older age. This may also lead to missing the level of the fistula during long performances of spinal angiography. Non-invasive imaging is, therefore, extremely helpful in characterizing these lesions and either replacing or allowing targeted conventional angiography of the SDAVFs. However, neither the location of the signal abnormality in conventional MRI nor the level of abnormal flow voids correlates with the level of the fistula. Thus, spinal DCE-MRA is an attractive tool for increasing the sensitivity and specificity of MR examination [5]. Recently, a number of papers have supported DCE-MRA as a useful tool in demonstrating the level of SDAVFs and predicting the location of the fistulas. Accuracy in detecting SDAVFs by dynamic MRA was reported to be between 92% and 100%, with a reasonable degree of success in identifying the level of the SDAVF (50—75%) [1,6—11]. However, these studies did not comprehensively investigate the delineation of the actual fistula and the exact angioarchitecture of SDAVFs at the lesion level. In addition, the reliability of this noninvasive technique for localizing SDAVFs has not been fully investigated and remains to be identified. In our study, DCE-MRA strongly supported the diagnosis of spinal vascular malformation in all cases. Specifically, DCE-MRA demonstrated an 80% accuracy for identifying the correct level and side of the SDAVFs. DCE-MRA did not give any false-positive or -negative results in our study. Patients 3 and 8 did not undergo successful DSA, but their surgical results confirmed the diagnostic accuracy of
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
Localization of spinal dural arteriovenous fistulas
5
Figure 1 A 78-year-old man who complained of progressive bilateral lower extremity weakness and sphincter dysfunction. A subtracted angiogram clearly demonstrates the fistulous point (straight arrow) and arterialized draining veins (arrowheads) of the SDAVF (A). MIP (B), coronal image (C), source image (E) and MPR image (F) of the DCE-MRA show the actual site of the fistula (straight arrow) and dilatation of the vessel (arrowheads).
CTA or DCE-MRA, and therefore, DSA was not essential for treatment. Using DCE-MRA, the fistulous connection can be traced by following the feeding intercostal or radiculomeningeal artery to the fistula. The small size of these structures and the typically slow flow through them render other forms of MR angiography (time-of-flight and phase-contrast) impractical. There are two types of fistula: microfistulas are small-sized lesions fed by one or multiple arterial feeders that drain into dilated veins; macrofistulas are high-flow shunts vascularized by enlarged arteries that drain primarily into giant ectatic veins [12]. McCutcheon et al. [13] found that the true fistulous connection in the dura is microvascular with multiple tangled and looped vessels. Therefore, spatial resolution remains the limiting factor in these cases. Comparison of the DCE-MRA and DSA images in cases of correctly localized SDAVFs showed that when the fistula was ‘‘visible’’ on DCE-MRA, it actually appeared as a short, hazy segment of enhancement that connected the radiculomedullary vein to the feeding artery. The area of enhancement likely represents the multiple tangled microscopic vessels that provide enough enhancement to overcome background noise and provide continuity among the larger vessels, even though the individual small vessels are not seen discretely (Figs. 1 and 2). In many cases
in which the fistula was not observed, the small, short ‘‘signal gap’’ between the arterial feeder and medullary vein was found to be the actual site of the fistula. This gap was, in these cases, only several millimeters at most in length. This method of fistula localization was further facilitated if the courses of the feeding artery and the draining vein were in line [14]. The level of the fistula also can be traced retrogradely by an engorged medullary vein to the neural foramen [15]. Mislocalization of the SDAVF on DCE-MRA occurred in two patients. The potential causes for mislocalization of the SDAVF include miscounting of the vertebral levels and occurrence of the SDAVF outside the chosen FOV. Most reports indicate SDAVF can have more then one feeders, and we observed that the SDAVFs consisted of multiple feeders in 27% of our patients. The AKA should also be visible, as an absence may indicate the presence of SDAVF-induced venous hypertension [16]. The DCE-MRA sequence should have excellent spatial resolution because of the small size of the structures in the spine (spinal cord 10—12 mm in diameter, spinal anterior artery 0.2—0.8 mm, AKA 0.5—1.0 mm) in normal anatomical conditions [17]. In six patients (40%), the artery of Adamkiewicz could be observed, and all were at the T9L2 level. The image quality of the DCE-MRA was sufficient
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
6
G. Zhou et al.
Figure 2 Forty-six-year-old man with a fistula site (straight arrow) at the level of T6-7 on the left side and draining veins (arrowheads) illustrated by preoperative DSA (A). MIP (B), coronal image(C), source image (E) and MPR image (F) of the DCE-MRA demonstrate the fistulous site (straight arrow) at the T6 vertebral body level.
to detect the AKA but was inferior to that of the DSA. The localization of AKA also reflected the detectability by our DCE-MRA. DCE-MRA also allows for a proper pretherapeutic overview of the regional and fistula level that will guide the angiographic procedure to avoid unnecessary catheterizations [18]. Our DCE-MRA studies benefited from the use of a highfield 3 T MRI unit. DCE-MRA at 3.0 T can clearly reveal the extent of spinal vascular malformations, feeding arteries, and fistulas. It is more sensitive to contrast agents and has a relative faster scanning speed compared with 1.5 T MR. Besides, 3.0 T yields a better background suppression and contrast-to-noise ratios, and this renders the depiction of the angioarchitecture of fistulas difficult to visualize using 1.5 T MR systems [19,20]. It should be noted that unlike studies in which the dose (0.1 mmol/kg bodyweight) were used routinely in clinical practice, the contrast used by us was perhaps higher than that reported for DCE-MRA. According to recent reports, nephrogenic systemic fibrosis (NSF) can develop in patients with renal insufficiency who undergo contrast-enhanced MR imaging, so use of a high dose of contrast has a disadvantage in terms of patient safety [21]. To avoid the adverse effects of renal fibrosis in our study, a 20-mL volume can fully meet the requirements. The main advantage of using double dose of contrast media with vascular remnant is a better temporal resolution as well as better conspicuity of small-size vessels. Double
dose of contrast medium leading to greater shortening of the T1 relaxation time and thus substantially increased signal intensity enhancement. In addition, a double dose of contrast medium and delayed imaging time allows more detailed evaluation of the shunt zone rather than merely identifying the arterial feeder. The disadvantage is the possible association between high dose contrast media and the risk of NSF in patients with severe renal insufficiency. Both DCE-MRA and CTA have proved to be valuable in predicting the site of shunt and are noninvasive in nature. Compared to conventional angiography, CTA has a reduced total administration of contrast medium and a very short scan time, which is convenient for weak patients. CTA also has a better vertebral visualization than DCE-MRA does. DCE-MRA eliminated patient exposure to radiation and iodinated-contrast agents. Besides, DCE-MRA scan has 3phase images, which could make the vascular shunt more easily observed. However, neither DCE-MRA nor CTA can provide the same high temporal and spatial resolution as readily as DSA can [22,23]. DCE-MRA may still have limitations in classifying DAV shunts according to the patterns of venous flow. Further improvements, in terms of spatial and temporal resolutions, must be achieved to increase the quality of the information obtained [17]. Diagnosing and localizing a vascular shunt could be more easily achieved by elevated temporal resolution and a 3-phase MRA images
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
Localization of spinal dural arteriovenous fistulas (arterial, arteriovenous, and delay phases) without venous contamination [4,7]. However, further study and comparison of this technique with other MR angiography techniques is needed to establish the best protocol in evaluating SDAVF. Nonetheless, it has become an important diagnostic tool in arteriovenous malformations of the spinal cord by detecting and localizing the arterial feeders and draining veins. MIP and MPR reconstructions are also crucial for the detection and analysis of vascular malformation [24].
Conclusion In conclusion, our specially designed DCE-MRA technique allows for reliable detection and localization of spinal dural arteriovenous shunts and their feeding arteries. However, considering that diagnosis relies mainly on good knowledge of the vascular anatomy, types of lesions, and technical parameters necessary to conduct high-quality DCE-MRA, larger studies are still needed to test its accuracy.
Disclosure of interest The authors declare that they have no competing interest.
Acknowledgment This study was supported by grants from the National Natural Science Foundation of China (Nos. 81471760). The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
References [1] Hetts SW, Moftakhar P, English JD, et al. Spinal dural arteriovenous fistulas and intrathecal venous drainage: correlation between digital subtraction angiography, magnetic resonance imaging, and clinical findings. J Neurosurg Spine 2012;16:433—40. [2] Geibprasert S, Pereira V, Krings T, et al. Dural arteriovenous shunts: a new classification of craniospinal epidural venous anatomical bases and clinical correlations. Stroke 2008;39:2783—94. [3] Zampakis P, Santosh C, Taylor W, et al. The role of non-invasive computed tomography in patients with suspected dural fistulas with spinal drainage. Neurosurgery 2006;58:686—94. [4] Ali S, Cashen TA, Carroll TJ, et al. Time-resolved spinal MR angiography: initial clinical experience in the evaluation of spinal arteriovenous shunts. AJNR Am J Neuroradiol 2007;28:1806—10. [5] Mull M, Nijenhuis RJ, Backes WH, et al. Value and limitations of contrast-enhanced MR angiography in spinal arteriovenous malformations and dural arteriovenous fistulas. AJNR Am J Neuroradiol 2007;28:1249—58. [6] Bink A, Berkefeld J, Wagner M, et al. Detection and grading of dAVF: prospects and limitations of 3 T MRI. Eur Radiol 2012;22:429—38. [7] Cao JB, Cui LL, Jiang XY, et al. Clinical application and diagnostic value of noninvasive spinal angiography in spinal vascular malformations. J Comput Assist Tomogr 2014;38:474—9.
7 [8] Kannath SK, Alampath P, Enakshy Rajan J, et al. Utility of 3D Space T2-weighted volumetric sequence in the localization of spinal dural arteriovenous fistula. J Neurosurg Spine 2016 [Epub ahead of print]. [9] Unsrisong K, Taphey S, Oranratanachai K. Spinal arteriovenous shunts: accuracy of shunt detection, localization, and subtype discrimination using spinal magnetic resonance angiography and manual contrast injection using a syringe. J Neurosurg Spine 2016;24:664—70. [10] Saindane AM, Boddu SR, Tong FC, et al. Contrast-enhanced time-resolved MRA for pre-angiographic evaluation of suspected spinal dural arterial venous fistulas. J Neurointerv Surg 2015;7:135—40. [11] Oda S, Utsunomiya D, Hirai T, et al. Comparison of dynamic contrast-enhanced 3 T MR and 64-row multidetector CT angiography for the localization of spinal dural arteriovenous fistulas. AJNR Am J Neuroradiol 2014;35:407—12. [12] Backes WH, Nijenhuis RJ. Advances in spinal cord MR angiography. AJNR Am J Neuroradiol 2008;29:619—31. [13] McCutcheon IE, Doppman JL, Oldfield EH. Microvascular anatomy of dural arteriovenous abnormalities of the spine: a microangiographic study. J Neurosurg 1996;84: 215—20. [14] Lindenholz A, TerBrugge KG, van Dijk JM, et al. The accuracy and utility of contrast-enhanced MR angiography for localization of spinal dural arteriovenous fistulas: the Toronto experience. Eur Radiol 2014;24:2885—94. [15] Amarouche M, Hart JL, Siddiqui A, et al. Time-resolved contrast-enhanced MR angiography of spinal vascular malformations. AJNR Am J Neuroradiol 2015;36:417—22. [16] Sasaki O, Yajima N, Ichikawa A, et al. Deterioration after surgical treatment of spinal dural arteriovenous fistula associated with spinal perimedullary fistula. Neurol Med Chir (Tokyo) 2012;52:516—20. [17] Condette-Auliac S, Boulin A, Roccatagliata L, et al. MRI and MRA of spinal cord arteriovenous shunts. J Magn Reson Imaging 2014;40:1253—66. [18] Morris JM, Kaufmann TJ, Campeau NG, et al. Volumetric myelographic magnetic resonance imaging to localize difficultto-find spinal dural arteriovenous fistulas. J Neurosurg Spine 2011;14:398—404. [19] Miller TR, Eskey CJ, Mamourian AC. Absence of abnormal vessels in the subarachnoid space on conventional magnetic resonance imaging in patients with spinal dural arteriovenous fistulas. Neurosurg Focus 2012;32:E15. [20] Vargas MI, Nguyen D, Viallon M, et al. Dynamic MR angiography (MRA) of spinal vascular diseases at 3 T. Eur Radiol 2010;20:2491—5. [21] Runge VM. Safety of the gadolinium-based contrast agents for magnetic resonance imaging, focusing in part on their accumulation in the brain and especially the dentate nucleus. Invest Radiol 2016;51:273—9. [22] Mine B, Pezzullo M, Roque G, et al. Detection and characterization of unruptured intracranial aneurysms: comparison of 3 T MRA and DSA. J Neuroradiol 2015;42:162—8. [23] Amarteifio E, Essig M, Böckler D, et al. Comparison of gadofosveset (Vasovist[® ]) with gadobenate dimeglumine (Multihance[® ])-enhanced MR angiography for high-grade carotid artery stenosis. J Neuroradiol 2015;42:236—44. [24] Petkova M, Gauvrit JY, Trystram D, et al. Three-dimensional dynamic time-resolved contrast- enhanced MRA using parallel imaging and a variable rate k-space sampling strategy in intracranial arteriovenous malformations. J Magn Reson Imaging 2009;29:7—12.
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
ARTICLE IN PRESS
Journal of Neuroradiology (2016) xxx, xxx—xxx
Available online at
ScienceDirect www.sciencedirect.com
ORIGINAL ARTICLE
Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T G. Zhou a, M.H. Li a, C. Lu b, Y.L. Yin c, Y.Q. Zhu a, X.E. Wei a, H.T. Lu a, Q.Q. Zheng d, W.W. Gao e,∗ a
Department of diagnostic and interventional radiology, Shanghai Jiao Tong university affiliated Sixth People’s hospital, Shanghai, China b School of radiology, Taishan medical college, Taian, Shandong, China c Department of anesthesiology, The Military general hospital of Beijing PLA, Beijing, China d Department of hematology, Shanghai Jiao Tong university affiliated Sixth People’s hospital, Shanghai, China e Department of neurosurgery, Shanghai Jiao Tong university affiliated Sixth People’s hospital, MS, No. 600, Yi Shan road, 200233 Shanghai, China
KEYWORDS Spinal dural arteriovenous fistulas; Dynamic contrast-enhanced MR angiography; Digital subtraction angiography; Angioarchitecture; Accuracy
Summary Objective: This study was undertaken to evaluate the accuracy of dynamic contrast-enhanced magnetic resonance angiography (DCE-MRA) in the precise location and demonstration of fistulous points in spinal dural arteriovenous fistulas (SDAVFs). Methods: Fifteen patients (14 men, 1 woman; age range: 40—78 years; mean: 55.5 years) harboring SDAVF who underwent preoperative DCE-MRA and spinal digital subtraction angiography (DSA) between January 2012 and January 2015 were evaluated retrospectively. Two reviewers independently evaluated the level and side of the arteriovenous fistula and feeding artery on 3 T DCE-MRA and DSA images. The accuracy of DCE-MRA was assessed by comparing its findings with those from DSA and surgery in each case. Results: All 15 patients underwent DCE-MRA and DSA. DSA was unsuccessful in two patients due to technical difficulties. All cases were explored surgically, guided by the DCE-MRA. Surgery confirmed that 14 AVF sites were located in the thoracic spine, 5 in the lumbar spine, and 1 in the cervical spine. The origin of the fistulas and feeding arteries was accurately shown by DCE-MRA in 11 of the 15 patients. DCE-MRA also detected dilated perimedullary veins in all 15 patients. Overall, DCE-MRA facilitated DSA catheterization in 10 cases. In six patients, the artery of Adamkiewicz could be observed. In 15 out of 20 fistulas (75%), both readers agreed
Abbreviations: SDAVF, Spinal dural arteriovenous fistula; DCE-MRA, Dynamic contrast-enhanced magnetic resonance angiography; DSA, Digital subtraction angiography; FOV, Field of view. ∗ Corresponding author. Tel.: +86 021 64844183; fax: +86 021 64844183. E-mail address: [email protected] (W.W. Gao). http://dx.doi.org/10.1016/j.neurad.2016.10.002 0150-9861/© 2016 Elsevier Masson SAS. All rights reserved.
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
2
G. Zhou et al. on the location on DCE-MRA images, and the coefficient of the interobserver agreement was 0.67 (95% confidence interval [CI], 0.16—0.87). In 13 of 16 shunts (75%), the DCE-MRA consensus findings and DSA findings coincided. The intermodality agreement was 0.77 (95% CI: 0.35—0.92). Conclusions: Our DCE-MRA studies benefited from the use of a high-field 3 T MR imaging unit and reliably detected and localized the SDAVF and feeding arteries. As experience with this technique grows, it may be possible to replace DSA with DCE-MRA if surgery is the planned treatment. © 2016 Elsevier Masson SAS. All rights reserved.
Introduction Spinal dural arteriovenous fistulas (SDAVF) can render devastating neurological consequences and lead to considerable morbidity. SDAVF is the most common type of spinal vascular malformation in which an arteriovenous connection is typically located within the dura mater of a nerve root sleeve, underneath the vertebral body pedicle in the neural foramen between the radicular artery and the medullary vein [1,2]. Diagnosis and treatment prior to irreversible cord ischemia or infarction are extremely important, since the symptoms at that point can be reversed. Accurately locating the fistula is of crucial importance before neurosurgery or superselective embolization. However, despite preoperative MRI and DSA, the precise localization and anatomic delineation of the fistula site remains challenging due to the complex angioarchitecture and small vascular caliber. The aim of the following article is to compare the agreement between intra-arterial DSA and DCE-MRA at 3.0 T in localizing the shunting point of SDAVFs.
Methods The protocol of this study was approved by the Ethical Committee of our institute. Between January 2012 and June 2015, a total of 15 patients (14 men and 1 woman; age range, 40—78 years; mean, 55.5 years) at a single institution were confirmed to have SDAVFs. All patients presented with congestive myelopathy (the mean length of the illness course was 6 months) and underwent MR imaging (MRI) and DSA. Our inclusion criteria were a diagnosis of SDAVF on the basis of spinal DSA scans, which was verified by surgery after spinal 3 T DCE-MRA. The fast DCE-MRA examinations were performed on an Achieva 3.0 T MRI system (Philips Healthcare, Amsterdam, The Netherlands). A testbolus (1 mL of contrast) technique positioned through the aorta at T10 vertebral body level was employed to evaluate the cycle time before examination. The contrast agent was injected through the antecubital veins using a highpressure syringe at a rate of 3 mL/s and a total volume of 20 mL (gadodiamide, 0.5 mmol/mL) based on the calculated cycle time using a test bolus injection. The technique was then implemented using a 3D radiofrequency-spoiled fast gradient-echo volume acquisition using the following parameters: TR/TE, 3.1 ms/1.1 ms; flip angle, 20◦ ; voxel size, 1.0 × 0.9 × 0.9 mm; field of view (FOV), 357 × 440 mm; four slabs (180 slices); SENSE factor, 2. The FOV was positioned to cover the entire T2 hyperintense cord and serpentine
flow voids. The placement of the FOV may also base on clinical suspicion of the location of the fistula. The data acquisition lasted 2 minutes and was performed during the contrast agent injection. The acquired image data sets were then transferred to a workstation (IntelliVue Guardian Early Warning Score; Philips Healthcare), in which reconstruction using maximum intensity projection and volume-rendering DCE-MRA images was performed using a 3D specialized software package (Volume Inspection, Philips Medical Systems, Amsterdam, The Netherlands). Fistulous points in these subjects were determined via maximum-intensity projection(MIP), volume rendering technique(VRT), multiplanar reformations (MPR) and their source images. Two neuroradiologists (W.-W. Gao and H.-T. Lu, with 25 and 10 years of experience in neuroimaging, respectively) were blinded to the subjects to analyze the images independently. They provided the diagnosis including the lesion range, feeding arteries, and fistulas. DSA was performed on a monoplanar unit (Axion Artis) with a 1024 × 1024 matrix and 17-20-cm FOV using conventional methods. The interval between DCEMRA and DSA studies ranged from 0—8 days (mean, 3 days). Two authors (W.-W. Gao and G. Zhou) compared the DCEMRA results with the DSA and surgical findings. Interobserver agreement on the angiographic findings was determined by calculating the coefficient. The accuracy of DCE-MRA was evaluated by checking whether the level and side of the AVF that it suggested corresponded with the DSA and surgical findings. All analyses were performed using the STATA 12 statistical package (StataCorp, College Station, TX, USA).
Results DCE-MRA and surgery were performed successfully in 15 patients. Intra-arterial DSA was unsuccessful in two patients due to technical difficulties, mainly related to coexisting aortic atherosclerosis and tortuosity. These two patients were treated surgically only on the basis of the DCE-MRA results. Surgery confirmed that 14 AVF sites were located in the thoracic spine, 5 in the lumbar spine, and 1 in the cervical spine (Tables 1 and 2). The locations of the feeders and fistulas identified on the DCE-MRA and DSA images are shown in Table 1. DCE-MRA demonstrated a precise fistula angioarchitectural configuration in 11 (11/15) cases. In four cases, the small shunts inside the nidus could not be individualized. The draining radicular vein could be followed in all cases from the fistula site to the coronal venous plexus around the spinal cord. DCE-MRA was effective in demonstrating the feeding artery, fistula site, and dilated perimedullary veins in most cases. In 11 (75%) patients, the
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
Consensus MRA shunt Site
DSA shunt Site
Surgery shunt Site
Level of feeder
No. of feeder
Location of AKA
1 2 3
40, M 54, M 41, M
T2-10 T4-12 T8-L1
T9-10, rt T11-12, rt Unsuccessful
T7-11 T9-12 T12-L1
T10, rt T9, rt T4, rt T6, lt T12, lt T12, rt T9, lt T12, lt
L1, lt T9, rt Cannot find
63, F 42, M 66, M
1 1 2
T10, rt L2, lt T12, lt
7
69, M
T6-L1
Cannot find
59, 40, 46, 60, 53, 54,
T7-11 T8-L2 T9-12 T6-L2 T11-L4 T11-L3
L3, lt T9, lt T5, lt T11, lt T12, rt L2, rt T8, lt T8, rt
2
8 9 10 11 12 13
1 1 1 1 1 2
Cannot Cannot Cannot T9, lt Cannot Cannot
14 15
68, M 78, M
T9-10, rt T11-12, rt T4, rt T9, lt L1, lt T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt T7-8, lt L3-4, rt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5
1 1 3
4 5 6
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, lt L3-4, rt T12, lt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5
1 1
Cannot find Cannot find
M M M M M M
T5-L1 T9-12
T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt Unsuccessful L3-4, rt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5
V2, rt Lateral sacral artery, lt
find find find find find
M: male; F: female; lt: left; rt: right; T: thoracic spine; L: lumbar spine; C: cervical spine; V: vertebral artery; AKA: Adamkiewicz artery.
ARTICLE IN PRESS
Range of cord Signal abnormality
+Model
Age (y), Sex
NEURAD-601; No. of Pages 7
Case no.
Summary of the MRA and DSA findings from 15 SDAVF patients.
Localization of spinal dural arteriovenous fistulas
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
Table 1
3
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
4
G. Zhou et al. Table 2
Summary of the localization of SDAVF at MRA and DSA.
Case no.
1 2 3
4 5 6 7 8 9 10 11 12 13 14 15
MRA shunt Site
Reader 1
Reader 2
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T12, lt L4, lt T9, rt T10-L1, lt T11-T12, lt C2, rt L4-5, lt
T9-10, lt T10-11, lt T4-5, rt T6-7, lt T11-T12, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T9, rt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
Interobserver agreement
Consensus MRA shunt Site
DSA shunt Site
Intermodality Agreement
16 (80%) = 0.79 [0.69—0.89]
T9-10, rt T11-12, rt T4-5, rt T6-7, lt T12-L1, lt T12, rt L2-3, rt T9-10, rt T12-L1, rt L3-4, lt T9-10, rt T7-8, rt L3-4, lt T12, lt L4, lt T9, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
T9-10, rt T11-12, rt Unsuccessful
12 (75%) = 0.73 [0.61—0.86]
T12, rt L2-3, rt T9-10, lt T12-L1, rt L3-4, lt T9-10, rt Unsuccessful L3-4, lt L3, lt L4, lt T7-8, rt T10-L1, lt T12-L1, lt C2, rt L4-5, lt
M: male; F: female; lt: left; rt: right; T: thoracic spine; L: lumbar spine; C: cervical spine; V: vertebral artery; AKA: Adamkiewicz artery.
DCE-MRA-derived level and side of the AVF corresponded with the level and side of the AVF and its feeding artery as determined by the DSA and surgical findings. In four cases, the AVF was fed by multiple feeders derived from different segmental arteries. Six of the 15 patients showed the origin of Adamkiewicz artery (AKA) via DCE-MRA at the T9-L2 level (Table 1). In 15 of 20 fistulas (75%), both readers agreed on the location on DCE-MRA images, and the coefficient of the interobserver agreement was 0.67 (95% confidence interval [CI], 0.16-0.87). In 13 of 16 shunts (75%), the DCE-MRA consensus findings and DSA findings coincided. The intermodality agreement was 0.77 (95% CI, 0.35—0.92) (Table 2).
Discussion To our knowledge, there is still a dearth of updated information regarding the accuracy of DCE-MRA in demonstrating the presence of spinal dural arteriovenous fistulas. This study, therefore, provides a retrospective series to assess the reliability of DCE-MRA in the depiction of the vasculature of SDAVFs. The findings in our report indicate that DCE-MRA can play a key role in the investigation of these lesions. The level and side of the shunt, feeding artery, draining veins, and the AKA, are key issues to be determined via this method [3]. The precise location and anatomic delineation of the fistula site remains challenging for MR angiography, because the fistula does not have a high flow or a large feeder lumen size [4]. DSA is the gold standard for diagnostic imaging of SDAVF. However, it is an invasive
method. Selective catheter angiography failed in the two cases because of the aortic atherosclerosis and tortuosity, as patients within this group present with an older age. This may also lead to missing the level of the fistula during long performances of spinal angiography. Non-invasive imaging is, therefore, extremely helpful in characterizing these lesions and either replacing or allowing targeted conventional angiography of the SDAVFs. However, neither the location of the signal abnormality in conventional MRI nor the level of abnormal flow voids correlates with the level of the fistula. Thus, spinal DCE-MRA is an attractive tool for increasing the sensitivity and specificity of MR examination [5]. Recently, a number of papers have supported DCE-MRA as a useful tool in demonstrating the level of SDAVFs and predicting the location of the fistulas. Accuracy in detecting SDAVFs by dynamic MRA was reported to be between 92% and 100%, with a reasonable degree of success in identifying the level of the SDAVF (50—75%) [1,6—11]. However, these studies did not comprehensively investigate the delineation of the actual fistula and the exact angioarchitecture of SDAVFs at the lesion level. In addition, the reliability of this noninvasive technique for localizing SDAVFs has not been fully investigated and remains to be identified. In our study, DCE-MRA strongly supported the diagnosis of spinal vascular malformation in all cases. Specifically, DCE-MRA demonstrated an 80% accuracy for identifying the correct level and side of the SDAVFs. DCE-MRA did not give any false-positive or -negative results in our study. Patients 3 and 8 did not undergo successful DSA, but their surgical results confirmed the diagnostic accuracy of
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
Localization of spinal dural arteriovenous fistulas
5
Figure 1 A 78-year-old man who complained of progressive bilateral lower extremity weakness and sphincter dysfunction. A subtracted angiogram clearly demonstrates the fistulous point (straight arrow) and arterialized draining veins (arrowheads) of the SDAVF (A). MIP (B), coronal image (C), source image (E) and MPR image (F) of the DCE-MRA show the actual site of the fistula (straight arrow) and dilatation of the vessel (arrowheads).
CTA or DCE-MRA, and therefore, DSA was not essential for treatment. Using DCE-MRA, the fistulous connection can be traced by following the feeding intercostal or radiculomeningeal artery to the fistula. The small size of these structures and the typically slow flow through them render other forms of MR angiography (time-of-flight and phase-contrast) impractical. There are two types of fistula: microfistulas are small-sized lesions fed by one or multiple arterial feeders that drain into dilated veins; macrofistulas are high-flow shunts vascularized by enlarged arteries that drain primarily into giant ectatic veins [12]. McCutcheon et al. [13] found that the true fistulous connection in the dura is microvascular with multiple tangled and looped vessels. Therefore, spatial resolution remains the limiting factor in these cases. Comparison of the DCE-MRA and DSA images in cases of correctly localized SDAVFs showed that when the fistula was ‘‘visible’’ on DCE-MRA, it actually appeared as a short, hazy segment of enhancement that connected the radiculomedullary vein to the feeding artery. The area of enhancement likely represents the multiple tangled microscopic vessels that provide enough enhancement to overcome background noise and provide continuity among the larger vessels, even though the individual small vessels are not seen discretely (Figs. 1 and 2). In many cases
in which the fistula was not observed, the small, short ‘‘signal gap’’ between the arterial feeder and medullary vein was found to be the actual site of the fistula. This gap was, in these cases, only several millimeters at most in length. This method of fistula localization was further facilitated if the courses of the feeding artery and the draining vein were in line [14]. The level of the fistula also can be traced retrogradely by an engorged medullary vein to the neural foramen [15]. Mislocalization of the SDAVF on DCE-MRA occurred in two patients. The potential causes for mislocalization of the SDAVF include miscounting of the vertebral levels and occurrence of the SDAVF outside the chosen FOV. Most reports indicate SDAVF can have more then one feeders, and we observed that the SDAVFs consisted of multiple feeders in 27% of our patients. The AKA should also be visible, as an absence may indicate the presence of SDAVF-induced venous hypertension [16]. The DCE-MRA sequence should have excellent spatial resolution because of the small size of the structures in the spine (spinal cord 10—12 mm in diameter, spinal anterior artery 0.2—0.8 mm, AKA 0.5—1.0 mm) in normal anatomical conditions [17]. In six patients (40%), the artery of Adamkiewicz could be observed, and all were at the T9L2 level. The image quality of the DCE-MRA was sufficient
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
6
G. Zhou et al.
Figure 2 Forty-six-year-old man with a fistula site (straight arrow) at the level of T6-7 on the left side and draining veins (arrowheads) illustrated by preoperative DSA (A). MIP (B), coronal image(C), source image (E) and MPR image (F) of the DCE-MRA demonstrate the fistulous site (straight arrow) at the T6 vertebral body level.
to detect the AKA but was inferior to that of the DSA. The localization of AKA also reflected the detectability by our DCE-MRA. DCE-MRA also allows for a proper pretherapeutic overview of the regional and fistula level that will guide the angiographic procedure to avoid unnecessary catheterizations [18]. Our DCE-MRA studies benefited from the use of a highfield 3 T MRI unit. DCE-MRA at 3.0 T can clearly reveal the extent of spinal vascular malformations, feeding arteries, and fistulas. It is more sensitive to contrast agents and has a relative faster scanning speed compared with 1.5 T MR. Besides, 3.0 T yields a better background suppression and contrast-to-noise ratios, and this renders the depiction of the angioarchitecture of fistulas difficult to visualize using 1.5 T MR systems [19,20]. It should be noted that unlike studies in which the dose (0.1 mmol/kg bodyweight) were used routinely in clinical practice, the contrast used by us was perhaps higher than that reported for DCE-MRA. According to recent reports, nephrogenic systemic fibrosis (NSF) can develop in patients with renal insufficiency who undergo contrast-enhanced MR imaging, so use of a high dose of contrast has a disadvantage in terms of patient safety [21]. To avoid the adverse effects of renal fibrosis in our study, a 20-mL volume can fully meet the requirements. The main advantage of using double dose of contrast media with vascular remnant is a better temporal resolution as well as better conspicuity of small-size vessels. Double
dose of contrast medium leading to greater shortening of the T1 relaxation time and thus substantially increased signal intensity enhancement. In addition, a double dose of contrast medium and delayed imaging time allows more detailed evaluation of the shunt zone rather than merely identifying the arterial feeder. The disadvantage is the possible association between high dose contrast media and the risk of NSF in patients with severe renal insufficiency. Both DCE-MRA and CTA have proved to be valuable in predicting the site of shunt and are noninvasive in nature. Compared to conventional angiography, CTA has a reduced total administration of contrast medium and a very short scan time, which is convenient for weak patients. CTA also has a better vertebral visualization than DCE-MRA does. DCE-MRA eliminated patient exposure to radiation and iodinated-contrast agents. Besides, DCE-MRA scan has 3phase images, which could make the vascular shunt more easily observed. However, neither DCE-MRA nor CTA can provide the same high temporal and spatial resolution as readily as DSA can [22,23]. DCE-MRA may still have limitations in classifying DAV shunts according to the patterns of venous flow. Further improvements, in terms of spatial and temporal resolutions, must be achieved to increase the quality of the information obtained [17]. Diagnosing and localizing a vascular shunt could be more easily achieved by elevated temporal resolution and a 3-phase MRA images
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002
+Model NEURAD-601; No. of Pages 7
ARTICLE IN PRESS
Localization of spinal dural arteriovenous fistulas (arterial, arteriovenous, and delay phases) without venous contamination [4,7]. However, further study and comparison of this technique with other MR angiography techniques is needed to establish the best protocol in evaluating SDAVF. Nonetheless, it has become an important diagnostic tool in arteriovenous malformations of the spinal cord by detecting and localizing the arterial feeders and draining veins. MIP and MPR reconstructions are also crucial for the detection and analysis of vascular malformation [24].
Conclusion In conclusion, our specially designed DCE-MRA technique allows for reliable detection and localization of spinal dural arteriovenous shunts and their feeding arteries. However, considering that diagnosis relies mainly on good knowledge of the vascular anatomy, types of lesions, and technical parameters necessary to conduct high-quality DCE-MRA, larger studies are still needed to test its accuracy.
Disclosure of interest The authors declare that they have no competing interest.
Acknowledgment This study was supported by grants from the National Natural Science Foundation of China (Nos. 81471760). The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
References [1] Hetts SW, Moftakhar P, English JD, et al. Spinal dural arteriovenous fistulas and intrathecal venous drainage: correlation between digital subtraction angiography, magnetic resonance imaging, and clinical findings. J Neurosurg Spine 2012;16:433—40. [2] Geibprasert S, Pereira V, Krings T, et al. Dural arteriovenous shunts: a new classification of craniospinal epidural venous anatomical bases and clinical correlations. Stroke 2008;39:2783—94. [3] Zampakis P, Santosh C, Taylor W, et al. The role of non-invasive computed tomography in patients with suspected dural fistulas with spinal drainage. Neurosurgery 2006;58:686—94. [4] Ali S, Cashen TA, Carroll TJ, et al. Time-resolved spinal MR angiography: initial clinical experience in the evaluation of spinal arteriovenous shunts. AJNR Am J Neuroradiol 2007;28:1806—10. [5] Mull M, Nijenhuis RJ, Backes WH, et al. Value and limitations of contrast-enhanced MR angiography in spinal arteriovenous malformations and dural arteriovenous fistulas. AJNR Am J Neuroradiol 2007;28:1249—58. [6] Bink A, Berkefeld J, Wagner M, et al. Detection and grading of dAVF: prospects and limitations of 3 T MRI. Eur Radiol 2012;22:429—38. [7] Cao JB, Cui LL, Jiang XY, et al. Clinical application and diagnostic value of noninvasive spinal angiography in spinal vascular malformations. J Comput Assist Tomogr 2014;38:474—9.
7 [8] Kannath SK, Alampath P, Enakshy Rajan J, et al. Utility of 3D Space T2-weighted volumetric sequence in the localization of spinal dural arteriovenous fistula. J Neurosurg Spine 2016 [Epub ahead of print]. [9] Unsrisong K, Taphey S, Oranratanachai K. Spinal arteriovenous shunts: accuracy of shunt detection, localization, and subtype discrimination using spinal magnetic resonance angiography and manual contrast injection using a syringe. J Neurosurg Spine 2016;24:664—70. [10] Saindane AM, Boddu SR, Tong FC, et al. Contrast-enhanced time-resolved MRA for pre-angiographic evaluation of suspected spinal dural arterial venous fistulas. J Neurointerv Surg 2015;7:135—40. [11] Oda S, Utsunomiya D, Hirai T, et al. Comparison of dynamic contrast-enhanced 3 T MR and 64-row multidetector CT angiography for the localization of spinal dural arteriovenous fistulas. AJNR Am J Neuroradiol 2014;35:407—12. [12] Backes WH, Nijenhuis RJ. Advances in spinal cord MR angiography. AJNR Am J Neuroradiol 2008;29:619—31. [13] McCutcheon IE, Doppman JL, Oldfield EH. Microvascular anatomy of dural arteriovenous abnormalities of the spine: a microangiographic study. J Neurosurg 1996;84: 215—20. [14] Lindenholz A, TerBrugge KG, van Dijk JM, et al. The accuracy and utility of contrast-enhanced MR angiography for localization of spinal dural arteriovenous fistulas: the Toronto experience. Eur Radiol 2014;24:2885—94. [15] Amarouche M, Hart JL, Siddiqui A, et al. Time-resolved contrast-enhanced MR angiography of spinal vascular malformations. AJNR Am J Neuroradiol 2015;36:417—22. [16] Sasaki O, Yajima N, Ichikawa A, et al. Deterioration after surgical treatment of spinal dural arteriovenous fistula associated with spinal perimedullary fistula. Neurol Med Chir (Tokyo) 2012;52:516—20. [17] Condette-Auliac S, Boulin A, Roccatagliata L, et al. MRI and MRA of spinal cord arteriovenous shunts. J Magn Reson Imaging 2014;40:1253—66. [18] Morris JM, Kaufmann TJ, Campeau NG, et al. Volumetric myelographic magnetic resonance imaging to localize difficultto-find spinal dural arteriovenous fistulas. J Neurosurg Spine 2011;14:398—404. [19] Miller TR, Eskey CJ, Mamourian AC. Absence of abnormal vessels in the subarachnoid space on conventional magnetic resonance imaging in patients with spinal dural arteriovenous fistulas. Neurosurg Focus 2012;32:E15. [20] Vargas MI, Nguyen D, Viallon M, et al. Dynamic MR angiography (MRA) of spinal vascular diseases at 3 T. Eur Radiol 2010;20:2491—5. [21] Runge VM. Safety of the gadolinium-based contrast agents for magnetic resonance imaging, focusing in part on their accumulation in the brain and especially the dentate nucleus. Invest Radiol 2016;51:273—9. [22] Mine B, Pezzullo M, Roque G, et al. Detection and characterization of unruptured intracranial aneurysms: comparison of 3 T MRA and DSA. J Neuroradiol 2015;42:162—8. [23] Amarteifio E, Essig M, Böckler D, et al. Comparison of gadofosveset (Vasovist[® ]) with gadobenate dimeglumine (Multihance[® ])-enhanced MR angiography for high-grade carotid artery stenosis. J Neuroradiol 2015;42:236—44. [24] Petkova M, Gauvrit JY, Trystram D, et al. Three-dimensional dynamic time-resolved contrast- enhanced MRA using parallel imaging and a variable rate k-space sampling strategy in intracranial arteriovenous malformations. J Magn Reson Imaging 2009;29:7—12.
Please cite this article in press as: Zhou G, et al. Dynamic contrast-enhanced magnetic resonance angiography for the localization of spinal dural arteriovenous fistulas at 3T. J Neuroradiol (2016), http://dx.doi.org/10.1016/j.neurad.2016.10.002