Eur Spine J DOI 10.1007/s00586-014-3348-1
ORIGINAL ARTICLE
Is the ‘‘snake-eye’’ MRI sign correlated to anterior spinal artery occlusion on CT angiography in cervical spondylotic myelopathy and amyotrophy? Zhengfeng Zhang • Honggang Wang
Received: 9 October 2013 / Revised: 26 April 2014 / Accepted: 26 April 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Objective The goal of this study was to identify anterior spinal artery (ASA) occlusion by CT angiography in cervical spondylotic myelopathy (CSM) and amyotrophy (CSA) with T2-weighted hyperintensity of MR image of documented small intramedullary high signal intensity known as ‘‘snake-eye appearance’’ (SEA). Method One hundred and six patients with CSM were admitted to the investigator group between June 2010 and June 2013. Intramedullary high signal intensity was found in 42 cases and was divided into two types, SEA and nonSEA. SEA was observed in 10 patients, including seven CSM patients and three CSA patients. All SEA patients were performed CT angiography of ASA after admission. Results The ASA was visualized in all 10 patients. ASA incomplete occlusion was found in one CSA patient and one CSM patient. No ASA occlusion was found in other CSA and CSM patients with SEA. Conclusion ASA occlusion is not commonly seen in CSM and CSA patients with SEA. Pathological changes about SEA in CSM and CSA have no close correlation with ASA occlusion, but may be with anterior radiculomedullary arteries. Keywords Cervical spondylotic myelopathy Cervical spondylotic amyotrophy CT angiography Anterior spinal artery Snake-eye appearance
Magnetic resonance imaging (MRI) of the cervical spine is essential for the preoperative evaluation of patients with cervical spondylotic myelopathy (CSM) and amyotrophy (CSA). The intramedullary high signal intensity on T2weighted MRI has been found in 41–97.2 % of patients with CSM [1]. Snake-eye appearance (SEA) is a unique neuroimaging finding characterized as nearly symmetrical round high signal intensity of the spinal parenchyma resembling the face of a snake demonstrated on axial T2weighted MR images [2, 3]. This condition is sometimes observed in CSM and CSA; however, its pathophysiological features are uncertain. It has been reported that edema, myelomalacia, gliosis, and inflammation are involved in intramedullary high signal intensity, which have suggested necrosis and ischemia [4]. The anterior spinal arteries (ASA) supply the anterior horn and the anterior part of the lateral column on the left or right side at each level of spinal cord, which may be theoretically compressed by herniated disc and resulted in spinal cord ischemia in CSM and CSA. SEA-related intramedullary high signal intensity is located in anterior horn, which is supplied by ASA. So we hypothesize that SEA-related intramedullary high signal intensity in CSM and CSA have a close correlation with ASA occlusion. The objective of this study was to identify ASA occlusion by CT angiography in CSM and CSA patients with SEA.
Methods Patients Z. Zhang (&) H. Wang Department of Orthopedics, Xinqiao Hospital, Third Military Medical University, 183 Xinqiao Street, Shapingba District, Chongqing 400037, China e-mail: [email protected]
Between June 2010 and June 2013, 106 patients with diagnosis of CSM were admitted to the investigator group (Z. Z.). Among them, intramedullary high signal intensity was
123
Eur Spine J Table 1 Clinical summary of patients Case no.
Sex/ Age
Level of compression
T2 cord hyperintensity
Location of T2 cord hyperintensity
Mode of myelopathy
ASA occlusion
Spinal canal sagittal diameter compression (%)
Type of operation
JOA score preop/ postop
1
M/44
C4/5 C5/6
C4–C6
Anterior horn
CSM
No
70
ACDF
15/17
2
M/71
C4/5 C5/6
C4–C6
Anterior horn
CSM
No
30
ACDF
15/16
3 4
F/52 M/38
C4/5 C3/4
C4/5 C3/4
Left anterior horn Anterior horn
CSM CSA
No Incomplete
40 10
ACDF None
16/17 16/16
5
M/65
C4/5 C5/6
C4/5
Anterior horn
CCM, CSA
No
80
ACDF
15/16
6
M/47
C4/5
C4/5
Right anterior horn
CSA
No
40
ACDF
16/17
7
M/58
C3/4
C3/4
Anterior horn
CCM
No
20
ACDF
16/17
8
M/43
C4/5
C5–C6
Anterior horn
CCM
Incomplete
20
ACDF
15/17
9
M/42
C4/5
C4/5
Anterior horn
CCM
No
80
ACDF
16/17
10
M/44
C4/5 C5/6
C4–C6
Anterior horn
CCM
No
60
ACDF
15/17
M male, F female, C cervical, CSM cervical spondylotic myelopathy, CSA cervical spondylotic amyotrophy, ACDF anterior cervical disctomy fusion
observed in 42 patients on axial T2-weighted MR imaging using a 1.5-tesla imager. The patterns of intramedullary high signal intensity were divided into two groups, SEA and nonSEA. SEA was defined as one left- and/or one right-sided small round or elliptical high signal intensity lesion in the central gray matter near the ventrolateral posterior column. Of the 42 patients with intramedullary high signal intensity, SEA was observed in 10 patients and non-SEA in 32. In 106 patients, three patients were identified as CSA according to the following fulfilled criteria: (1) a Nurick grade of 0 or 1 (no gait impairment); (2) symptoms and signs present weakness and wasting of upper limb muscles as difficulty in shoulder abduction, positive arm-drop sign, or positive wristdrop sign; and (3) without sensory or lower limb involvement. All three CSA patients presented SEA in MRI. The clinical characteristics of the SEA patients are summarized in Table 1. In these 10 patients (9 males and 1 female), the distribution of cervical level was from C3 to C6; the duration from onset until surgery ranged from 1 to 6 months; following up ranged from 4 to 16 months (mean, 8.3 months). All patients underwent plain cervical radiographs, CT scans, and 1.5-Tesla MRI. T1-weighted and T2-weighted MRIs were performed to define cord compression, spinal canal sagittal diameter compression proportion, and the extent of intramedullary high signal intensity. Nine of 10 patients were performed anterior decompression and fusion surgery, except one CSA patient who presented no cord compression on MRI. Preoperative and postoperative neurological statuses were evaluated using JOA scoring system and Hirabayashi formula [5]. CT angiography of ASA With the institutional review board approval of the research protocol, CT angiography of ASA was performed
123
on the SEA patients. One or 2 days after the patients underwent MRI, were scanned with 64-slice MDCT scanners (LightSpeed 64, GE Medical System, Milwaukee, WI, USA). All CT scans were obtained with 0.5 s rotation, 0.625 mm nominal detector widths, pitch of 0.984, 120 kV, and 480 mA. Transverse sections were reconstructed with 50 % overlap relative to the effective section thickness of 0.6 mm. Iohexol (Amersham Health, Princeton, NJ, USA) was administered through an antecubital vein with a dose of 120–150 ml (350 mgI/ml) at a rate of 5 ml/s. The scan delay was determined by a preliminary 20 ml test injection at the level of basilar artery. All observations were made retrospectively by two neuroradiologists. To examine the ASA transverse sections, multiplanar reformations (MPR), curved planar reformation (CPR), and thin-slab (2–4 mm) maximum intensity projections (MIP) were generated and displayed on a workstation (Advantage Workstation 4.3, GE Medical Systems) with window and level settings selected to maximize arterial to background discrimination. The ASA was identified by maintaining the MIP-slab parallel to the anterior surface of the spinal cord at each vertebral level assessed from C1 to C7. An enhanced artery on the midline ventral surface of the spinal cord was interpreted as the ASA. At the same time, an artery originating from the aorta and coursing through the intervertebral foramen to join the ASA in a hairpin configuration was interpreted as the Adamkiewicz artery (AKA). Co-visualizations of ASA and AKA were used to identify ASA; covisualizations of veins were not the criteria for identifying ASA. Another criterion was that the ASA originated from vertebral artery in skull base. Inter-rater reliability, in the form of consensual identification of ASA by two radiologists, was required.
Eur Spine J
Fig. 1 Case 4. A 38-year-old man with CSA at C3/4. a Sagittal T2weighted MRI showing cord hyperintensity at C3/4 (arrows). b, c Transverse T2-weighted MRI showing the snake eyes changes in anterior horn of cord at the C3–C4 level (arrows). d, e Sagittal view
(d) and coronal view (f) of CT angiography showing suspected ASA incomplete occlusion at C3–C4 (arrows). (f–h) Coronal view (f, g) and sagittal view (h) of MR angiography showing ASA incomplete occlusion at C3–C4 (arrows)
Results
(80 %), no ASA occlusion was found, even for some cases with 80 % spinal canal sagittal diameter compression.
ASA occlusion Neurological status The clinical data are summarized in Table 1. The ASA was visualized in all 10 SEA patients. The course of the ASA extended from C1 to C7 in sagittal, coronal and transverse view. The ASA runs along the entire length of the anterior surface of the spinal cord. Anterior radiculomedullary arteries (ARAs), ASA brand which supply the cord throughout its length by anastomoses branching upward and downward, were not visualized for the angiography, because ARAs are very little or even no flow in two directions. The Adamkiewicz artery (AKA) was also visualized in most cases. For three CSA cases, ASA incomplete occlusion was found in one patient (case 4) (Fig. 1). For seven CSM cases, one case (case 8) presented ASA incomplete occlusion (Fig. 2). For other eight CSM and CSA patients
Preoperative and postoperative neurological statuses of the patients are given in Table 1. All seven CSM patients had undergone anterior cervical disctomy/corpectomy fusion (ACDF/ACCF) and did not encounter complications. Six patients with JOA score 15 or 16 improved to 17 after surgery. One patient (case 2) only improved JOA score from 15 to 16, but not 17, after 14 months following up. For three CSA patients, two patients presented cervical compression and underwent ACDF. The 65 years of age patient (case 5) who presented both CSM and CSA improved neurological status from 15 to 16 of JOA score after 12 months following up. Another patient (case 6) improved JOA score from 16 to 17 after surgery. The third patient (illustrative case 4) presented no cord compression
123
Eur Spine J
Fig. 2 Case 8. A 43-year-old man with CSM at C4/5. a Sagittal T2weighted MRI reveals that the cervical spinal cord was compressed at C4/5. Spinal cord hyperintensity is showing from C5–C6 (arrow). b Transverse T2-weighted MRI showing spinal cord compression at
the C4/5. c Transverse T2-weighted MRI showing the SEA at the C5– C6 (arrow). d–i Sagittal to obliquesagittal views of CT angiography showing ASA incomplete occlusion at C5 (arrows)
on MRI. So the patient did not undergo surgery and had no neurological improvement.
shoulder for 5 years. The strength of the supraspinatus was 4/5, the deltoid was 4/5, the biceps brachii was 4/5, the extensor carpi radialis longus and brevis was 5/5, and the triceps was 5/5. The right supraspinatus, deltoid, and biceps showed slightly atrophy. There were no other neurological abnormalities, no dysfunction of the cranial nerves, no hyperreflexia of the upper or lower extremities, and no sensory deficits. He stated that, at the age of 33, he was unable to flex or abduct his shoulder immediately after a
Illustrative cases Case 4 (CSA with ASA incomplete occlusion) (Fig. 1) A 38-year-old man was admitted to orthopedic department complaining of weakness of flexion and abduction of right
123
Eur Spine J
neck sprain following a fall to the ground. The patient did not go to the hospital. No record about spinal cord injury or compression of MRI was available at the time. He recovered his strength of shoulder for 6 months gradually without any medicine. Sagittal T2-weighted MRI demonstrated cord hyperintensity at C3/4, transverse T2-weighted MRI revealed the snake eye changes in anterior horn of cord at the C3–C4 level. The patient was diagnosed as CSA according to the clinical presentation and MRI. CT angiography presented ASA incomplete occlusion at C3/4. Because ASA visualization of CT angiography was not very clear to make sure it was incomplete occlusion, an MR angiography was performed to visualize the ASA and found the same result (Fig. 1). The patient did not take any medical care because of long history and no compression of spinal cord on MR images. His atrophy of the supraspinatus, deltoid, and biceps had not improved after 15 months following up. Case 8 (CSM with ASA incomplete occlusion) (Fig. 2) A 43-year-old man was admitted with limb weakness and numbness for 3 months. Neurological examination revealed that the strength of right limbs was 4/5; deep tendon reflexes were increased in right limbs; the Babinski sign was positive; sensations were almost normal. T2weighted MR images demonstrated that the cervical spinal cord was compressed at the right C4/5 intervertebral levels. However, The SEA was showing at the C5–C6, but not at C4/5. The patient was diagnosed as CSM. CT angiography presented ASA incomplete occlusion at C5 and normal ASA at C4/5. After ACDF was performed, the patient improved JOA score from 15 to 16 after 13 months following up.
Discussion ASA visualization and spinal cord ischemia The blood supply of the cervical spinal cord is derived from the vertebral artery and segmental vessels, which coalesce to form the ASA and the two posterior spinal arteries. The ASA (diameter, 0.2–0.8 mm) extends the length of the spinal cord and distributes blood to the anterior two-thirds of the spinal cord [6]. ARAs supply the cord throughout its length by anastomoses and courses through a typical hairpin turn to the ASA. The posterior spinal arteries are usually paired, course on the posterolateral surface of the spinal cord along its entire length, and may occasionally be discontinuous. Theoretically, visualization of ARAs is an ideal method to characterize spinal cord blood supply. However, they can only occasionally be
depicted in vivo by catheter angiography at present, because ARAs and posterior spinal arteries are\0.5 mm in diameter and even no flow in either direction [7]. So, to some extent, visualization of ASA is feasible to depict spinal cord blood supply and spinal cord pathological ischemia. Catheter, CT angiography and MR angiography are the high-resolution imaging techniques available for diagnosing, localizing, and classifying spinal vascular lesions [8]. Catheter angiography remains the gold standard to define the spinal cord vasculature, but it has several major drawbacks including invasive technique, ionizing radiation exposure, risk for major complications, and only be performed by experts [9, 10]. MR angiography benefits include having a strong background of suppression techniques, which allow the depiction of vessels smaller than the voxel size at acquisition [11]. However, MR angiography does not allow for visualization of the spinal cord and bone anatomy. Imaging is not only required for the lesion but it is also equally important for locating spinal vessels. This is why we chose CTA for visualizing ASA, although its disadvantages include inherent exposure to ionizing radiation and the requirement of administering a potentially nephrotoxic iodine contrast agent. MR angiography was chosen to confirm ASA occlusion only if CT angiography presented blur images. The contrast injection parameters in this paper are based on the preliminary experiments and related literature [12, 13]. Pathological changes in SEA and ASA occlusion MRI-documented SEA has suggested that its presence indicated an irreversible cystic necrosis rather than a reversible lesion of the spinal cord by clinical study and histological examination [14]. It also has been considered that patients with SEA have mechanical compression and secondary disturbance of spinal blood flow. It is likely that the hypoperfusion in the distribution of the anterior spinal artery is the cause of the ischemia of a significant longitudinal extent of the ventral horn resulting in a multisegmental damage to the cord, and wasting and weakness of the muscles [15]. Some authors suggested that venous congestion is the cause of the gray matter involvement of the cervical cord [14, 16]. It also reported postmortem findings of severe infarction of the gray matter with mild damage to the pyramidal tract in the cervical cord of a patient with distal CSA [17]. However, no ASA infarct image was reported in the patients with SEA. For seven CSM patients with SEA in our series, ASA incomplete occlusion was found in one patient, but not with 80 % spinal canal sagittal diameter compression cases. In our previous study, ASA infarct or occlusion is not commonly seen in CSM patients with spinal canal
123
Eur Spine J
sagittal diameter compression less than 80 % and sever blunt cervical spinal cord injury patients. It indicates that the spinal cord is more easily compressed than the ASA in its anterior surface [18, 19]. Although ischemia is a key factor and key contributor to the secondary pathogenesis for CSM [20], ARAs pathological injury including occluded and compressed, but not ASA occlusion, is possible reason for SEA. For visualization of ARAs, we failed through catheter, CT angiography and MR angiography (data not shown). Advancing imaging technique will definitely further improve the possibility to non-invasively visualize intradural spinal cord arteries and veins for spinal cord. For three CSA patients with SEA in our series, ASA incomplete occlusion was found in one patient, although transverse T2-weighted MRI shows the snake eye changes in anterior horn of cord in all three patients. Spinal cord infarction with selective gray matter involvement has been previously described [21]. Ischemic damage to the anterior horns has been considered to be due to the higher hypoxic susceptibility of the a-motoneurons which are located in the terminal territory of the sulcal arteries [22]. It is suggested CSA and anterior spinal artery syndrome are not separate diseases, because ASA can cause the hypoperfusion of the anterior horn throughout the multisegmental areas [15]. Although some authors attributed the motor loss to ischemic damage of the anterior horn cells [22, 23], another pathogenesis of CSA is ventral root compression [24]. So the term ‘‘anterior spinal artery syndrome’’ has not been used in reports about CSA. For the damage to the anterior horn of the CSA, we considered it may be result from ARAs infarct but not ASA infarct. The limitation and degree of cord ischemia may be the key points to differentiate CSA with cord compression and anterior spinal artery syndrome with cervical spondylosis.
intensity forming SEA may play an important role in causing segmental weakness and atrophy of the upper limbs because of the association with significant neuronal loss in the anterior horn. Recovery of neurological function in cases of SEA is thus unlikely despite successful decompressive surgery. After ACDF was performed to the cord compression in our series, neurological statuses demonstrated some degree of improvement, especially for CSM. We suggested that ASA reperfusion and decompression may be the reason for the improvement and prognosis of neurological status for the CSM and CSA patients. Although there has been pathological changes about SEA have not shown a close correlation with ASA occlusion, the limitations of this preliminary study include visualization of ARAs, patient population, and CT angiography and MRI taking during following up. Future advances in different imaging techniques will be combined to exponentially increase the knowledge gained from these types of studies.
Conclusions ASA occlusion in cervical spondylosis can be identified by CT angiography of ASA. ASA incomplete occlusion was found in two CSA and CSM patients with SEA, but not in all patients with SEA. Thus pathological changes about SEA in CSM and CSA do not have a close correlation with ASA occlusion, but may be ARAs. SEA may act as a prognostic indicator for postoperative recovery. Conflict of interest
The authors declare no conflict of interest.
References Neurological prognosis Neurological recovery after operation is to some extent related to various factors such as age at surgery, preoperative duration and severity of myelopathy, degree and extent of the cord compression, and the transverse diameter or deformity of the spinal cord. The intramedullary high signal intensity on T2-weighted MRI is associated with poor postoperative neurologic outcome in CSM patients [25, 26]. The cervical instability at the narrowest canal, large range of motion and segmental kyphosis were considered to correlate with this high signal intensity area and to be adverse prognostic factors [27, 28]. Mizuno et al. [14] have concluded that patients with SEA experienced a poor outcome after surgery compared with those in whom MR imaging demonstrated non-SEA or absent intramedullary high signal intensity. Multilevel intramedullary high signal
123
1. Yone K, Sakou T, Yanase M, Ijiri K (1992) Preoperative and postoperative magnetic resonance image evaluations of the spinal cord in cervical myelopathy. Spine 17:S388–S392 2. Al-Mefty O, Harkey LH, Middleton TH, Smith RR, Fox JL (1988) Myelopathic cervical spondylotic lesions demonstrated by magnetic resonance imaging. J Neurosurg 68:212–222 3. Takahashi M, Yamashita Y, Sakamoto Y, Kojima R (1989) Chronic cervical cord compression: clinical significance of increased signal intensity on MR images. Radiology 173:219–224 4. Yagi M, Ninomiya K, Kihara M, Horiuchi Y (2010) Long-term surgical outcome and risk factors in patients with cervical myelopathy and a change in signal intensity of intramedullary spinal cord on magnetic resonance imaging. J Neurosurg Spine 12(1): 59–65 5. Hirabayashi K, Miyakawa J, Satomi K, Maruyama T, Wakano K (1981) Operative results and postoperative progression of ossification among patients with ossification of cervical posterior longitudinal ligament. Spine 6:354–364 6. Thron A (2002) Vascular Anatomy of the Spine. Oxford University Press, Oxford
Eur Spine J 7. Krauss WE (1999) Vascular anatomy of the spinal cord. Neurosurg Clin N Am 10:9–15 8. Than KD, Sangala JR, Wang AC, Gandhi D, La Marca F, Park P (2013) The current status and recent advances in high-resolution imaging of spinal vascular malformations. J Clin Neurosci 20(1):66–71 9. Savader SJ, Williams GM, Trerotola SO, Perler BA, Wang MC, Venbrux AC et al (1993) Preoperative spinal artery localization and its relationship to postoperative neurologic complications. Radiology 189:165–171 10. Kieffer E, Fukui S, Chiras J, Koskas F, Bahnini A, Cormier E (2002) Spinal cord arteriography: a safe adjunct before descending thoracic or thoracoabdominal aortic aneurysmectomy. J Vasc Surg 35:262–268 11. Nijenhuis RJ, Leiner T, Cornips EM, Wilmink JT, Jacobs MJ, van Engelshoven JM et al (2004) Spinal cord feeding arteries at MR angiography for thoracoscopic spinal surgery: feasibility study and implications for surgical approach. Radiology 233:541–547 12. Ou P, Schmit P, Layouss W, Sidi D, Bonnet D, Brunelle F (2007) CT angiography of the artery of Adamkiewicz with 64-section technology: first experience in children. AJNR 28:216–219 13. Zhao SH, Logan L, Schraedley P, Rubin GD (2009) Assessment of the anterior spinal artery and the artery of Adamkiewicz using multi-detector CT angiography. Chin Med J (Engl) 122:145–149 14. Mizuno J, Nakagawa H, Inoue T, Hashizume Y (2003) Clinicopathological study of ‘‘snake-eye appearance’’ in compressive myelopathy of the cervical spinal cord. J Neurosurg 99(2 Suppl):162–168 15. Shibuya R, Yonenobu K, Yamamoto K, Kuratsu S, Kanazawa M, Onoue K et al (2005) Acute arm paresis with cervical spondylosis: three case reports. Surg Neurol 63:220–228 16. Tsuboi Y, Tokumarau Y, Hirayama K (1995) [Clinical difference between ‘‘proximal’’ and ‘‘distal’’ type of cervical spondylotic amyotrophy]. Rinsho Shinkeigaku 35:147–152 Jpn 17. Ebara S, Yonenobu K, Fujiwara K, Yamashita K, Ono K (1988) Myelopathy hand characterized by muscle wasting: a different type of myelopathy hand in patients with cervical spondylosis. Spine 13:785–791
18. Zhang Z, Wang H (2013) CT angiography of anterior spinal artery in cervical spondylotic myelopathy. Eur Spine J 22:2515–2519 19. Zhang Z, Wang H, Zhou Y, Wang J (2013) Computed tomographic angiography of anterior spinal artery in acute cervical spinal cord injury. Spinal Cord 51:442–447 20. Baron EM, Young WF (2007) Cervical spondylotic myelopathy: a brief review of its pathophysiology, clinical course, and diagnosis. Neurosurgery 60:S35–S41 21. Kepes JJ (1965) Selective necrosis of spinal cord grey matter: a complication of dissecting aneurysm of the aorta. Acta Neuropathol 4:29–298 22. Gilles FH, Nag D (1971) Vulnerability of human spinal cord in transient cardiac arrest. Neurology 21:833–839 23. Herrick MK, Mills PE Jr (1971) Infarction of spinal cord. Two cases of selective gray matter involvement secondary to asymptomatic aortic disease. Arch Neurol 24:228–241 24. Keegan JJ (1965) The cause of dissociated motor loss in the upper extremity with cervical spondylosis; a case report. J Neurosurg 23:528–536 25. Vedantam A, Rajshekhar V (2013) Does the type of T2-weighted hyperintensity influence surgical outcome in patients with cervical spondylotic myelopathy? A review. Eur Spine J 22:96–106 26. Uchida K, Nakajima H, Takeura N, Yayama T, Guerrero AR, Yoshida A, Sakamoto T, Honjoh K, Baba H (2013) Prognostic value of changes in spinal cord signal intensity on magnetic resonance imaging in patients with cervical compressive myelopathy. Spine J. doi:10.1016/j.spinee.2013.09.038 27. Yagi M, Ninomiya K, Kihara M, Horiuchi Y (2010) Long-term surgical outcome and risk factors in patients with cervical myelopathy and a change in signal intensity of intramedullary spinal cord on magnetic resonance imaging. J Neurosurg Spine 12:59–65 28. Oshima Y, Seichi A, Takeshita K, Ono T, Baba S, Morii J, Oka H, Kawaguchi H, Nakamura K, Tanaka S (2012) Natural course and prognostic factors in patients with mild cervical spondylotic myelopathy with increased signal intensity on T2-weighted magnetic resonance imaging. Spine 37(22):1909–1913
123
ORIGINAL ARTICLE
Is the ‘‘snake-eye’’ MRI sign correlated to anterior spinal artery occlusion on CT angiography in cervical spondylotic myelopathy and amyotrophy? Zhengfeng Zhang • Honggang Wang
Received: 9 October 2013 / Revised: 26 April 2014 / Accepted: 26 April 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Objective The goal of this study was to identify anterior spinal artery (ASA) occlusion by CT angiography in cervical spondylotic myelopathy (CSM) and amyotrophy (CSA) with T2-weighted hyperintensity of MR image of documented small intramedullary high signal intensity known as ‘‘snake-eye appearance’’ (SEA). Method One hundred and six patients with CSM were admitted to the investigator group between June 2010 and June 2013. Intramedullary high signal intensity was found in 42 cases and was divided into two types, SEA and nonSEA. SEA was observed in 10 patients, including seven CSM patients and three CSA patients. All SEA patients were performed CT angiography of ASA after admission. Results The ASA was visualized in all 10 patients. ASA incomplete occlusion was found in one CSA patient and one CSM patient. No ASA occlusion was found in other CSA and CSM patients with SEA. Conclusion ASA occlusion is not commonly seen in CSM and CSA patients with SEA. Pathological changes about SEA in CSM and CSA have no close correlation with ASA occlusion, but may be with anterior radiculomedullary arteries. Keywords Cervical spondylotic myelopathy Cervical spondylotic amyotrophy CT angiography Anterior spinal artery Snake-eye appearance
Magnetic resonance imaging (MRI) of the cervical spine is essential for the preoperative evaluation of patients with cervical spondylotic myelopathy (CSM) and amyotrophy (CSA). The intramedullary high signal intensity on T2weighted MRI has been found in 41–97.2 % of patients with CSM [1]. Snake-eye appearance (SEA) is a unique neuroimaging finding characterized as nearly symmetrical round high signal intensity of the spinal parenchyma resembling the face of a snake demonstrated on axial T2weighted MR images [2, 3]. This condition is sometimes observed in CSM and CSA; however, its pathophysiological features are uncertain. It has been reported that edema, myelomalacia, gliosis, and inflammation are involved in intramedullary high signal intensity, which have suggested necrosis and ischemia [4]. The anterior spinal arteries (ASA) supply the anterior horn and the anterior part of the lateral column on the left or right side at each level of spinal cord, which may be theoretically compressed by herniated disc and resulted in spinal cord ischemia in CSM and CSA. SEA-related intramedullary high signal intensity is located in anterior horn, which is supplied by ASA. So we hypothesize that SEA-related intramedullary high signal intensity in CSM and CSA have a close correlation with ASA occlusion. The objective of this study was to identify ASA occlusion by CT angiography in CSM and CSA patients with SEA.
Methods Patients Z. Zhang (&) H. Wang Department of Orthopedics, Xinqiao Hospital, Third Military Medical University, 183 Xinqiao Street, Shapingba District, Chongqing 400037, China e-mail: [email protected]
Between June 2010 and June 2013, 106 patients with diagnosis of CSM were admitted to the investigator group (Z. Z.). Among them, intramedullary high signal intensity was
123
Eur Spine J Table 1 Clinical summary of patients Case no.
Sex/ Age
Level of compression
T2 cord hyperintensity
Location of T2 cord hyperintensity
Mode of myelopathy
ASA occlusion
Spinal canal sagittal diameter compression (%)
Type of operation
JOA score preop/ postop
1
M/44
C4/5 C5/6
C4–C6
Anterior horn
CSM
No
70
ACDF
15/17
2
M/71
C4/5 C5/6
C4–C6
Anterior horn
CSM
No
30
ACDF
15/16
3 4
F/52 M/38
C4/5 C3/4
C4/5 C3/4
Left anterior horn Anterior horn
CSM CSA
No Incomplete
40 10
ACDF None
16/17 16/16
5
M/65
C4/5 C5/6
C4/5
Anterior horn
CCM, CSA
No
80
ACDF
15/16
6
M/47
C4/5
C4/5
Right anterior horn
CSA
No
40
ACDF
16/17
7
M/58
C3/4
C3/4
Anterior horn
CCM
No
20
ACDF
16/17
8
M/43
C4/5
C5–C6
Anterior horn
CCM
Incomplete
20
ACDF
15/17
9
M/42
C4/5
C4/5
Anterior horn
CCM
No
80
ACDF
16/17
10
M/44
C4/5 C5/6
C4–C6
Anterior horn
CCM
No
60
ACDF
15/17
M male, F female, C cervical, CSM cervical spondylotic myelopathy, CSA cervical spondylotic amyotrophy, ACDF anterior cervical disctomy fusion
observed in 42 patients on axial T2-weighted MR imaging using a 1.5-tesla imager. The patterns of intramedullary high signal intensity were divided into two groups, SEA and nonSEA. SEA was defined as one left- and/or one right-sided small round or elliptical high signal intensity lesion in the central gray matter near the ventrolateral posterior column. Of the 42 patients with intramedullary high signal intensity, SEA was observed in 10 patients and non-SEA in 32. In 106 patients, three patients were identified as CSA according to the following fulfilled criteria: (1) a Nurick grade of 0 or 1 (no gait impairment); (2) symptoms and signs present weakness and wasting of upper limb muscles as difficulty in shoulder abduction, positive arm-drop sign, or positive wristdrop sign; and (3) without sensory or lower limb involvement. All three CSA patients presented SEA in MRI. The clinical characteristics of the SEA patients are summarized in Table 1. In these 10 patients (9 males and 1 female), the distribution of cervical level was from C3 to C6; the duration from onset until surgery ranged from 1 to 6 months; following up ranged from 4 to 16 months (mean, 8.3 months). All patients underwent plain cervical radiographs, CT scans, and 1.5-Tesla MRI. T1-weighted and T2-weighted MRIs were performed to define cord compression, spinal canal sagittal diameter compression proportion, and the extent of intramedullary high signal intensity. Nine of 10 patients were performed anterior decompression and fusion surgery, except one CSA patient who presented no cord compression on MRI. Preoperative and postoperative neurological statuses were evaluated using JOA scoring system and Hirabayashi formula [5]. CT angiography of ASA With the institutional review board approval of the research protocol, CT angiography of ASA was performed
123
on the SEA patients. One or 2 days after the patients underwent MRI, were scanned with 64-slice MDCT scanners (LightSpeed 64, GE Medical System, Milwaukee, WI, USA). All CT scans were obtained with 0.5 s rotation, 0.625 mm nominal detector widths, pitch of 0.984, 120 kV, and 480 mA. Transverse sections were reconstructed with 50 % overlap relative to the effective section thickness of 0.6 mm. Iohexol (Amersham Health, Princeton, NJ, USA) was administered through an antecubital vein with a dose of 120–150 ml (350 mgI/ml) at a rate of 5 ml/s. The scan delay was determined by a preliminary 20 ml test injection at the level of basilar artery. All observations were made retrospectively by two neuroradiologists. To examine the ASA transverse sections, multiplanar reformations (MPR), curved planar reformation (CPR), and thin-slab (2–4 mm) maximum intensity projections (MIP) were generated and displayed on a workstation (Advantage Workstation 4.3, GE Medical Systems) with window and level settings selected to maximize arterial to background discrimination. The ASA was identified by maintaining the MIP-slab parallel to the anterior surface of the spinal cord at each vertebral level assessed from C1 to C7. An enhanced artery on the midline ventral surface of the spinal cord was interpreted as the ASA. At the same time, an artery originating from the aorta and coursing through the intervertebral foramen to join the ASA in a hairpin configuration was interpreted as the Adamkiewicz artery (AKA). Co-visualizations of ASA and AKA were used to identify ASA; covisualizations of veins were not the criteria for identifying ASA. Another criterion was that the ASA originated from vertebral artery in skull base. Inter-rater reliability, in the form of consensual identification of ASA by two radiologists, was required.
Eur Spine J
Fig. 1 Case 4. A 38-year-old man with CSA at C3/4. a Sagittal T2weighted MRI showing cord hyperintensity at C3/4 (arrows). b, c Transverse T2-weighted MRI showing the snake eyes changes in anterior horn of cord at the C3–C4 level (arrows). d, e Sagittal view
(d) and coronal view (f) of CT angiography showing suspected ASA incomplete occlusion at C3–C4 (arrows). (f–h) Coronal view (f, g) and sagittal view (h) of MR angiography showing ASA incomplete occlusion at C3–C4 (arrows)
Results
(80 %), no ASA occlusion was found, even for some cases with 80 % spinal canal sagittal diameter compression.
ASA occlusion Neurological status The clinical data are summarized in Table 1. The ASA was visualized in all 10 SEA patients. The course of the ASA extended from C1 to C7 in sagittal, coronal and transverse view. The ASA runs along the entire length of the anterior surface of the spinal cord. Anterior radiculomedullary arteries (ARAs), ASA brand which supply the cord throughout its length by anastomoses branching upward and downward, were not visualized for the angiography, because ARAs are very little or even no flow in two directions. The Adamkiewicz artery (AKA) was also visualized in most cases. For three CSA cases, ASA incomplete occlusion was found in one patient (case 4) (Fig. 1). For seven CSM cases, one case (case 8) presented ASA incomplete occlusion (Fig. 2). For other eight CSM and CSA patients
Preoperative and postoperative neurological statuses of the patients are given in Table 1. All seven CSM patients had undergone anterior cervical disctomy/corpectomy fusion (ACDF/ACCF) and did not encounter complications. Six patients with JOA score 15 or 16 improved to 17 after surgery. One patient (case 2) only improved JOA score from 15 to 16, but not 17, after 14 months following up. For three CSA patients, two patients presented cervical compression and underwent ACDF. The 65 years of age patient (case 5) who presented both CSM and CSA improved neurological status from 15 to 16 of JOA score after 12 months following up. Another patient (case 6) improved JOA score from 16 to 17 after surgery. The third patient (illustrative case 4) presented no cord compression
123
Eur Spine J
Fig. 2 Case 8. A 43-year-old man with CSM at C4/5. a Sagittal T2weighted MRI reveals that the cervical spinal cord was compressed at C4/5. Spinal cord hyperintensity is showing from C5–C6 (arrow). b Transverse T2-weighted MRI showing spinal cord compression at
the C4/5. c Transverse T2-weighted MRI showing the SEA at the C5– C6 (arrow). d–i Sagittal to obliquesagittal views of CT angiography showing ASA incomplete occlusion at C5 (arrows)
on MRI. So the patient did not undergo surgery and had no neurological improvement.
shoulder for 5 years. The strength of the supraspinatus was 4/5, the deltoid was 4/5, the biceps brachii was 4/5, the extensor carpi radialis longus and brevis was 5/5, and the triceps was 5/5. The right supraspinatus, deltoid, and biceps showed slightly atrophy. There were no other neurological abnormalities, no dysfunction of the cranial nerves, no hyperreflexia of the upper or lower extremities, and no sensory deficits. He stated that, at the age of 33, he was unable to flex or abduct his shoulder immediately after a
Illustrative cases Case 4 (CSA with ASA incomplete occlusion) (Fig. 1) A 38-year-old man was admitted to orthopedic department complaining of weakness of flexion and abduction of right
123
Eur Spine J
neck sprain following a fall to the ground. The patient did not go to the hospital. No record about spinal cord injury or compression of MRI was available at the time. He recovered his strength of shoulder for 6 months gradually without any medicine. Sagittal T2-weighted MRI demonstrated cord hyperintensity at C3/4, transverse T2-weighted MRI revealed the snake eye changes in anterior horn of cord at the C3–C4 level. The patient was diagnosed as CSA according to the clinical presentation and MRI. CT angiography presented ASA incomplete occlusion at C3/4. Because ASA visualization of CT angiography was not very clear to make sure it was incomplete occlusion, an MR angiography was performed to visualize the ASA and found the same result (Fig. 1). The patient did not take any medical care because of long history and no compression of spinal cord on MR images. His atrophy of the supraspinatus, deltoid, and biceps had not improved after 15 months following up. Case 8 (CSM with ASA incomplete occlusion) (Fig. 2) A 43-year-old man was admitted with limb weakness and numbness for 3 months. Neurological examination revealed that the strength of right limbs was 4/5; deep tendon reflexes were increased in right limbs; the Babinski sign was positive; sensations were almost normal. T2weighted MR images demonstrated that the cervical spinal cord was compressed at the right C4/5 intervertebral levels. However, The SEA was showing at the C5–C6, but not at C4/5. The patient was diagnosed as CSM. CT angiography presented ASA incomplete occlusion at C5 and normal ASA at C4/5. After ACDF was performed, the patient improved JOA score from 15 to 16 after 13 months following up.
Discussion ASA visualization and spinal cord ischemia The blood supply of the cervical spinal cord is derived from the vertebral artery and segmental vessels, which coalesce to form the ASA and the two posterior spinal arteries. The ASA (diameter, 0.2–0.8 mm) extends the length of the spinal cord and distributes blood to the anterior two-thirds of the spinal cord [6]. ARAs supply the cord throughout its length by anastomoses and courses through a typical hairpin turn to the ASA. The posterior spinal arteries are usually paired, course on the posterolateral surface of the spinal cord along its entire length, and may occasionally be discontinuous. Theoretically, visualization of ARAs is an ideal method to characterize spinal cord blood supply. However, they can only occasionally be
depicted in vivo by catheter angiography at present, because ARAs and posterior spinal arteries are\0.5 mm in diameter and even no flow in either direction [7]. So, to some extent, visualization of ASA is feasible to depict spinal cord blood supply and spinal cord pathological ischemia. Catheter, CT angiography and MR angiography are the high-resolution imaging techniques available for diagnosing, localizing, and classifying spinal vascular lesions [8]. Catheter angiography remains the gold standard to define the spinal cord vasculature, but it has several major drawbacks including invasive technique, ionizing radiation exposure, risk for major complications, and only be performed by experts [9, 10]. MR angiography benefits include having a strong background of suppression techniques, which allow the depiction of vessels smaller than the voxel size at acquisition [11]. However, MR angiography does not allow for visualization of the spinal cord and bone anatomy. Imaging is not only required for the lesion but it is also equally important for locating spinal vessels. This is why we chose CTA for visualizing ASA, although its disadvantages include inherent exposure to ionizing radiation and the requirement of administering a potentially nephrotoxic iodine contrast agent. MR angiography was chosen to confirm ASA occlusion only if CT angiography presented blur images. The contrast injection parameters in this paper are based on the preliminary experiments and related literature [12, 13]. Pathological changes in SEA and ASA occlusion MRI-documented SEA has suggested that its presence indicated an irreversible cystic necrosis rather than a reversible lesion of the spinal cord by clinical study and histological examination [14]. It also has been considered that patients with SEA have mechanical compression and secondary disturbance of spinal blood flow. It is likely that the hypoperfusion in the distribution of the anterior spinal artery is the cause of the ischemia of a significant longitudinal extent of the ventral horn resulting in a multisegmental damage to the cord, and wasting and weakness of the muscles [15]. Some authors suggested that venous congestion is the cause of the gray matter involvement of the cervical cord [14, 16]. It also reported postmortem findings of severe infarction of the gray matter with mild damage to the pyramidal tract in the cervical cord of a patient with distal CSA [17]. However, no ASA infarct image was reported in the patients with SEA. For seven CSM patients with SEA in our series, ASA incomplete occlusion was found in one patient, but not with 80 % spinal canal sagittal diameter compression cases. In our previous study, ASA infarct or occlusion is not commonly seen in CSM patients with spinal canal
123
Eur Spine J
sagittal diameter compression less than 80 % and sever blunt cervical spinal cord injury patients. It indicates that the spinal cord is more easily compressed than the ASA in its anterior surface [18, 19]. Although ischemia is a key factor and key contributor to the secondary pathogenesis for CSM [20], ARAs pathological injury including occluded and compressed, but not ASA occlusion, is possible reason for SEA. For visualization of ARAs, we failed through catheter, CT angiography and MR angiography (data not shown). Advancing imaging technique will definitely further improve the possibility to non-invasively visualize intradural spinal cord arteries and veins for spinal cord. For three CSA patients with SEA in our series, ASA incomplete occlusion was found in one patient, although transverse T2-weighted MRI shows the snake eye changes in anterior horn of cord in all three patients. Spinal cord infarction with selective gray matter involvement has been previously described [21]. Ischemic damage to the anterior horns has been considered to be due to the higher hypoxic susceptibility of the a-motoneurons which are located in the terminal territory of the sulcal arteries [22]. It is suggested CSA and anterior spinal artery syndrome are not separate diseases, because ASA can cause the hypoperfusion of the anterior horn throughout the multisegmental areas [15]. Although some authors attributed the motor loss to ischemic damage of the anterior horn cells [22, 23], another pathogenesis of CSA is ventral root compression [24]. So the term ‘‘anterior spinal artery syndrome’’ has not been used in reports about CSA. For the damage to the anterior horn of the CSA, we considered it may be result from ARAs infarct but not ASA infarct. The limitation and degree of cord ischemia may be the key points to differentiate CSA with cord compression and anterior spinal artery syndrome with cervical spondylosis.
intensity forming SEA may play an important role in causing segmental weakness and atrophy of the upper limbs because of the association with significant neuronal loss in the anterior horn. Recovery of neurological function in cases of SEA is thus unlikely despite successful decompressive surgery. After ACDF was performed to the cord compression in our series, neurological statuses demonstrated some degree of improvement, especially for CSM. We suggested that ASA reperfusion and decompression may be the reason for the improvement and prognosis of neurological status for the CSM and CSA patients. Although there has been pathological changes about SEA have not shown a close correlation with ASA occlusion, the limitations of this preliminary study include visualization of ARAs, patient population, and CT angiography and MRI taking during following up. Future advances in different imaging techniques will be combined to exponentially increase the knowledge gained from these types of studies.
Conclusions ASA occlusion in cervical spondylosis can be identified by CT angiography of ASA. ASA incomplete occlusion was found in two CSA and CSM patients with SEA, but not in all patients with SEA. Thus pathological changes about SEA in CSM and CSA do not have a close correlation with ASA occlusion, but may be ARAs. SEA may act as a prognostic indicator for postoperative recovery. Conflict of interest
The authors declare no conflict of interest.
References Neurological prognosis Neurological recovery after operation is to some extent related to various factors such as age at surgery, preoperative duration and severity of myelopathy, degree and extent of the cord compression, and the transverse diameter or deformity of the spinal cord. The intramedullary high signal intensity on T2-weighted MRI is associated with poor postoperative neurologic outcome in CSM patients [25, 26]. The cervical instability at the narrowest canal, large range of motion and segmental kyphosis were considered to correlate with this high signal intensity area and to be adverse prognostic factors [27, 28]. Mizuno et al. [14] have concluded that patients with SEA experienced a poor outcome after surgery compared with those in whom MR imaging demonstrated non-SEA or absent intramedullary high signal intensity. Multilevel intramedullary high signal
123
1. Yone K, Sakou T, Yanase M, Ijiri K (1992) Preoperative and postoperative magnetic resonance image evaluations of the spinal cord in cervical myelopathy. Spine 17:S388–S392 2. Al-Mefty O, Harkey LH, Middleton TH, Smith RR, Fox JL (1988) Myelopathic cervical spondylotic lesions demonstrated by magnetic resonance imaging. J Neurosurg 68:212–222 3. Takahashi M, Yamashita Y, Sakamoto Y, Kojima R (1989) Chronic cervical cord compression: clinical significance of increased signal intensity on MR images. Radiology 173:219–224 4. Yagi M, Ninomiya K, Kihara M, Horiuchi Y (2010) Long-term surgical outcome and risk factors in patients with cervical myelopathy and a change in signal intensity of intramedullary spinal cord on magnetic resonance imaging. J Neurosurg Spine 12(1): 59–65 5. Hirabayashi K, Miyakawa J, Satomi K, Maruyama T, Wakano K (1981) Operative results and postoperative progression of ossification among patients with ossification of cervical posterior longitudinal ligament. Spine 6:354–364 6. Thron A (2002) Vascular Anatomy of the Spine. Oxford University Press, Oxford
Eur Spine J 7. Krauss WE (1999) Vascular anatomy of the spinal cord. Neurosurg Clin N Am 10:9–15 8. Than KD, Sangala JR, Wang AC, Gandhi D, La Marca F, Park P (2013) The current status and recent advances in high-resolution imaging of spinal vascular malformations. J Clin Neurosci 20(1):66–71 9. Savader SJ, Williams GM, Trerotola SO, Perler BA, Wang MC, Venbrux AC et al (1993) Preoperative spinal artery localization and its relationship to postoperative neurologic complications. Radiology 189:165–171 10. Kieffer E, Fukui S, Chiras J, Koskas F, Bahnini A, Cormier E (2002) Spinal cord arteriography: a safe adjunct before descending thoracic or thoracoabdominal aortic aneurysmectomy. J Vasc Surg 35:262–268 11. Nijenhuis RJ, Leiner T, Cornips EM, Wilmink JT, Jacobs MJ, van Engelshoven JM et al (2004) Spinal cord feeding arteries at MR angiography for thoracoscopic spinal surgery: feasibility study and implications for surgical approach. Radiology 233:541–547 12. Ou P, Schmit P, Layouss W, Sidi D, Bonnet D, Brunelle F (2007) CT angiography of the artery of Adamkiewicz with 64-section technology: first experience in children. AJNR 28:216–219 13. Zhao SH, Logan L, Schraedley P, Rubin GD (2009) Assessment of the anterior spinal artery and the artery of Adamkiewicz using multi-detector CT angiography. Chin Med J (Engl) 122:145–149 14. Mizuno J, Nakagawa H, Inoue T, Hashizume Y (2003) Clinicopathological study of ‘‘snake-eye appearance’’ in compressive myelopathy of the cervical spinal cord. J Neurosurg 99(2 Suppl):162–168 15. Shibuya R, Yonenobu K, Yamamoto K, Kuratsu S, Kanazawa M, Onoue K et al (2005) Acute arm paresis with cervical spondylosis: three case reports. Surg Neurol 63:220–228 16. Tsuboi Y, Tokumarau Y, Hirayama K (1995) [Clinical difference between ‘‘proximal’’ and ‘‘distal’’ type of cervical spondylotic amyotrophy]. Rinsho Shinkeigaku 35:147–152 Jpn 17. Ebara S, Yonenobu K, Fujiwara K, Yamashita K, Ono K (1988) Myelopathy hand characterized by muscle wasting: a different type of myelopathy hand in patients with cervical spondylosis. Spine 13:785–791
18. Zhang Z, Wang H (2013) CT angiography of anterior spinal artery in cervical spondylotic myelopathy. Eur Spine J 22:2515–2519 19. Zhang Z, Wang H, Zhou Y, Wang J (2013) Computed tomographic angiography of anterior spinal artery in acute cervical spinal cord injury. Spinal Cord 51:442–447 20. Baron EM, Young WF (2007) Cervical spondylotic myelopathy: a brief review of its pathophysiology, clinical course, and diagnosis. Neurosurgery 60:S35–S41 21. Kepes JJ (1965) Selective necrosis of spinal cord grey matter: a complication of dissecting aneurysm of the aorta. Acta Neuropathol 4:29–298 22. Gilles FH, Nag D (1971) Vulnerability of human spinal cord in transient cardiac arrest. Neurology 21:833–839 23. Herrick MK, Mills PE Jr (1971) Infarction of spinal cord. Two cases of selective gray matter involvement secondary to asymptomatic aortic disease. Arch Neurol 24:228–241 24. Keegan JJ (1965) The cause of dissociated motor loss in the upper extremity with cervical spondylosis; a case report. J Neurosurg 23:528–536 25. Vedantam A, Rajshekhar V (2013) Does the type of T2-weighted hyperintensity influence surgical outcome in patients with cervical spondylotic myelopathy? A review. Eur Spine J 22:96–106 26. Uchida K, Nakajima H, Takeura N, Yayama T, Guerrero AR, Yoshida A, Sakamoto T, Honjoh K, Baba H (2013) Prognostic value of changes in spinal cord signal intensity on magnetic resonance imaging in patients with cervical compressive myelopathy. Spine J. doi:10.1016/j.spinee.2013.09.038 27. Yagi M, Ninomiya K, Kihara M, Horiuchi Y (2010) Long-term surgical outcome and risk factors in patients with cervical myelopathy and a change in signal intensity of intramedullary spinal cord on magnetic resonance imaging. J Neurosurg Spine 12:59–65 28. Oshima Y, Seichi A, Takeshita K, Ono T, Baba S, Morii J, Oka H, Kawaguchi H, Nakamura K, Tanaka S (2012) Natural course and prognostic factors in patients with mild cervical spondylotic myelopathy with increased signal intensity on T2-weighted magnetic resonance imaging. Spine 37(22):1909–1913
123
