Myelopathies of unknown origin (MUO) constitute a challenge for the clinician, due to the absence of image findings that explain the symptoms of the patients.
The role of spinal angiography (SA) in MUO has not been studied before in the medical literature.
MethodsRetrospective study of 93 SA performed in a neurointerventional reference center in a ten-year period (January 2013–December 2022). Suspected, angiopraphic and final diagnosis for each patient were studied and compared.
ResultsSA constituted an 0.8% of the total neurointerventional procedures performed (93 out of 10997). Suspected diagnosis were: dural fistula 39 patients (41.9%), vertebral tumours 19 (20.4%), MUO 12 (12.9%), spinal arteriovenous malformation 9 (9.6%), spontaneous epidural hematoma 8 (8.6%), spinal stroke 2 (2.2%), angioma 2 (2.2%), cavernoma 1 (1.1%), y spinal artery aneurism 1 (1.1%). Angiographic diagnoses in patients with MUO were: no abnormal findings in 11 patients (91,7%) and one cervical dural fistula dural (8.3%). Final diagnosis in patients with MUO without abnormal findings in the SA were: spinal muscular atrophy, transverse myelitis, syringomyelia, sciatic neuropathy, chronic back pain, lumbosacral neuropathy, simulated paraplegia, lumbosacral plexopathy, somatoform disorder, autoinmune myelopathy and L4/L5 discal herniation.
ConclusionsIn this series of 93 patients, MUO were the third more frequent indication for SA, only after dural fistula and vertebral tumours. Other indications were spinal arteriovenous malformations, spinal epidural hematomas, cavernomas, spinal artery aneurism and spinal stroke. SA seems to be a useful tool in the management of some cases of MUO.
Las mielopatías de origen desconocido (MOE) constituyen un grupo de patologías de difícil manejo, dada la ausencia de hallazgos en las pruebas de imagen no invasivas que justifiquen la clínica. La utilidad de la arteriografía medular (AM) en este contexto no ha sido estudiada antes.
MétodosEstudio retrospectivo de las 93 AM diagnósticas realizadas en un centro regional de neurointervencionismo a lo largo de un periodo de diez años (Enero 2013-Diciembre 2022). Se analizaron los diagnósticos de sospecha, diagnósticos angiográficos y diagnósticos finales de los pacientes.
ResultadosLas AM constituyeron un 0.8% del total de procedimientos neurointervencionistas (93 de 10997). Diagnósticos de sospecha: fístula dural 39 pacientes (41.9%), tumor vertebral 19 (20.4%), MOD 12 (12.9%), malformacion arteriovenosa medular 9 (9.6%), hematoma epidural espontáneo 8 (8.6%), infarto medular 2 (2.2%), angioma 2 (2.2%), cavernoma 1 (1.1%), y aneurisma de arteria espinal 1 (1.1%). Diagnósticos angiográficos en los pacientes con MOD: estudio sin hallazgos patológicos en 11 casos (91,7%) y fístula dural en 1 paciente (8.3%). Diagnósticos finales en los pacientes con MOD: atrofia muscular espinal, mielitis transversa, siringomielia, neuropatía del ciático, lumbociática crónica, radiculopatía crónica lumbosacra, paraplejia psicógena, plexopatía lumbosacra, trastorno somatomorfo, mielopatía autoinmune y hernia discal L4/L5.
ConclusionesEn esta serie de 93 AM, las MOD constituyen el tercer motivo de petición más frecuente en la práctica diaria, detrás de las fístulas durales y los tumores vertebrales. Otras indicaciones son hematomas epidurales, malformaciones arteriovenosas, cavernomas, aneurisma de arteria espinal e ictus espinal. La AM puede ser una herramienta útil en el manejo de algunos pacientes con MOD.
Myelopathy is one of the main indications for spinal cord neuroimaging, particularly MRI and CT. Although most causes of spinal injury can be diagnosed or confirmed with these techniques, a significant proportion of cases cannot be explained by neuroimaging and are classified as myelopathy of unknown origin (MUO).1 One possible cause of MUO misdiagnosis is spinal vascular malformations (SVM), especially dural fistulas, which are the most frequent type of SVM and one of the most complex to diagnose with MRI and/or MR angiography, particularly in the absence of marked perimedullary venous engorgement or spinal cord oedema.2–6 In this context, spinal angiography (SA) is a potentially useful diagnostic tool, although no study to date has analysed its applicability in MUO. We analysed the SA studies performed at a regional referral centre for interventional neuroradiology over a period of 10 years, with a view to determining the real usefulness of this technique in the management of MUO, confirming its indications in other conditions, and determining its diagnostic yield.7–14
Material and methodsWe conducted a retrospective, descriptive study of SA procedures performed at Hospital Clínico Universitario Virgen de la Arrixaca between January 2013 and December 2022. This hospital is the only referral centre for interventional neuroradiology in the region of Murcia, serving a population of approximately 1.5 million. It is equipped with 2 biplane angiography suites where all neurovascular procedures are performed, and staffed by a team of 6 interventional neuroradiologists. The team covers both suites during daytime hours, and is available on-call for urgent neurovascular care (code stroke, aneurysmal subarachnoid haemorrhage) 365 days a year.
We reviewed all SA studies performed at our centre during the study period. We selected those performed for diagnostic purposes, recording both the suspected diagnosis and the angiographic diagnosis. In cases where SA showed no pathological findings, data on patients’ clinical course were gathered from the medical records and the final diagnosis was recorded.
Spinal angiography protocol. Following femoral artery puncture and placement of an introducer sheath connected to a heparinised flush system, angiography is performed either as selective regional angiography of the area of interest (spinal or vertebral tumours) or as complete spinal angiography, which involves catheterisation of the vertebral, subclavian, intercostal, lumbar, and sacral arteries (other conditions). The procedure can be performed under sedation or general anaesthesia, depending on the degree of patient cooperation and the duration of the procedure, and occasionally needs to be performed in 2 sessions due to its long duration. The catheters used vary in size (typically 4F or 5F) and curve complexity to adapt to each patient’s anatomy and the region under study (JB©, SideWinder© 1 or 2, Vertebral©, Cobra© 1 or 2, etc). Injections into segmental or lumbar arteries are usually performed manually using 1-mL syringes, and apnoea may be induced to prevent artefacts. When a dural fistula is suspected, initial injection into the aortic arch and/or the abdominal aorta is frequently performed using a multi-holed catheter (Pigtail©) with 30-40 mL of contrast medium, aiming to determine its approximate location.
Suspected diagnosis. The suspected diagnosis, as indicated by the requesting specialist (neurologist or neurosurgeon), was recorded: spinal dural fistula, spinal arteriovenous malformation, vertebral tumour, spinal cord infarction, epidural haematoma, spinal aneurysm, angioma, cavernoma, or MUO.
Angiographic diagnosis. The following final diagnoses were established after SA:
- 1
Spinal dural fistula: direct communication between a radiculomeningeal artery and a radiculomedullary vein, without the normal intermediary capillary network.
- 2
Spinal intradural fistula: direct communication between a spinal or radiculopial artery and a radiculomedullary vein, without the normal intermediary capillary network.
- 3
Spinal arteriovenous malformation: communication between branches of spinal arteries and a spinal vein through a nidus located within the intramedullary or perimedullary parenchyma.
- 4
Spinal aneurysm: saccular or dysplastic dilation at some point along the trajectory of a spinal artery.
- 5
Anterior spinal artery occlusion: abrupt interruption of blood flow in the anterior spinal artery.
- 6
Vertebral tumour: hypervascular lesion presenting blush at the level of the vertebral body.
Final diagnosis. In patients whose SA study detected no pathological findings, electronic medical records were reviewed to gather data on the clinical course and final diagnosis.
ResultsBetween January 2013 and December 2022, a total of 10 997 neurointerventional procedures were performed at our centre, 132 of which were SA studies (1.2%). Of these, 93 were performed for diagnostic purposes and 39 for therapeutic purposes. For the purposes of this study, only diagnostic SA studies were included (0.8%).
Among these, 22 (23.6%) focused specifically on the lesion site (vertebral tumours, spinal cord infarction, and spinal aneurysm) and the remaining 71 (76.3%) studied the whole spinal cord. The mean volume of contrast medium used was 160.86 mL. The mean duration of the procedure was 87.6 min.
One patient (1.1%) presented a self-limiting episode of severe lumbar pain and extremely painful paraesthesia in both legs following selective contrast injection into a left T10 segmental artery. Symptoms resolved gradually after the catheter was withdrawn from the arterial ostium. No other complications were recorded.
As shown in Table 1, the suspected diagnoses were as follows: spinal dural fistula in 39 patients (41.9%), vertebral tumour in 19 (20.4%), MUO in 12 (12.9%), spinal arteriovenous malformation in 9 (9.6%), spontaneous epidural haematoma in 8 (8.6%), spinal cord infarction in 2 (2.2%), angioma in 2 (2.2%), cavernoma in one (1.1%), and spinal artery aneurysm in one (1.1%).
Suspected, angiographic, and final diagnoses.
| Angiographic diagnosis → n (%) Suspected diagnosis ↓ n (%) | Spinal dural fistula | Vertebral tumour | Spinal arteriovenous malformation | Anterior spinal artery occlusion | Spinal intradural fistula | Spinal artery aneurysm | No pathological findings / Final diagnosis |
|---|---|---|---|---|---|---|---|
| Spinal dural fistula: 39 (41.9%) | 27 (69.2%) | 1 (2.6%) | 1 (2.6%) | 10 (25.6%) / Multiple sclerosis (3), unknown origin (2), myelitis, spinal cord infarction, spondylolisthesis L5-S1, lumbar spinal stenosis, spinal collaterals due to Takayasu arteritis | |||
| Vertebral tumour: 19 (20.4%) | 19 (100%) | ||||||
| Myelopathy of unknown origin: 12 (12.9%) | 1 (8.3%) | 11 (91.7%) / Spinal muscular atrophy, transverse myelitis, syringomyelia, sciatic neuropathy, chronic sciatica and lumbar pain, chronic lumbosacral radiculopathy, psychogenic paraplegia, lumbosacral plexopathy, somatic symptom disorder, autoimmune myelopathy, L4-L5 spinal disc herniation | |||||
| Spinal arteriovenous malformation: 9 (9.6%) | 7 (77.8%) | 2 (22.2%) / Cavernoma (2) | |||||
| Epidural haematoma: 8 (8.6%) | 8 (100%) / Spontaneous epidural haematoma (8) | ||||||
| Spinal cord infarction: 2 (2.2%) | 1 (50%) | 1 (50%) / Spinal cord infarction | |||||
| Angioma: 2 (2.2%) | 2 (100%) / Cavernoma (2) | ||||||
| Spinal artery aneurysm: 1 (1.1%) | 1 (100%) | ||||||
| Cavernoma: 1 (1.1%) | 1 (100%) / Cavernoma |
Among patients with suspected spinal dural fistula, the angiographic diagnosis was: dural fistula in 27 patients (69.2%), no pathological findings in 10 (25.6%), spinal arteriovenous malformation in one (2.6%), and spinal intradural fistula in one (2.6%).
In all cases of suspected vertebral tumour, the diagnosis was confirmed by SA.
Among patients referred for MUO, angiographic diagnosis was no pathological findings in 11 (91.7%) and cervical dural fistula in one (8.3%) (Fig. 1). The final diagnosis of the patients showing no pathological findings on SA are listed in Table 1.
Cervical spinal dural fistula. (A) T2-weighted MRI study (sagittal plane) showing absence of alterations in the cervical spinal cord. B-D) 3D reconstruction (B) and digital subtraction angiography images in the anteroposterior (C) and lateral projections (D) showing the location of the fistula (red arrow) and the associated venous drainage (blue arrows).
Among patients with suspected spinal arteriovenous malformations, angiographic diagnosis was spinal arteriovenous malformation in 7 (77.8%) and no pathological findings in 2 (22.2%). The latter 2 patients were ultimately diagnosed with cavernomas.
In all patients with suspected spontaneous epidural haematoma, SA revealed no pathological findings, and follow-up studies revealed no underlying vascular lesions.
In the 2 patients with suspected spinal cord infarction, SA revealed anterior spinal artery occlusion in one case and no pathological findings in the other. A follow-up MRI scan confirmed spinal cord infarction in the latter patient.
In the 2 patients with suspected angioma, SA revealed no pathological findings; both were ultimately diagnosed with cavernoma.
In the patient with suspected spinal artery aneurysm, SA confirmed the diagnosis. In the patient with suspected cavernoma, SA showed no pathological findings.
DiscussionThe term MUO refers to a set of symptoms of spinal cord involvement that cannot be explained by routine laboratory tests or standard imaging studies, most commonly MRI. According to different series,1,15–17 MUO may account for up to 40% of all myelopathies. Its management is challenging, often requiring serial imaging studies and, occasionally, other complementary tests to rule out entities included in the differential diagnosis, such as SA as a tool for the screening of SVM. However, no previous studies have analysed the frequency and real value of SA in this context.
The main finding from our series is that MUO was the third most frequent indication for SA (12.9%), after dural fistulas (41.9%) and vertebral tumours (20.4%). This suggests that SA is frequently requested in patients with MUO, and is a useful tool in the management of this entity.
In most cases, its usefulness lies in its ability to definitively rule out the presence of an underlying vascular pathology (91.7%); alongside complementary tests and clinical follow-up, this enabled the treating neurologist to establish a final diagnosis in all patients (Table 1).
Furthermore, in one patient (8.3%), SA detected an underlying cervical dural fistula (Fig. 1). Cervical dural fistulas are uncommon, as most dural fistulas affect the thoracic and lumbar segments; however, they can present clinically as progressive myelopathy (as is also the case with complicated intracranial dural arteriovenous fistulas, or Cognard type V fistulas), and should therefore be considered in the differential diagnosis of MUO.18,19
Beyond its application in MUO, our series of 93 patients confirms that the main indication for SA in real clinical practice is the diagnostic confirmation of dural fistulas and vertebral tumours (41.9% and 20.4%, respectively), with high diagnostic yield in both cases (69.2% and 100%), in addition to its usefulness in treatment planning. However, its role in the diagnosis of vascular malformations that are detectable with MRI but are angiographically silent, such as cavernomas (2 cases) or so-called angiomas (one case), remains uncertain, as it is unclear whether it provides relevant information (none in our patients). Similarly, in spinal cord infarction, SA rarely demonstrates flow interruption in the anterior spinal artery. Interestingly, for spontaneous epidural haematoma, a frequent indication for SA in our series (8.6%), an underlying cause was never identified. Given the large volume of contrast used, radiation exposure, and potential complications of SA, its indication in these conditions should be carefully evaluated.
In our series, morbidity associated with SA was low, with only one patient developing self-limiting sensory alterations (pain, numbness, dysaesthesia) in the lumbar region and both legs following selective contrast injection into a left T10 segmental artery. This low rate of complications has also been reported in similar series, suggesting that diagnostic SA performed at referral centres is a reasonably safe procedure.
To our knowledge, this is the second largest series in the literature analysing the indications and diagnostic yield of SA, after the study published by Chen and Gailloud.6 The most recent reviews on this topic3,4 include smaller series that focus on specific diseases,6–15 and do not analyse the actual indications for SA in a centre with an interventional neuroradiology unit, nor its diagnostic yield for each condition. Our study provides additional data to the scarce literature available.
This study has some methodological limitations, including its retrospective design; however, its purpose was to reflect the different indications for SA in real-world radiological practice and to evaluate its usefulness in the management of MUO, a topic that has received little attention in the literature.
