Intradural spinal cord arteriovenous shunts in a personal series of 210 patients: novel classification with emphasis on anatomical disposition and angioarchitectonic distribution, related to spinal cord histogenetic units

Katsuhiro MizutaniDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;
Department of Neurosurgery, Keio University School of Medicine, Shinjuku, Tokyo; and
Department of Neurosurgery, Ashikaga Red Cross Hospital, Ashikaga, Tochigi, Japan

Search for other papers by Katsuhiro Mizutani in
jns
Google Scholar
PubMed
Close
 MD, PhD
,
Arturo ConsoliDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Arturo Consoli in
jns
Google Scholar
PubMed
Close
 MD
,
Federico Di MariaDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Federico Di Maria in
jns
Google Scholar
PubMed
Close
 MD
,
Stéphanie Condette AuliacDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Stéphanie Condette Auliac in
jns
Google Scholar
PubMed
Close
 MD
,
Anne BoulinDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Anne Boulin in
jns
Google Scholar
PubMed
Close
 MD
,
Oguzhan CoskunDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Oguzhan Coskun in
jns
Google Scholar
PubMed
Close
 MD
,
Julie GratieuxDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Julie Gratieux in
jns
Google Scholar
PubMed
Close
 MD
, and
Georges RodeschDepartment of Diagnostic and Interventional Neuroradiology, Hôpital Foch, Suresnes, Hauts-de-Seine, France;

Search for other papers by Georges Rodesch in
jns
Google Scholar
PubMed
Close
 MD, PhD
Free access

OBJECTIVE

Few classifications of intradural spinal arteriovenous shunts (ID-SAVSs) have considered their anatomical localization in relation to their phenotype and angioarchitectonics. The authors propose another vision of ID-SAVSs allowing a reappraised classification based on analysis of the anatomical disposition, angioarchitecture, and histogenetic location of these vascular malformations.

METHODS

The radiological and clinical records of 210 patients with ID-SAVSs were retrospectively reviewed, considering their localization, vascular architectonics, and correlation with the 5 histogenetic units of the spinal cord. Among these, 183 files with complete data allowed precise analysis of the ID-SAVSs.

RESULTS

Among these 183 files (162 and 21 cases with single and multiple lesions, respectively), different entities were identified: 13 pial macro arteriovenous fistulas (MAVFs), 92 pial micro arteriovenous fistulas (mAVFs), 33 superficial pial niduses, and 69 intramedullary niduses. Thirteen sulcal shunts (either fistulas or niduses) were considered subtypes of pial lesions. Among the 21 multiple cases, 11 were monomyelomeric while 10 were multimyelomeric. Pial lesions, either fistulas or niduses, were dominantly vascularized by pial arteries (anterior or posterior depending on the localization of the shunt) and occasionally (except for MAVFs) by transmedullary arteries. Pial niduses occasionally extended into the funiculus by recruiting intrinsic veins or by extension of the nidus itself inside the white matter. Intramedullary niduses were always vascularized by both centrifugal and centripetal feeders, respectively, from sulcal arteries (SAs) and pial arteries. Sulcal lesions are pial lesions located within the ventral median sulcus and vascularized by SAs and veins. Single or multiple ID-SAVSs can be part of various syndromes such as hereditary hemorrhagic telangiectasia, Parkes-Weber, RASA1, CLOVES, and spinal arteriovenous metameric syndromes. Histogenetic analyses revealed a specific distribution of each ID-SAVS in the 5 histogenetic units of the spinal cord: intramedullary niduses were found almost equally from cervical to thoracic units, while MAVFs and mAVFs were mostly found from thoracic to postcrural ones. Pial niduses showed intermediate features between intramedullary and fistulous lesions and were mostly distributed from brachial to crural segments.

CONCLUSIONS

Precise analysis of the anatomical disposition of ID-SAVSs in relation to functional histogenetic units allows a better understanding of these lesions and improved therapeutic management.

ABBREVIATIONS

ASA = anterior spinal artery; ASV = anterior spinal vein; AVM = arteriovenous malformation; CLOVES = congenital lipomatous overgrowth, vascular malformations, epidermal nevis, spinal/skeletal anomalies/scoliosis; DSA = digital subtraction angiography; HHT = hereditary hemorrhagic telangiectasia; ID-SAVS = intradural spinal arteriovenous shunt; MAVF = macro arteriovenous fistula; mAVF = micro arteriovenous fistula; PSA = posterior spinal artery; PSV = posterior spinal vein; PWS = Parkes-Weber syndrome; SA = sulcal artery; SAMS = spinal arteriovenous metameric syndrome.

OBJECTIVE

Few classifications of intradural spinal arteriovenous shunts (ID-SAVSs) have considered their anatomical localization in relation to their phenotype and angioarchitectonics. The authors propose another vision of ID-SAVSs allowing a reappraised classification based on analysis of the anatomical disposition, angioarchitecture, and histogenetic location of these vascular malformations.

METHODS

The radiological and clinical records of 210 patients with ID-SAVSs were retrospectively reviewed, considering their localization, vascular architectonics, and correlation with the 5 histogenetic units of the spinal cord. Among these, 183 files with complete data allowed precise analysis of the ID-SAVSs.

RESULTS

Among these 183 files (162 and 21 cases with single and multiple lesions, respectively), different entities were identified: 13 pial macro arteriovenous fistulas (MAVFs), 92 pial micro arteriovenous fistulas (mAVFs), 33 superficial pial niduses, and 69 intramedullary niduses. Thirteen sulcal shunts (either fistulas or niduses) were considered subtypes of pial lesions. Among the 21 multiple cases, 11 were monomyelomeric while 10 were multimyelomeric. Pial lesions, either fistulas or niduses, were dominantly vascularized by pial arteries (anterior or posterior depending on the localization of the shunt) and occasionally (except for MAVFs) by transmedullary arteries. Pial niduses occasionally extended into the funiculus by recruiting intrinsic veins or by extension of the nidus itself inside the white matter. Intramedullary niduses were always vascularized by both centrifugal and centripetal feeders, respectively, from sulcal arteries (SAs) and pial arteries. Sulcal lesions are pial lesions located within the ventral median sulcus and vascularized by SAs and veins. Single or multiple ID-SAVSs can be part of various syndromes such as hereditary hemorrhagic telangiectasia, Parkes-Weber, RASA1, CLOVES, and spinal arteriovenous metameric syndromes. Histogenetic analyses revealed a specific distribution of each ID-SAVS in the 5 histogenetic units of the spinal cord: intramedullary niduses were found almost equally from cervical to thoracic units, while MAVFs and mAVFs were mostly found from thoracic to postcrural ones. Pial niduses showed intermediate features between intramedullary and fistulous lesions and were mostly distributed from brachial to crural segments.

CONCLUSIONS

Precise analysis of the anatomical disposition of ID-SAVSs in relation to functional histogenetic units allows a better understanding of these lesions and improved therapeutic management.

In Brief

This study aims to classify intradural spinal arteriovenous shunts (ID-SAVSs) into 5 types of lesions, based on the precise anatomical disposition, angioarchitectonics, and spinal histogenetic segmentations. Each type of lesion is associated with specific genetic nonhereditary and hereditary syndromes, specific angioarchitectonics, and a specific distribution in spinal histogenetic segmentation. The current classification will facilitate a better understanding of the pathophysiology and etiology of ID-SAVSs and better therapeutic management.

Intradural spinal arteriovenous shunts (ID-SAVSs) are rare and correspond to several different entities. Various classifications have been introduced since 19711 allowing distinctions between fistulas and nidus-type lesions,2–10 and further subdividing them into different types based on their gross morphology.3–6,8,9 Such descriptions are most often subjectively made and frequently lack anatomical analyses of the architectonics of the shunts. Furthermore, the location of an ID-SAVS is always reported in relation to the 3 spine levels opposite from where the shunt is localized: cervical, thoracic, and lumbar. This concept does not take into consideration the fact that the spinal cord is, in fact, formed by 5 histogenetic segments that correspond to the functional segments existing in the vertebrate spinal cord.11,12 The latter data could, however, allow an improved understanding of genetic and functional influences on each type of ID-SAVS.

In this article, we integrate this recent knowledge to propose an alternative overview of ID-SAVSs, suggesting a reappraised objective classification of ID-SAVSs based on precise anatomical dispositions, angioarchitecture, and histogenetic functional neuronal segmentation of the spinal cord to facilitate comprehension of the pathophysiology of ID-SAVSs.

Methods

Patients

This study focused exclusively on intradural arteriovenous shunts affecting the spinal cord, and the analysis does not include those affecting solely nerve roots or the filum terminale. We retrospectively reviewed 210 consecutive patients harboring ID-SAVSs who were treated or underwent consultation at Hôpital Foch from March 2002 to July 2019. Due to the unavailability of clinical or neuroradiological data, 27 patients were excluded and 183 patients were ultimately analyzed (86 men and 97 women; average age at onset 26.5 ± 14.6 years, range 1–77 years). The local institutional ethics committees approved this retrospective review and waived informed consent (given the study design) under the ethical standards of the 1964 Declaration of Helsinki and its later amendments.

Types of ID-SAVSs and Vascular Architecture

All patients were evaluated using both MR images and digital subtraction angiography (DSA), including 3D views and cone-beam CT reconstructions whenever possible. Review of these images allowed classification of ID-SAVSs depending on their morphology, gross architecture, and anatomical disposition (Fig. 1). ID-SAVSs were thus separated into 3 subtypes, as previously reported:5 macro arteriovenous fistulas (MAVFs; Fig. 1B), micro arteriovenous fistulas (mAVFs; Fig. 1C and D), and nidus-type malformations (Fig. 1EH). Arteriovenous fistulas are direct communications between arteries and veins. If a large fistula primarily drains into a giant venous ectasia due to its high-flow nature, the lesion is an MAVF. A fistula draining through a smaller shunt in congested spinal cord veins is an mAVF. A nidus-type lesion corresponds to a cluster of vessels interposed between the arterial feeders and the venous drainage. Some of these lesions can evolve over time and modify their initial aspect, mainly because of sprouting or nonsprouting angiogenesis.5,13 Careful analysis of the architecture may then detect such changes, especially after comparison with the initial angiographic data.

FIG. 1.
FIG. 1.

Schematic drawings of the phenotypic aspects of some of the various ID-SAVSs encountered in our experience. A: Spinal cord normal arterial anatomy without arteriovenous shunts. B: Dorsal pial MAVF. C and D: Pial mAVFs: ventral (arrow) without transmedullary arteries or draining veins (C) and dorsal (double arrow) with transmedullary supply and veins (D). E–G: Pial nidus AVMs: without (E) and with (F) nidal funicular extension, or with (G) intrinsic venous congestion (venous funicular extension). H: Intramedullary nidus AVM. I–K: Sulcal pial MAVF (I), mAVF (J), and nidus (K). Red represents arteries and arterialized veins, and purple represents veins. Copyright Katsuhiro Mizutani. Published with permission. Figure is available in color online only.

Each vascular malformation was further analyzed according to its angioarchitectonics: this allowed clear identification not only of the type of lesion but also of its localization according to the vascular territories. The arterial and venous anatomy has been precisely reported by Thron14 and Lasjaunias et al.15 The anterior spinal artery (ASA) anatomically gives off primary branches (sulcal arteries [SAs] or pial arteries); secondary branches coming from the latter ones (perforating centrifugal and centripetal tributaries, respectively) penetrate into the cord and vascularize the gray or white matter with sometimes transmedullary connections. The posterior spinal artery (PSA) diverges into primary pial branches with its contralateral homolog, participating in the constitution of the network covering the dorsal surface of the cord. Coronal arteries are superficial pial vessels linking the posterior and anterior circulations. Perforating centripetal arteries arise from both these pial circulations. Apart from these extrinsic arterial networks, intrinsic longitudinal and axial anastomoses also allow interterritorial connections via transmedullary or transfunicular anastomoses14 (Fig. 1A). Unlike the arterial disposition, the venous drainage of the cord is independent of the gray or white matter territory; it is mostly drained by intrinsic centrifugal (radial) veins that secondarily reach extrinsic pial veins either directly or through transmedullary veins,14,16 and then anterior spinal veins (ASVs) or posterior spinal veins (PSVs).

The multiplicity of ID-SAVSs is defined by the existence of more than one lesion localized either on only one single cord myelomere or on several myelomeres separated from one another by normal cord tissue. Multiple shunts can thus be monomyelomeric (if existing on the same cord segment) or multimyelomeric (if located on different myelomeres). Coexisting syndromes (spinal arteriovenous metameric syndrome [SAMS; or Cobb syndrome],5,17 Parkes-Weber syndrome [PWS],5,18 CLOVES [congenital lipomatous overgrowth, vascular malformations, epidermal nevis, spinal/skeletal anomalies/scoliosis] syndrome,19,20 hereditary hemorrhagic telangiectasia [HHT],5 or RASA1 mutation) with ID-SAVSs were also investigated.

Functional Histogenetic Segmentation of the Spinal Cord and ID-SAVSs

The histogenetic unit is a concept involving embryological neuronal subdivision regulated by specific gene expression. Each histogenetic unit contains several different myelomeres and has a regional genetic influence on the organization of its vascular network and neuronal tissues, which results in distinctive anatomical and functional features in each segment. Accordingly, histogenetic approaches to ID-SAVSs could further elucidate their etiology and pathophysiology, which has not been performed previously to the best of our knowledge. According to the literature,11,12 the cord can be subdivided into 5 functional/neuronal segments that can harbor ID-SAVSs: cervical, brachial, thoracic, crural, and postcrural, corresponding to the myelomere levels C1–4, C5–T1, T2–L1, L2–S1, and S2–5, respectively (Supplemental Fig. 1).21

Results

In the current study, we identified 207 ID-SAVS lesions in 183 patients consisting of 5 entities: 13 pial MAVFs, 92 pial mAVFs, 33 superficial pial niduses, 69 intramedullary niduses, and, as a subtype of pial lesions, 13 sulcal lesions (either fistulas or niduses).

Single ID-SAVS Lesions

We found 162 patients with a single lesion: 10 MAVFs, 56 mAVFs, 32 pial niduses, and 64 intramedullary niduses. Thirteen sulcal lesions were identified: 1 MAVF, 5 mAVFs, and 7 niduses. The average ages were relatively younger in patients with MAVFs (22.2 years old) and intramedullary niduses (20.6 years old) compared to patients with pial niduses (27.5 years old) and mAVFs (33.9 years old). Female dominancy was observed across all lesions except for mAVFs (Table 1).

TABLE 1.

Single ID-SAVS lesions

VariablePial LesionsIntramedullary Nidus
MAVFmAVFNidus
No. of cases (n = 162)10563264
Age range in yrs3–436–771–542–52
 Average22.233.927.520.6
 Median19.531.524.520
Females/males (% F)7/3 (70.0)26/30 (46.4)20/12 (62.5)36/28 (56.3)
Association w/ other syndromes, n (%)
 HHT3 (30.0)000
 SAMS08 (14.3)2 (6.3)13 (20.3)
 PWS0001 (1.6)
 CLOVES1 (10.0)2 (3.6)1 (3.1)0

MAVF, mAVF, and pial nidus lesions include sulcal lesions of each type.

Angioarchitectonics of ID-SAVSs

The above-mentioned 207 lesions were analyzed: the 162 single ones and 45 other shunts constituting the multiple lesions in 21 patients. For this study, these latter shunts were analyzed individually. The angioarchitectonics of all these ID-SAVSs is summarized in Table 2.

TABLE 2.

Angioarchitectonics of ID-SAVS lesions

VariableSuperficial Pial Lesions (n = 138)Intramedullary Nidus (n = 69)
Nonsulcal Pial (n = 125)Sulcal Type (n = 13)
MAVF (n = 12)mAVF (n = 87)Pial Nidus (n = 26)MAVF (n = 1)mAVF (n = 5)Nidus (n = 7)
Feeders, n (%)
 ASA11 (91.7)59 (67.8)20 (76.9)1 (100)5 (100)7 (100)69 (100)
 PSA10 (83.3)73 (83.9)19 (73.1)01 (20.0)2 (28.6)68 (98.6)
 SA07 (8.0)4 (15.4)1 (100)5 (100)7 (100)69 (100)
 Centrifugal transmedullary supply from SA07 (8.0)4 (15.4)000NA
 Centripetal supply from pial arteries (from either ASA or PSA)00000069 (100)
Nidus, n (%)
 Funicular extension007 (26.9)000NA
Drainers, n (%)
 ASV10 (83.3)55 (63.2)19 (73.1)1 (100)5 (100)7 (100)55 (79.7)
 PSV9 (75.0)70 (80.5)18 (69.2)01 (20.0)2 (28.6)52 (75.4)
 Intrinsic venous drainage02 (2.3)8 (30.8)000NA

NA = not applicable.

MAVFs and mAVFs are pial lesions: the shunts are located at the surface of the cord and never inside the cord parenchyma (Supplemental Figs. 2 and 3). Sulcal ID-SAVSs22 correspond to a subtype of pial lesion as they are also superficial extracordal, localized within the ventral sulcus of the cord (Supplemental Fig. 4). All superficial pial arteriovenous shunts were vascularized by pial arteries of the region affected by the vascular malformation. Sulcal pial lesions (1 MAVF, 5 mAVFs, and 7 niduses) were vascularized only by SAs.

Sometimes dorsal mAVFs were also supplied by transmedullary arteries arising from SAs (n = 7, 8.0% of nonsulcal mAVFs). These arteries are then in balance with the pial circulation; analysis of their trajectory allowed identification of the shunting area, and thus the exact localization of the arteriovenous communication. They drained primarily into homologous pial veins but also secondarily spread toward neighboring territories by congestion of extrinsic or intrinsic veins (2 mAVFs, 2.3%).

Niduses were more complex entities to analyze, but 3D and cone-beam CT allowed distinction of several subtypes. True intramedullary niduses (n = 69; Fig. 2) were always located within the cord, affecting both gray and white matter. They were vascularized by both centripetal and centrifugal arteries. Because of their central position, they drained into ASVs and/or PSVs without predominance. Superficial pial niduses (n = 33, 26 nonsulcal and 7 sulcal; Fig. 3) were always located on the surface of the cord. Among them, some niduses remained strictly at the surface of the cord and did not invade the cord parenchyma (n = 26), while other superficial niduses encroached on the white matter funiculi (n = 7, 26.9% of nonsulcal pial niduses; Table 2) either by real nidus network expansion from the surface of the cord (Supplemental Fig. 5) or by only myelomeric draining veins bulging into the cord (Fig. 4) and thus mimicking an intramedullary nidus. The vascular supply of superficial niduses mainly relied on primary pial branches of either the ASA or PSA; transmedullary feeders were noted in 4 cases (15.4%; Table 2). Superficial niduses first drained into pial extrinsic veins according to the same patterns seen in arteriovenous fistulas. Sulcal shunts drained into sulcal veins before secondarily joining the ASVs22 (Fig. 1 I–K, Supplemental Fig. 4).

FIG. 2.
FIG. 2.

A–C: Hemorrhagic intramedullary nidus with false aneurysm (asterisk in F) explored by MRI on sagittal (Sag; A) and axial (Ax; B) T2-weighted images and axial contrast-enhanced cone-beam CT (CBCT; C). Intramedullary nidus embedded in the left hemicord enlarging the C2 myelomere of the cervical histogenetic unit. D–F: Left vertebral artery (VA) angiography in anteroposterior (AP; D) and lateral (Lat; E and F) views. Both ASAs and PSAs vascularize the nidus via centrifugal feeders (from SAs) and centripetal feeders (from pial arteries [PAs]). The nidus is completely embedded inside the cord (B and C). Note the PA (double arrow) arising from the radicular segment of the radiculomedullary artery running over the anterolateral aspect of the cord and participating in the vascularization of the nidus by centripetal arteries.

FIG. 3.
FIG. 3.

Superficial pial nidus without funicular extension. C4–5 myelomere of cervical and brachial histogenetic units. A–C: Sagittal (Sg; A) and axial (B) T2-weighted MRI views show that the nidus is located on the left anterolateral surface of the cord and displaces the cord that is slightly rotated toward the right (B), as also confirmed by CBCT angiography, which shows the superficial localization (C). D and E: Angiography of the left vertebral artery (AP view, D; lateral view, E) confirms that the lesion is vascularized by pial branches of the ASA running at the surface of the left anterior funiculi and draining cranially in the ASV.

FIG. 4.
FIG. 4.

Pial nidus with venous funicular extension. A–D: MR images (sagittal, A; axial, B) show the superficial pial localization of the nidus extending into the right posterior funiculi, confirmed by fusion images of CBCT and MRI (C and D). E–G: Left VA angiograms (AP view, E; lateral views, F and G) confirm the superficial pial nidus mainly vascularized by pial feeders from a PSA. The late-phase lateral view of the VA angiogram (G) shows the important congestion of the intraparenchymal intrinsic veins, draining secondarily in the PSVs and ASV, which corresponds to the extension of the lesions seen in the fusion images (C and D). Figure is available in color online only.

Multiple ID-SAVS Lesions

Multiple ID-SAVSs were detected in 21 patients (11.5%). Sixteen patients had multiple lesions of the same type (13 cases with 2 mAVFs, 1 case with 4 mAVFs, and 2 cases with 2 intramedullary niduses). Five patients had different types of lesions (3 cases with an MAVF + an mAVF, 1 case with a pial nidus + an mAVF, and 1 with an intramedullary nidus with 2 mAVFs). Accordingly, a total of 45 lesions (3 MAVFs, 36 mAVFs, 1 pial nidus, and 5 intramedullary niduses) were thus seen in 21 patients. We found 11 monomyelomeric multiple ID-SAVSs and 10 multimyelomeric ID-SAVS cases (Supplemental Fig. 6). The findings concerning multiple lesions are summarized in Table 3.

TABLE 3.

Multiple ID-SAVS lesions

Multiple Cases (n = 21)Monomyelomeric (n = 11)Multimyelomeric (n = 10)
Age range in yrs8–586–38
 Mean28.320.7
 Median27.021.0
Females/males (% F)4/7 (36.4)4/6 (40.0)
No. of lesions
 2 (n = 19)118
 3 (n = 1)01
 4 (n = 1)01
Histogenetic segment distribution, n
 Cervical10
 Brachial32*
 Thoracic47*
 Crural11
 Postcrural11
Type of lesion, n (%)
 Multiple mAVFs8 (72.7)6 (60.0)
 Multiple intramedullary niduses02 (20.0)
 MAVF + mAVF2 (18.2)1 (10.0)
 Pial nidus + mAVF1 (9.1)0
 Intramedullary nidus + mAVF01 (10.0)
Associated syndromes, n (%)
 SAMS1 (9.1)4 (40.0)
 PWS01 (10.0)
RASA1 mutation1 (9.1)0

One case had 2 multimyelomeric intramedullary niduses located in brachial and thoracic segments.

Syndromic Association

The ID-SAVSs seen in HHT (n = 3) were always single MAVFs. In PWS (n = 2), only intramedullary nidus-type lesions were found (1 single and 1 multimyelomeric type). RASA1 mutation was identified in 1 patient with multiple monomyelomeric shunts (MAVFs and mAVFs). SAMS (n = 28) was associated with all types of ID-SAVSs except MAVFs. Among them, 5 multiple cases were identified (4 multimyelomeric and 1 monomyelomeric). CLOVES syndrome (n = 4) was associated with single MAVF, mAVF, and pial nidus lesions.

ID-SAVSs and Functional Segmentation of the Spinal Cord

This disposition is summarized in Table 4. MAVFs and mAVFs affected mostly thoracic, crural, and postcrural regions. Intramedullary niduses were almost equally distributed from cervical to thoracic segments (about 30% in each) but less so in crural segments (7.2%). Pial niduses were relatively less seen in cervical units (6.1%) and more in crural ones (18.2%) as compared with intramedullary lesions. There was no specific distribution of multiple monomyelomeric or multimyelomeric lesions as compared to single ones in the various histogenetic units (Table 3).

TABLE 4.

Distribution of ID-SAVSs

VariableMAVF (n = 13)mAVF (n = 92)Pial Nidus (n = 33)Intramedullary Nidus (n = 69)
Classic vertebral column segmentation, n (%)
 Cervical2 (15.4)22 (23.9)15 (45.5)39 (56.5)
 Thoracic9 (69.2)51 (55.4)18 (54.5)32 (46.4)
 Lumbar3 (23.1)20 (21.7)1 (3.0)0
Histogenetic functional segmentation, n (%)
 Cervical1 (7.7)8 (8.7)2 (6.1)20 (29.0)
 Brachial1 (7.7)15 (16.3)12 (36.4)22 (31.9)
 Thoracic4 (30.8)32 (34.8)12 (36.4)26 (37.7)
 Crural5 (38.5)20 (21.7)6 (18.2)5 (7.2)
 Postcrural2 (15.4)18 (19.6)1 (3.0)0

Because some lesions were distributed in more than one region, the sum of each ID-SAVS type exceeds 100%. MAVF, mAVF, and pial nidus lesions include sulcal lesions of each type.

SAMS was found across all the segments without any predominance. In PWS, 1 patient with right upper-limb involvement had multiple intramedullary niduses at the thoracic and brachial segments of the cord, while another patient with left upper-limb involvement had multiple intramedullary niduses at the brachial segment. CLOVES syndrome was associated with crural (n = 2) and postcrural (n = 2) lesions. RASA1 mutation was found in 1 brachial lesion (multiple monomyelomeric MAVFs and mAVFs). HHT was associated with thoracic, crural, and postcrural MAVFs (n = 1 for each region).

Discussion

Previous Classifications

Many classifications of ID-SAVSs have been reported,2–8 but there is still a lack of consensus between the neurosurgical and neuroradiological communities concerning their analysis and practical use. These classifications often describe the shunts in a subjective way (i.e., “glomus,” “juvenile”), are not always consistent with the vascular anatomy of ID-SAVSs, and do not fully cover every clinical entity of ID-SAVSs. Among these classifications, Merland and colleagues3,23 in 1987 and 1993, and Barrow et al. in 199424 subdivided intradural fistulas according to their size, velocity, and number of arterial feeders. Such descriptions preclude precise localization of the shunt. In 2002, Rodesch et al.5 proposed a classification of ID-SAVSs, taking into consideration their morphological aspects and architecture as well as their relation to genetic hereditary and nonhereditary syndromic conditions. Spetzler and colleagues4,6 subdivided spinal cord shunts into extradural, intradural dorsal, and intradural ventral fistulas, and extradural-intradural, intramedullary, and conus medullaris niduses. Intramedullary niduses were further divided into compact and diffuse lesions. This classification adds to the confusion, as it is not always consistent with the anatomical disposition of the lesions. We believe that thanks to the recent neuroscientific and anatomical knowledge gained by high-level MR, 3D, and cone-beam CT angiographies, with fusion of these images, ID-SAVSs can now be analyzed and reappraised differently. The experience that we have gained over recent years and the number of patients studied have led us to propose an updated objective classification based on the gross morphology of the shunt and its angioarchitectonic distribution. It is indeed the functional arterial distribution related to the vascular territories of the cord that allows us to pinpoint the location of the shunt.

Spinal Cord Histogenetic Units and ID-SAVS

Recent papers have reported that the spinal cord is not a homogeneous organ but is segmented into histogenetic units.11,12 The mammalian spinal cord is divided into 6 segments made up of several myelomeres, and each segment has a specific function: 1) the cervical segment, characterized by the branchial motor neuron of accessory nerves and the phrenic nucleus motor neuron; 2) the brachial segment, corresponding to the cervical enlargement of the cord and characterized by the motor neuron of the upper limbs; 3) the thoracic segment, characterized by preganglionic sympathetic neurons; 4) the crural segment, corresponding to the lumbar enlargement of the cord and characterized by the motor neuron of the lower limbs; 5) the postcrural segment, characterized by the preganglionic parasympathetic neurons; and 6) the caudal segment, characterized by the tail muscle neurons. Humans lack this latter segment, and thus their cord is made up of 5 histogenetic units. Each has a specific genetic background determined by the Hox gene family.11,12 In our series, not all types of ID-SAVSs are seen in every functional histogenetic segment of the spinal cord (Supplemental Fig. 7), which suggests that each type of lesion develops under the influence of a specific gene set. The natural history and architecture of the shunts might also potentially depend on their histogenetic localization. The genes present in each histogenetic unit may indeed interact differently with the vasculature of such segments and thus be responsible for the development of specific types of lesions in particular localizations, as appears to be the case in the patients of our series.

Types of Lesions and Architectonics

Our series shows that “nidus” is not always a synonym of “intramedullary.” A precise analysis of these entities allowed us to distinguish superficial pial niduses from intramedullary ones. Some articles have already reported the existence of two nidus entities,5,25 but have not yet separately analyzed them in detail. Only niduses localized centrally in the cord, extending into both gray and white matter, should be considered as true intramedullary lesions. They recruit all the potential territorial feeders, resulting in vascularization by both centrifugal and centripetal feeders via anterior or posterior pial arteries. Pial niduses are located at the surface of the cord and are vascularized by primary branches of pial arteries. In a similar way, sulcal and gyral brain arteriovenous malformations (AVMs) are also different entities.26–28 The former is within the sulcus, is thus extracerebral superficial, and is vascularized mainly by the pial vascular networks of the brain surface, which is equivalent to the cord superficial pial nidus. The latter is embedded within the brain parenchyma and is vascularized by perforating vessels that reach the lesion, as is an intramedullary lesion. The two types of lesions have different angioarchitectonic dispositions and carry different risks during treatment.26–28 Similarly, it is important to differentiate intramedullary and superficial niduses; their management hazards (surgical or endovascular) have different clinical consequences. As our series shows, superficial niduses can encroach on the cord either by real extension of the nidus or just by intrinsic venous congestion. The so-called surgical “subpial resection technique”29 is best applied to the latter anatomical dispositions. The illustrative drawing in the paper by Velat and colleagues29 represents this superficial nidus; a true intramedullary nidus would not be cured by this technique, which would have still left its deeper part vascularized by the intrinsic vasculature.

If intramedullary niduses are quite easy to diagnose because of their vascular architectonics, the subtypes of superficial niduses can be more difficult to recognize. One may think that it is inappropriate to consider that superficial nidus lesions with funicular extension are still superficial pial shunts even if they are partially embedded inside the cord. Understanding the microvascular anatomy of the cord allows clarification: the vascular territory of pial arteries also comprises the white matter funiculi because of their centripetal arteries. Accordingly, if we consider a pial lesion to be superficial (be it an mAVF or a nidus), a funicular extension (nidal or purely venous) can also be considered as part of it. However, this extension should usually be limited to the white matter; if the gray matter is reached, centrifugal arteries would also definitely drain the lesion (which would then be intramedullary). Intrinsic vessels coming from transmedullary or transfunicular feeders can occasionally participate in the vascular architectonics of superficial pial shunts, but in a minor way after a long perforating trajectory. This might depend on the flow or size of the arteriovenous shunt and the anatomical balance between pial and transmedullary arteries and veins. Superficial pial lesions drain mainly into pial veins, but intrinsic drainage could also be seen in some cases for the same microanatomical reasons. The fact that such intrinsic vascularization was never seen in our series in MAVFs was intriguing; it could be related to the compression of the cord by the giant venous ectasia, which would not allow development or visualization of the intrinsic vessels. A specific vascular disposition related to MAVFs in a particular histogenetic unit remains another debatable hypothesis. A superficial pial shunt (nidus or mAVF) that drains into intrinsic veins, or a nidus with funicular extension, can angiographically mimic an intramedullary lesion at first sight. Careful analysis of the architecture using MRI, 2D-DSA, 3D rotational angiography, cone-beam CT, and MR angiography fused images helps to make a precise diagnosis and allows improved (i.e., less risky) therapeutic management. Sulcal arteriovenous shunts being embedded in the ventral sulcus of the cord are similar to brain sulcal arteriovenous shunts. These anteriorly located challenging lesions are difficult to manage surgically and are best treated by embolization.22

Multiplicity and Associated Syndromes

Multiple shunts are related to metameric dispositions or influenced by systemic genetic disorders.5,17,30 It is, however, difficult to consider that one disposition is only related to an embryologic failure and the other to a genetic problem, and that these dispositions are independent from one another. It may be more appropriate to consider that all of these vascular dispositions are influenced by the interaction of genes in the cord histogenetic units and their corresponding surrounding areas, in relation to the migration of the abnormal precursor cells.7 SAMS, the so-called Cobb syndrome, is the best known of these multiple lesions.5,17,30 In the full spectrum of SAMS, vascular malformations are present in the neuronal tissue, bone, epidural space, paraspinal soft tissues, muscles, subcutaneous tissues, and skin at the same metameric level. A partial spectrum is more frequently observed, with vascular lesions often affecting the vertebra (in total or in part) and the corresponding myelomere. SAMS affecting more than one neighboring myelomere can also be seen: 4 (14.3%) of 28 SAMS cases had multimyelomeric ID-SAVSs.

ID-SAVS can be correlated with a limb vascular malformation, as in PWS or CLOVES syndrome. PWS is characterized by capillary malformations and arteriovenous fistulas of a limb, occurring in bones and soft tissues and creating its overgrowth.31 The association of ID-SAVSs with PWS has already been described in previous reports.5,32 PWS is now recognized as one of the phenotypes of capillary malformation–AVM, also related to the RASA1 mutation.33 In our series, PWS was associated with intramedullary nidus-type lesions (1 single and 1 multimyelomeric multiple). A RASA1 mutation was also identified in 1 patient not affected by a limb vascular malformation. This 11-year-old girl presented with skin capillary malformations (also present in her father) and multiple ID-SAVSs (associated with a monomyelomeric MAVF + mAVFs). CLOVES syndrome is due to a PIK3CA mutation34 and is characterized by vascular anomalies, congenital lipomatous overgrowth, epidermal nevi, and spinal or skeletal anomalies.35 Klippel-Trenaunay syndrome was initially considered to be associated with ID-SAVSs, but recent studies have referred this association to CLOVES syndrome.19,20 In the limited number of CLOVES patients of our series, the limb vascular anomalies were associated with single superficial pial fistulas or niduses. An MAVF was noted in one exceptional CLOVES patient with peculiar architecture in which high-flow shunts vascularized by multiple tortuous and stenotic arteries ended in a giant venous ectasia. No specific genetic testing was performed in this patient. To explain the broad extension of all such lesions, it has been advocated that a two-hit mechanism at relatively early fetal stages impacting a large number of cells is responsible for these multimyelomeric features36 and dispositions.

HHT phenotypes are regulated by three different genes: Endoglin,37 acting for HHT type 1; ALK-1,38 acting for HHT type 2; and SMAD4,39 acting for juvenile polyposis/hemorrhagic telangiectasia syndrome. Endothelial cells lacking these genes fail the appropriate remodeling of vessels and form abnormal vessels and connections, resulting in vascular malformations.40 One mutation in these genes alone does not produce the vascular malformation; two-hit insults,41 secondary local inflammation, hypoxia, or cell injury after the genetic mutation40 is necessary to develop the vascular malformation in HHT. Compared to the brain,42 in which arteriovenous shunts can be either single or multiple micro- or macrofistulas or niduses, ID-SAVSs in HHT in our series were always single MAVFs.5 The origin of this disposition is currently not precisely known. Two factors, however, could be considered in the development of the various vascular malformative phenotypes seen in HHT: a timeline effect43,44 (niduses are common in the brains of adult patients, while fistulas [especially MAVFs] are more frequent in the pediatric population44) and a regional gene expression. The brain45 and spinal cord11,12 are indeed histogenetically segmented, and each segment develops under a specific gene set. Accordingly, the genetic mutations in HHT influenced by the specific genetic background in each histogenetic segment may develop different types of vascular malformation. In our series, each phenotype of ID-SAVSs (MAVF, mAVF, pial nidus, and intramedullary nidus) showed a specific distribution in the segmented spinal cord, which was also consistent in both single and multiple lesions, emphasizing the influence of the histogenetic background on the types of lesions. We thus consider that in the spinal cord, these local and time factors may play a key role in the development of single MAVFs in patients with HHT.

Conclusions

The current analysis based on the precise anatomical disposition and angioarchitectonics allows an objective classification of ID-SAVSs into 5 phenotypes (Tables 1 and 2), namely, MAVFs, mAVFs, pial niduses (with or without funicular extension), intramedullary lesions, and sulcal lesions (as a subtype of pial lesions). These lesions can be single or multiple (mono- or multimyelomeric). Multiplicity is frequently related to a metameric disposition. The descriptions made initially by our group5 have to be completed and updated, as we have now shown that ID-SAVSs can be part of more complex syndromes such as genetic nonhereditary (SAMS, PWS, and CLOVES syndrome) and genetic hereditary (HHT and RASA1) disorders. Each syndrome appears to be associated with specific types of ID-SAVSs (Supplemental Fig. 7). Superficial vascular malformations can be located on the ventral, dorsal, or lateral aspect of the cord; analysis of the arterial supply allows one to precisely pinpoint the localization of the shunting area. The venous drainage can have various patterns not necessarily correlated with the exact localization of the ID-SAVS because of the rich intrinsic and extrinsic venous networks. This important venous congestion related to such anatomical reasons can give rise to false interpretations of diffuse niduses and be responsible for remote clinical symptoms. ID-SAVSs have specific distributions in histogenetic segmentations that are controlled by peculiar gene expression. These results suggest that regional and genetic influences play key roles in the development of these lesions. We believe that such a current reappraised classification is anatomically, histogenetically, and angiographically important to understand the pathophysiology and etiology of ID-SAVSs. It also may allow more accurate therapeutic management.

Acknowledgments

We would like to pay tribute to Professor Anton Valavanis and Professor Luis Puelles for having introduced the concept of cerebral histogenetic units in neuroradiology, which represents, in our opinion, one of the major advancements in the understanding of brain vascular malformations. We also thank Mrs. Polly Gobin for the English-language editing.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Mizutani, Rodesch. Acquisition of data: Mizutani, Rodesch. Analysis and interpretation of data: Mizutani, Consoli, Rodesch. Drafting the article: Mizutani. Critically revising the article: Consoli, Di Maria, Rodesch. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Mizutani. Administrative/technical/material support: Rodesch. Study supervision: Consoli, Di Maria, Condette Auliac, Boulin, Coskun, Gratieux, Rodesch.

Supplemental Information

Online-Only Content

Supplemental material is available with the online version of the article.

References

  • 1

    Di Chiro G, Doppman JL, Ommaya AK. Radiology of spinal cord arteriovenous malformations. Prog Neurol Surg. 1971;4:329354.

  • 2

    Heros RC, Debrun GM, Ojemann RG, et al.. Direct spinal arteriovenous fistula: a new type of spinal AVM. Case report. J Neurosurg. 1986;64(1):134139.

  • 3

    Mourier KL, Gobin YP, George B, et al.. Intradural perimedullary arteriovenous fistulae: results of surgical and endovascular treatment in a series of 35 cases. Neurosurgery. 1993;32(6):885891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Spetzler RF, Detwiler PW, Riina HA, Porter RW. Modified classification of spinal cord vascular lesions. J Neurosurg. 2002;96 (2)(suppl):145156.

  • 5

    Rodesch G, Hurth M, Alvarez H, et al.. Classification of spinal cord arteriovenous shunts: proposal for a reappraisal—the Bicêtre experience with 155 consecutive patients treated between 1981 and 1999. Neurosurgery. 2002;51(2):374380.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Kim LJ, Spetzler RF. Classification and surgical management of spinal arteriovenous lesions: arteriovenous fistulae and arteriovenous malformations. Neurosurgery. 2006;59 (5)(suppl 3):S195S201, S3S13.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Rosenblum B, Oldfield EH, Doppman JL, Di Chiro G. Spinal arteriovenous malformations: a comparison of dural arteriovenous fistulas and intradural AVM’s in 81 patients. J Neurosurg. 1987;67(6):795802.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Anson JA, Spetzler RF. Classification of spinal arteriovenous malformations and implications for treatment. BNI Q. 1992;8(2):28.

  • 9

    Krings T, Thron AK, Geibprasert S, et al.. Endovascular management of spinal vascular malformations. Neurosurg Rev. 2010;33:1.

  • 10

    Lee YJ, Terbrugge KG, Saliou G, Krings T. Clinical features and outcomes of spinal cord arteriovenous malformations: comparison between nidus and fistulous types. Stroke. 2014;45(9):26062612.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Philippidou P, Dasen JS. Hox genes: choreographers in neural development, architects of circuit organization. Neuron. 2013;80(1):1234.

  • 12

    Watson C, Paxinos G, Kayalioglu G. The Spinal Cord: A Christopher and Dana Reeve Foundation Text and Atlas. Elsevier/Academic Press; 2009.

    • Search Google Scholar
    • Export Citation
  • 13

    Rodesch G, Hurth M, Alvarez H, et al.. Spinal cord intradural arteriovenous fistulae: anatomic, clinical, and therapeutic considerations in a series of 32 consecutive patients seen between 1981 and 2000 with emphasis on endovascular therapy. Neurosurgery. 2005;57(5):973983.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Thron AK. Vascular Anatomy of the Spinal Cord: Radioanatomy as the Key to Diagnosis and Treatment. Springer International Publishing; 2016.

    • Search Google Scholar
    • Export Citation
  • 15

    Lasjaunias PL, Berenstein A, terBrugge KG. Surgical Neuroangiography: 1 Functional Anatomy of Craniofacial Arteries. Springer; 2001.

  • 16

    Miyasaka K, Asano T, Ushikoshi S, et al.. Vascular anatomy of the spinal cord and classification of spinal arteriovenous malformations. Interv Neuroradiol. 2017;6(Suppl 1):195198.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Matsumaru Y, Pongpech S, Laothamas J, et al.. Multifocal and metameric spinal cord arteriovenous malformations. Review of 19 cases. Interv Neuroradiol. 1999;5(1):2734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18

    Namba K, Nemoto S. Parkes Weber syndrome and spinal arteriovenous malformations. AJNR Am J Neuroradiol. 2013;34(9):E110E112.

  • 19

    Alomari AI, Orbach DB, Mulliken JB, et al.. Klippel-Trenaunay syndrome and spinal arteriovenous malformations: an erroneous association. AJNR Am J Neuroradiol. 2010;31(9):16081612.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20

    Alomari AI, Chaudry G, Rodesch G, et al.. Complex spinal-paraspinal fast-flow lesions in CLOVES syndrome: analysis of clinical and imaging findings in 6 patients. AJNR Am J Neuroradiol. 2011;32(10):18121817.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    O’Rahilly R, Müller F, Carpenter S, Swenson R. The spinal cord and meninges. In: Basic Human Anatomy: A Regional Study of Human Structure. Saunders; 1983:566.

    • Search Google Scholar
    • Export Citation
  • 22

    Roccatagliata L, Kominami S, Krajina A, et al.. Spinal cord arteriovenous shunts of the ventral (anterior) sulcus: anatomical, clinical, and therapeutic considerations. Neuroradiology. 2017;59(3):289296.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23

    Gueguen B, Merland JJ, Riche MC, Rey A. Vascular malformations of the spinal cord: intrathecal perimedullary arteriovenous fistulas fed by medullary arteries. Neurology. 1987;37(6):969979.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24

    Barrow DL, Colohan ART, Dawson R. Intradural perimedullary arteriovenous fistulas (type IV spinal cord arteriovenous malformations). J Neurosurg. 1994;81(2):221229.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Rodesch G, Hurth M, Tadie M, et al.. Intrinsic spinal cord arteriovenous malformations. Neuroradiol J. 2015;22(1 Suppl):115128.

  • 26

    Beneš V, Bradáč O. Brain Arteriovenous Malformations: Pathogenesis, Epidemiology, Diagnosis, Treatment and Outcome. Springer; 2017.

  • 27

    Valavanis A, Pangalu A, Tanaka M. Endovascular treatment of cerebral arteriovenous malformations with emphasis on the curative role of embolisation. Schweiz Arch Neurol Psychiatr. 2004;155(7):341347.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Valavanis A. The role of angiography in the evaluation of cerebral vascular malformations. Neuroimaging Clin N Am. 1996;6(3):679704.

  • 29

    Velat GJ, Chang SW, Abla AA, et al.. Microsurgical management of glomus spinal arteriovenous malformations: pial resection technique: clinical article. J Neurosurg Spine. 2012;16(6):523531.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 30

    Niimi Y, Uchiyama N, Elijovich L, Berenstein A. Spinal arteriovenous metameric syndrome: clinical manifestations and endovascular management. AJNR Am J Neuroradiol. 2013;34(2):457463.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 31

    Mulliken JB, Young AE. Vascular Birthmarks: Hemangiomas and Malformations. Saunders; 1988.

  • 32

    Niimi Y, Ito U, Tone O, et al.. Multiple spinal perimedullary arteriovenous fistulae associated with the Parkes-Weber syndrome. A case report. Interv Neuroradiol. 1998;4(2):151157.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Revencu N, Boon LM, Mulliken JB, et al.. Parkes Weber syndrome, vein of Galen aneurysmal malformation, and other fast-flow vascular anomalies are caused by RASA1 mutations. Hum Mutat. 2008;29(7):959965.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 34

    Kurek KC, Luks VL, Ayturk UM, et al.. Somatic mosaic activating mutations in PIK3CA cause CLOVES syndrome. Am J Hum Genet. 2012;90(6):11081115.

  • 35

    Sapp JC, Turner JT, van de Kamp JM, et al.. Newly delineated syndrome of congenital lipomatous overgrowth, vascular malformations, and epidermal nevi (CLOVE syndrome) in seven patients. Am J Med Genet A. 2007;143A(24):29442958.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 36

    Revencu N, Boon LM, Mendola A, et al.. RASA1 mutations and associated phenotypes in 68 families with capillary malformation-arteriovenous malformation. Hum Mutat. 2013;34(12):16321641.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 37

    McAllister KA, Grogg KM, Johnson DW, et al.. Endoglin, a TGF-β binding protein of endothelial cells, is the gene for hereditary haemorrhagic telangiectasia type 1. Nat Genet. 1994;8(4):345351.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 38

    Johnson DW, Berg JN, Baldwin MA, et al.. Mutations in the activin receptor-like kinase 1 gene in hereditary haemorrhagic telangiectasia type 2. Nat Genet. 1996;13(2):189195.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 39

    Gallione CJ, Repetto GM, Legius E, et al.. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet. 2004;363(9412):852859.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 40

    Marchuk DA, Srinivasan S, Squire TL, Zawistowski JS. Vascular morphogenesis: tales of two syndromes. Hum Mol Genet. 2003;12(Spec No 1):R97R112.

  • 41

    Snellings DA, Gallione CJ, Clark DS, et al.. Somatic mutations in vascular malformations of hereditary hemorrhagic telangiectasia result in bi-allelic loss of ENG or ACVRL1. Am J Hum Genet. 2019;105(5):894906.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 42

    Krings T, Kin H, Power S, et al.. Neurovascular manifestations in hereditary hemorrhagic phenotype correlations. AJNR Am J Neuroradiol. 2015;36(5):863870.

  • 43

    Krings T, Ozanne A, Chng SM, et al.. Neurovascular phenotypes in hereditary haemorrhagic telangiectasia patients according to age. Review of 50 consecutive patients aged 1 day-60 years. Neuroradiology. 2005;47(10):711720.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 44

    Mahadevan J, Ozanne A, Yoshida Y, et al.. Hereditary haemorrhagic telangiectasia cerebrospinal localization in adults and children. Review of 39 cases. Interv Neuroradiol. 2004;10(1):2735.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 45

    Puelles L. A segmental morphological paradigm for understanding vertebrate forebrains. Brain Behav Evol. 1995;46(4-5):319337.

  • Collapse
  • Expand

Composite figure based on findings from the article by Salas-Vega et al. (pp 864–870), which show that elective lumbar laminectomies occurring later in the workweek are associated with longer hospital stays (in days).

  • View in gallery
    FIG. 1.

    Schematic drawings of the phenotypic aspects of some of the various ID-SAVSs encountered in our experience. A: Spinal cord normal arterial anatomy without arteriovenous shunts. B: Dorsal pial MAVF. C and D: Pial mAVFs: ventral (arrow) without transmedullary arteries or draining veins (C) and dorsal (double arrow) with transmedullary supply and veins (D). E–G: Pial nidus AVMs: without (E) and with (F) nidal funicular extension, or with (G) intrinsic venous congestion (venous funicular extension). H: Intramedullary nidus AVM. I–K: Sulcal pial MAVF (I), mAVF (J), and nidus (K). Red represents arteries and arterialized veins, and purple represents veins. Copyright Katsuhiro Mizutani. Published with permission. Figure is available in color online only.

  • View in gallery
    FIG. 2.

    A–C: Hemorrhagic intramedullary nidus with false aneurysm (asterisk in F) explored by MRI on sagittal (Sag; A) and axial (Ax; B) T2-weighted images and axial contrast-enhanced cone-beam CT (CBCT; C). Intramedullary nidus embedded in the left hemicord enlarging the C2 myelomere of the cervical histogenetic unit. D–F: Left vertebral artery (VA) angiography in anteroposterior (AP; D) and lateral (Lat; E and F) views. Both ASAs and PSAs vascularize the nidus via centrifugal feeders (from SAs) and centripetal feeders (from pial arteries [PAs]). The nidus is completely embedded inside the cord (B and C). Note the PA (double arrow) arising from the radicular segment of the radiculomedullary artery running over the anterolateral aspect of the cord and participating in the vascularization of the nidus by centripetal arteries.

  • View in gallery
    FIG. 3.

    Superficial pial nidus without funicular extension. C4–5 myelomere of cervical and brachial histogenetic units. A–C: Sagittal (Sg; A) and axial (B) T2-weighted MRI views show that the nidus is located on the left anterolateral surface of the cord and displaces the cord that is slightly rotated toward the right (B), as also confirmed by CBCT angiography, which shows the superficial localization (C). D and E: Angiography of the left vertebral artery (AP view, D; lateral view, E) confirms that the lesion is vascularized by pial branches of the ASA running at the surface of the left anterior funiculi and draining cranially in the ASV.

  • View in gallery
    FIG. 4.

    Pial nidus with venous funicular extension. A–D: MR images (sagittal, A; axial, B) show the superficial pial localization of the nidus extending into the right posterior funiculi, confirmed by fusion images of CBCT and MRI (C and D). E–G: Left VA angiograms (AP view, E; lateral views, F and G) confirm the superficial pial nidus mainly vascularized by pial feeders from a PSA. The late-phase lateral view of the VA angiogram (G) shows the important congestion of the intraparenchymal intrinsic veins, draining secondarily in the PSVs and ASV, which corresponds to the extension of the lesions seen in the fusion images (C and D). Figure is available in color online only.

  • 1

    Di Chiro G, Doppman JL, Ommaya AK. Radiology of spinal cord arteriovenous malformations. Prog Neurol Surg. 1971;4:329354.

  • 2

    Heros RC, Debrun GM, Ojemann RG, et al.. Direct spinal arteriovenous fistula: a new type of spinal AVM. Case report. J Neurosurg. 1986;64(1):134139.

  • 3

    Mourier KL, Gobin YP, George B, et al.. Intradural perimedullary arteriovenous fistulae: results of surgical and endovascular treatment in a series of 35 cases. Neurosurgery. 1993;32(6):885891.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Spetzler RF, Detwiler PW, Riina HA, Porter RW. Modified classification of spinal cord vascular lesions. J Neurosurg. 2002;96 (2)(suppl):145156.

  • 5

    Rodesch G, Hurth M, Alvarez H, et al.. Classification of spinal cord arteriovenous shunts: proposal for a reappraisal—the Bicêtre experience with 155 consecutive patients treated between 1981 and 1999. Neurosurgery. 2002;51(2):374380.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 6

    Kim LJ, Spetzler RF. Classification and surgical management of spinal arteriovenous lesions: arteriovenous fistulae and arteriovenous malformations. Neurosurgery. 2006;59 (5)(suppl 3):S195S201, S3S13.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Rosenblum B, Oldfield EH, Doppman JL, Di Chiro G. Spinal arteriovenous malformations: a comparison of dural arteriovenous fistulas and intradural AVM’s in 81 patients. J Neurosurg. 1987;67(6):795802.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8

    Anson JA, Spetzler RF. Classification of spinal arteriovenous malformations and implications for treatment. BNI Q. 1992;8(2):28.

  • 9

    Krings T, Thron AK, Geibprasert S, et al.. Endovascular management of spinal vascular malformations. Neurosurg Rev. 2010;33:1.

  • 10

    Lee YJ, Terbrugge KG, Saliou G, Krings T. Clinical features and outcomes of spinal cord arteriovenous malformations: comparison between nidus and fistulous types. Stroke. 2014;45(9):26062612.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Philippidou P, Dasen JS. Hox genes: choreographers in neural development, architects of circuit organization. Neuron. 2013;80(1):1234.

  • 12

    Watson C, Paxinos G, Kayalioglu G. The Spinal Cord: A Christopher and Dana Reeve Foundation Text and Atlas. Elsevier/Academic Press; 2009.

    • Search Google Scholar
    • Export Citation
  • 13

    Rodesch G, Hurth M, Alvarez H, et al.. Spinal cord intradural arteriovenous fistulae: anatomic, clinical, and therapeutic considerations in a series of 32 consecutive patients seen between 1981 and 2000 with emphasis on endovascular therapy. Neurosurgery. 2005;57(5):973983.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Thron AK. Vascular Anatomy of the Spinal Cord: Radioanatomy as the Key to Diagnosis and Treatment. Springer International Publishing; 2016.

    • Search Google Scholar
    • Export Citation
  • 15

    Lasjaunias PL, Berenstein A, terBrugge KG. Surgical Neuroangiography: 1 Functional Anatomy of Craniofacial Arteries. Springer; 2001.

  • 16

    Miyasaka K, Asano T, Ushikoshi S, et al.. Vascular anatomy of the spinal cord and classification of spinal arteriovenous malformations. Interv Neuroradiol. 2017;6(Suppl 1):195198.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Matsumaru Y, Pongpech S, Laothamas J, et al.. Multifocal and metameric spinal cord arteriovenous malformations. Review of 19 cases. Interv Neuroradiol. 1999;5(1):2734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18

    Namba K, Nemoto S. Parkes Weber syndrome and spinal arteriovenous malformations. AJNR Am J Neuroradiol. 2013;34(9):E110E112.

  • 19

    Alomari AI, Orbach DB, Mulliken JB, et al.. Klippel-Trenaunay syndrome and spinal arteriovenous malformations: an erroneous association. AJNR Am J Neuroradiol. 2010;31(9):16081612.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20

    Alomari AI, Chaudry G, Rodesch G, et al.. Complex spinal-paraspinal fast-flow lesions in CLOVES syndrome: analysis of clinical and imaging findings in 6 patients. AJNR Am J Neuroradiol. 2011;32(10):18121817.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    O’Rahilly R, Müller F, Carpenter S, Swenson R. The spinal cord and meninges. In: Basic Human Anatomy: A Regional Study of Human Structure. Saunders; 1983:566.

    • Search Google Scholar
    • Export Citation
  • 22

    Roccatagliata L, Kominami S, Krajina A, et al.. Spinal cord arteriovenous shunts of the ventral (anterior) sulcus: anatomical, clinical, and therapeutic considerations. Neuroradiology. 2017;59(3):289296.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23

    Gueguen B, Merland JJ, Riche MC, Rey A. Vascular malformations of the spinal cord: intrathecal perimedullary arteriovenous fistulas fed by medullary arteries. Neurology. 1987;37(6):969979.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24

    Barrow DL, Colohan ART, Dawson R. Intradural perimedullary arteriovenous fistulas (type IV spinal cord arteriovenous malformations). J Neurosurg. 1994;81(2):221229.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Rodesch G, Hurth M, Tadie M, et al.. Intrinsic spinal cord arteriovenous malformations. Neuroradiol J. 2015;22(1 Suppl):115128.

  • 26

    Beneš V, Bradáč O. Brain Arteriovenous Malformations: Pathogenesis, Epidemiology, Diagnosis, Treatment and Outcome. Springer; 2017.

  • 27

    Valavanis A, Pangalu A, Tanaka M. Endovascular treatment of cerebral arteriovenous malformations with emphasis on the curative role of embolisation. Schweiz Arch Neurol Psychiatr. 2004;155(7):341347.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Valavanis A. The role of angiography in the evaluation of cerebral vascular malformations. Neuroimaging Clin N Am. 1996;6(3):679704.

  • 29

    Velat GJ, Chang SW, Abla AA, et al.. Microsurgical management of glomus spinal arteriovenous malformations: pial resection technique: clinical article. J Neurosurg Spine. 2012;16(6):523531.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 30

    Niimi Y, Uchiyama N, Elijovich L, Berenstein A. Spinal arteriovenous metameric syndrome: clinical manifestations and endovascular management. AJNR Am J Neuroradiol. 2013;34(2):457463.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 31

    Mulliken JB, Young AE. Vascular Birthmarks: Hemangiomas and Malformations. Saunders; 1988.

  • 32

    Niimi Y, Ito U, Tone O, et al.. Multiple spinal perimedullary arteriovenous fistulae associated with the Parkes-Weber syndrome. A case report. Interv Neuroradiol. 1998;4(2):151157.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Revencu N, Boon LM, Mulliken JB, et al.. Parkes Weber syndrome, vein of Galen aneurysmal malformation, and other fast-flow vascular anomalies are caused by RASA1 mutations. Hum Mutat. 2008;29(7):959965.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 34

    Kurek KC, Luks VL, Ayturk UM, et al.. Somatic mosaic activating mutations in PIK3CA cause CLOVES syndrome. Am J Hum Genet. 2012;90(6):11081115.

  • 35

    Sapp JC, Turner JT, van de Kamp JM, et al.. Newly delineated syndrome of congenital lipomatous overgrowth, vascular malformations, and epidermal nevi (CLOVE syndrome) in seven patients. Am J Med Genet A. 2007;143A(24):29442958.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 36

    Revencu N, Boon LM, Mendola A, et al.. RASA1 mutations and associated phenotypes in 68 families with capillary malformation-arteriovenous malformation. Hum Mutat. 2013;34(12):16321641.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 37

    McAllister KA, Grogg KM, Johnson DW, et al.. Endoglin, a TGF-β binding protein of endothelial cells, is the gene for hereditary haemorrhagic telangiectasia type 1. Nat Genet. 1994;8(4):345351.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 38

    Johnson DW, Berg JN, Baldwin MA, et al.. Mutations in the activin receptor-like kinase 1 gene in hereditary haemorrhagic telangiectasia type 2. Nat Genet. 1996;13(2):189195.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 39

    Gallione CJ, Repetto GM, Legius E, et al.. A combined syndrome of juvenile polyposis and hereditary haemorrhagic telangiectasia associated with mutations in MADH4 (SMAD4). Lancet. 2004;363(9412):852859.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 40

    Marchuk DA, Srinivasan S, Squire TL, Zawistowski JS. Vascular morphogenesis: tales of two syndromes. Hum Mol Genet. 2003;12(Spec No 1):R97R112.

  • 41

    Snellings DA, Gallione CJ, Clark DS, et al.. Somatic mutations in vascular malformations of hereditary hemorrhagic telangiectasia result in bi-allelic loss of ENG or ACVRL1. Am J Hum Genet. 2019;105(5):894906.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 42

    Krings T, Kin H, Power S, et al.. Neurovascular manifestations in hereditary hemorrhagic phenotype correlations. AJNR Am J Neuroradiol. 2015;36(5):863870.

  • 43

    Krings T, Ozanne A, Chng SM, et al.. Neurovascular phenotypes in hereditary haemorrhagic telangiectasia patients according to age. Review of 50 consecutive patients aged 1 day-60 years. Neuroradiology. 2005;47(10):711720.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 44

    Mahadevan J, Ozanne A, Yoshida Y, et al.. Hereditary haemorrhagic telangiectasia cerebrospinal localization in adults and children. Review of 39 cases. Interv Neuroradiol. 2004;10(1):2735.

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 45

    Puelles L. A segmental morphological paradigm for understanding vertebrate forebrains. Brain Behav Evol. 1995;46(4-5):319337.

Metrics

All Time Past Year Past 30 Days
Abstract Views 702 234 0
Full Text Views 721 563 49
PDF Downloads 645 438 33
EPUB Downloads 0 0 0