Normal Vascular Anatomy of the Spinal Cord
The vascular anatomy of the spinal cord is divided into extrinsic and intrinsic spinal arteries and veins. The intrinsic arteries are further divided into central and peripheral arterial systems, which are supplied by central arteries anteriorly, and the pial network. More specifically, the central system supplies the anterior two thirds of the spinal cord through the anterior spinal artery (ASA). The posterior spinal artery (PSA) consists of the peripheral system and supplies the posterior component of the spinal cord. Overlaps between these two systems are found only in the terminal branches, which are not true anastomoses; in the inner white matter; and at the periphery of the gray matter. The extrinsic arteries are metameric in nature, integrating tissues with one source of supply.
Anterior Spinal Artery
The ASA supplies the anterior two thirds of the spinal cord. Rostrally it originates from the vertebral arteries, before they join to form the basilar artery. The ASA travels through the anterior median fissure, and its diameter decreases as it proceeds rostrally. The diameter of the ASA becomes consistent once it reaches the thoracic region. The size variation is anatomically explicit: the ASA gets progressively smaller until it joins with the artery of Adamkiewicz, at which point it becomes very prominent. Lastly, the terminal branches of the ASA allow it to form several anastomoses ( Fig. 67-1 ).
Posterior Spinal Artery
The vertebral artery gives off the paired posterior spinal arteries (PSA); however, the PSA sometimes originates from a posterior radicular artery. The arteries proceed rostrally and travel posterolaterally; they terminate near the end of the spinal cord after providing several branches to posterior rootlets of the cauda equina. The PSA is fed by the 10 to 20 ipsilateral posterior radiculomedullary arteries. Nevertheless, in a few instances, a single radiculomedullary artery does supply both posterior spinal arteries. Anatomically, the PSA has a characteristic single-vessel appearance; however, it can form several anastomosing channels. The PSA has a plexiform design, and it can become so small that it is difficult to see ( Fig. 67-2 ; see also Fig. 67-1 ).
Pial Arterial Plexus
The pial arterial plexus is formed by the surface anastomoses of the ASA and the PSA systems, and it is responsible for blood supply of the spinal cord surface. It supplies the peripheral sections of the spinal cord and includes the posterior horns and substantia gelatinosa. The pial arterial plexus branches infiltrate the dorsal midline of the spinal cord and then proceed inward in a perpendicular fashion. The plexus is fed on the lateral surfaces by the posterior radicular arteries, which are also responsible for supplying the dura mater, spinal ganglia, PSA, and the nerve roots (see Fig. 67-1 ).
Radicular Arteries
Radicular arteries originate from segmental vessels that arise from such larger tributaries as the aorta or the subclavian artery. The 31 pairs of radicular arteries are responsible for supplying numerous structures that include the dura mater, spinal ganglia, PSA, and ASA. Radicular arteries are classified into those that divide into 1) arteries that do not reach the dura of the spinal cord; 2) arteries that penetrate the dura but end early; and 3) radiculomedullary arteries that actually vascularize the spinal cord. It is important to make a note about the artery of Adamkiewicz, also known as the arteria radicularis magna or the great radicular artery, which is the largest radiculomedullary artery, with a diameter of 1.0 to 1.3 mm. Nearly 80% of the time, it is present on the left side, anastomoses with the ASA, and then divides into small ascending and large descending branches (see Fig. 67-1 ).
Central Arteries
The central arteries arise from the ASA and pierce the anterior median fissure to enter the spinal cord. The 210 central arteries form a centrifugal system that supplies the middle of the spinal cord, the central sulcus, anterior and posterior gray horns, and periphery of the white matter. Once it reaches the anterior gray matter, it divides into short ascending and descending branches that supply the edges of the gray matter. The number of central arteries present at specific segments of the spinal cord also differs. The central arteries are most numerous in the cervical and lumbosacral region and are least numerous in the thoracic region. Lastly, the acute angle formed by these arteries in the lumbosacral region permits for a wider perfusion of the spinal cord.
Veins of the Spinal Cord
The veins of the spinal cord follow a pattern similar to the arteries of the spinal cord. The intrinsic spinal cord veins are formed by two groups. First is the longitudinal anteromedian group, which collects into the central veins. The second group is collected through radial veins to coronal veins. Three anterior and posterior spinal veins drain the spinal cord and are in turn drained by the anterior and posterior radicular veins. The internal vertebral venous plexus integrates the venous drainage in the epidural space; this plexus communicates with the dural sinuses and external vertebral venous plexus.
The spinous Batson venous plexus is valveless, which allows blood to pass into the systemic venous system. This network provides an easy route of dissemination for metastases. This can be quite significant in cases of prostate cancer, in which increased intraabdominal pressure facilitates spread of metastatic tumor in the vertebrae, brain, or skull through this venous plexus.
Capillaries of the Spinal Cord
The capillaries of the spinal cord differ according to location. The capillary bed is at least five times denser in the gray matter of the spinal cord than in the white matter, because other arterial structures provide the white matter. It is greatest near the site of cell bodies, which reflects their increased metabolic requirements. The capillary beds of the white matter are robust, as are the nerve fiber routes. When the capillary density of the white matter alone is compared with the transition zone of the gray and the white matter, the latter has been shown to have a higher density.
Hemangioblastoma
Genetics
Hemangioblastoma is sporadic in nature or is inherited in an autosomal dominant fashion with von Hippel–Landau (VHL) syndrome. The VHL gene is located at the locus of 3p25-p26. Mutations such as point and frameshift mutations and large deletions have been identified within the VHL gene. A mutation within codon 238 is specifically implicated in retinal and CNS hemangioblastoma. A loss of function mechanism of a single tumor suppressor gene has also been attributed in development of hemangioblastoma pathology, and several molecular factors have been associated with hemangioblastoma. Increased hypoxia-induced factor (HIF) and vascular endothelial growth factor (VEGF) transcripts have been found in ocular hemangioblastoma. Increased expression of VEGF in part is responsible for tumor growth through abundant neovascularization. Other HIF-induced molecules—including transforming growth factor (TGF), platelet-derived growth factor (PDGF), fibroblast growth factor (FGF), and epidermal growth factor (EGF)—are increased in hemangioblastoma and are all factors associated with angiogenesis. Presence of erythropoietin (Epo) and Epo receptor (EpoR) also alludes to the presence of involvement of progenitor cells in VHL-associated lesions.
The origin of the stromal, neoplastic component of hemangioblastoma continues to elude researchers; however, immunohistochemical studies have consistently identified the following epitopes: neuron-specific enolase, neural cell adhesion molecule (CD56), and vimentin. The vascular endothelial cells in hemangioblastoma are also characterized by expression of von Willebrand factor, platelet–endothelial cell adhesion molecules, and Weibel–Palade bodies.
Epidemiology
Spinal cord hemangioblastomas ( Fig. 67-3 ; see also Fig. 67-2 ) are rare lesions that represent 1% to 3% of all intramedullary spinal cord tumors, and men are affected twice as often as women. Central nervous system (CNS) hemangioblastomas are present in 21% to 72% of patients with VHL disease, and approximately 40% are located in the spinal cord. Multiple lesions may be present in 25% to 33% of patients with CNS hemangioblastomas, although this figure may underestimate the actual incidence, because not all patients undergo a full workup for VHL disease. Spinal hemangioblastomas are intramedullary (75%) or have extramedullary or intradural extension (10% to 15%). Approximately 96% of spinal hemangioblastomas are located posterior to the dentate ligament. Extradural hemangioblastomas are rare and may arise from the vertebral bodies. By location, 50% of spinal hemangioblastomas occur in the thoracic cord, 40% are in the cervical cord, and 6% are in the lumbar region. Hemangioblastomas have also been rarely reported in the conus medullaris, filum terminale, nerve roots, and peripheral nerves. Typical age at presentation is between 40 and 50 years in patients with sporadic hemangioblastomas (lesions are more often intracranial); patients with VHL disease are usually seen in their late twenties to early thirties.
Hemangioblastomas usually present with nonspecific signs of intramedullary mass and syrinx, and syrinx is seen in approximately 50% to 70% of patients. Initial symptoms can be divided into three major groups: 1) sensory changes, which occur in 38.9% of patients, mostly as numbness and involvement of posterior columns; 2) weakness in 27.8%; and 3) pain in 33% (in these cases the tumor frequently extends into or originates from the dorsal root entry zone). Patients may also come to medical attention with signs of myelopathy and urinary incontinence. Surprisingly, spontaneous hemorrhage from spinal hemangioblastomas is quite rare; although the majority of patients had subarachnoid hemorrhage (SAH), intramedullary hemorrhage was less common.
Pathology
The typical spinal cord hemangioblastoma usually enlarges the cord, is well demarcated, and consists of a highly vascular nodule with an associated cyst; leptomeningeal vessels are prominent. Histologically, these tumors are composed of an intricate vascular network of irregular and often dilated capillaries with intervening stromal cells. These stromal cells can produce erythropoietin, resulting in erythrocytosis. Immunostaining for epithelial markers is negative for hemangioblastoma; these markers are important when differentiating between hemangioblastoma and metastatic renal cell carcinoma, which may also develop in patients with VHL syndrome. A recent study showed high Ki67 activity in intramedullary–extramedullary hemangioblastomas, whereas the Ki67 activity was less than 1% in intramedullary lesions.
Imaging
Dilated, tortuous feeding arteries and draining pial veins can be seen on a myelogram in approximately 50% of cases. Angiography demonstrates a highly vascular mass with dense vascular blush and draining vessels, which can mimic an AVM. Preoperative embolization is a valid option. Magnetic resonance imaging (MRI) findings are consistent with diffuse cord expansion with high signal intensity on T2-weighted imaging with prominent foci of high-velocity signal loss. Cyst formation and syrinx are seen in 50% to 70% of cases. The tumor nodule is strongly enhanced with contrast administration, and intraoperative use of indocyanine green angiography facilitates lesion delineation to ensure completeness of resection.
Surgical Considerations
Progressive neurologic deterioration caused by mass effect of the tumor and enlarging syrinx and acute neurologic deficit caused by hemorrhage are indications to surgically intervene. In patients with VHL disease, lesions may be multiple, and it is very important to pinpoint the deficit to the particular symptomatic location. Ultimately, multiple surgical interventions may be needed in these patients to treat the disease over their lifetime. Hemangioblastomas are considered in the discussion of vascular spinal cord malformations, because they often behave like AVMs during surgical resection. Presurgical embolization can be implemented to reduce the risk of intraoperative bleeding.
Surgical Technique
- ▪
The patient is positioned depending on the location of hemangioblastoma, and the appropriate approach is performed to extend one level above and one level below the margins of the tumor. Bone removal (laminectomy, laminoplasty, or corpectomy) must be adequate to allow exposure of tumor margins along with associated feeding and draining vessels.
- ▪
The dura mater is incised in the midline, elevated, and retracted the entire length of the exposure with preservation of the arachnoid membrane. Sharp or blunt “tearing” techniques can be used to extend the dural opening. Cotton pads or balls are sometimes used to protect the underlying spinal cord during dural opening; recall that spinal dura has only one layer, unlike cranial dura. Cottonoid strips can be packed into the lateral paraspinal gutters to maintain a bloodless operative field, and dural leaflets are tacked up to adjacent muscles or drapes with 4-0 braided nylon suture.
- ▪
The microscope is brought into the operative field, and the arachnoid is sharply dissected from the surface of the hemangioblastoma and associated vessels. In general, the tumor is approached in much the same manner as an AVM, and special attention is paid to feeding and draining vessels.
- ▪
Pial vessels that cross the margin of the tumor at its junction with the pia mater are coagulated using bipolar cautery at a low setting and are sharply divided to clearly expose the margin of the tumor at the pial surface. Sensory rootlets embedded into the tumor may be dissected free or interrupted if the tumor is to be completely resected.
- ▪
The plane of dissection is developed in a circumferential manner using bipolar cautery, microscissors, and small cottonoid strips. The tumor capsule is normally prominent. It is important that dissection be performed in a completely bloodless field, so that each feeding and draining vessel can be distinguished from en passant vessels and interrupted as it reaches the surface of tumor capsule. Again, surgical technique mirrors that used for AVM resection.
- ▪
Traction on the spinal cord, including “tenting,” should be avoided while reflecting the poles of the tumor.
- ▪
Bipolar electrocautery must be used judiciously and at low voltage to avoid thermal injury to adjacent neural tissue. If bleeding occurs from the tumor capsule, coagulation often makes it worse. Hemostasis can be obtained by application of a variety of hemostatic agents, such as Gelfoam soaked in thrombin.
- ▪
Piecemeal resection of the tumor often causes vigorous bleeding; thus it should not be attempted unless the tumor is large and cannot otherwise be safely removed. In this scenario, meticulous coagulation and hemostasis are imperative. A portion of the tumor can be removed to afford additional exposure.
- ▪
The operative bed is directly inspected to make certain no tumor remains and that hemostasis is complete.
- ▪
Dural closure is performed in a watertight manner with monofilament 4-0 or 5-0 suture on a tapered needle. Some surgeons apply fibrin glue over the suture line.
- ▪
Multiple-layer closure of the wound is performed in a standard fashion.
Outcomes
Hemangioblastomas can be safely removed without significant new postoperative deficit. Approximately 96% of patients will remain unchanged or will improve neurologically, and 4% will worsen. A recent National Institutes of Health (NIH) study published on surgical outcomes after hemangioblastoma resection demonstrated the following :
- ▪
Location of the tumor anterior to the dentate ligament carries a higher risk of new postoperative neurologic deficit.
- ▪
Likelihood of new permanent postoperative neurologic deficit increases with lesions larger than 500 mm.
- ▪
Cysts associated with hemangioblastoma diminish or resolve in almost all patients. Presence of a cyst preoperatively does not alter the surgical outcome, and further surgical manipulations on the tumor cyst during resection are not needed.
Spinal Cord Cavernous Malformations
Genetics
Cavernous malformations have a strong genetic predisposition, and some familial and sporadic forms have incomplete penetrance and variable expressivity. Nearly 150 types of mutations are reported, and frameshift and nonsense mutations are the most prominent. Sporadic cases account for 80% of the cavernous malformations, with an incidence of 1 in 200, and these cases demonstrate great locus and allelic heterogeneity. Several rare causes have also been reported, such as a balanced translocation between chromosomes 3 and X in a female with skewed X inactivation.
Epidemiology
Cavernous malformations ( Fig. 67-4 ) can be considered neoplastic lesions based on their features and growth pattern. These lesions can occur sporadically or in a familial pattern and have an identifiable genetic abnormality with an autosomal dominant pattern of inheritance and incomplete penetrance. Spinal cord cavernous malformations represent 5% to 12% of all spinal cord vascular malformations and 3% to 15% of all cavernous malformations that occur in the CNS. There is slight female predominance, and symptomatic presentation and diagnosis usually occur in the fourth decade of life. The thoracic cord is affected more often than the cervical, and lesions in the conus medullaris and cauda equina are rare.
Clinical Presentation and Natural History
The clinical course of spinal cord cavernous malformations is variable. Patients can develop acute symptoms attributed to hemorrhage, or they may come to medical attention with stepwise deterioration, which can mimic demyelinating disorders. Acute presentation is characterized by pain that corresponds to the level of the cavernous malformation and neurologic deterioration that can occur over several days. This is different from the typical hemorrhage caused by an AVM of the spinal cord, which is typically more acute, and neurologic deficit is concomitant with the onset of pain. Initial hemorrhage from a cavernous malformation can cause paraplegia or quadriplegia, although incomplete neurologic deficit followed by some degree of recovery, which is rarely complete, is more common. In untreated lesions, repeated hemorrhages may occur months to years after the initial hemorrhage. A more subtle presentation can occur when the lesion is primarily localized on the dorsal aspect of the spinal cord, and patients initially complain of intermittent paresthesias. Radiculopathy is more common with lesions in the dorsal root entry zone. With the widespread use of MRI, cavernous malformations are often discovered at an early symptomatic stage or even while asymptomatic.
Pathology
Spinal cord cavernous malformations are identical in appearance and histopathology to intracranial cavernous malformations, and grossly they may be described as soft and spongy with a dark blue to red-brown hue. Cavernous malformations are usually well circumscribed, and hemosiderin staining of the surrounding tissues as a result of repeat bleeding can clearly define the plane of dissection. This discoloration is sometimes the only visual clue that a cavernous malformation may be located under the pial surface. Microscopically, cavernous malformations consist of endothelium-lined channels filled with blood with no intervening brain tissue. Vessel walls lack elastic and muscular layers, and calcifications are rare. A gliotic, often hemosiderin-laden plane usually is evident around the malformation. Tonguelike extensions of the cavernous malformation can extend into the surrounding gliotic plane, and this should be kept in mind during resection to achieve complete excision.
Imaging
Findings on MRI include signs of hemorrhage in different stages of blood product degradation with a mixture of high- and low-intensity signals. The typical appearance of a cavernous malformation is an inhomogeneous high-intensity signal on both T1- and T2-weighted images with a surrounding dark ring of hemosiderin and appearing hypointense on T1- and T2-weighted images. Enhancement is not typical for cavernous malformations. Unlike their intracranial counterparts, the diagnosis of spinal cord cavernous malformations with MRI is not always straightforward, especially with small lesions. The classic “popcorn” appearance is not always seen in spinal cord cavernous malformations. In some cases, spinal MRI, particularly T2-weighted images, can be misleading for surgical planning when trying to estimate where the malformation comes closest to the pial surface. Malformations that appear to be located superficially on MRI may be found to lie deeper in the spinal cord during surgical exploration. Nevertheless, MRI represents an invaluable imaging technique, compared with more traditional imaging modalities, and with emergence of more powerful MR scanners, the quality of spinal cord imaging is rapidly improving.
Angiography has very little value in diagnosing these angiographically occult lesions. The angiogram may demonstrate a venous anomaly associated with a cavernous malformation; the cavernous malformation, not the venous anomaly, is felt to be the source of recurrent hemorrhage. The venous anomaly represents an anatomic variant that should be preserved during surgical resection, because it provides venous drainage to the surrounding normal tissues. Preoperative embolization is not an option with cavernous malformations.
Surgical Considerations
The increasing experience with surgical excision of intramedullary spinal cord cavernous malformations and the high probability of neurologic deterioration if these are left untreated have expanded the role for surgical treatment. Studies clearly demonstrate that progression of neurologic symptoms in patients with spinal cord cavernous malformations is the rule rather than the exception. Neurologic outcome is most dependent on the preoperative neurologic status of the patient, and best outcomes are achieved in patients with good neurologic status preoperatively.
Given its small cross-sectional area and high eloquence, the spinal cord is unlikely to tolerate even minor expansions from hemorrhage or from growth of the malformation, and this is an important consideration. Modern microsurgical technique can provide good outcomes with an acceptable level of postoperative morbidity in patients with spinal cord cavernous malformations. A recent study showed the effectiveness of a carbon dioxide laser in spinal cord cavernous malformation resection; the laser allows the surgeon to perform delicate myelotomies safely and to shrink cavernous malformations away from eloquent spinal cord tissue. Surgery may be recommended for appropriate candidates with symptomatic lesions, especially when the cavernous malformation extends to the pial surface. However, this decision is significantly more difficult in patients with asymptomatic or minimally symptomatic lesions and in patients with deep-seated lesions. In these cases, recommendations for radical surgical resection should be tailored to each individual case. Young patients and patients with large lesions are the most appropriate candidates in this group, because they are most likely to experience long-term benefit from early surgical intervention.
Surgical Technique
Dorsally Located Lesions ( Fig. 67-5 )
- ▪
Preoperative localization is an important part of surgical planning and can be done using techniques such as external skin marking or image guidance with fiducial application.
- ▪
Most cavernous malformations can be exposed and resected via a posterior approach. Laminectomy or laminoplasty should provide adequate exposure for dorsally located lesions. Laminoplasty has been recommended for cervical or upper thoracic lesions to prevent postsurgical kyphotic deformity.
- ▪
Laminoplasty in patients without significant degenerative disease or spinal cord expansion can be performed with a pneumatic drill and a footplate attachment with laminar cuts on both sides. Ligamentous structures are sharply divided, and the laminae and spinous processes are removed en bloc over the levels of interest. Absolute hemostasis should be obtained before opening the dura.
- ▪
The intraoperative microscope is brought into the operative field, and the dura mater is incised in the midline with preservation of the underlying arachnoid layer as described in previous sections. The dural edges are tacked up to the drapes or paraspinous muscles using 4-0 braided nylon sutures.
- ▪
The arachnoid is opened sharply in the midline, and the edges are secured to the ipsilateral dural leaflet.
- ▪
The spinal cord is examined under high magnification; malformations that extend toward the pial surface may be visible at the surface, and in other cases, blue or red-brown discoloration of the spinal cord caused by hemosiderin deposits will be visible and will point to the location of the malformation. Image guidance or intraoperative ultrasound could be used to localize those lesions that leave no clues as to their location.
- ▪
A two-point method is used to determine the optimal entry point and trajectory through the spinal cord to the cavernous malformation: a line is drawn through the center of the lesion to the point where it comes closest to the surface. For deeper lesions, this technique is modified in the spinal cord to avoid eloquent tracts and to take advantage of better tolerated avenues of approach (see Fig. 67-5 ).
- ▪
For deep-seated lesions, myelotomy is performed under high magnification, either through the dorsal median sulcus or along the dorsal root entry zone, whichever offers a better trajectory to the malformation.
- ▪
Care must be taken to avoid damaging the adjacent normal spinal cord parenchyma, and sharp dissection with judicious use of bipolar electrocautery is the standard of care. Myelotomies should be parallel to fiber tracts on the long axis of the spinal cord to minimize damage.
- ▪
Resection of the lesion is performed using microcurettes and gentle suction aspiration. Handheld suction devices with thumb apertures offer controlled suction strength, which is critical to avoid injury to surrounding tissues. Typically, lesions will be removed in a piecemeal fashion, although some can be resected en bloc. Although not truly encapsulated, cavernous malformations have a well-defined gliotic plane that separates them from the surrounding spinal cord.
- ▪
Bleeding is seldom a problem with cavernous malformations because of their low-flow nature, and hemostasis should be accomplished using hemostatic agents and gentle compression. Bipolar cautery use should be avoided unless absolutely necessary, and when used at all, it should be set to a low power. Venous draining anomalies are often associated with cavernous malformations and should be preserved, because they may provide venous drainage for adjacent eloquent tissues.
- ▪
After hemostasis is obtained, careful inspection of the resection bed under high magnification is imperative to identify and further resect small “tongues” of the cavernous malformation that may extend into the adjacent tissue. Incompletely resected lesions can recur and hemorrhage; therefore every attempt should be made to resect these lesions fully during the first surgery.
- ▪
Dura is closed in a watertight fashion as described in previous sections. A multilayer wound closure is performed using standard techniques.
Related posts:
Craniovertebral Junction Instabilities and Surgical Fixation Techniques
Surgical Approaches to the Craniovertebral Junction in Rheumatoid Arthritis
Anterior and Posterior Cervicothoracic Junction Stabilization Techniques
Percutaneous and Endoscopic Diskectomy
Sacral Screw Fixation and Plating Techniques
Surgical Management of Gunshot Wounds to the Spine
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
