Surgical Management of Intramedullary Spinal Cord Tumors in Adults

Published on 13/03/2015 by admin

Filed under Neurosurgery

Last modified 22/04/2025

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 5 (1 votes)

This article have been viewed 2901 times

Chapter 186 Surgical Management of Intramedullary Spinal Cord Tumors in Adults

The neurosurgical literature on intramedullary spinal cord tumors (IMSCTs) contains many case reports and few large series, even for tumors of glial origin, which are the most numerous.115 As a matter of fact, these lesions are relatively rare and occur in any age group. Indeed, IMSCTs account for 2% to 4% of all central nervous system tumors in adults and 15% of all primary intradural tumors in adults.1618 Ependymomas are the most common tumors in adults and astrocytomas are the most common in children.11,19 Hemangioblastomas and cavernomas represent special entities and require specific strategies.20,21 Other tumors of nonglial origin are still more exceptional. It is well known that IMSCTs have no typical clinical presentation.

At present, magnetic resonance imaging (MRI) is the best and, in most cases, the only examination to perform in investigating these cases. Although MRI can be a highly accurate diagnostic tool, it does not always provide accurate differentiation between ependymomas and astrocytomas. Evoked potentials, both sensory and motor, are now standard intraoperative monitoring tools used during the surgery of these lesions.2224 The ultrasonic aspirator is now used routinely and provides significant assistance.

Whenever possible, we aim at achieving complete removal of spinal cord benign tumors irrespective of the histologic types; however, the surgical treatment of intramedullary spinal cord tumors is not routine surgery. The operative duration is often very long and the procedure is always delicate and technically difficult. That is why neurosurgeons specializing in this field of neurosurgery are not numerous.

Anatomy

The spinal cord is located entirely within the spinal canal. Its rostral end is in continuity with the caudal portion of the medulla, in front of the middle of the atlas anterior arch, to the upper border of the C1 nerve root. The spinal cord has the shape of a roughly cylindrical stem, is ventrally and dorsally slightly flattened, and whitish. It is 42 to 45 cm long and 1 cm wide in adults. It has an average weight of 30 g. It presents two enlargements. The cervical enlargement is 10 cm long, extending from the C4 to the T1 vertebral levels. The lumbar enlargement is 8 cm long from T9 to T12 and in continuity with the conus medullaris, which tapers off at the level of the L1–2 disc space into the filum terminale, an atrophic remnant of the caudal segment of the embryonic spinal cord.

The ventral surface is marked by a ventral fissure that runs along the entire length of the spinal cord. This 2- to 3-mm deep fissure splits the ventral aspect of the spinal cord into two symmetrical ventral columns 2 to 3 mm wide, the lateral borders of which give rise to the ventral roots. The anterior spinal artery runs along the ventral aspect of the cord but not in the anterior median fissure. The lateral surface of the cord contains a lateral column located between the entrance of the dorsal roots and the exit of the ventral roots. On the dorsal surface there is also a dorsal medial sulcus. Although it is not a fissure, it is also possible to separate its edges to visualize the sulcocommissural arteries, which are clearly identified under the microscope.

The spinal cord consists of gray matter surrounded with white matter. The gray matter has a typical H shape in cross section and is characterized by a vestigial central or ependymal canal, which runs the entire length of the spinal cord and is a remnant of a larger embryonic central canal, which is nearly always completely obliterated in adults by ependymal cells or neuroglial clusters. Sometimes this vestigial canal persists over a few millimeters in length, but it lies in the central substantia gelatinosa and is lined with ependymal cells. Such an ependymal canal becomes visible on MRI in the shape of a “split central cavity” without pathologic meaning. Otherwise, the vestigial canal may be dilated in hydrosyringomyelic cavities or in the satellite cysts of intramedullary tumors.

The spinal meninges differ from those of the brain owing to the presence of a thicker pia mater attached to the inner dural surface by the dentate ligaments. The medial border of each is adherent to the lateral column, all along the spinal cord. The lateral border of each ligament is free, with the exception of the areas adjacent to the roots. These are thick serrations whose apices are attached to the dura between the overlying and underlying root sheaths. The arachnoid consists of a dense impermeable superficial layer adjoining the dura and of fenestrated dorsal septa that run from the superficial layer of the arachnoid to the pial surface of the spinal cord. That is why the cord is strengthened by the meninges without interference with the free circulation of the cerebrospinal fluid in the subarachoid space. The spinal dura encloses the spinal cord and the cauda equina from the foramen magnum to the sacrum. The diameter of the dural tube is smaller than that of the spinal canal but is much larger than that of the spinal cord. The dura, which forms a cylindrical sheath, is separated from the spinal canal by the epidural space, containing fat and the epidural venous plexuses. That is why the spinal cord is protected by the meninges and the cerebrospinal fluid and can be slightly mobilized within the spinal canal.

The spinal cord is vascularized by the anterior spinal artery, which arises from the vertebral arteries, in the upper cervical region and by the pial anastomotic network supplied by the radiculospinal and radiculopial arteries, which run with spinal nerves. Because of their size, two ventral radiculospinal arteries have been distinguished: the artery of the lumbar enlargement (Adamkiewicz’s artery), which runs with a spinal nerve on the left side in 75% to 85% of cases and between T9 and T12 in 75% of cases, and the artery of the cervical enlargement, which follows the course of a nerve root between C4 and C8. The anterior spinal artery has a mean diameter of 200 to 500 μm. The posterior blood supply is provided by discontinuous arteries of smaller size (100 to 200 μm). The pial network and the radially penetrating arteries supply the white matter, and the central or sulcocommissural arteries arising from the anterior spinal artery supply the gray matter. Finally, the territory of the anterior spinal artery includes the anterior two thirds of the cord, and the remaining posterior third is supplied by dorsal vessels. The lack of anastomosis between central arteries and the pial network partially justifies the reputation of the midthoracic spinal cord as being “surgically fragile.”

Venous drainage takes place first via the intrinsic vessels, which drain in turn into the pial veins. The anterior spinal vein lies dorsal to the artery. The posterior dorsal vein, which is often very large (400 to 1000 μm), has a winding pattern, particularly in the thoracic region, and zigzags from one posterior column to the other over the posterior median sulcus.

The surgical approach of the intramedullary tumors depends on the anatomy of the cord and its vessels. The anterior approach to the ventral fissure is blocked by the anterior spinal artery and the branches arising from it. The posterior approach is not blocked by arteries or veins, but there is no open fissure on this side of the spinal cord. It is also possible to open the dorsal medial sulcus by separating its edges without damaging the dorsal columns and their vessels.

Incidence, Types, and Prognosis

IMSCTs can be classified in three main groups: tumors of glial origin, tumors of nonglial origin, and pseudotumors.

Tumors of Glial Origin

The group of tumors of glial origin is by far mainly represented by ependymomas and astrocytomas.20,25 Others, such as oligodendriogliomas, as rare.26

Ependymomas

Spinal cord ependymomas (Fig. 186-1) constitute the most common IMSCTs in adult patients. They can reach a considerable size because they are slow-growing processes. The majority of spinal cord ependymomas have a good prognosis because most of the time they can be removed completely and are grade II according to the World Health Organization (WHO) classification. In the literature, the rate of complete removal varies from 69% to 97%.3,15,27 The functional prognosis is particularly satisfying as far as patients who are operated on when their preoperative condition are still good. Indeed, according to the literature, 48% to 75% are stabilized after the procedure, 10% and 40% improve, and between 9% and 15% deteriorate.3,11,15,28,29 This highlights the need for prompt surgery in case of neurologic deterioration.3,11,15,28,29 The 5- and 10-year survival rates reach respectively between 83% to 96% and 80% to 91%.11,15,27,2931 Incomplete resection is considered the most important factor predicting recurrence.3,32,33

image image image

FIGURE 186-1 Grade II cervical ependymoma. A to E, Preoperative magnetic resonance images (MRIs). A, Sagittal T2-weighted image. B, Sagittal T1-weighted image. C, Sagittal contrast-enhanced T1-weighted image. D, Axial contrast-enhanced T1-weighted image. E, Axial T2-weighted image. F to M, Perioperative views. F, Division of the pia mater above the medial sulcus. G, Midline approach with posterior columns separated like a book.Figure 186-1, cont’dH to J, The tumor is progressively resected by dissection of the tumor–spinal cord interface and debulking with an ultrasonic aspirator. K, Great care must be taken when reaching its anterior aspect to prevent any damage to the anterior spinal artery, which is often close to the lesion. L, View after complete resection of the lesion. M, The spinal cord has been closed. N to W, Early (N to R) and 9-month (S to W) postoperative MRIs comparable to A to E. Used with permission from Hôpital Erasme, Université Libre de Bruxelles.

Astrocytomas

Spinal cord astrocytomas (Fig. 186-2) represent more than 80% of the IMSCTs in children,11 and in adults astrocytomas is the second most common type of tumor.11 The fraction of pilocytic astrocytomas varies greatly in large studies, from more than 30% to more than 60%.11,15 The percentage of anaplastic lesions or glioblastomas is higher in adults (25% to 30%) than in children (10% to 17%).1,5,8,11,3436 Unfortunately, unlike with ependymomas, radical resection is most of the time not possible at surgery because by nature grades II to IV tumors display an infiltrative behaviour. Surgery consists then in removing as much tumor as possible or, in some circumstances, in performing only biopsy or decompression with duraplasty. The reported rates of complete and subtotal removal vary respectively between 11% to 31% and 21% to 62%.11,15,37 Pilocytic astrocytomas and the surgeon’s experience are factors associated with higher rates of complete resection.11,15 In low-grade astrocytomas the 5- and 10-year survival rates rise about 77% to 82%, but this rate drops to 27% and 14% for malignant ones.11,38 Low spinal level (conus), malignant grade, and adult age are considered important factors in relation with higher tumor recurrence rate.11 The amount of resection has not been found in relation with the progression-free survival or the overall survival in many series when tumor debulking has been carried out,6,11,16,35,38,39 whereas it was for others in low-grade astrocytomas.5

Gangliogliomas

Gangliogliomas are very uncommon lesions mainly reported as case reports and a few series.10,40 These tumors develop mainly in children and young adults.10 In a large series of 56 patients, complete or subtotal resection has been obtained in 82% and 18% of the cases with 5-year actuarial survival rate of 88% and progression-free survival rate of 67%.10

Tumors of Nonglial Origin

Within the group of tumors of nonglial origin, hemangioblastomas and cavernomas are the most common. These are benign vascular tumors, being characteristically well delineated and sometimes multifocal. Other lesions such as metastasis, epidermoids, dermoid cysts, lipomas, intramedullary schwannomas, primitive neuroectodermal tumors (PNET), and teratomas can also be enumerated.11,25,39 Besides these rare tumors, we have also encountered lymphomas and neuroglial cysts.

Hemangioblastomas

Hemangioblastomas (Figs. 186-3 and 186-4) represent almost 2% to 15% of IMSCTs in some series.11,15,21,24,41,42 These highly vascular lesions consist of two main components: large vacuolated stromal cells, which have been identified as the neoplastic cell of origin, and a rich capillary network.43 Hemangioblastomas are often observed as an encapsulated lesion abutting the pia, especially at the posterior or posterolateral aspect of the spinal cord nearby the dorsal root entry zone, but they may be also observed anteriorly or, rarely, are purely intramedullary.11,21,42 Radicular arteries or anterior branches provide the blood supply to these lesions.44 Associated cyst and syrinx, present in 80% to 90% of cases, and edema can be responsible for significant neurologic morbidity.21

The treatment of choice is en bloc surgical resection, achieving excellent neurological outcome. Not all patients require surgery: patients with incidental asymptomatic solitary lesions may simply be followed.21 Surgery is indicated as soon as symptoms develop or sequential MRI demonstrates tumor or cyst growth.11 The timing of surgery remains nevertheless a matter of debate for patients with von Hippel-Lindau disease and multiple lesions.42 Some authors advocate operating on asymptomatic patients if radiologic progression is observed, before significant neurologic deficits occur.42 The follow-up of patients with von Hippel-Lindau disease must be carried out yearly by spinal and brain MRI and also for associated disease such as pheochromocytomas and for renal and pancreatic cancers.11,45

Cavernomas

Intramedullary cavernomas (Fig. 186-5) accounted for 3% to 16% of IMSCTs in large series.11,15,24,46 Cavernous malformations are well-delineated lesions composed of closely packed, capillary-like vessels, without intervening brain or spinal tissue.46 Their clinical presentation may be variable: Stepwise neurologic deterioration explained by repeated hemorrhages, slow progressive deterioration induced by small hemorrhages and increasing gliosis, or acute onset with rapid or gradual deterioration due to major bleeding.4651 Checking the entire central nervous system is recommended by some authors when an intraspinal cavernoma is discovered.46 In fact as many as 40% of patients with a spinal cavernoma harbor a coexisting intracranial lesion.52

In our opinion, removal of intramedullary cavernomas is indicated in two conditions: if the lesion is symptomatic or if the lesion is easily accessible, meaning located posteriorly and abutting the pial surface, even in asymptomatic patients. The strategy is controversial for asymptomatic patients,47,48 but other authors defend the same philosophy.11,53 The annual rate of hemorrhage must be taken into account: It ranges in the literature from 1.4% to 4.5% in cases of symptomatic lesions but rises even to 66% in case of previous bleeding.51,54 Most lesions can be totally removed, and postoperatively most patients are satisfied by clinical improvement. A better postoperative outcome can be obtained if symptoms last for less than 3 years.51

Associated Pathologies

Most IMSCTs develop sporadically; only a minority are associated with genetic diseases. Hemangioblastomas are a component of von Hippel-Lindau disease; ependymomas and astrocytomas are elements of neurofibromatosis types 1 and 2.11,21,42,56,57 Overall, IMSCTs are noted in these genetic disorders in respectively 20% to 40% and 19% of the cases.21,58 Hamartomas may be accompanied by a different kind of spinal dysraphism such as dermal sinus associated with dermoid cysts.11,5962

Imaging

MRI is nowadays the diagnostic modality of choice in management of IMSCTs. It is the most effective and sensitive technique for the detection of an intramedullary spinal cord lesion.1,25 Any tumoral infiltration produces spinal cord enlargement. Conversely, an enlarged spinal cord is not necessarily caused by a neoplastic process, particularly when enlargement is limited to one or two vertebral segments. The interpretation of the images allows the clinician to know the location and the extension of the lesion, to distinguish between cystic and solid tumoral components, and to propose a histologic diagnostic formulation. However, apart from lesions with specific appearances such as hemangioblastomas, lipomas, or pseudotumors like dermoid cysts, epidermoid cysts, and cavernomas, there are few tumor-specific MRI characteristics.

Ependymomas are observed typically at the center of the spinal cord because they arise from ependymal cells lining the central canal.6466 On MRI, their signal can be variable on T1-weighted images and hypersignal on T2-weighted images. Typically, ependymomas appear as well-circumscribed lesions and enhance vividly and homogeneously after gadolinium injection.25 Four patterns of contrast-enhancement have been determined by Miyazawa:67 homogeneous pattern, heterogeneous pattern, heterogeneous with cyst wall enhancement, and enhancing nodule on cyst wall. Associated satellite cysts are found in 60% of the cases.11,25 On T2 sequences, it is interesting to look for the caplike appearance, corresponding to hyposignals at both poles of the solid part of the tumor. Although this finding is not pathognomonic, it was observed in our experience in 30% of ependymomas and 12% of hemangioblastomas. We have not observed this sign in astrocytomas. It has been correlated histologically and intraoperatively with areas of chronic hemorrhage with hemosiderin deposits.25

Unlike ependymomas, intramedullary spinal cord astrocytomas can be strongly suspected when the tumoral image is eccentric, when the contrast medium enhancement is heterogeneous, and especially if the tumor image has poorly defined borders. On MRI, astrocytomas appear most of the time hypointense on T1-weighted images and hyperintense on T2-weighted images.25 This last sequence is advocated to determine the exact extension of an astrocytoma whose borders are not well demarcated owing to its infiltrative nature. A cystic component is depicted in 27% to 42% 11,25 of the cases and associated satellites and hydromyelia in about half of the cases.25 Grade II astrocytomas are not enhancing after gadolinium administration.25 Conversely, pilocytic astrocytomas enhance intensely, as do high-grade lesions that appear more heterogeneous with necroticocystic areas.25 Gradient echo T2-weighted imaging is a useful sequence for determining hemorrhagic zones within the lesion.25 However, as in the case of ependymomas, there is no MRI-pathognomonic appearance in astrocytomas. Insofar as the only possibility of cure results from gross total resection at the first operation, it appears that all IMSCTs with appropriate imaging appearance should be considered to be ependymomas until proved otherwise.

In contrast to glial tumors, MRI has a high degree of accuracy with nonglial tumors. The nodule of a hemangioblastoma appears classically iso- to hypointense on T1-weighted images and iso- to hyperintense on T2-weighted images.25 Brisk enhancement is evident after gadolinium injection. Proton density imaging and T2-weighted imaging are appropriate for detailing enlarged feeding arteries, rich tumoral vascular network, and dilated draining veins.25 An associated cyst has been found in 88% of the cases in large series.11 Angio MRI using contrast-enhanced bolus three-dimensional T1-weighted sequences can identify some feeding arteries and large draining veins. Angiography is considered only when preoperative embolization is required to reduce the hemorrhagic risk during operation and should be performed immediately prior to surgery.

Cavernous malformations have a characteristic appearance on MRI; the lesion is composed of mixed signal intensities, surrounded by a hemosiderin hypointense ring. Completing the preoperative workup by a brain MRI is advocated because associated lesions are often noted and reinforce the diagnosis.25

The surgical strategy must be determined by analyzing MRI data in the three usual planes (sagittal, axial, and coronal) before and after the injection of gadolinium. It is also essential, before a surgical procedure, to localize the lesion within the spinal cord with regard to its lateralization, depth, and extension.

Differential Diagnosis

Several pathologies must be known in order to avoid unnecessary surgery. Ruling out an intramedullary tumor and determining the diagnosis is sometimes possible based on crucial information given by circumstances, associated conditions, and timing of development. Indeed, intramedullary bacterial, fungal, or parasitic abscesses are unusual in the absence of systemic disease and, unlike IMSCTs, a rapid clinical deterioration is observed.68 An abrupt onset can also lead one to suspect a medulla infarction.69

Multiple sclerosis can mimic an intramedullary tumor, especially active lesions that enhance on MRI.39 In the case of any doubt, one has to screen the complete central nervous system for other lesions by MRI. If no other lesion is discovered, the diagnosis of demyelinating or inflammatory disease cannot be excluded definitively because some patients present with a solitary lesion; on the other hand, detecting other lesions can help in defining the diagnosis.11

Cerebrospinal fluid sampling looking for oligoclonal bands and serum markers can also be relevant.39 If doubt persists, a control MRI has to be planned a few weeks or months later. Indeed, the signal characteristics of the lesion on MRI can change in the case of demyelinating or inflammatory disease, but this delay is too short to be associated with changes in the case of a neoplastic process.11

Spinal cord edema and rimlike contrast enhancement can also be encountered in cases of subacute necrotizing myelopathy. In this pathology, the spinal cord atrophies over time.68 The same evolution can also appear following post-traumatic intramedullary hypersignal that initially can appear to be a tumor. In a study of patients operated on for neoplasic-like lesions, reported by Lee and colleagues,70 the most consistent characteristic found for differentiating non-neoplastic lesions from neoplastic ones was the absence of or minimal spinal cord expansion on preoperative MRI.70

Intraoperative Electrophysiologic Monitoring

Intraoperative monitoring of evoked potentials is a standard monitoring tool for surgical treatment of spinal cord lesions. Monitoring of sensory-evoked and motor-evoked potentials (SEPs and MEPs) is used to assess the functional integrity of the spinal cord during surgical procedures.22 The perioperative analysis of evoked potentials recording provides invaluable information about the functional status of the spinal cord and shows the abnormalities of the sensory and motor pathways, transient or not, that are used to guide the extent of surgical removal of the tumor. At the end, the monitoring allows one to predict the occurrence of postoperative neurologic deficit. Carrying out intraoperative electrophysiologic monitoring is technically difficult but useful during the operative procedure. This technique requires strict collaboration with the neurosurgeon and the electrophysiology team.

One limitation of SEP monitoring is that the evoked responses may be absent or attenuated preoperatively and cannot be monitored intraoperatively. The evoked potentials may be lost during the procedure, but it is essential to know if the abnormalities are transient or permanent. Dorsal column injury is currently detected during the operation. Irreversible abnormalities predict postoperative sensory deficits. Conversely, the absence of dorsal column conduction changes is a good indicator. However, motor function may be damaged without changes in the SEP intraoperative recording. That is why monitoring both SEPs and MEPs is useful every time these waveforms are monitorable. The same limitation concerns MEP monitoring: patients with severe preoperative motor deficit cannot be monitored. Changes caused by nonsurgical factors are easily recognized by the anesthesiologist. Changes observed during the procedure have to be understood insofar as warning to the surgeon, which allows a change in the surgical approach before the motor pathways are injured. However, in patients with a monitorable preoperative curve of MEPs, a reduction of amplitude that remains until the end of the procedure can predict a postoperative transient motor deficit.

A benefit of monitoring both SEPs and MEPs is the demonstration that surgery of spinal cord tumors can be followed by complete restoration of spinal cord sensory and motor functions. However, we cannot assert that monitoring is necessary for successful removal of intramedullary tumors. We are conscious that the best chance for good long-term results is a complete surgical resection of the tumor during the first procedure.

Operative Procedure

General anesthesia is managed taking into account the requirements of intraoperative neurophysiologic monitoring.1,22,71 The surgical procedure is performed while the SEPs and MEPs are recorded. Halogenated volatile anesthetics should be avoided because they modify the SEPs. General anesthesia is established with intravenous opioids and the continuous administration of propofol without bolusing drugs during the operative procedure. MEPs cannot be monitored under complete neuromuscular blockade. Therefore, short-acting muscle relaxants are used to facilitate intubation and to avoid potentially dangerous movements by the patient. These drugs are not routinely readministered or are kept at a minimum, allowing MEPs to be monitored throughout an often lengthy surgical procedure.

When the bladder catheter has been placed, the patient is turned to the prone position on bolster pillows, freeing the abdomen and thorax from pressure. This is the most widely used position for all tumor locations. Head immobilization can be achieved with three-point fixation for procedures above T6, to prevent inadvertent cervical flexion and to prevent pressure sores on the face. We no longer use the sitting position for cervical intramedullary spinal cord tumors because of safety issues and for ergonomic reasons, especially for lengthy surgical procedures.

A midline incision is made, centered at the level of the lesion but extending one level above and below it. It is not necessary to extend the opening over cysts. Bone removal should provide sufficient access to the solid part of the tumor. This step is delicate; either laminectomy or laminotomy may be performed. Hemilaminectomy, proposed for small, mainly extramedullary spinal cord tumors, is inappropriate for surgery of IMSCTs because midline exposure is required. Laminectomy, if carried out gently and patiently, with removal of small pieces of bone, avoids any damage to the adjacent spinal cord and preserves the medial facet joints, decreasing the risk for postoperative kyphosis. When an extensive laminectomy or laminotomy is to be performed, we keep one posterior arch intact at every fifth to sixth vertebra. It is possible to work underneath it, and it is of great assistance for the stabilization of the spine.

Except for pediatric patients, an osteoplastic laminotomy is unnecessary for two reasons. First, when the tumoral spinal cord is very large, in contact with the laminae, the cord and the roots are more threatened by the surgical instruments during laminotomy than during simple laminectomy. Then, osteoplastic laminotomy can result in a very effective fusion, with a solid block of bone forming the posterior border of the spinal canal. That is to be considered if repeat surgery is contemplated at a later date. For pediatric patients, a laminotomy is performed as a rule, and in infants, unilateral incision of the soft tissues can be performed with surgery through a unilateral laminotomy.

Hemostasis has to be meticulous, and large moistened cottonoids are placed along the symmetrical retracted muscular masses to prevent epidural oozing and to establish a dry and very clean field before opening the dura. Surgical strips are placed in the epidural space, compressing the epidural veins. At this stage, intraoperative ultrasonography may be helpful in locating the solid and cystic areas of the tumor, but it is being used less and less.

Ependymomas and Astrocytomas

A midline opening of the dura is performed under magnification from the operating microscope and is extended cranially and caudally to expose the whole tumoral enlargement. Care is taken to not open the arachnoid layer. Dural traction sutures are placed. The arachnoid membrane is opened separately with microscissors and delicately freed from the posterior or lateral spinal cord. Fine sutures identify the sides of the opened arachnoid layer and prepare this layer for closure. Careful inspection of the spinal cord can reveal subpial color modification by the tumor. A SEP electrode is then placed on the dorsal columns and maintained in this position without applying pressure. The tumoral portion of the spinal cord is often enlarged, swollen, smooth, and tense and more or less vascularized. Gentle evaluation of its consistency will confirm the location of solid and cystic areas.

The dorsal median sulcus is then identified under high-power magnification. It appears as a distinct median raphe over which the very tortuous posterior spinal vein runs. Sometimes this sulcus is easily recognized; in other conditions it is identified only by the convergence of vessels toward the midline. The vessels of varying size running vertically over the dorsal columns are dissected and mobilized laterally to expose the posterior sulcus, sacrificing the smallest possible number of vessel branches bridging the two columns and trying to spare all the thinnest arterial or venous vessels in the sulcocommissural region. However, the problem is not always simple in an eccentrically placed tumor in an enlarged and rotated spinal cord. Such distortion can make midline identification difficult or even impossible when the definition of the posterior median sulcus is lost. Sometimes, the location of the true midline has to be evaluated in relation to the posterior roots on both sides, but this can prove impossible owing to the asymmetrical distortion of the cord or the adherence of the posterior columns. In such cases, the midline is identified above and below this region, and then the two openings may join.

In our practice, the midline surgical approach is used except when the lesion is located in one dorsal column and is apparent on the surface without any cortical mantle or in the rare case of an exophytic tumor. We do not incise the tissue and we disagree with those authors who recommend laser myelotomy. We prefer to open the spinal cord by spreading the dorsal columns apart with microscissors and microdissectors and then carefully retracting them with warm saline-moistened cottonoids. The surgical field is extended over the entire length of the solid portion of the tumor and continued to expose the rostral and caudal cysts, if these are present. The opening of the spinal cord must allow exposure of the poles of the lesion and the cyst walls. Pial sutures improve the surgical exposure and reduce the severity of repeated trauma due to dissection. This can be accomplished using a fine suture without any tension to hold the median pia mater and dura mater close together.

One next exposes a sufficient portion of the tumor to obtain a biopsy sample with forceps and scissors but without coagulation. This is immediately followed by histologic examination. Careful hemostasis can now be carried out before proceeding with surgery. This examination sometimes provides a great deal of information when the limits between the tumor and the spinal cord are not clear. Information suggesting an infiltrating or malignant tumor, or both, is crucial in deciding whether tumor removal should be continued.

Tumor removal begins by reducing the volume of the tumor with an ultrasonic aspirator. The tumor is debulked before looking for a cleavage plane. Intratumoral resection is performed from inside to outside, and this is sometimes facilitated by the presence of a cyst or an intratumoral hematoma. After strict control of hemostasis, dissection can be started laterally, on the side on which resection proves easiest. This dissection is performed by the tip of the microforceps with a gentle traction on the tumor against the countertraction provided by the pial sutures. If the tumor is not too friable, or if there is a capsule, it can be grasped, allowing visualization of a dissectible plane. However, common sense and patience are necessary. If there is any difficulty, we prefer to move the microscope to another area and to come back later; for example, to leave one pole and go to the other one. The same policy is adopted when SEPs start to alter. Often, the color is helpful in recognizing the tumor; for example, purple-blue or brown in recognizing ependymoma and distinguishing it from the normal white spinal cord tissue. In our experience, a clear cleavage plane can be found in most ependymomas except in the rare case of a malignant variant. The final objective is total removal of the tumor, which can be performed in most cases. However, the absence of a plane of dissection, particularly in an infiltrating tumor, malignant or not, requires the surgeon to be cautious and avoid continuing tumor removal, which may be dangerous and useless.

Most intramedullary spinal cord tumors have a vascular pedicle arising from the anterior spinal artery. Some large tumors separate both sides of the spinal cord, resulting in a true diastematomyelia with a high risk for injuring the anterior spinal artery and catastrophic operative results. For the detection of residual tumor, we have observed that bleeding spontaneously stops when tumor removal is macroscopically complete. In case of doubt, ultrasonography may be useful. However, in the absence of a polar cyst, it is not always easy to distinguish between the filiform end of the tumor and the increasingly dense fibrous bend into which it merges. The fibrous gliotic bend should be cut where it enters the center of the cord, but we recommend that the last portion of resected tissue undergo histologic examination.

At the end of the procedure it is necessary to inspect the wall of the cyst or cysts adjacent to the tumor bed. When normal spinal cord tissue can be seen through a transparent cyst wall, surgery can be ended, because the cyst wall adjacent to ependymoma does not contain tumor. After the tumor is removed, the dorsal columns are released from pial traction and carefully reapproximated. We like to approximate the cord with fine interrupted pial sutures whenever possible. The arachnoid may also be partially reconstituted if it was originally preserved. The dura is closed in a watertight fashion, always without tension. If the spinal cord remains expanded, because possible residual tumor or edema is present, a duraplasty is performed with fascia, which we prefer to all foreign material. If laminotomy has been performed, the bone is returned to its place, avoiding compression of the spinal cord, with or without internal fixation. After laminectomy, the bone gap can be partly filled with Surgicel. A nonsuction drain is inserted into the subfascial or subcutaneous space.

Hemangioblastomas and Cavernomas

In rare cases in which the lesion is found entirely inside the spinal cord, it has to be accessed through the midline, as described earlier. But, most of the time, hemangioblastomas and cavernomas develop essentially superficially. The lesion may then be observed directly abutting the posterior or posterolateral surface of the spinal cord after dura opening and can therefore be straightforwardly approached.72 Whenever one is present, we always take advantage of a huge and tense syringomyelic cyst by gently withdrawing the fluid with a 22-gauge needle. This maneuver collapses the spinal cord and facilitates the access to the solid tumoral nodule.73

In the case of huge hemangioblastomas, the surgeon must be particularly cautious to avoid any coagulation of the draining vein at the beginning of the procedure. Indeed, interrupting this venous outflow will induce hyperpressure inside the malformation and provoke uncontrollable tumoral bleeding. If the lesion is located on the anterolateral aspect of the spinal cord, its access is possible by dividing one or two dentate ligament attachments and gently rotating the spinal cord with 6-0 silk sutures applied on the ligaments. In this way, most anterolateral lesions may be discovered through the transparent pia mater.20,74 In hemangioblastomas, biopsy or tumor debulking is never performed in order to prevent massive bleeding that will compromise the surgical procedure.

Under microscopic magnification, tumoral limits have to be found. An area with few vessels represents our starting dissection point. The cleavage plane remains distinct as long as bleeding does not interfere with the dissection. Slight tumor retraction can be obtained by gentle coagulation on its surface to facilitate detachment of the lesion from the cord. This maneuver must be completed with nonadherent bipolar forceps or under fluid washing. Repeated coagulations and divisions of vascular connections allow a progressive devascularization until the draining vein is reached. It is only at the end of the procedure that this vein can be interrupted safely. In fact, during this devascularization process, the draining vein appears more and more blue, reflecting the suppression of arterial feeders, as in the surgical treatment of arteriovenous malformations. In the case of huge cystic cavities adjacent to hemangioblastomas, removal of the tumor nodule with opening of the cyst is definitively sufficient without any need for shunting.20

As with hemangioblastomas, cavernomas should be removed most of the time en bloc. Nevertheless, in rare circumstances, reducing the volume of large cavernomas with the ultrasonic aspirator can be necessary. Then, the lesion is shrunk by gentle coagulation on its surface to decrease the pressure on the dissection plane. In some huge cavernomas, it might even be mandatory to use strategies combining different approaches (posterior and anterolateral) in the same surgical procedure.

In hemangioblastomas and cavernomas, we leave open the spinal cord opening but we close the arachnoid layer to prevent the spinal cord from tethering against the dura.

Outcome

Complete removal of the lesion is the goal of surgical treatment of IMSCTs, and this should be performed whenever possible. A tumor cannot be described as “unresectable” if removal has not been attempted. This recommendation concerns all histologic types, astrocytomas included.

Although our results for the surgery of IMSCTs have been published previously,1,20,71 we reiterate our main conclusions here.

The evaluation of surgical results with IMSCTs requires analysis of postoperative functional outcome related to the patient’s functional preoperative state. This evaluation must take into account both motor and sensory functions, as proposed by McCormick in a neurologic grading scale, which is useful for evaluating patients before and after surgery.1

When early postoperative function was compared with preoperative function at the time of discharge, we have observed that the condition of each patient was at least slightly worse for a few days but improves in a good number of patients. With rare exceptions, patients do not recover from severe preoperative neurologic deficits, and those patients who do not have disabling deficits after surgery are those who exhibited few or no deficits prior to operation. In our present series of 440 patients, we have observed 49 patients improving after removal of their tumor. The postoperative outcome based at 3 months after surgery is strongly correlated with the preoperative one. We have observed according to the preoperative McCormick’s grade that:

Our results confirm that we cannot insist strongly enough on the fact that IMSCTs should be operated on at clinical grade I or II for achieving the best results. The postoperative quality of life depends definitively on the preoperative neurologic status.

When patients wake up in the recovery room, they invariably experience discomfort with diffuse hyperesthesia and paresthesias that can last several days. Because of the separation of posterior columns, they have deep sensory deficits that can disturb early postoperative rehabilitation, but in most cases recovery is observed within 6 to 12 weeks. The severity of dysesthesia and pain of various origins is a well-known phenomenon in the early management of intramedullary tumors. Often, at some time after surgery, there are new or worsened dysesthetic complaints of variable severity. Diffuse and permanent paresthesias can reduce the patient’s quality of life, and pharmacologic treatment is not very effective. These unpleasant sensations are fortunately often less severe and resolve spontaneously within a few months. However, the paresthesias might not completely disappear and are an annoyance that patients eventually have to adapt to. Generally, sensory deficits change little after the third postoperative month, whereas motor function continues to improve at least until the end of the first postoperative year. Finally, even in the best circumstances, the patient’s long-term neurologic functional condition will be the same as prior to surgery.

Sphincter dysfunction, principally urinary difficulties and sexual difficulties, cannot be dissociated from the paraplegic or tetraplegic picture with which they are associated.

After the first postoperative year, any increase in motor deficit and any decrease in the patient’s functional capacity suggest tumor recurrence usually due to regrowth of a malignant tumor.

If spinal deformity is one of the earliest symptoms leading to the discovery of an intramedullary tumor, particularly in children, postoperative spinal deformities present before surgery are not exacerbated by surgery.5,75 The sometimes extensive laminectomies required for the surgical treatment of intramedullary tumors may be responsible for severe, but fortunately rather rare, disorders of spinal stability. Four factors are responsible for these severe postoperative deformities: young age, the presence of preoperative spinal deformity, laminectomy involving at least six vertebrae (especially if it includes C2), and malignant neoplasm or adjunctive radiotherapy, or both. In the great majority of cases, postoperative spinal deformations such as increased lordosis or scoliosis or, most often, kyphosis or kyphoscoliosis of varying degrees of severity, are spontaneously stabilized within a few months.

Postoperative Strategy

Within 2 days of surgery, all our patients undergo an early MRI examination. This is particularly relevant for demonstrating any residual tumor and for serving as a basis for the assessment of subsequent lesion progression or recurrence. During the follow-up, we recommend an annual outpatient visit after MRI assessment. Even when the tumor has been completely removed, subsequent examinations are still necessary to exclude any tumor recurrence.

When complete tumor removal during the initial operation has not been accomplished owing to timidity or caution on the part of the surgeon, we recommend performing a second operation to attempt a total resection.

We are definitively convinced that there is absolutely no indication for radiation therapy in benign IMSCTs even after incomplete removal, recurrence, or progression. Our personal opinion is based on a follow-up longer than 5 years in 193 patients surgically treated either for a low-grade ependymoma or a low-grade astrocytoma, without adjunctive radiotherapy. In 90.2% of 122 ependymomas, we were able to resect the lesion completely. The rate of complete removal drops in astrocytomas, but 40.8% of 71 were nonetheless resected completely as ependymomas. When residual tumor is known to remain and is confirmed on postoperative imaging, we recommend repeating MRI evaluation at more frequent intervals during the first postoperative years. We only observed two recurrent ependymomas after 18 and 19 years and three recurrent astrocytomas after, respectively 5, 6, and 7 years in those completely removed tumors. In fact, even low-grade astrocytomas partially removed have a very indolent evolution. Seventeen remain stable in spite of partial removal; a few show a slow MRI evolution without any clinical consequences.

When facing a recurrence or regrowth of a known residue, it is always possible to repeat surgery, if clinically required.

Contrariwise, we have faced very difficult situations when performing additional surgery on several patients sent from other departments who had previously received radiation therapy after biopsy or partial removal. All have been worsened after our surgery, a result opposite to those who did not receive radiotherapy. When we have to face an astrocytoma progression with no possibility to reoperate, we always start with chemotherapy (temozolamide) because radiotherapy is not comparable to chemotherapy.

We only recommend, as have others,76 radiation therapy in malignant gliomas. In the rare case of malignant ependymoma, postoperative radiotherapy may be performed because palliation is based more on compassion than on proof of efficacy. For patients who presented with a malignant astrocytoma, the results also remain poor. Irrespective of the treatments, repeat surgery, radiotherapy, or chemotherapy, the disease is usually fatal within 1 to 3 years.

Key References

Balériaux D., Gultasli N. Intradural spinal tumors. In: Van Goethem J.W.M., van den Hauwe L., Parizel P.M. Spinal Imaging. Berlin: Springer-Verlag; 2007:417-460.

Brotchi J. Intrinsic spinal cord tumor resection. Neurosurgery. 2002;50:1059-1063.

Brotchi J., Bruneau M., Lefranc F., et al. Surgery of intraspinal cord tumors. Clin Neurosurg. 2006;53:209-216.

Brotchi J., Fischer G. Spinal cord ependymomas. Neurosurg Focus. 1998;4:e2.

Chang U.K., Choe W.J., Chung S.K., et al. Surgical outcome and prognostic factors of spinal intramedullary ependymomas in adults. J Neurooncol. 2002;57:133-139.

Constantini S., Miller D.C., Allen J.C., et al. Radical excision of intramedullary spinal cord tumors: surgical morbidity and long-term follow-up evaluation in 164 children and young adults. J Neurosurg. 2000;93:183-193.

Cooper P.R. Outcome after operative treatment of intramedullary spinal cord tumors in adults: intermediate and long-term results in 51 patients. Neurosurgery. 1989;25:855-859.

Ferrante L., Mastronardi L., Celli P., et al. Intramedullary spinal cord ependymomas—a study of 45 cases with long-term follow-up. Acta Neurochir (Wien). 1992;119:74-79.

Fisher G., Brotchi J. Intramedullary Spinal Cord Tumors. New York: Thieme; 1996.

Garces-Ambrossi G.L., McGirt M.J., Mehta V.A., et al. Factors associated with progression-free survival and long-term neurological outcome after resection of intramedullary spinal cord tumors: analysis of 101 consecutive cases. J Neurosurg Spine. 2009;11:591-599.

Guidetti B., Mercuri S., Vagnozzi R. Long-term results of the surgical treatment of 129 intramedullary spinal gliomas. J Neurosurg. 1981;54:323-330.

Harrop J.S., Ganju A., Groff M., et al. Primary intramedullary tumors of the spinal cord. Spine. 2009;34:S69-S77.

Houten J.K., Cooper P.R. Spinal cord astrocytomas: presentation, management and outcome. J Neurooncol. 2000;47:219-224.

Hsu F.P.K., Clatterbuck R.E., Kim L.J., Spetzler R.F. Intramedullary spinal cord cavernous malformations. Oper TechNeurosurg. 2003;6:32-40.

Jallo G.I., Freed D., Epstein F.J. Spinal cord gangliogliomas: a review of 56 patients. J Neurooncol. 2004;68:71-77.

Jallo G.I., Freed D., Zareck M., et al. Clinical presentation and optimal management for intramedullary cavernous malformations. Neurosurg Focus. 2006;21:e10.

Klekamp J., Samii M. Intramedullary tumors. In: Klekamp J., Samii M. Surgery of Spinal Tumors. Berlin: Springer-Verlag; 2007:20-131.

Lefranc F., Brotchi J. Surgical strategy in spinal cord hemangioblastomas. Operative Techniques in Neurosurgery. 2003;6:24-31.

McCormick P.C., Torres R., Post K.D., et al. Intramedullary ependymoma of the spinal cord. J Neurosurg. 1990;72:523-532.

Milano M.T., Johnson M.D., Sul J., et al. Primary spinal cord glioma: a surveillance, epidemiology, and end results database study. J Neurooncol. 2010;98(1):83-92.

Raco A., Esposito V., Lenzi J., et al. Long-term follow-up of intramedullary spinal cord tumors: a series of 202 cases. Neurosurgery. 2005;56:972-981. discussion 972-981

Schwartz T.H., McCormick P.C. Intramedullary ependymomas: clinical presentation, surgical treatment strategies and prognosis. J Neurooncol. 2000;47:211-218.

Schwartz T.H., McCormick P.C. Intramedullary spinal cord tumors. Special issue. J Neurooncol. 2000;47:187-317.

Van Velthoven V., Reinacher P.C., Klisch J., et al. Treatment of intramedullary hemangioblastomas, with special attention to von Hippel-Lindau disease. Neurosurgery. 2003;53:1306-1313. discussion 1313-1314

Zevgaridis D., Medele R.J., Hamburger C., et al. Cavernous haemangiomas of the spinal cord. A review of 117 cases. Acta Neurochir (Wien). 1999;141:237-245.

Numbered references appear on Expert Consult.

References

1. Fisher G., Brotchi J. Intramedullary Spinal Cord Tumors. New York: Thieme; 1996.

2. Schwartz T.H., McCormick P.C. Intramedullary spinal cord tumors. Special issue. J Neurooncol. 2000;47:187-317.

3. Chang U.K., Choe W.J., Chung S.K., et al. Surgical outcome and prognostic factors of spinal intramedullary ependymomas in adults. J Neurooncol. 2002;57:133-139.

4. Abdel-Wahab M., Etuk B., Palermo J., et al. Spinal cord gliomas: a multi-institutional retrospective analysis. Int J Radiat Oncol Biol Phys. 2006;64:1060-1071.

5. Constantini S., Miller D.C., Allen J.C., et al. Radical excision of intramedullary spinal cord tumors: surgical morbidity and long-term follow-up evaluation in 164 children and young adults. J Neurosurg. 2000;93:183-193.

6. Cooper P.R. Outcome after operative treatment of intramedullary spinal cord tumors in adults: intermediate and long-term results in 51 patients. Neurosurgery. 1989;25:855-859.

7. Garces-Ambrossi G.L., McGirt M.J., Mehta V.A., et al. Factors associated with progression-free survival and long-term neurological outcome after resection of intramedullary spinal cord tumors: analysis of 101 consecutive cases. J Neurosurg Spine. 2009;11:591-599.

8. Guidetti B., Mercuri S., Vagnozzi R. Long-term results of the surgical treatment of 129 intramedullary spinal gliomas. J Neurosurg. 1981;54:323-330.

9. Harrop J.S., Ganju A., Groff M., et al. Primary intramedullary tumors of the spinal cord. Spine. 2009;34:S69-S77.

10. Jallo G.I., Freed D., Epstein F.J. Spinal cord gangliogliomas: a review of 56 patients. J Neurooncol. 2004;68:71-77.

11. Klekamp J., Samii M. Intramedullary tumors. In: Klekamp J., Samii M. Surgery of Spinal Tumors. Berlin: Springer-Verlag; 2007:20-131.

12. McGirt M.J., Goldstein I.M., Chaichana K.L., et al. Extent of surgical resection of malignant astrocytomas of the spinal cord: outcome analysis of 35 patients. Neurosurgery. 2008;63:55-60. discussion 60-61

13. Milano M.T., Johnson M.D., Sul J., et al. Primary spinal cord glioma: a surveillance, epidemiology, and end results database study. J Neurooncol. 2010;98(1):83-92.

14. Minehan K.J., Brown P.D., Scheithauer B.W., et al. Prognosis and treatment of spinal cord astrocytoma. Int J Radiat Oncol Biol Phys. 2009;73:727-733.

15. Raco A., Esposito V., Lenzi J., et al. Long-term follow-up of intramedullary spinal cord tumors: a series of 202 cases. Neurosurgery. 2005;56:972-981. discussion 972-981

16. Stein B.M., McCormick P.C. Intramedullary neoplasms and vascular malformations. Clin Neurosurg. 1992;39:361-387.

17. Osborn A.G. Tumors, cysts and tumorlike lesions of the spine and spinal cord. In: Osborn A.G., editor. Diagnostic Neuroradiology. Saint Louis: Mosby; 1994:876-918.

18. Cooper P.R., Epstein F. Radical resection of intramedullary spinal cord tumors in adults. Recent experience in 29 patients. J Neurosurg. 1985;63:492-499.

19. Constantini S., Epstein F. Pediatric intraspinal tumors. In: Choux M., Di Rocco R.C., Hockley A., Walker M. Pediatric Neurosurgery. London: Churchill-Livingstone; 1999:601-613.

20. Brotchi J. Intrinsic spinal cord tumor resection. Neurosurgery. 2002;50:1059-1063.

21. Mandigo C.E., Ogden A.T., Angevine P.D., et al. Operative management of spinal hemangioblastoma. Neurosurgery. 2009;65:1166-1177.

22. Kothbauer F.K. Intraoperative neurophysiologic monitoring for intramedullary spinal cord tumor surgery. Operative Techniques in Neurosurgery. 2003;6:2-68.

23. Morota N., Deletis V., Constantini S., et al. The role of motor evoked potentials during surgery for intramedullary spinal cord tumors. Neurosurgery. 1997;41:1327-1336.

24. Matsuyama Y., Sakai Y., Katayama Y., et al. Surgical results of intramedullary spinal cord tumor with spinal cord monitoring to guide extent of resection. J Neurosurg Spine. 2009;10:404-413.

25. Balériaux D., Gultasli N. Intradural spinal tumors. In: Van Goethem J.W.M., van den Hauwe L., Parizel P.M. Spinal Imaging. Berlin: Springer-Verlag; 2007:417-460.

26. Guppy K.H., Akins P.T., Moes G.S., et al. Spinal cord oligodendroglioma with 1p and 19q deletions presenting with cerebral oligodendrogliomatosis. J Neurosurg Spine. 2009;10:557-563.

27. Ferrante L., Mastronardi L., Celli P., et al. Intramedullary spinal cord ependymomas—a study of 45 cases with long-term follow-up. Acta Neurochir (Wien). 1992;119:74-79.

28. Brotchi J., Fischer G. Spinal cord ependymomas. Neurosurg Focus. 1998;4:e2.

29. Lonjon M., Goh K.Y., Epstein F.J. Intramedullary spinal cord ependymomas in children: treatment, results and follow-up. Pediatr Neurosurg. 1998;29:178-183.

30. Shirato H., Kamada T., Hida K., et al. The role of radiotherapy in the management of spinal cord glioma. Int J Radiat Oncol Biol Phys. 1995;33:323-328.

31. Sgouros S., Malluci C.L., Jackowski A. Spinal ependymomas—the value of postoperative radiotherapy for residual disease control. Br J Neurosurg. 1996;10:559-566.

32. Vijayakumar S., Estes M., Hardy R.W.Jr., et al. Ependymoma of the spinal cord and cauda equina: a review. Cleve Clin J Med. 1988;55:163-170.

33. Rawlings C.E.3rd, Giangaspero F., Burger P.C., et al. Ependymomas: a clinicopathologic study. Surg Neurol. 1988;29:271-281.

34. Epstein F.J., Farmer J.P. Pediatric spinal cord tumor surgery. Neurosurg Clin N Am. 1990;1:569-590.

35. Houten J.K., Cooper P.R. Spinal cord astrocytomas: presentation, management and outcome. J Neurooncol. 2000;47:219-224.

36. Innocenzi G., Raco A., Cantore G., et al. Intramedullary astrocytomas and ependymomas in the pediatric age group: a retrospective study. Childs Nerv Syst. 1996;12:776-780.

37. Kim M.S., Chung C.K., Choe G., et al. Intramedullary spinal cord astrocytoma in adults: postoperative outcome. J Neurooncol. 2001;52:85-94.

38. Jallo G.I., Danish S., Velasquez L., et al. Intramedullary low-grade astrocytomas: long-term outcome following radical surgery. J Neurooncol. 2001;53:61-66.

39. McCormick P.C., Stein B.M. Intramedullary tumors in adults. Neurosurg Clin N Am. 1990;1:609-630.

40. Park C.K., Chung C.K., Choe G.Y., et al. Intramedullary spinal cord ganglioglioma: a report of five cases. Acta Neurochir (Wien). 2000;142:547-552.

41. Murota T., Symon L. Surgical management of hemangioblastoma of the spinal cord: a report of 18 cases. Neurosurgery. 1989;25:699-707. discussion 708

42. Van Velthoven V., Reinacher P.C., Klisch J., et al. Treatment of intramedullary hemangioblastomas, with special attention to von Hippel-Lindau disease. Neurosurgery. 2003;53:1306-1313. discussion 1313-1314

43. Böhling T.P.K., Haltia M.J., Alitalo K., Neumann H.P.H. Von Hippel-Lindau disease and capillary haemangioblastoma. In: Kleihues P.C.W., editor. Tumours of the Nervous System. Pathology and Genetics. Lyon: IARCPress; 2000:223-226.

44. Hurth M. [Intraspinal hemangioblastomas.]. Neurochirurgie. 1975;21(Suppl 1):1-136.

45. Michels V.V. Investigative studies in von Hippel-Lindau disease. Neurofibromatosis. 1988;1:159-163.

46. Jallo G.I., Freed D., Zareck M., et al. Clinical presentation and optimal management for intramedullary cavernous malformations. Neurosurg Focus. 2006;21:e10.

47. Cantore G., Delfini R., Cervoni L., et al. Intramedullary cavernous angiomas of the spinal cord: report of six cases. Surg Neurol. 1995;43:448-451. discussion 451-452

48. Cristante L., Hermann H.D. Radical excision of intramedullary cavernous angiomas. Neurosurgery. 1998;43:424-430. discussion 430-431

49. Deutsch H., Jallo G.I., Faktorovich A., et al. Spinal intramedullary cavernoma: clinical presentation and surgical outcome. J Neurosurg. 2000;93:65-70.

50. Ogilvy C.S., Louis D.N., Ojemann R.G. Intramedullary cavernous angiomas of the spinal cord: clinical presentation, pathological features, and surgical management. Neurosurgery. 1992;31:219-229. discussion 229-230

51. Zevgaridis D., Medele R.J., Hamburger C., et al. Cavernous haemangiomas of the spinal cord. A review of 117 cases. Acta Neurochir (Wien). 1999;141:237-245.

52. Cohen-Gadol A.A., Jacob J.T., Edwards D.A., et al. Coexistence of intracranial and spinal cavernous malformations: a study of prevalence and natural history. J Neurosurg. 2006;104:376-381.

53. Ojemann R.G., Crowell R.M., Ogilvy C.S. Management of cranial and spinal cavernous angiomas (honored guest lecture). Clin Neurosurg. 1993;40:98-123.

54. Sandalcioglu I.E., Wiedemayer H., Gasser T., et al. Intramedullary spinal cord cavernous malformations: clinical features and risk of hemorrhage. Neurosurg Rev. 2003;26:253-256.

55. Varron L., Broussolle C., Candessanche J.P., et al. Spinal cord sarcoidosis: report of seven cases. Eur J Neurol. 2009;16:289-296.

56. Lee M., Rezai A.R., Freed D., et al. Intramedullary spinal cord tumors in neurofibromatosis. Neurosurgery. 1996;38:32-37.

57. Schwartz T.H., McCormick P.C. Intramedullary ependymomas: clinical presentation, surgical treatment strategies and prognosis. J Neurooncol. 2000;47:211-218.

58. Parsa A.T., Fiore A.J., McCormick P.C., et al. Genetic basis of intramedullary spinal cord tumors and therapeutic implications. J Neurooncol. 2000;47:239-251.

59. Barkovich A.J., Edwards M., Cogen P.H. MR evaluation of spinal dermal sinus tracts in children. AJNR Am J Neuroradiol. 1991;12:123-129.

60. Gok A., Bayram M., Coskun Y., et al. Unusual malformations in occult spinal dysraphism. Turk J Pediatr. 1995;37:391-397.

61. Shen W.C., Chiou T.L., Lin T.Y. Dermal sinus with dermoid cyst in the upper cervical spine: case note. Neuroradiology. 2000;42:51-53.

62. Peter J.C., Sinclair-Smith C., de Villiers J.C. Midline dermal sinuses and cysts and their relationship to the central nervous system. Eur J Pediatr Surg. 1991;1:73-79.

63. McCormick P.C., Torres R., Post K.D., et al. Intramedullary ependymoma of the spinal cord. J Neurosurg. 1990;72:523-532.

64. Fine M.J., Kricheff II, Freed D., et al. Spinal cord ependymomas: MR imaging features. Radiology. 1995;197:655-658.

65. Parizel P.M., Baleriaux D., Rodesch G., et al. Gd-DTPA-enhanced MR imaging of spinal tumors. AJR Am J Roentgenol. 1989;152:1087-1096.

66. Sun B., Wang C., Wang J., et al. MRI features of intramedullary spinal cord ependymomas. J Neuroimaging. 2003;13:346-351.

67. Miyazawa N., Hida K., Iwasaki Y., et al. MRI at 1.5 T of intramedullary ependymoma and classification of pattern of contrast enhancement. Neuroradiology. 2000;42:828-832.

68. Schwartz T.H., McCormick P.C. Non-neoplastic intramedullary pathology. Diagnostic dilemna: to Bx or not to Bx. J Neurooncol. 2000;47:283-292.

69. Andrews B.T., Kwei U., Greco C., et al. Infarct of the conus medullaris simulating a spinal cord tumor: case report. Surg Neurol. 1991;35:139-142.

70. Lee M., Epstein F.J., Rezai A.R., et al. Nonneoplastic intramedullary spinal cord lesions mimicking tumors. Neurosurgery. 1998;43:788-794. discussion 794-795

71. Brotchi J., Bruneau M., Lefranc F., et al. Surgery of intraspinal cord tumors. Clin Neurosurg. 2006;53:209-216.

72. Roonprapunt C., Silvera V.M., Setton A., et al. Surgical management of isolated hemangioblastomas of the spinal cord. Neurosurgery. 2001;49:321-327. discussion 327-328

73. Lefranc F., Brotchi J. Surgical strategy in spinal cord hemangioblastomas. Operative Techniques in Neurosurgery. 2003;6:24-31.

74. Hsu F.P.K., Clatterbuck R.E., Kim L.J., Spetzler R.F. Intramedullary spinal cord cavernous malformations. Oper TechNeurosurg. 2003;6:32-40.

75. Houten J.K., Weiner H.L. Pediatric intramedullary spinal cord tumors: special considerations. J Neurooncol. 2000;47:225-230.

76. Benes V.3rd, Barsa P., Benes V.Jr., et al. Prognostic factors in intramedullary astrocytomas: a literature review. Eur Spine J. 2009;18:1397-1422.

77. Resche F., Moisan J.P., Mantoura J., et al. Haemangioblastoma, haemangioblastomatosis, and von Hippel-Lindau disease. Adv Tech Stand Neurosurg. 1993;20:197-304.