Intradural Extramedullary Tumors

The compartment. Intradural extramedullary (IDEM) tumours lie within the dural sac but outside the spinal cord parenchyma. They account for 55-75% of all intradural spinal tumours and are predominantly benign.

The patterns. Meningioma shows a strong female predominance (80%) and a thoracic predilection. Schwannoma affects males and females equally and occurs at any spinal level, often cervical or lumbar. Myxopapillary ependymoma of the filum terminale is the most common primary tumour of the conus and cauda equina region.

Schwannoma. Arises from the Schwann cells of a spinal nerve root, usually a single one, and grows eccentrically from it as an encapsulated mass. The nerve fibres are displaced around the capsule rather than infiltrated, which is why the nerve can often be preserved at surgery. Malignant transformation is extremely rare in sporadic schwannomas but may occur in NF2.

• Antoni A areas: cellular, organised palisading (Verocay bodies)
• Antoni B areas: loose, myxoid, less cellular
• S-100 protein positive

Meningioma. Arises from arachnoid cap cells in the spinal meninges, most commonly at the posterolateral aspect of the dura, and grows as a typically well-circumscribed mass with a broad dural attachment. It may calcify (psammoma bodies), and the dural tail seen on MRI reflects reactive dural thickening. There is an association with female sex and hormonal factors; progesterone receptors are often positive.

• Meningothelial: the most common spinal subtype
• Psammomatous: with calcifications
• Transitional

View larger

Neurofibroma. Contains Schwann cells, fibroblasts and collagen, and the nerve fibres pass through the tumour rather than around it. That infiltrative architecture is the practical difference from schwannoma: complete resection usually requires sacrifice of the nerve root. S-100 marks the Schwann cell component and CD34 the fibroblast component.

• Localised: a single nerve, similar to schwannoma
• Plexiform: "bag of worms" appearance, pathognomonic of NF1
• Diffuse: poorly circumscribed

Myxopapillary ependymoma. A distinct subtype arising from the ependymal cells of the filum terminale: well-circumscribed and encapsulated, with papillary architecture over myxoid cores, GFAP positive. Its mucin content makes it T2 hyperintense, and it may contain haemorrhagic areas. The histology is benign, but violating the capsule at surgery increases the risk of CSF seeding and recurrence.

Pain. The most common initial symptom: localised back pain in 70%, or radicular pain following the distribution of a nerve root, and it may precede any neurological deficit by months to years. Night pain, and pain that is worse on lying down, is the discriminator. Mechanical back pain is typically relieved by recumbency, so the patient whose pain wakes them or drives them out of bed to sit up has earned an MRI. The proposed mechanism is the rise in epidural venous pressure and the loss of the distraction of daytime activity when supine.

The trap. Because pain can precede any deficit by months to years, these patients are frequently treated as mechanical back pain through several cycles of physiotherapy before anyone images them.

Deficit. Sensory change (numbness, paraesthesiae), progressive motor weakness and gait disturbance; bladder and bowel dysfunction is late.

Each tumour presents in its own way.

• Schwannoma: radicular pain in a dermatomal distribution, sensory loss in the affected root, and motor weakness if the motor root is involved
• Meningioma: often progressive myelopathy with gait disturbance and spasticity, Brown-Sequard syndrome if the compression is lateral, and may have minimal radicular symptoms
• Myxopapillary ependymoma: low back pain and cauda equina symptoms (bowel and bladder, saddle anaesthesia), often with a long history before diagnosis

Examination. Cord compression gives upper motor neuron signs below the lesion: spasticity, hyperreflexia, clonus and a positive Babinski. Root compression gives lower motor neuron signs at the affected level: weakness in a myotomal distribution, hyporeflexia and muscle atrophy.

View larger

MRI is the gold standard, and each sequence answers a different question. T1 defines the anatomical relationships and gives the pre-contrast baseline; tumours are usually isointense to cord. T2 or STIR is the key differentiating sequence, schwannoma hyperintense and meningioma iso- or hypointense, and shows cord oedema. All IDEM tumours enhance after gadolinium: schwannoma heterogeneously, meningioma homogeneously and with a dural tail. Image the whole spine to exclude skip lesions, additional tumours and NF2.

View larger

Schwannoma. Iso- or hypointense to cord on T1 and hyperintense on T2, round and eccentric to the nerve root, with heterogeneous enhancement that may have non-enhancing cystic areas. Foraminal extension and bone erosion mark the dumbbell tumour. The tumour-to-fat signal intensity ratio on T2 is significantly higher in schwannoma than in meningioma and differentiates the two reliably.

Meningioma. Isointense on T1 and iso- or hypointense on T2, with avid homogeneous enhancement from a broad dural base, a lateral or posterolateral position, and a ginkgo leaf shape. Calcification may be present and is dark on all sequences. The dural tail is suggestive but not pathognomonic.

Neurofibroma. T2 hyperintense like a schwannoma, but with a target sign of central low T2 signal; plexiform lesions have the bag-of-worms appearance, and the tumours may be multiple in NF1.

Myxopapillary ependymoma. T2 hyperintense and well-circumscribed at the filum terminale, with homogeneous enhancement; haemorrhage shows as T1 bright areas.

Drop metastases. Multiple nodular lesions coating the nerve roots. Consider a primary CNS tumour (medulloblastoma, ependymoma).

View larger

CT shows the bone erosion of a dumbbell tumour and the calcification of a meningioma, and is used for pre-operative planning. Blood tests are the routine pre-operative workup; consider genetic testing if NF is suspected.

The standard operation is posterior. Most IDEM tumours are approached this way, with intraoperative neurophysiological monitoring (SSEP, MEP) recommended throughout. Laminectomy is the standard approach and laminoplasty the alternative that may reduce post-laminectomy kyphosis; whether the lamina is removed or replaced, and whether instrumentation is added, is decided on the stability grounds set out below.

1. Posterior midline incision
2. Laminectomy or laminoplasty
3. Durotomy (midline or paramedian)
4. Tumour identification
5. Microsurgical resection
6. Watertight dural closure
7. Bone replacement if laminoplasty

Schwannoma. The goal is gross total resection with the nerve preserved where possible, and it is achievable in 90%+. Identify the functioning fascicles with stimulation, peel the tumour from the nerve capsule, and sacrifice the parent root only if necessary; a sensory root is often tolerated. Long-term control after gross total resection is excellent.

View larger

View larger

Dumbbell tumours. The approach follows the Eden type. Stage the surgery when the extraforaminal component is large, and in the cervical spine plan for the vertebral artery.

View larger

Meningioma. The goal is gross total resection with the dural attachment dealt with, achievable in 85-95% of spinal meningiomas. Identify the dural attachment, dissect circumferentially from the cord, and excise the involved dura with primary closure or a graft; where excision is not possible, coagulate the dural margin. Simpson grading predicts recurrence:

• Grade I: GTR + dural excision + bone removal
• Grade II: GTR + coagulation of the dural attachment
• Grade III: GTR without dural treatment
• Grade IV: subtotal resection
• Grade V: decompression only

Myxopapillary ependymoma. The goal is en bloc resection without violating the capsule, preserving the nerve roots where possible. The prognosis after gross total resection is excellent, capsule violation increases the risk of CSF seeding, and en bloc removal is preferred over piecemeal. After a gross total resection the patient is observed; after a subtotal resection or a violated capsule, radiotherapy. Follow-up MRI is annual.

What the largest series shows. Weber's 183 patients: 10-year overall survival 92.4% and 10-year progression-free survival 61.2%. Quote both, because this tumour rarely kills and frequently returns. Treatment failed in about a third (31.7%), mostly locally (26.8%), but distant spinal failure in 9.3% and intracranial in 6.0% is why the whole neuraxis is imaged, not just the lumbar spine. Extent of resection and adjuvant radiotherapy were independent prognostic factors, as was age: patients under 36 were more likely to progress.

En bloc is the aim, not the outcome that matters. The outcome that separates good from catastrophic is whether the resection was total or partial, and whether an incomplete resection received adequate radiotherapy, not whether the specimen came out in one piece.

View larger

Observation suits small asymptomatic tumours, slow-growing tumours followed with serial MRI, and elderly or medically unfit patients. Radiation is rarely used for benign IDEM tumours: consider it for recurrent or incompletely resected meningioma, stereotactic radiosurgery for small recurrences, and as primary treatment if the patient is not a surgical candidate.

The typical spinal meningioma. Most are soft, posterolateral, WHO Grade I lesions that are gross totally resected in 85-95 percent of cases. A minority behave very differently, and recognising them preoperatively changes the consent and the surgical plan: a ventral calcified thoracic meningioma carries a higher recurrence and complication rate than the typical posterolateral lesion.

View larger

View larger

The approach itself can destabilise the spine. Resecting an IDEM tumour requires a laminectomy or laminoplasty, and post-laminectomy kyphosis (a swan-neck deformity in the cervical spine) is the chief late mechanical problem, so deciding when to add instrumented fusion is part of operative planning. Laminoplasty, replacing the lamina rather than removing it, preserves the posterior tension band and may reduce kyphosis.

View larger

Early. CSF leak is the most common, at 5-10%; the others are wound infection, a new neurological deficit in 2-5%, and haematoma.

Late. Post-laminectomy kyphosis, tumour recurrence, chronic pain and arachnoiditis.

Preventing the CSF leak. Meticulous dural closure, a dural graft if primary closure is not possible, fibrin sealant augmentation, and a watertight closure in layers. If a leak occurs:

1. Wound care
2. Lumbar drain for 3-5 days
3. Re-exploration if conservative measures fail

View larger

View larger

Recurrence. The conventional figures, by extent of resection:

Read this table as orders of magnitude, not as sourced figures. The schwannoma, meningioma and neurofibroma rows are the conventional teaching estimates and no single series establishes them. For myxopapillary ependymoma the best figure available is Weber's 10-year progression-free survival quoted under Management, so roughly two in five progress by ten years; that series does not publish a clean GTR-versus-STR recurrence split, so quoting one would be inventing it.

View larger

Global Epidemiology

IDEM tumors comprise roughly 55-75% of all intradural spinal tumors. Incidence of primary spinal tumours overall is about 0.5-2.5 per 100,000 person-years. Schwannoma and meningioma together account for the majority; spinal meningioma shows a 3:1 to 4:1 female predominance with a thoracic predilection, while nerve sheath tumours are evenly distributed by sex and across spinal levels. Myxopapillary ependymoma is the commonest tumour of the conus/filum and presents in young adults.

Society Guidance & Frameworks (side by side)

There is no single arthroplasty-style registry for IDEM tumours; outcome evidence derives from institutional series, the Rare Cancer Network (myxopapillary ependymoma) and national cancer/CNS-tumour registries (e.g. CBTRUS in the US) which inform incidence and survival rather than implant survival.

High- vs Limited-Resource Practice Variation

In well-resourced centres, microsurgery with the operating microscope, intra-operative neurophysiological monitoring (SSEP/MEP/D-wave), ultrasonic aspiration and ready access to MRI surveillance underpin gross total resection rates comparable to international centres of excellence. In limited-resource settings, delayed presentation with established myelopathy is common, monitoring may be unavailable, and emphasis shifts to timely decompression and watertight closure to limit CSF leak. Genetic testing and dedicated NF2/schwannomatosis multidisciplinary clinics are concentrated in tertiary referral networks; telemedicine and regional referral pathways help extend specialist input where local capacity is limited.

1. Laminectomy vs laminoplasty vs instrumented fusion. Whether laminoplasty or routine instrumentation reduces post-operative deformity remains unsettled. Most series reserve fusion for multilevel laminectomy (3 or more levels), the cervicothoracic/thoracolumbar junctions, pre-existing deformity, and paediatric patients in whom post-laminectomy kyphosis is far more common.

2. Nerve root sacrifice in nerve sheath tumors. Traditional teaching preserves the parent root; contemporary data (Vandenbulcke et al., 2023) suggest the involved root in foraminal/dumbbell tumors is frequently non-functional and can be sacrificed to achieve gross total resection with a low risk of clinically significant deficit. Intra-operative stimulation to confirm a non-eloquent root is the pragmatic compromise.

3. Extent of dural handling for meningioma. Simpson Grade I (dural excision and duraplasty) lowers recurrence versus Grade II (coagulation only) but raises the risk of CSF leak and cord injury when the dura is ventral. Many surgeons accept Grade II for ventral/ventrolateral thoracic meningiomas where aggressive dural resection is hazardous.

4. Role of radiosurgery. Stereotactic radiosurgery is increasingly used for residual/recurrent meningioma and schwannoma and for NF2 tumors, but long-term spinal cord tolerance data are limited; surgery remains first-line for symptomatic compressive lesions.

5. Surveillance versus treatment of small asymptomatic tumors. Incidental small IDEM tumors (especially in NF2) are often observed with serial MRI; growth rate and symptom development, rather than size alone, drive the decision to operate.

1. Takashima H, Takebayashi T, Yoshimoto M, et al. Differentiating spinal intradural-extramedullary schwannoma from meningioma using MRI T2 weighted images. Br J Radiol. 2018;91(1092):20180262. doi:10.1259/bjr.20180262. PMID 30052467.
2. Seppala MT, Haltia MJ, Sankila RJ, et al. Long-term outcome after removal of spinal schwannoma: a clinicopathological study of 187 cases. J Neurosurg. 1995;83(4):621-6. doi:10.3171/jns.1995.83.4.0621. PMID 7674010.
3. Gottfried ON, Gluf W, Quinones-Hinojosa A, et al. Spinal meningiomas: surgical management and outcome. Neurosurg Focus. 2003;14(6):e2. doi:10.3171/foc.2003.14.6.2. PMID 15669787.
4. Weber DC, Wang Y, Miller R, et al. Long-term outcome of patients with spinal myxopapillary ependymoma: treatment results from the MD Anderson Cancer Center and institutions from the Rare Cancer Network. Neuro Oncol. 2015;17(4):588-95. doi:10.1093/neuonc/nou293. PMID 25301811.
5. Vandenbulcke A, D'Onofrio GF, Capo G, et al. Sacrifice of involved nerve root during surgical resection of foraminal and/or dumbbell spinal neurinomas. Brain Sci. 2023;13(1):109. doi:10.3390/brainsci13010109. PMID 36672090.