Spine
Spinal Tumors: Compartment First, Then the Short List
Fellowship-level neuroradiology teaching by E. Brooke Schrickel, MD. Open this topic in the interactive reading room →
Watch · concise explainerVisual summary
Core clinical idea
There are three useful compartments: extradural, intradural–extramedullary, and intramedullary. Decide which one you are in on the axial images before naming a tumor. An extradural mass effaces epidural fat and often arises from bone; an intradural–extramedullary mass displaces the cord; an intramedullary mass expands the cord from within.
Bottom line
First say what the lesion does to the cord: displaces it, expands it, or compresses it from outside the dura. That single observation turns an unhelpful long differential into a practical short list.
Core workstation questions
- Which compartment is this on axial postcontrast and T2 images: extradural, intradural–extramedullary, or intramedullary?
- Is the cord displaced, expanded from within, or compressed by disease that effaces epidural fat and/or arises from bone?
- Extradural: is there a vertebral-body or posterior-element lesion, pathologic fracture, epidural extension, foraminal/paraspinal disease, or multilevel involvement?
- Intradural–extramedullary: does nerve-root continuity or foraminal/dumbbell extension favor schwannoma, or does a broad dural base, thoracic location, and calcification favor meningioma?
- At the conus or cauda equina, should the short list expand to myxopapillary ependymoma, paraganglioma, or leptomeningeal/drop metastasis?
- Intramedullary: is the lesion central and circumscribed with hemorrhage/cap sign and a syrinx, or eccentric and infiltrative?
- Is there an avid mural nodule with feeding vessels and disproportionate edema/syrinx suggesting hemangioblastoma?
- In a patient with cancer, is a small avid intramedullary lesion with extensive edema a metastasis rather than a primary cord tumor?
- Before calling tumor, do the expansion and enhancement pattern fit a mimic such as inflammatory myelitis, demyelination, infarct, or vascular malformation?
What changes reporting / management
- Lead with the compartment and exact level. Then describe craniocaudal length, degree of cord displacement or expansion, cord edema, syrinx/cyst, hemorrhage, enhancement pattern, multiplicity, and any leptomeningeal disease.
- For extradural neoplasm, metastasis is the default starting point. Report osseous origin, pathologic fracture, epidural and foraminal extension, cord signal abnormality, and the maximal Bilsky ESCC grade; urgent high-grade compression should be communicated.
- Schwannoma usually follows a nerve root and may widen a foramen or form a dumbbell mass. Meningioma more often has a broad dural attachment, homogeneous enhancement, thoracic location, and calcification. A dural tail supports but does not prove meningioma.
- Cauda-equina masses need a level-specific differential: myxopapillary ependymoma and paraganglioma join nerve-sheath tumor; multiplicity or nodular root coating raises leptomeningeal metastatic disease.
- In adults, ependymoma is commonly central, symmetric, and relatively circumscribed, often with hemorrhage/cap sign or an associated syrinx. Astrocytoma is more often eccentric, infiltrative, and longer segment, but overlap is substantial.
- Hemangioblastoma often presents as a vividly enhancing nodule with prominent feeding/draining vessels and edema or syrinx out of proportion to tumor size. Multiplicity should prompt syndromic context such as VHL.
- Intramedullary metastasis should be considered when a small enhancing lesion produces extensive edema, especially with known systemic cancer or concurrent brain, osseous, or leptomeningeal disease.
- Recommend additional imaging only when it answers a defined question: CT for mineralization or osseous detail; whole-spine/neuraxis imaging for multiplicity, suspected CSF dissemination, syndromic disease, or an unexplained syrinx—not automatically for every mass.
Practical traps
- Reading only the sagittal sequence. Compartment localization usually becomes obvious on axial images.
- Calling every enhancing intradural mass a meningioma because of a dural tail; schwannoma, metastasis, and other lesions can mimic it.
- Assuming an adult intramedullary tumor must be ependymoma. Astrocytoma, hemangioblastoma, metastasis, and non-neoplastic mimics remain pattern- and context-dependent.
- Using one sign as a diagnosis: cap sign, foraminal widening, calcification, and homogeneous enhancement are helpful tendencies, not absolutes.
- Failing to grade and communicate epidural cord compression in metastatic disease.
- Letting tumor-pattern teaching obscure an emergency mimic: epidural abscess or hematoma can occupy the same extradural space and require immediate action.
Teaching pearls
- Displaced cord = think intradural–extramedullary; expanded cord = intramedullary; effaced epidural fat/osseous origin = extradural.
- Extradural tumor is metastatic until the imaging and clinical context give you a better answer.
- Foraminal/dumbbell favors schwannoma; broad dural base and calcification favor meningioma.
- Adult intramedullary: central/circumscribed/hemorrhagic favors ependymoma; eccentric/infiltrative favors astrocytoma.
- A tiny avid nodule with too much edema or syrinx is the hemangioblastoma pattern.
- A small intramedullary lesion with too much edema in a patient with cancer is a metastasis pattern.
- The useful report is not just a tumor name: compartment, level, compression, cord signal, bone stability, and multiplicity determine what happens next.
Teaching visuals
Source lectures
- Radiopaedia: Intradural extramedullary spinal tumors
- Radiopaedia: Epidural spinal cord compression scale
- Koeller & Shih, Radiographics 2019: Intradural Extramedullary Spinal Neoplasms
- Ottenhausen et al., Frontiers in Surgery 2019: Intradural spinal tumors in adults
- Moghaddamjou et al., Frontiers in Radiology 2022: Imaging of metastatic epidural spinal cord compression
- Samartzis et al., Global Spine Journal 2015: Intramedullary spinal cord tumors
Educational material for radiology residents and neuroradiology fellows. Nothing here drives individual patient care, and it contains no patient data.
