Devoted to diagnostic and interventional spine imaging and therapeutics


Examining Intramedullary Spinal Cord Metastases in Patients with Systemic Malignancies 2012

General Spine

Maya, E, Hartman, MD
Erica, Pollack, MD, Non ASSR Member
Sasan, Karimi, MD, Non ASSR Member
Eric, Lis, MD, Non ASSR Member
George, Krol, MD, Non ASSR Member
Robert, J, Young, MD, Non ASSR Member



Intramedullary spinal cord metastases are uncommon complications of systemic malignancies. The purpose of this study is to describe the clinical and imaging characteristics of intramedullary metastases, and to examine their relationship with leptomeningeal metastases.

Methods & Materials

We retrospectively searched a departmental database at a tertiary care cancer center for patients with systemic cancers who underwent MRI total spines over an 8 month period. A total of 22 patients had intramedullary enhancing mass lesions consistent with metastatic disease. The diagnosis was established by pathology, cytopathology and/or follow up; with only pathologically proven, clinically definite, or clinically probable intramedullary metastases included in the analysis. A total of 10 MRI characteristics were rated: homogeneous enhancement, well circumscribed margins, T2 hyperintense changes, hemorrhage, cystic/necrotic changes, length >1 vertebral segment, central location, posterior location, diffusion restriction, and associated leptomeningeal metastases. Chart and imaging review was also performed to evaluate for brain, bone, and systemic metastases.


The primary study cohort had a median age of 59 years with 10 males and 12 females. Breast cancer was the most common systemic malignancy (n=9), followed by melanoma, prostate and colon (tied at n=2). All intramedullary enhancing lesions showed T2 hyperintense changes, with the next most common abnormalities consisting of well circumscribed margins in 73% and homogeneous enhancement in 64%. Hemorrhage and cystic/necrotic changes were rare at 5% each. Concurrent brain and systemic metastases were found in 64% and 73%, respectively. When comparing breast to non-breast intramedullary metastases, breast cancer more commonly showed length >1 vertebral segment (56% vs. 23%, odds ratio [OR]=4.2), and leptomeningeal metastases with cord invasion and bone metastases (both 89% vs. 62%, OR=5). Pure intramedullary metastases in the absence of leptomeningeal disease were found in 27% of all patients, but in only 11% of breast cancers as compared to 38% of non-breast cancers (OR=.2).


Nearly three-quarters of the intramedullary spinal cord metastases in our series were due to direct cord invasion by leptomeningeal metastases. Pure intramedullary metastases without leptomeningeal metastases occurred more commonly in non-breast cancers. Further work is necessary to better understand the patterns of spread of metastatic disease to the spinal cord and leptomeninges, and to determine better diagnostic and prognostic imaging tools, in order to improve patient outcomes.

References/Financial Disclosures

1. Lee SS, Kim MK, Sym SJ, et al. Intramedullary spinal cord metastases: a single-institution experience. J Neurooncol 2007;84:85-89 2. Hrabalek L. Intramedullary spinal cord metastases: review of the literature. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub 2010;154:117-122 3. Kosmas C, Koumpou M, Nikolaou M, et al. Intramedullary spinal cord metastases in breast cancer: report of four cases and review of the literature. J Neurooncol 2005;71:67-72 Figure Legends Figs 1,2. Sagittal (1) and axial (2) contrast T1-weighted images in a patient with esophageal carcinoma show mild leptomeningeal metastases along the cervicothoracic cord surface with direct intramedullary invasion at the dorsal T1-T2 and T3-T4 levels. Figs 3,4. Sagittal (3) and axial (4) contrast T1-weighted images in a patient with non-small cell lung carcinoma show a solid and cystic/necrotic intramedullary enhancing mass at the T3-T4 level. The patient did not have leptomeningeal metastases, although she did have brain, bone and other systemic metastases (not shown). Disclosures All authors declare that they have no financial disclosures.