The use of computed tomography of the spine to identify patients at high risk for epidural metastases.

The usefulness of spinal computed tomography (CT) in predicting the presence of epidural tumor was evaluated in cancer patients undergoing CT myelography for suspected epidural tumor. Two hundred ninety two vertebral levels were evaluated in 30 patients. Spinal CT demonstrated cortical disruption surrounding the epidural space from metastatic cancer in 109 vertebrae. Eighty-five (78%) of these vertebral levels had tumor extension into the adjacent epidural space. The incidence of epidural tumor adjacent to vertebrae which had normal spinal CT or metastatic tumor without cortical disruption was 11%. Eighty-six percent of the epidural tumor adjacent to these vertebrae were a result of craniocaudal tumor extension in the epidural space from adjacent vertebral levels with cortical disruption. Twenty-one of 23 patients (91%) with cortical disruption at more than one vertebral level on spinal CT had epidural tumor. Synchronous noncontiguous epidural lesions were observed in 38% of patients with epidural tumor. Spinal CT is an important diagnostic test in determining which patients are at high risk for epidural tumor. Myelography should be performed in all patients with suspected epidural tumor to accurately define the full extent of tumor.