Although magnetic resonance imaging (MRI) may be more sensitive than myelography with CT scanning in detecting spinal metastatic disease (SMD), the limitations of these MRI examinations have not been defined. The feasibility of obtaining early postoperative MRI to screen for SMD, and the type and frequency of imaging abnormalities were evaluated in this prospective study. Twenty-seven patients (78% with nondisseminating neoplasms or nonneoplastic diseases) were studied with postoperative spinal MRI without and with contrast enhancement, obtained within 24 h of craniotomy. Procedural problems with immediate spinal MRI were identified. Physician monitoring of the examination did not guarantee an adequate examination, and small abnormalities were missed or incompletely evaluated at the time of the MRI. Surprisingly, abnormal signal from intraspinal gas, blood, or blood breakdown products was not seen. In the absence of metastatic disease, spinal meningeal enhancement was absent. True SMD was detected, but nonneoplastic focal contrast enhancement (false positive) occurred in 11%. An unusual nonneoplastic subdural cerebrospinal fluid enhancement was seen in 3 of 13 (23%) of all patients undergoing posterior fossa craniectomies. This enhancement could hide small metastases or could be confused for SMD. Although immediate postoperative spinal MRI is potentially effective as a screening tool for the detection of SMD, we have identified some significant limitations of this technique.