Detection of cerebrospinal fluid metastasis: CT myelography or MR?

PURPOSE To determine the sensitivity of contrast MR versus myelography followed by CT in the detection of cerebrospinal fluid metastases in children with primary central nervous system tumors. METHODS Thirty-three patients who had primary central nervous system malignancies had spinal MR with gadolinium within 2 weeks of a myelogram followed by CT. MR technique included T1-weighted image sequences of the entire spine with and without gadolinium. CT scans were routinely performed at T-12 to L-2, L-4 to S-1, and foramen magnum to C-2. All studies were reviewed blindly; the number, character, and location of all metastases was recorded and the results of the two studies compared. Cerebrospinal fluid cytologic findings were recorded for each patient, and compared with the results of the imaging studies. RESULTS Seven of the 33 patients had metastases detected; metastases were seen on both MR and myelography followed by CT. However, MR showed 24 lesions and myelography followed by CT showed only 15. When a lesion was seen on both MR and myelography followed by CT, the MR was usually more convincing. Superficial lesions seen on MR sometimes would be missed on myelography followed by CT. Both MR and myelography followed by CT were quite sensitive in the detection of small lesions (2 to 3 mm) when present on spinal nerve roots. Whereas MR showed multiple lesions not seen on myelography followed by CT, CT failed to show any metastases not seen on MR. Imaging studies showed metastases in 3 patients who had normal cytologic findings. CONCLUSIONS MR shows significantly more cerebrospinal fluid metastases than myelography followed by CT.