Comparison of radionuclide bone scans and magnetic resonance imaging in detecting spinal metastases.

A retrospective comparison was made between 99mTc-MDP bone scans and corresponding spine MR images in 35 patients who had complementary studies within 2 mo. Bone scans were performed with planar imaging of the entire body and MRI was performed with a 1.5 tesla signal scanner using standard techniques with T1- and T2-weighted images. There were 18 male and 17 female patients diagnosed with cancer prior to these studies. Cancer diagnoses included 14 prostate, 12 breast, 1 bladder, 2 renal, 2 lung, 1 each of esophagus, melanoma, myeloma and adenocarcinoma of unknown primary cancer. Of the regions compared, 69 were positive for bony metastases by MRI and 63 regions by bone scans. Thirty-eight regions were concordantly positive and 56 regions concordantly negative. No patients with entirely positive bone scans were negative by MRI, but one patient was entirely positive by MRI but negative by a bone scan. At least one region was discordantly read in 21 patients. Distribution of positive regions was similar on bone scan and MRI. The greatest number and proportion of discordant readings occurred in the lumbar regions and more frequently in patients with prostate cancer. Considering its widespread availability and the ease of performing a whole-body survey for metastasis, radionuclide bone scanning remains the study of choice for initial evaluation of patients with cancer. However, MRI is an excellent complementary technique when bone scan findings are inadequate for answering clinical questions. MRI appears to be quite sensitive and probably more specific for metastasis in certain locations of the spine.