We retrospectively reviewed spinal surveys performed at our institution from June to Novermber 2009. All patients were known to have spinal metastases at the time of imaging by prior pathologic confirmation. All spinal surveys employed sagittal T1, STIR and T2 weighted sequences. A contrast enhanced sequence was obtained if there were no contraindications to gadolinium based agents. In addition, an echo-planar (EPI) diffusion weighted sequence (DWI) was obtained (3-6 directions, B value 500-800 ) in the sagittal plane.

Studies were consensus reviewed twice by 2 experienced radiologists and each lesion was graded separately. In the first review, only the T1, STIR and (if available) contrast enhanced sequences were reviewed, and each lesion was graded on a scale of 1-4 (1 = missed lesion that was visible only in retrospect on second review; 2 = equivocal lesion; 3 = probable lesion; 4 = definite lesion). Subsequently the DWI sequence was reviewed in conjunction with conventional sequences. Lesions were again graded on a scale of 1-4 based on additional information garnered from their diffusion characteristics. In addition conspicuity on diffusion was compared to conventional sequences.