PURPOSE: To determine the possible mechanism by which an area of high signal intensity appears on T1-weighted MR images adjacent to a vacuum cleft in intervertebral disks. MATERIALS AND METHODS: We analyzed a total of 14 disks in nine patients in whom a vacuum cleft with T1-signal hyperintensity was observed. Lesions were present from T11-12 to L5-S1 using a 1.5-T whole-body imager, sagittal spine-echo T1-weighted and gradient-echo images (flip angle, 20 'and 60 ) were obtained. In order to identify the vacuum cleft, using plain radiographs in all patients and CT scans in two were also obtained. A 3% agar-gel block containing empty slits to form a magnetic susceptibility difference, a phantom was designed. The air spaces were 1.6 mm in thickness, 25 mm in width, and 20 to 25 mm in depth with 1.6-mm spacing. RESULTS: In all patients, vacuum clefts were confirmed by plain radiographs and CT scans. At the level containing air, T1-weighted images (both spin-echo and gradient-echo) showed a signal void resulting from the intervertebral disk vacuum cleft. A hyperintense band adjacent to the vacuum cleft was, however, observed. A gradient-echo image with a 60 'flip angle showed a brighter signal intensity than one with a 20 'angle. Our phantom study gave the same results. CONCLUSION: The magnetic susceptibility artifact may be responsible for the T1-signal hyperintensity observed adjacent to the vacuum cleft in intervertebral disks. In addition, in order to generate signal hyperintensity, the desiccating disk material must contain a certain amount of water molecules.