Ninety-eight consecutive patients with 344 collapsed vertebrae underwent conventional and/or digital radiography and MRI. Vertebral collapse was due to osteopenia (16 cases), trauma (17 cases), and vertebral osteonecrosis (3 cases). Other causes were: spondylodiscitis (9 patients), primary neoplasm (4 patients), metastases (37 patients), and hemomyelopathies (7 patients); 5 patients bore vertebral angiomas. Sixty-three patients also underwent CT and 25 bone scintigraphy. As far as collapsed vertebral endplates are concerned, according to a previous classification, outcomes were divided into 4 groups. Type 1 (focal concave collapse) was observed in 10% of benign lesions and in 1% of malignant ones; type 2 (diffuse concave collapse) in 24% of benign and in 16% of malignant lesions; type 3 (focal collapse with an acute angle) in 11% of benign and malignant lesions, and type 4 (diffuse collapse with an acute angle) in 11% of benign and in 26% of malignant lesions. Neural arch involvement was observed in 3% of benign lesions and in 16% of malignant ones; paravertebral soft-tissue involvement in 6% of benign and in 29% of malignant lesions; vertebral canal involvement in 11% of benign lesions and in 20% of malignant ones. The intervertebral disk proximal to vertebral collapse was more involved in benign lesions (24%) than in malignant ones (8%). MR signal followed 4 main patterns: low signal on T1-weighted images and high signal on proton-density and T2-weighted images (2% of benign lesions, 49% of malignant ones), low signal on all sequences (7% of benign lesions and 25% of malignant ones), isointense signal on all sequences (50% of benign and 21% of malignant lesions), and high signal on all sequences (41% of benign lesions and 0% of malignant ones).