PURPOSE: To compare the value of MRI sequences with dynamic MRI measurements (d-MRT) for the assessment of risk of lumbar vertebral fractures in patients with multiple myeloma. MATERIAL AND METHODS: In 10 patients with multiple myeloma a sagittal T 1 -, T 2 -weighted spin-echo and an inversion-recovery sequence were performed. For dynamic measurements a fast gradient-echo sequence (turbo fast low angle shot 2 D) with machine-controlled Gd-DTPA administration was used. The presence of bone marrow abnormalities was determined and additionally the value of the highest signal increase (amplitude Alpha) was calculated for each of the 50 vertebral bodies. The subsequent development of vertebral fractures was assessed by MRI at a mean time interval of 6.2 months after the initial d-MRI. The pattern of marrow involvement and the amplitudes of the vertebral bodies that collapsed in the observation period were statistically compared with those of the vertebral bodies that did not collapse. RESULTS: During the follow-up period newly or progressive fractures occurred in 6 of 10 patients (7 of 50 vertebral bodies). The degree of pathological signal changes that preceded fractures was not significantly different (p > 0.05) from that of the other contemporary lesions identified in vertebral bodies that did not collapse in follow-up. In contrast, the amplitude of vertebrae that collapsed (A: 33.1 +/- 8.2) was significantly higher (p < 0.0001) than the amplitude of vertebrae that did not collapse (A: 16.7 +/- 4.2). The amplitude was a reliable predictor of vertebrae that collapsed in all cases. CONCLUSION: In contrast to the analysis of marrow lesions detected with non-dynamic MRI, this study suggests that the perfusion-parameter amplitude in dynamic contrast-enhanced MRI is a potentially relevant value to predict the risk of vertebral fractures in patients with multiple myeloma.