The clinical utility of magnetic resonance imaging (MRI) in judging therapeutic response of bone metastases was evaluated in 18 patients with advanced breast cancer. Treatment efficacy was assessed by MRI and conventional methods such as plain radiograph, bone scan, pain and analgesic scale, and serum CA15-3. The response by MRI was evaluated mainly on T1-weighted sequences by measuring the volume of the bone lesion and soft tissue component. The patient was assumed to be a conventional responder if a complete or partial response was observed in any of the conventional methods described above. Response was most concordant between plain radiographs and MRI findings (91%, 10/11, 95% confidence interval [CI]: 58.7-99.8). The rate of concordance was 61% (11/18, 95% CI 35.8-82.7) for all conventional methods and MRI. MRI revealed response in four patients in whom progressive disease was observed by bone scan and the marker response was not measurable. This pilot study suggests that posttherapy evaluation with MRI may provide useful clinical information in breast cancer patients with bone metastases and may be a valuable adjunct to conventional methods with conflicting results.