The purpose of this prospective study was to define the value of bone scintigraphy (BS), bone marrow scintigraphy (BMS) and the new fast spin-echo (FSE) magnetic resonance imaging (MRI) sequences in screening for bone metastases in patients with solid malignant tumours. It was our particular interest to classify patients into a group with and a group without bone metastases, and not only to compare the absolute number of metastases detected by each method. Thirty-two patients were examined using technetium-99m dicarboxy propane diphosphonate bone scintigraphy, 99Tc-labelled monoclonal anti-granulocyte antibodies for bone marrow scintigraphy and 1.5 T MRI using T1-weighted and FSE T2-weighted sequences. Against a reference standard obtained by re-evaluation of all clinical and imaging data 1 year after prospective BS, BMS and MRI had been performed, the three imaging modalities were falsely positive in two, eight and two cases and falsely negative in zero and four cases, respectively. BMS was falsely positive in eight patients because of vertebral marrow degeneration which caused photopenic defects which could not be differentiated from metastases. MRI showed these lesions to unequivocally contain fat. BMS and MRI were falsely negative in four cases because of the limited field of examination. In our study the key factor in classifying a patient as bone M1 or M0 was the possibility of surveying the entire skeleton, as is the case in BS, and not that MRI had a higher sensitivity compared to BS when analysis was on a lesion-by-lesion basis.(ABSTRACT TRUNCATED AT 250 WORDS)