BACKGROUND When musculoskeletal pain is persistent or out of proportion to clinical findings, diagnostic imaging is often used to evaluate for the presence of occult pathologic conditions. OBJECTIVE To examine the efficacy of bone scintigraphy for identifying occult malignancy in such patients. METHODS Bone scans of 491 patients with new or recurrent complaints of musculoskeletal pain and no previously known malignancy were retrospectively reviewed. The findings of each scan were classified into 1 of the following categories: (A) no evidence of malignancy; (B) equivocal (1-3 lesions of uncertain significance); or (C) probable metastatic disease (> 3 lesions without trauma or other benign explanation). All diagnoses of malignancy established within 6 months of the bone scan were then determined from review of radiological and clinical records. RESULTS The distribution of scan interpretations was as follows: category A, 386 (79%); category B, 84 (17%); and category C, 21 (4%). Among the 181 patients younger than 50 years, 161 (89%) had category A and 20 (11%) had category B scans. Of the 310 patients aged 50 years or older, 226 (73%) had scans in category A, 64 (21%) in category B, and 21 (7%) in category C. Thirty patients (6%) had malignancy involving bone, 29 among those aged 50 years or older. In patients aged 50 years or older, 10 (16%) of 64 with equivocal scan findings (category B) and 19 (90%) of 21 with widespread abnormalities (category C) had malignancy involving bone. Initial radiographs of symptomatic sites showed lytic or blastic bone abnormalities suggestive of malignancy in 16 (59%) of the 29 older patients in whom this diagnosis was confirmed. CONCLUSIONS In patients aged 50 years or older, the yield of bone scintigraphy for identification of occult malignancy (29/310; 9%) is sufficient to justify its use in the investigation of the cause of problematic bone or musculoskeletal symptoms.