Seventy-three patients with various underlying disorders that could potentially alter distribution of bone marrow underwent In-111 labeled autologous leukocyte and Tc-99m SC marrow imaging to exclude musculoskeletal infection. There were 22 cases of infection among the 73 patients studied. When interpreted in isolation, labeled leukocyte images were considered positive for infection when uptake of labeled cells in the region of interest exceeded uptake in the corresponding contralateral side. The sensitivity, specificity, and accuracy of the study using this criterion were 86%, 12%, and 34%, respectively. When interpreted in conjunction with sulfur colloid marrow images, studies were considered positive for infection when uptake in the region of interest on leukocyte and marrow images was spatially incongruent. The sensitivity, specificity, and accuracy of this dual tracer modality were 100%, 94%, and 96%, respectively. Three patients with infection and abnormal leukocyte/marrow images underwent repeat imaging after appropriate therapy. Images were interpreted as normal in two of them; both were infection free at subsequent surgery. Images of the third patient were interpreted as consistent with persistent infection, and operative cultures grew out Staphylococcus aureus. In summary, combined leukocyte/marrow imaging is a highly accurate method for diagnosing musculoskeletal infection and overcomes certain inherent limitations of labeled leukocyte imaging alone.