To assess the value of imaging by 0.1 T MR and by 99mTc-HMPAO-labeled leukocytes in confirming skeletal infection in patients with soft-tissue infections and/ or bone pathology. Methods: Thirty-nine anatomical sites (35 patients) with suspected bone infection were prospectively imaged with 0.1 T MR and 99mTc-HMPAO-labeled leukocytes. Thirty-two infected areas were confirmed: 12 osteomyelitis (out of which 3 were spondylitis) and 27 soft-tissue infections (both bone and soft-tissue infection in 7 areas). Results: MR imaging showed 3 1 true-positive, 3 true-negative, 4 false-positive and one false-negative diagnosis of infection and scintigraphy 27, 7, 0 and 5 respectively. The sensitivity of MR for osteomyelitis was 100% (12/12) and of scintigraphy 42% (5/12), peO.01. The specificity of MR and of scintigraphy for osteomyelitis were 81% (22/27) and 93% (25/27) respectively. The sensitivity of MR for soft-tissue infection was 96% (26/27) and specificity 75% (9112). The correspoding figures for scintigraphy were 85% (23/27) and 100% (12/12). M R and scintigraphy were concordant with respect to the final diagnosis in 28/39 (72%) sites and discordant in 10 (26%). In one patient with Charcot osteoarthropathy a false-positive finding was found by both methods. MR detected all 3 cases of spondylitis, scintigraphy none. Nonpyogenic inflammations and neuroarthropathic joints were indistinguishable from infection by MR. Conclusion: Combined imaging with MR and 99mTc-labeled leukocytes is recommended in diagnostically complicated bone infections except for spondylitis where MR is the method of choice. Congruent positive findings are highly suggestive of infection, the extent of which can be determined. Congruent negative results exclude infection.