The non-invasive diagnosis of musculoskeletal infections remains a challenge. Recent studies have indicated that fluorine-18 fluorodeoxyglucosepositron emission tomography (FDG-PET) is a highly accurate imaging technique in selected patient groups with infected total hip replacement. The present study analyses the diagnostic accuracy in a consecutive series of patients with suspected musculoskeletal infections.

METHODS: 163 consecutive patients with suspected periprosthetic infections (40 THR, 46 TKR), discitis (22) or a suspected infection involving the peripheral skeleton (55) were studied with FDG-PET. In this retrospective study two independent nuclear medicine physicians interpreted the images solely based on the information provide at the time of investigation. The final diagnosis was based on histopathological studies or microbiological culture or on clinical findings after at least twelve months of follow-up.

Results: Based on the final composite assessment, 21/40 patients with THR, 15/46 with TKR, 22/55 with suspected infection in the peripheral skeleton and 10/22 with suspected discitis had infection. FDG-PET identified correctly 68/76 infections (sensitivity 89.5%) and demonstrated a negative predictive value of 81/87 (specificity 93.1%). FDG-PET was of different diagnostic value at different sites with sensitivity and specificity for suspected infections of THR (100/81.3), TKR (81.8/85.7) infections of the peripheral skeleton (90.9/100) and discitis (100/100).

DISCUSSION AND Conclusions: FDG-PET is highly accurate for the evaluation of musculoskeletal infections. While it correctly identified all patients with suspected discitis, it seems also be reliable to rule out infected THR. However, the specificity in suspected infections of THR and TKR is lower due to granulomatous tissue caused by wear-induced polyethylene particles in aseptic loosening.