Background: The whole-blood interferon-gamma release assay (QuantiFERON-TB Gold [QFT-G]: Cellestis, Carnegie, Victoria, Australia) has been studied primarily for the use of diagnosing active pulmonary tuberculosis (TB) or latent TB. In the present study, the usefulness of QFT-G was evaluated for the diagnosis of extra-pulmonary tuberculosis (EP-TB). Methods: From June 2006 to February 2009, we evaluated the usefulness of QFT-G in patients (n=65) suspected with EP-TB, retrospectively. The diagnostic sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the QFT-G assay were analyzed. Results: EP-TB was diagnosed in 33 (51%) participants. The overall sensitivity, specificity, PPV, and NPV of the QFT-G assay for EP-TB were 78%, 79%, 81%, and 77%, respectively. Of the 33 with EP-TB, 14 (42%) were diagnosed with TB pleurisy, 7 (21%) with TB lymphadenitis, 7 (21%) with intestinal TB, and 5 (15%) with EP-TB in other sites. In subgroup analyses according by site of infection, the QFT-G showed 86% sensitivity, 64% specificity, and 78% NPV in TB pleurisy. On the other hand, the sensitivity, specificity, and NPV of the assay were 71%, 83% and 71%, respectively in TB lymphadenitis, and 86%, 100% and 88%, respectively in intestinal TB. Among the patients with suspected alternative site EP-TB, the sensitivity, specificity, and NPV of the assay were 50%, 80% and 67%, respectively. Conclusion: The QFT-G assay showed moderate diagnostic accuracy in EP-TB. However, negative QFT-G assay does not exclude EP-TB because of the low NPV of this assay.