The tuberculin skin test (TST) and interferon gamma (IFN-${gamma}$) release assay (IGRA) have been widely used for diagnosis of latent tuberculosis infection (LTBI). In order to overcome limitations of current LTBI diagnostic methods, the development of a novel molecular assay which is able to measure the IFN-${gamma}$ messenger RNA (mRNA) expression level after stimulation with Mycobacterium tuberculosis (MTB) specific antigen was recently developed. The ability of a molecular assay to detect MTB infection was similar to commercial IGRA however, the optimal incubation time for stimulating IFN-${gamma}$ was not yet established. Therefore, in this study the direct comparisons of MTB Ag stimulation times (4 and 24 hrs) were performed for diagnosis of MTB infection. Data showed that the coincident rate between QFT-GIT IFN-${gamma}$ ELISA and IFN-${gamma}$ RT-PCR (4 hrs) was 88.35% and that of QFT-GIT and IFN-${gamma}$ RT-PCR (24 hrs) was 70.85%. Based on a receiver operating characteristic (ROC) curve, the 4 hrs-MTB specific Ag stimulation time for IFN-${gamma}$ RT-PCR had the significant P value, 95% CI value, and AUC (P < 0.0001, 95% CI=0.82 to 1.02, and AUC=0.9214) in comparison with 24 hrs-MTB specific Ag stimulation time (P = 0.009, 95% CI=0.06 to 0.94, and AUC=0.7711). These results show that 4-hr was the most optimal MTB Ag stimulation time for performing IFN-${gamma}$ RT-PCR. Although semi-quantitative RT-PCR had a few analytical limitations, it might be useful as an alternative molecular diagnostic method for detecting MTB infection.