Differentiation of sarcoidosis from tuberculosis using real-time PCR assay for the detection and quantification of Mycobacterium tuberculosis.

BACKGROUND The clinical and pathological features of sarcoidosis and tuberculosis may mimic each other, and when the caseous necrosis is not seen in tuberculosis tissue, differentiation is not easy. OBJECTIVE This study evaluates the ability of real-time PCR quantification and sets the quantitive value to differentiate sarcoidosis from TB. METHODS Formalin-fixed and paraffin-embedded sections of biopsy samples, from 104 patients with sarcoidosis, 31 patients with tuberculosis, and 55 controls with other respiratory diseases (26 with nonspecific lymphadenitis and 29 with emphysema bullae), were collected to amplify insertion sequence IS986 of Mycobacterium tuberculosis (MTB) genome by real-time quantitative PCR. The diagnostic performance of qualitative and quantitative analysis of real-time quantitative PCR was assessed by receiver-operating characteristic (ROC) curves. RESULTS MTB DNA was detected in 20 of the 104 sarcoidosis samples and 7 of the 55 control samples, but was detected in all of the 31 tuberculosis samples. The numbers of MTB genomes were 0-4.71x10(3) copies per ml in sarcoidosis samples, 1.58x10(2)-5.43x10(7) copies per ml in tuberculosis samples and 0-1.02x10(3) copies per ml in controls with quantitative analysis. Receiver-operating characteristic (ROC) curves showed that MTB genome quantification had greater diagnostic performance than MTB genome qualitation in discriminating patients with sarcoidosis from those with tuberculosis (area under the ROC curves: 0.994 vs 0.904, P<0.001). The sensitivity and specificity of qualitative analysis were 100% and 80.8% respectively. At cutoff value of 1.14x10(3) copies per ml for MTB genome quantification, the sensitivity was 96.8% and specificity was 98.1%. CONCLUSIONS The real time PCR quantification is a valuable test for differentiation between sarcoidosis and tuberculosis, and the MTB genome copies number of 1.14x10(3) copies per ml should be preferred as quantitative cutoff value for the differentiation.