Diagnostic accuracy of metagenomic next-generation sequencing for active tuberculosis in clinical practice at a tertiary general hospital

BACKGROUND: To evaluate the diagnostic accuracy of metagenomic next-generation sequencing (mNGS) for active tuberculosis (TB). METHODS: We retrospectively collected 820 samples at Zhongshan Hospital, Fudan University in Shanghai, China, between 1 April 2017 and 31 March 2018. They were classified into TB cases (125, 15.2%) and NOT TB cases (695, 84.8%) according to the clinical diagnosis. Specimens were evaluated by a regular clinical microbiological assay and mNGS performed in parallel. RESULTS: Sixty-one confirmed TB cases and 64 clinical TB cases were included. The overall sensitivity of mNGS was 49.6% [95% confidence interval (95% CI), 40.6–58.6%], and the specificity was 98.3% (95% CI, 96.9–99.1%), with peak sensitivities of 88.9% (95% CI, 50.7–99.4%) for lung tissue, 55.0% (95% CI, 32.0–76.2%) for bronchoalveolar lavage fluid (BALF), and 50.0% (95% CI, 32.8–67.2%) for serous fluids. The overall sensitivity of mNGS was superior to that of the culture assay (35.2%, 95% CI, 27.0–44.3), but no superior sensitivity for sputum was observed in mNGS compared with the culture assay (mNGS: 52.3%, 95% CI, 31.1–72.6%; culture: 60.9%, 95% CI, 38.8–79.5%). In clinical TB cases, mNGS detected additional positive results (40.6%, 26/64). mNGS reduced the turnaround time from 2–6 weeks to 32–36 hours. CONCLUSIONS: mNGS may be a promising technology for the early auxiliary diagnosis of active TB, especially sputum-negative pulmonary TB (PTB) and tuberculous serous effusion.