Clinical metagenomics for pathogen detection in lower respiratory infections: a diagnostic study.

BACKGROUND: Metagenomic next-generation sequencing (mNGS) enables comprehensive detection of all potential pathogens in a sample. However, its diagnostic performance in various clinical settings requires further validation through real-world data.

METHODS: This retrospective study included 186 patients from the First Affiliated Hospital of Zhejiang University School of Medicine who underwent mNGS testing on respiratory samples. We compared mNGS with traditional culture method using the same samples and also assessed its performance against conventional microbiological testing combinations (CMTs) for the same patients. Additionally, we analyzed the diagnostic performance of mNGS in different disease states (immunosuppressive status and mechanical ventilation).

RESULTS: The positivity rate of mNGS was 81.2 % (151/186). In lower respiratory tract infections (LRTIs), mNGS successfully detected 84.6 % (137/162) of bacteria, 89.0 % (65/73) of fungi, 100.0 % (72/72) of viruses, 88.9 % (16/18) of Mycobacterium tuberculosis, and 100.0 % (9/9) of non-tuberculous mycobacteria. Mixed infections were the most common infection type in LRTIs in this study (69/145, 47.6 %). The detection rate of mixed infections by mNGS was significantly higher than that of CMTs (91.3 % vs 43.5 %, P < 0.01). In the comparison based on a composite LRTI diagnostic standard, mNGS showed significantly higher sensitivity than paired culture (89.0 % vs 32.4 %, P < 0.01) and CMTs (89.0 % vs 57.2 %, P < 0.01), but lower specificity (46.3 % vs 87.8 % and 46.3 % vs 82.9 %, P < 0.01). In non-mechanically ventilated patients, mNGS maintained high sensitivity (87.0 % vs 96.7 %, P = 0.19), while culture and CMTs showed significant sensitivity decline (P < 0.01).

CONCLUSION: mNGS demonstrates superior diagnostic performance for LRTIs compared to CMTs.