Optimization of decision thresholds for Mycobacterium tuberculosis can effectively improve the performance of mNGS in tuberculosis diagnosis.

Background: Pulmonary tuberculosis (TB) diagnosis remains challenging due to limitations in traditional methods. This study aimed to optimize the metagenomic next-generation sequencing (mNGS) threshold for Mycobacterium tuberculosis complex (MTBC) detection and evaluate its efficacy compared to standard diagnostic approaches.

Methods: A total of 264 bronchoalveolar lavage fluid (BALF) samples were collected from patients with suspected pulmonary TB at Yongkang First People's Hospital between January 2022 and June 2023. After excluding patients with incomplete data, 59 clinically confirmed TB patients and 111 with non-tuberculous conditions were enrolled. mNGS data were analyzed to calculate reads per million (RPM) for MTBC, and thresholds of 0.02, 0.05, and 0.10 RPM were evaluated for diagnostic efficacy using clinical diagnosis as the gold standard.

Results: The area under the receiver operating characteristic (ROC) curve (AUC) for mNGS in diagnosing TB at RPM thresholds of ≥0.02, ≥0.05, and ≥0.10 were 0.881, 0.873, and 0.814, respectively. The optimal detection threshold was found at RPM ≥ 0.05. Comparative analysis showed mNGS (AUC = 0.873) outperformed routine culture (0.718), PCR (0.741), and Xpert (0.763). Combining mNGS with these methods improved AUC values to 0.837, 0.868, and 0.873, respectively.

Conclusion: Optimizing the mNGS threshold to ≥0.05 significantly enhances MTBC detection in pulmonary TB. Combining mNGS with traditional methods further improves diagnostic efficacy, suggesting a potential role for mNGS in clinical TB management.