Hierarchical integration of mNGS, PCR, and other conventional methods for precision TB diagnostics

作者：Yating Zhao, Longting Du, Junli Song, Wei Sun, Yili Chen, Xuegao Yu, Hao Huang, Gang Huang, Enpu Huang, Ni Wang, Shu An, Lu Ai, Peisong Chen · 发表于：Microbiology Spectrum · 年份：2025 · DOI：10.1128/spectrum.01931-25 · 被引用次数：3 · 研究领域：Tuberculosis Research and Epidemiology、Diagnosis and treatment of tuberculosis、Pneumonia and Respiratory Infections

ABSTRACT This study systematically compared the diagnostic accuracy of seven assays for detecting the Mycobacterium tuberculosis complex, including metagenomic next-generation sequencing (mNGS), droplet digital polymerase chain reaction, real-time quantitative polymerase chain reaction, EasyNAT MTC, GeneXpert MTB/RIF, interferon-gamma release assay (IGRA), and acid‒fast staining (AFS). We try to select appropriate combinations of tuberculosis (TB) detection methods for regions with varying levels of medical resources, based on sensitivity, cost-effectiveness, and operational feasibility. A retrospective analysis was conducted on 141 samples collected from patients with suspected active TB at The First Affiliated Hospital of Sun Yat-sen University between April 2022 and April 2024. Among these samples, there were 100 cases assigned to the case group and 41 cases to the control group, based on the tuberculosis diagnostic criteria. Historical data for Xpert, IGRA, and AFS were collected, and parallel experiments using mNGS, droplet digital PCR (ddPCR), real-time quantitative polymerase chain reaction (RT-qPCR), and EasyNAT were conducted on all samples. Diagnostic performance was evaluated by comparing it with the final clinical diagnoses. Sensitivity, specificity, positive predictive value, negative predictive value, and receiver operating characteristic (ROC) curve analysis were conducted, along with DeLong tests for statistical comparison. Compared with the final clinical dia...