结核分枝杆菌靶向下一代测序技术在快速检测、物种鉴别、耐药性分析及预后评估中的临床应用价值：一项多中心回顾性队列研究

Abstract

BackgroundTuberculosis (TB) remains a global health threat, which is complicated by challenges in distinguishing TB from nontuberculous mycobacteria (NTM) infections and drug resistance. Conventional diagnostics (including culture, Xpert MTB/RIF, AFB smear, T-SPOT.TB, etc.) are limited by inadequate differentiation of TB and NTM, incomplete resistance profiling, and prolonged turnaround times.MethodsThis retrospective multicenter study enrolled 130 patients with suspected mycobacterial infection between August 2023 and June 2024 at two tertiary hospitals in China. Patients were classified as TB (n=91), NTM (n=18), Other (n=17) and Mixed infection (n=4) based on a composite reference standard (final clinical diagnosis) established independently of targeted next-generation sequencing (tNGS) results. Performance of tNGS versus conventional methods was evaluated across different sample types. Diagnostic performance, drug-resistance detection, therapeutic impact and clinical outcomes were analyzed.ResultsThe overall agreement of tNGS detection with clinical diagnoses of pathogens was 80.9% in sputum and 71.3% in bronchoalveolar lavage fluid (BALF), exceeding conventional methods (which remained below 55%). Critically, tNGS effectively distinguished TB from NTM, with diagnostic sensitivity of 84.0% (95% CI, 75.0%–90.8%) for TB and 90.5% (95% CI, 69.6%–98.8%) for NTM, and specificity of 100.0% (95% CI, 90.3%–100.0%) for TB and 99.1% (95% CI, 95.0%–100.0%) for NTM. tNGS simultaneously detected resistance-associated variants relevant to 11 antimycobacterial drugs within approximately 24 hours. Treatment regimens for 52 patients (41.3%) were adjusted following tNGS reporting. Of these, 82.7% showed subsequent clinical improvement, however, these outcome data remain descriptive and do not establish causality.ConclusionstNGS enables rapid, accurate, and comprehensive diagnosis of mycobacterial infections, reliably differentiates TB from NTM, and provides drug-resistance profil