Development and clinical validation of a multiplex PCR-based targeted next-generation sequencing panel for simultaneous DNA/RNA pathogen detection in respiratory tract infections: a comprehensive performance evaluation
Abstract Objectives Lower respiratory tract infections (LRTI) remain a leading cause of global morbidity and mortality, with approximately 62 % of cases lacking definitive etiological diagnoses via conventional methods. This study aimed to develop and validate a multiplex PCR-based targeted next-generation sequencing (mp-tNGS) panel for comprehensive pathogen detection in respiratory samples. Methods We conducted a combined retrospective (n=209) and prospective (n=256) study involving 465 bronchoalveolar lavage fluid (BALF) samples from patients with suspected LRTI. The mp-tNGS panel was designed to detect 277 clinically relevant pathogens, integrating simultaneous DNA/RNA co-detection and single-nucleotide polymorphism (SNP)-based contamination control. Diagnostic performance was evaluated against conventional microbiological tests (CMTs) and metagenomic next-generation sequencing (mNGS). Results The mp-tNGS assay revealed a limit of detection (LoD) of 25 copies/mL for bacteria/fungi and 50 copies/mL for viruses, with acceptable precision (intrabatch CV=9.42 %; interbatch CV=26.03 %). A novel SNP-based bioinformatics pipeline achieved exceptional discriminative performance (AUC=0.992) for distinguishing true infections from contamination. Compared with CMTs and mNGS, mp-tNGS showed superior sensitivity (93.59 %, 95 % CI: 90.81–95.73 %), particularly for viruses (AUC=0.92 vs. 0.845). The assay reliably identified fastidious pathogens (e.g., Pneumocystis jirovecii and Legionella pneumophila ) and predicted antimicrobial resistance (97.08 % concordance with phenotypic testing). Conclusions We developed an mp-tNGS panel with diagnostic performance comparable to that of mNGS. The panel integrates broad pathogen coverage, rapid turnaround time, machine learning-based false-positive reduction, and concordance with phenotypic susceptibility testing, positioning it as a promising molecular diagnostic tool for LRTI.
Authors
- Dandan Tie
- Juan Long (ORCID: https://orcid.org/0009-0003-8567-2331)
- Li Jiang (ORCID: https://orcid.org/0009-0000-7152-2154)
- Xiaoyan Tao
- Junli Wang
- Jie Zhang
- Shenhui Yang
Institutions
- University of Electronic Science and Technology of China (CN)
- Sichuan Provincial Hospital of Traditional Chinese Medicine (CN)
Publication Details
- Journal
- Clinical Chemistry and Laboratory Medicine (CCLM)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.1515/cclm-2026-0540
- Primary Topic
- Bacterial Identification and Susceptibility Testing
- Type
- article
- Field-Weighted Citation Impact
- 0.00