Enhancing lower respiratory tract infection diagnosis: implementation and clinical assessment of multiplex PCR-based and hybrid capture-based targeted next-generation sequencing.

Background: Shotgun metagenomic next-generation sequencing (mNGS) is widely used to detect pathogens in bronchoalveolar lavage fluid (BALF). However, mNGS is complex and expensive. This study explored the feasibility of targeted next-generation sequencing (tNGS) in distinguishing lower respiratory tract infections in clinical practice.

Methods: We used 229 retrospective BALF samples to establish thresholds and diagnostic values in a prospective cohort of 251 patients. After target pathogen selection, primer and probe design, optimization experiments, and bioinformatics analysis, multiplex PCR-based tNGS (mp-tNGS) and hybrid capture-based tNGS (hc-tNGS), targeting 198 and 3060 pathogens (DNA and RNA co-detection workflow) were established and performed.

Findings: mp-tNGS and hc-tNGS took 10.3 and 16 h, respectively, with low sequencing data sizes of 0.1 M and 1 M reads, and test costs reduced to a quarter and half of mNGS. The LoDs of mp-tNGS and hc-tNGS were 50-450 CFU/mL. mp-tNGS and hc-tNGS were highly accurate, with 86.5% and 87.3% (vs. 85.5% for mNGS) sensitivities and 90.0% and 88.0% (vs. 92.1% for mNGS) specificities. tNGS detection rates for casual pathogens were 84.3% and 89.5% (vs. 88.5% for mNGS), significantly higher than conventional microbiological tests (P < 0.001). In seven samples, tNGS detected Pneumocystis jirovecii, a fungus not detected by mNGS. Whereas mNGS detected six samples with filamentous fungi (Rhizopus oryzae, Aureobasidium pullulans, Aspergillus niger complex, etc.) which missed by tNGS. The anaerobic bacteria as pathogen in eight samples was failed to detect by mp-tNGS.

Interpretation: tNGS may offer a new, broad-spectrum, rapid, accurate and cost-effective approach to diagnosing respiratory infections.

Funding: National Natural Science Foundation of China (81625014 and 82202535).