Comprehensive evaluation of metagenomic next‐generation sequencing for pathogen identification in patients with suspected infection: Establishment of a standardized validation workflow

Abstract Metagenomic next‐generation sequencing (mNGS) enables unbiased identification of pathogens directly from clinical specimens. Despite its growing application in the diagnosis of infectious diseases, comprehensive validation of mNGS performance remains limited. This study systematically evaluated the analytical and clinical performance of mNGS through a standardized validation workflow, including analytical evaluation using a reference panel of 12 representative pathogens, in silico simulations involving 108 pathogens, evaluation of diagnostic performance in a retrospective cohort of 499 patients, and assessment of clinical impact in a prospective cohort of 142 patients. Our analyses revealed that mNGS achieved 100% qualitative concordance across both intra‐run and inter‐batch testing. The limit of detection (LoD) ranged from 10.0 to 269.1 copies/mL. No cross‐reactivity was observed between closely related species. Moreover, strong linearity supported its robustness and semiquantitative capability. In silico simulations further confirmed its high performance across a broader range of microorganisms. For the culture‐based comparison, mNGS demonstrated a markedly shorter laboratory turnaround time (28.6 h vs. 136.6 h, p < .001) and a higher positive detection rate (82.7% vs. 47.1%, p < .001) than microbial culture. Using composite clinical diagnosis as the reference standard, mNGS exhibited superior detection sensitivity (78.6% vs. 45.2%, p < .001) and overall accuracy (79.8% vs. 53.7%, p < .001) compared to conventional mictobiological testing (CMT). In the prospective cohort, mNGS contributed to antimicrobial treatment decisions in 101 of 120 infected patients. These findings demonstrate that mNGS may serve as a sensitive adjunct to CMT, particularly when interpreted within a standardized quality‐control and clinical adjudication framework.