Antimicrobial resistance is a rigorous health issue around the world. Because of the short turn-around-time and broad pathogen spectrum, culture-independent metagenomic next-generation sequencing (mNGS) is a powerful and highly efficient tool for clinical pathogen detection. The increasing question is whether mNGS is practical in the prediction of antimicrobial susceptibility. This review summarizes the current mNGS-based antimicrobial susceptibility testing technologies. The critical determinants of mNGS-based antibacterial resistance prediction have been comprehensively analyzed, including antimicrobial resistance databases, sequence alignment tools, detection tools for genomic antimicrobial resistance determinants, as well as resistance prediction models. The clinical challenges for mNGS-based antibacterial resistance prediction have also been reviewed and discussed.
Objective To perform rapid antimicrobial susceptibility testing (RAST) on positive blood cultures of Enterobacterales using a total laboratory automation (TLA) system following both Clinical and Laboratory Standards Institute (CLSI) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) standards, and to evaluate the two RAST methods. Methods Positive blood culture bottles growing Enterobacterales [54 Escherichia coli (E. coli) and 60 Klebsiella pneumoniae (K. pneumoniae) isolates] were collected at West China Hospital of Sichuan University between April and August 2022. CLSI RAST (8 and 16 h) and EUCAST RAST (4, 6, and 8 h) were performed using the TLA system, and results were compared with Vitek 2 antimicrobial susceptibility testing. Results CLSI RAST demonstrated lower categorical agreement with Vitek 2 (E. coli: 66.7% at 8 h, 81.9% at 16 h; K. pneumoniae: 72.8% at 8 h, 84.0% at 16 h) and tended to overcall resistance. EUCAST RAST showed increasing zone readability over time and high categorical agreement with Vitek 2 (E. coli: 97.1%, 96.2%, and 96.1% at 4, 6, and 8 h, respectively; K. pneumoniae: 96.1%, 97.1%, and 97.9% at 4, 6, and 8 h, respectively), as well as low error rates (major errors <3%). With the exception of ciprofloxacin, both the readability and categorical agreement for each antimicrobial agent against E. coli at 8 h were favorable (>90%). Similarly, for K. pneumoniae, the readability and categorical agreement for each antimicrobial agent at 8 h also demonstrated good performance (>90%). Conclusions According to CLSI-M52 criteria, EUCAST RAST at 4, 6, and 8 h shows equivalent performance to Vitek 2. Except for ciprofloxacin against E. coli, all antimicrobials tested at 8 h exhibit good readability and categorical agreement (>90%). CLSI RAST has lower categorical agreement and a higher tendency to interpret isolates as resistant.