Key points are not available for this paper at this time.
Diagnosing antimicrobial resistance (AMR) remains critical for improving patient survival rates and treatment outcomes. Current antibiotic susceptibility tests (AST) suffer prolonged turnaround times, necessitating a minimum of 24 h for results. Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy emerges as a promising phenotypic testing method in bacteriology due to its rapid chemical characterization capability. Here, we present an innovative approach utilizing ATR-FTIR spectroscopy for rapid AMR assessment, distinguishing between methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA). Our approach focuses on detecting early markers of effective antibiotic action and using these to predict resistance profiles. To identify the earliest time for detection, five MSSA and five MRSA strains were subjected to oxacillin exposure for up to 2 h. We observed discernible molecular changes arising in MSSA as early as 1 h after exposure to oxacillin, which were absent in MRSA strains. Bands at 1624 and 1515 cm–1 were identified as markers of positive drug response in MSSA using principal component analysis (PCA) and were associated with peptidoglycan precursor accumulation upon transpeptidation inhibition. To develop predictive models for determining resistance profiles, we implemented ML-based modeling of the spectral data, reflective of the oxacillin-induced chemical composition changes in MSSA and MRSA. Partial least squares discriminant analysis (PLS-DA) and support vector machines classification (SVM-C) algorithms produced the best results, achieving 100% consistency with minimum inhibitory concentration (MIC) classification. Our models were independently validated by blind testing with 35 clinical strains and demonstrated 100% agreement with resistance profiling determined by MIC. Our study underscores the potential of ATR-FTIR spectroscopy for rapid and accurate AMR assessment, with the capacity to revolutionize diagnostics in combating antibiotic resistance.
Kochan et al. (Mon,) studied this question.