Rapid Bacterial Identification and Antibiotic Susceptibility Testing through Interferometry-based Surface Topography Measurement

Antimicrobial resistance poses a critical global health threat. For many bacterial infections, such as bacteremia, treatment can fail due to the time it takes to identify appropriate antibiotics. Current antibiotic susceptibility testing (AST) methods require 8-72 hours after isolation of infecting bacterial strains from parient samples, often forcing clinicians to initially prescribe broad-spectrum antibiotics that may be ineffective and contribute to the evolution of antimicrobial resistance. Crucially, AST results are interpreted differently depending on the species of the pathogen; the CLSI breakpoint category must be identified before AST results can be returned to a clinician. Thus, AST also requires bacterial species identification, and the faster these data are available, the faster patients can receive effective treatment. To address these needs, we demonstrate a rapid, comprehensive approach using white-light interferometry to measure bacterial population topography for simultaneous antibiotic susceptibility testing and pathogen identification. This method extracts biophysically relevant features from nanometer-precision measurements of the surface structure of bacterial populations grown on antibiotic-containing agar plates. Using machine learning classification of topographic features, we achieved 95\% accuracy in determining bacterial genus at 4 hr, and correctly determined if a given strain was resistant or susceptible to an antibiotic with 97\% accuracy. We observed that topographic information played a crucial role in the high accuracies. These data reveal that the use of topography to provide species identification and antibiotic susceptibility represents a significant advancement toward rapid, personalized antimicrobial therapy, potentially reducing treatment failures and slowing resistance development.