Background: Antimicrobial susceptibility testing (AST) is central to clinical microbiology and antimicrobial stewardship. Conventional AST methods, though standardized, are time-consuming and subject to inter-observer variability. Artificial intelligence (AI) has emerged as a potential tool to enhance standardization and efficiency in AST interpretation.
Objectives: This study evaluated the performance of AI-assisted interpretation of disk diffusion AST compared with conventional manual interpretation in a routine clinical microbiology laboratory.
Methods: A laboratory based retrospective analytical study was conducted on 300 non-duplicate bacterial isolates. AST was performed using Kirby–Bauer disk diffusion following CLSI guidelines. High-resolution images of AST plates were analysed using an AI-assisted image interpretation system. Results were compared with manual readings by experienced microbiologists. Categorical agreement, error rates, and interpretation time were assessed.
Results: Overall categorical agreement between AI-assisted and manual interpretation was 96.2%. Very major errors, major errors, and minor errors were 1.2%, 1.5%, and 3.8%, respectively. AI-assisted interpretation significantly reduced interpretation time while maintaining acceptable accuracy.
Conclusion: AI-assisted AST interpretation can serve as a reliable adjunct to conventional methods, improving efficiency and consistency while preserving the essential role of microbiologist oversight.
Keywords: Artificial intelligence, antimicrobial susceptibility testing, disk diffusion, clinical microbiology, automation, antimicrobial resistance
How to cite this article: Askar S, Artificial Intelligence in Antimicrobial Susceptibility Testing: Transforming Clinical Microbiology from Plates to Prediction. Int J Drug Deliv Technol. 2026;16(4s): 998-1001; DOI: 10.25258/ijddt.16.4s.117
Source of support: Nil
Conflict of interest: None