Researchers from the Oxford Martin Programme on Antimicrobial Resistance Testing at the College of Oxford have unveiled a breakthrough in a rapid diagnostic test capable of delivering results within just 30 days to commemorate World Anti-Microbial Awareness Week.
Utilizing fluorescence microscopy and artificial intelligence (AI), the team identified antimicrobial resistance (AMR) in their research published in Communications Biology. Their approach hinges on training deep-learning models to assess bacterial cellular images and detect potential fundamental alterations in cells following antibiotic exposure. The method exhibited notable efficacy on a per-cell basis across various antibiotics, achieving a minimum accuracy of 80%.
The team envisions the future application of their innovation in determining the resistance of scientific sample cells to a broad spectrum of antibiotics.
The method underwent validation using diverse medical E. coli isolates, each showcasing varying degrees of ciprofloxacin resistance. The deep-learning models proved to be at least ten times faster than established gold standard scientific techniques in effectively identifying antibiotic resistance.
With the goal of enhancing their methodology for swifter and more extensive adoption in clinical settings, the researchers aim to adapt it for different bacteria strains and medications. The University-led Global Research on Antimicrobial Resistance (GRAM) Project reported approximately 1.3 million deaths attributed to AMR in 2019.
The prevalent growth of bacterial colonies in the presence of antibiotics constitutes a pivotal aspect of current screening methods. Traditional tests, which are sluggish and often span several days, are utilized to gauge the resistance of bacteria to various medications.
This delay poses a significant risk when dealing with critical infections like fevers that necessitate immediate attention. Consequently, healthcare providers are frequently compelled to choose between administering a combination of antibiotics effective against multiple bacterial strains or prescribing specific medications based on clinical judgment. Misguided prescriptions may exacerbate infections, leading to the need for additional antibiotic therapies and potentially fostering increased community-wide antibiotic resistance.
The researchers highlight that the expeditious nature of their approach, if further refined, could facilitate targeted antibacterial interventions, thereby curtailing treatment durations, minimizing side effects, and ultimately impeding the escalation of AMR.
Dr. Piers Turner, a postdoctoral researcher affiliated with the University’s Department of Physics and the Oxford Martin Programme on Antimicrobial Resistance Testing, co-authored the study.
The groundbreaking strides achieved by our team in promptly identifying antibiotic resistance are particularly significant in an era where the public health threat posed by antimicrobial resistance looms large. This technology has the potential to enhance the precision and timeliness of infectious disease treatment decisions, potentially saving lives.
Aleksander Zagajewski, a doctoral candidate from the Department of Physics at the university, also contributed to the study.
As our arsenal against bacterial infections faces a ticking clock, we anticipate that our state-of-the-art diagnostics will pave the way for a new era of targeted therapies for the most critically ill patients.
The research was generously supported by experts across various university departments, including the Department of Physics, Nuffield Departments of Medicine, Sir William Dunn School of Pathology, and Women’s Reproductive Health. Additionally, specialists from the Department of Microbiology and Infectious Diseases at the Oxford University Hospitals NHS Foundation Trust were involved in the study.
