How can new non-antibiotic therapeutic approaches and AI help us tackle antimicrobial resistance?

Artificial intelligence and non-antibiotic therapeutic approaches could play a role in tackling antimicrobial resistance according to new research.

Dr Darren Ting, BHP Clinician Scientist Fellow from the School of Infection, Inflammation and Immunology and Honorary Consultant Ophthalmologist (at Birmingham and Midland Eye Centre), has recently led an international team of experts to review the state of play in the field of antimicrobial resistance (AMR) and looked specifically at the emerging non-antibiotic therapeutic approaches and the potential role of artificial intelligence in tackling AMR.

This work was produced in collaboration with authors from several world-leading institutes, including University of Cambridge, University of Pennsylvania, University of British Columbia, National Laboratory of Biotechnology (Hungary), Duke-NUS, and Singapore Eye Research Institute.

The review, published in Lancet Microbe this week, highlights the problem of evolving AMR, with 5 million deaths associated with bacterial AMR in 2019. Beyond high mortality rates, AMR also results in high morbidity, prolonged hospital admission and increased health-care costs.

The authors emphasise the need for an multifaceted approach to tackling this global health challenge, including:

• understanding mechanisms and drivers at both individual and population levels,
• AMR surveillance,
• antimicrobial stewardship or judicious use of antibiotics including considered drug choice, dosage and length of prescriptions,
• improved infection control and
• the development of novel non-antibiotic antimicrobial therapeutics, including antimicrobial peptides, bacteriophages, monoclonal antibodies, antibody-antibiotic conjugate, and gene therapy.

Power of AI

Homing in on the potential of artificial intelligence (AI) driven technologies and the availability of big data to address AMR, the authors point to a number of areas that could be enhanced by AI.

Antibiotic susceptibility testing could be sped up by machine learning. Instead of taking 24-28 hours to determine bacterial growth, as traditional testing methods would, AI has the potential to predict antibiotic susceptibility much quicker by analysing whole-genome sequencing data.

Machine learning also offers great promise for large-scale AMR surveillance programmes, particularly in lower and middle-income countries. AI also shows potential in creating models that guide the use of specific, narrow-spectrum antibiotics, reducing the need for second-line treatments and inappropriate antibiotic use when compared to clinician-prescribed treatments.

Furthermore, AI-driven efforts can dramatically accelerate antibiotic discovery and development, reducing the process from years to days.

The review calls for fostering of collaborative and sustainable research efforts to effectively mitigate the threat of AMR and safeguard the future of modern medicine, global health and the economy.