Cambridge researchers have shown that artificial intelligence (AI) can be used to identify drug-resistant infections, significantly reducing the time taken to make the correct diagnosis. The team showed that an algorithm could be trained to correctly identify drug-resistant bacteria from microscopy images alone.
Antimicrobial resistance is a growing global health problem meaning that many infections are becoming difficult to treat, with fewer treatment options available. This also increases the possibility of some infections becoming untreatable in the near future.
One of the challenges facing healthcare workers is the ability to rapidly differentiate between organisms that can be treated with first-line drugs and those that are resistant to treatment. Traditional testing can take several days, requiring the bacteria to be cultured, tested against various antimicrobial treatments, and analyzed by a laboratory technician or machine. This delay often results in patients being treated with inappropriate medications, which can lead to more serious outcomes and, potentially, increase drug resistance.
In research published in nature communicationA team led by researchers from the lab of Professor Stephen Baker at the University of Cambridge developed a machine-learning tool capable of identifying them from microscopy images. Salmonella Typhimurium bacteria that are resistant to the first-line antibiotic ciprofloxacin â even without testing the bacteria against the drug.
S. Typhimurium causes gastrointestinal illness and typhoid-like illness in severe cases, whose symptoms include fever, fatigue, headache, nausea, abdominal pain, and constipation or diarrhea. In severe cases, it can be life threatening. While infections can be treated with antibiotics, bacteria are becoming increasingly resistant to many antibiotics, making treatment more complex.
The team used high-resolution microscopy to examine S. Typhimurium isolates exposed to increasing concentrations of ciprofloxacin and identified the five most important imaging characteristics to differentiate between resistant and susceptible isolates.
They then trained and tested a machine-learning algorithm to recognize these features using imaging data from 16 samples.
The algorithm was able to correctly predict in each case whether the bacteria were sensitive or resistant to ciprofloxacin without the need for exposure to the drug. This was the case when the isolates were cultured for only six hours, compared to the usual 24 hours for culturing a sample in the presence of antibiotics.
Dr Tuan-Anh Tran, who worked on this research as a PhD student at the University of Oxford and now works at the University of Cambridge, said: âS. Typhimurium bacteria that are resistant to ciprofloxacin have several notable differences from bacteria that are still sensitive to the antibiotic. Although an expert human operator may be able to identify some of these, they alone will not be sufficient to confidently differentiate resistant and susceptible bacteria.
“The beauty of the machine learning model is that it can identify resistant bacteria based on certain subtle features on microscopy images that human eyes cannot detect.”
To analyze a sample using this approach, it will still be necessary to isolate the bacteria from the sample â for example a blood, urine or stool sample. However, because bacteria do not need to be tested against ciprofloxacin, it means the entire process can be reduced from several days to a few hours.
Although there are limits to how practical and cost-effective this particular approach will be, the team says it theoretically shows how powerful artificial intelligence can be in the fight against antimicrobial resistance.
Dr Sushmita Sridhar, who started the project while she was a PhD student in the Department of Medicine at the University of Cambridge and is now a postdoc at the University of New Mexico and Harvard School of Public Health, said: “Given that this approach resolution imaging, this is not yet a solution that can be easily deployed everywhere. But it shows real promise in that by capturing only a few parameters about the shape and structure of bacteria, it can allow us to Can provide sufficient information to predict drug resistance.
The team now aims to work on larger collections of bacteria to create a more robust experimental set that could speed up the identification process even further and allow them to identify resistance to ciprofloxacin and other antibiotics in different species of bacteria. Is.
Sridhar said: âWhat will be really important, especially for the clinical context, will be to be able to take a complex sample â for example blood or urine or sputum â and identify susceptibility and resistance directly from that . This is a much more complex problem and has not really been solved even in clinical diagnosis in any hospital. If we could find a way to do this, we could reduce the time it takes to identify drug resistance and at a much lower cost. This could be truly transformative.â
Source: Cambridge University