Sepsis, a life-threatening condition with a high mortality rate, demands rapid diagnosis and treatment. Researchers at Seoul National University in South Korea have developed a novel clinical laboratory test that uses artificial intelligence (AI) to diagnose sepsis and determine antimicrobial susceptibility significantly faster than traditional methods.
The new test, described in a recent publication in Nature, aims to reduce the time required for antimicrobial susceptibility testing (AST), a crucial step in guiding antibiotic treatment. Sepsis affects over 40 million people globally each year, with mortality rates ranging from 20% to 50%, leading to over 10 million deaths annually, according to Sunghoon Kwon, PhD, professor at Seoul National University and senior author of the study.
How the AI-Powered Test Works
The process begins with a blood sample from the patient. Researchers attach peptide-coated magnetic nanoparticles to the sample, which bind to pathogens. Magnets are then used to collect the bacteria, and their DNA is extracted, amplified, and analyzed to identify the microbes present. The pathogens are exposed to antibiotics, and an AI algorithm evaluates their growth patterns to predict the most effective treatments. This entire process can be completed in approximately 12 hours, significantly reducing the 30 to 40 hours required by traditional AST methods.
"The principle is simple," said Kwon in a Nature podcast. "We have a magnetic nanoparticle. The surface of the magnetic nanoparticle we coat in a peptide that can capture the bacteria."
Clinical Performance and Impact
The team evaluated the test's performance on 190 hospital patients with suspected sepsis. The test achieved 100% accuracy in identifying bacterial species and demonstrated 96.2% efficiency in capturing Escherichia coli and 91.5% efficiency in capturing Staphylococcus aureus. These results suggest the potential for significantly faster and more accurate diagnosis and treatment of sepsis.
The current clinical protocols for selecting optimal antibiotic therapy are slow, often taking two to three days to complete due to existing blood culture procedures. The new test bypasses the need for blood cultures, directly analyzing pathogens from the blood sample. The microbial load in patient blood is extremely low, ranging between 1 and 100 colony-forming units (CFU) ml−1, vastly outnumbered by blood cells, according to the study authors.
Addressing Antimicrobial Resistance
Beyond faster diagnosis, the researchers believe their test can help prevent antimicrobial resistance (AMR) by enabling more targeted antibiotic prescriptions. By quickly identifying the specific pathogens and their susceptibility to different antibiotics, clinicians can avoid broad-spectrum antibiotic use, which contributes to the development of resistant bacteria.
While further research, studies, and regulatory approval are needed, the South Korean scientists estimate that the test could be ready for clinical use within two to three years. This innovation represents a significant step forward in the fight against sepsis and the global effort to combat antimicrobial resistance.
