Summary: Sepsis requires fast and accurate diagnosis, and a new AI-based antimicrobial susceptibility testing method promises rapid and precise predictions of antimicrobial resistance directly from patient blood samples.
Takeaways:
- Sepsis accounts for 1.7 million hospitalizations and 350,000 deaths annually in the U.S., with mortality risk increasing by up to 8% every hour without effective treatment.
- The Keynome gAST system uses machine learning algorithms to analyze bacterial whole genomes from blood samples, bypassing the need for culture growth and enabling quick, accurate predictions of antimicrobial resistance.
- This AI-based method could revolutionize sepsis diagnosis and treatment, reducing hospital stays and saving lives, though further studies are needed to validate the findings across larger sample sizes.
Sepsis is a life-threatening infection complication and accounts for 1.7 million hospitalizations and 350,000 deaths annually in the U.S. Fast and accurate diagnosis is critical, as mortality risk increases up to 8% every hour without effective treatment. However, the current diagnostic standard is reliant on culture growth, which typically takes 2-3 days. Doctors may choose to administer broad-spectrum antibiotics until more information is available for an accurate diagnosis, but these can have limited efficacy and potential toxicity to the patient.
AI-Based Antimicrobial Susceptibility Testing
In a study presented at ASM Microbe, a team from Day Zero Diagnostics unveiled a novel approach to antimicrobial susceptibility testing using artificial intelligence (AI). Their system, Keynome gAST, or genomic Antimicrobial Susceptibility Test, bypasses the need for culture growth by analyzing bacterial whole genomes extracted directly from patient blood samples. The interim findings are based on studies that collected samples from 4 Boston-area hospitals.
Unlike traditional methods that rely on known resistance genes, the machine learning algorithms autonomously identify drivers of resistance and susceptibility based on data from a continuously growing large-scale database of more than 75,000 bacterial genomes and 800,000 susceptibility test results (48,000 bacterial genomes and 450,000 susceptibility test results at the time of this study). This allows for rapid and accurate predictions of antimicrobial resistance, revolutionizing sepsis diagnosis and treatment.
Further reading: Nasal Microbiota is a Potential Sepsis Diagnostic Biomarker
“The result is a first-of-its-kind demonstration of comprehensive and high-accuracy antimicrobial susceptibility and resistance predictions on direct-from-blood clinical samples,” says Jason Wittenbach, PhD, director of Data Science at Day Zero Diagnostics and lead author on the study. “This represents a critical demonstration of the feasibility of rapid machine learning-based diagnostics for antimicrobial resistance that could revolutionize treatment, reduce hospital stays and save lives.”
The researchers say that further study is needed, given the limited sample size, but the findings could contribute to significant advancements in patient outcomes amid the rising threat of antimicrobial resistance and the need for rapid diagnosis and treatment of sepsis.
Funding for this research was provided in part by the Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator (CARB-X).
