AI Helps Develop Life-Saving Sepsis Test

A team of Canadian researchers used artificial intelligence to develop a blood test and portable device capable of detecting the onset of sepsis faster and more accurately, thus preventing recovering patients from succumbing to the potentially life-threatening condition.

“This demonstrates the immense value of AI in analyzing extremely complex data to identify the important genes for predicting sepsis and writing an algorithm that predicts sepsis risk with high accuracy,” says the study’s co-author Dr Bob Hancock, University of British Columbia professor of microbiology and immunology and CEO of Sepset.

Sepsis is the body’s extreme reaction to an infection, leading the immune system to start attacking the body’s organs and tissues. Moreover, it is difficult to predict sepsis because the early symptoms are nonspecific, and current diagnostic tests take 18 hours, delaying the treatment. Consequently, the chance of death increases by 8% per hour. The biotechnology company Sepset put blood samples from more than 3,000 hospital patients with suspected sepsis through machine learning to identify a six-gene expression signature that predicts sepsis nine times out of ten, hours ahead of a formal diagnosis. By adding 248 blood samples using Reverse Transcription Polymerase Chain Reaction, or RT-PCR, a common hospital laboratory technique, the test detected early-stage sepsis with 94% accuracy before the condition of patients worsens.