Machine learning, AI used to rapidly detect sepsis

Technique was 97% accurate in identifying which sepsis endotypes patient had

Florida Woman Loses Limbs During Pregnancy Due To Sepsis

Sepsis strikes roughly 2 million people each year and is the cause of one in three hospital deaths in the United States, according to the Centers for Disease Control and Prevention.

“Sepsis is the body’s extreme response to an infection and, if not identified and treated quickly, may lead to serious medical consequences and death,” the report states. “Sepsis can occur at any age, but infants, people with chronic conditions, people with weakened immune systems, and older adults are at high risk. In 2019, there were 201,092 deaths in the United States involving sepsis, with three-fourths of those deaths occurring among persons aged 65 and over.”

A groundbreaking advance by researchers in the Hancock Lab and the department of microbiology and immunology at the University of British Columbia might lower those numbers, however.

“This new technique dissects the dysfunctional immune responses involved in sepsis like never before, providing new insights into the biological processes involved in sepsis of any type, including that from COVID-19,” said Arjun Baghela, a graduate student in the Hancock Lab who led the analysis. “People don’t know much about sepsis, but in 2020, the numbers of deaths from life-threatening sepsis are likely much higher than one in five, since pretty much everyone who has died from COVID-19 has actually died from sepsis.”

According to the researchers, it usually takes a day or two before doctors can be sure a patient has sepsis. But for every hour that treatment is delayed, they wrote, the risk of death increases by as much as 7.6%, highlighting the need for rapid detection.

“Typically, a patient arrives in the emergency room feeling profoundly ill, with a bunch of symptoms that are fairly non-specific,” Bob Hancock, a UBC Killam professor in the department of microbiology and immunology, said in a release by the university. “The physician looks at that patient if they have an aggregate of symptoms and says, “This is a patient that might have sepsis,” but only if they have some certainty can they start to treat them immediately. They’re in a bit of a ‘look-and-see’ game for the first 24–48 hours.”

However, using machine learning, or artificial intelligence, the researchers were able to identify sets of genes that predict whether a patient will acquire severe sepsis, and could make sense of the five distinct ways sepsis manifests itself.

The AI technique was 97% accurate in identifying which of the five endotypes of sepsis occurred in each patient. These biomarkers also worked in the ICU, where it was shown that one endotype was particularly deadly, with a mortality rate of 46%.

The researchers said the technology for measuring gene expression is already in hospitals, and the technique can be performed within two hours of admission to the emergency room. Their study was published Monday in EBioMedicine.

For more content like this, sign up for the Pulse newsletter here.