AI Sepsis Diagnostic Tool Gets FDA Authorization

— EHR-embedded software algorithm uses biomarkers, clinical data

MedicalToday
FDA APPROVED Sepsis ImmunoScore over a photo of the application on an iPad.

The FDA granted marketing authorization for the Sepsis ImmunoScore, an artificial intelligence (AI) software tool for rapid diagnosis and risk assessment of sepsis, Prenosis announced Wednesday.

Integrated into electronic health records, the AI tool uses a combination of 22 individual biomarkers and patient parameters that provide diagnostic and predictive information about the presence or progression of sepsis within 24 hours of patient assessment in emergency department (ED) or hospital settings. The software generates a risk score and four discrete risk categories, informing clinicians about the patient's risk of deterioration, in-hospital mortality, and escalation of care within 24 hours to ICU admission.

"FDA authorization of a sepsis diagnostic tool with significant predictive power is a landmark event for people that could ever be at risk of sepsis at some point in their lives," Bobby Reddy, Jr., PhD, co-founder and CEO of Prenosis, said in a .

Authorization for the AI diagnostic tool was achieved through the FDA's de novo pathway and supported by a in which researchers prospectively collected samples from 1,400 adult patients in EDs suspected of sepsis to evaluate the machine-learning algorithm.

The AI tool accurately identified sepsis, and a higher score predicted less favorable outcomes regarding discharge time, 30-day mortality, and 30-day inpatient re-admission. Predicted risk scores for patients were higher for those with septic shock and infection and organ dysfunction compared with those with sepsis but without those conditions (P<0.0001 for both).

According to the CDC, 1.7 million people in the U.S. develop sepsis every year and at least 350,000 die from the condition. Of people who die in the hospital, one in three has sepsis. People at highest risk for sepsis are adults over the age of 65 and those with severe recent illness, are immunocompromised, or have chronic medical conditions such as diabetes, cancer, lung disease, and kidney disease.

Although early treatment with antibiotics improves outcomes in sepsis, timely initiation of treatment remains an issue in clinical practice, due to difficulties in recognizing sepsis in patients with vague symptoms and the need for thorough evaluation before beginning treatment with antibiotics.

Prenosis said the Sepsis ImmunoScore was designed using its proprietary Immunix platform, a biobank and dataset consisting of more than 100,000 blood samples from over 25,000 patients.

"Similar to how sequencing technology enabled the precision medicine revolution in cancer, the platform "has the potential to unlock valuable insights to enable the creation of precision therapeutics guided by AI diagnostics," Reddy said. "To date, this type of approach has been used predominantly in healthcare outside of emergency departments and hospitals."

  • author['full_name']

    Katherine Kahn is a staff writer at , covering the infectious diseases beat. She has been a medical writer for over 15 years.