A study published in the New England Journal of Medicine AI (NEJM AI) has validated the Sepsis ImmunoScore™, the first FDA-authorized AI diagnostic tool for sepsis, demonstrating its high accuracy in diagnosing sepsis and predicting key adverse patient outcomes. The study underscores the potential of AI to transform sepsis management by providing clinicians with trustworthy predictive information, enabling a shift from reactive to proactive care.
The peer-reviewed study, titled "FDA-Authorized AI/ML Tool for Sepsis Prediction: Development and Validation," revealed that the Sepsis ImmunoScore™ accurately diagnoses sepsis and simultaneously predicts critical outcomes, including in-hospital mortality, length of stay before discharge, ICU admission, need for mechanical ventilation, and vasopressor use. This marks the first instance of an FDA-authorized tool for sepsis capable of predicting all these critical outcomes concurrently.
Predictive Power of Sepsis ImmunoScore
The Sepsis ImmunoScore™ integrates multidimensional inputs—vital signs, lab data, and sepsis-specific biomarkers—into a comprehensive AI-driven tool. This approach provides more reliable diagnostic information and predicts future patient outcomes more effectively than single biomarkers like Procalcitonin (PCT), which have limitations in predicting future developments.
According to Nathan I. Shapiro, MD, MPH, Professor of Emergency Medicine at Harvard Medical School and corresponding author of the paper, the Sepsis ImmunoScore™ is the first FDA-authorized sepsis AI algorithm that embeds in the EMR and aids providers in identifying and assessing patients with sepsis.
Addressing a Critical Need in Sepsis Management
Sepsis, an overactive immune response to infection, poses a significant burden on the U.S. healthcare system, costing billions annually and claiming at least 350,000 lives each year—more than all cancers combined. The Sepsis ImmunoScore™ aims to address the challenges clinicians face in the early and accurate prediction of adverse outcomes due to this rapidly progressing condition.
Bobby Reddy, Jr., Prenosis Co-Founder and CEO, emphasized that the Sepsis ImmunoScore™ offers critical insights to improve patient outcomes by delivering FDA-validated AI results with transparency and explainability for clinicians.
Development and Validation
The Sepsis ImmunoScore™ received FDA authorization via the De Novo pathway in April 2024. It provides a combination of diagnostic and predictive information previously unavailable in a legally marketed device for sepsis. Unlike early warning systems that often produce high rates of false positives, the Sepsis ImmunoScore™ delivers actionable and trustworthy insights, working in concert with a clinician's suspicion of serious infection to inform decision-making.
The Sepsis ImmunoScore™ was developed using Prenosis' Immunix™ platform, which facilitates precision medicine in acute care. Over a decade of collaboration with more than ten U.S. partner hospitals, Prenosis used Immunix™ to build a proprietary biobank and dataset of over 113,000 blood samples from more than 28,000 patients. This dataset, merging biological data with clinical information from Electronic Medical Records, enables the development of AI algorithms to detect rapid immune responses and recommend personalized therapies in real time in hospitals and emergency departments.
Akhil Bhargava, the study's lead author and Prenosis' Senior Data Scientist, noted that the study confirmed AI, when validated, can be a reliable, comprehensive tool to assist providers with identifying patients at high risk of sepsis and predicting sepsis-related adverse outcomes.
