A machine learning and centrifugal microfluidics platform for bedside prediction of sepsis
Publication Date
Publication Journal
Author(s)
Lidija Malic, Peter G. Y. Zhang, Pamela J. Plant, Liviu Clime, Christina Nassif, Dillon Da Fonte, Evan E. Haney, Byeong-Ui Moon, Victor Mun-Sing Sit, Daniel Brassard,MaxenceMounier, Eryn Churcher,James T. Tsoporis, Reza Falsafi, Manjeet Bains, Andrew Baker,Uriel Trahtemberg, Ljuboje Lukic, John C. Marshall, Matthias Geissler, Robert E. W. Hancock, Teodor Veres, Claudia C. dos Santos
Sepsis, a dysfunctional immune response to infection that results in multi-organ failure, is a leading cause of death, with mortality rates that skyrocket alongside delay in diagnosis and treatment. Using machine learning to analyze blood samples from suspected sepsis patients, the authors identified a six-gene expression signature that reflects the immune cell reprogramming characteristic of the disease, providing a biomarker of early-stage sepsis. The team also developed a portable blood testing device for this "Sepset" signature, enabling faster detection of sepsis at the bedside.