Sepsis, a life-threatening complication of bacterial infection, leads to millions of worldwide deaths requires significant time and resources to diagnose. This disease is associated with very high mortality rates, making early detection crucial for treatment.
Researchers have investigated direct clinical evaluation by using dark field imaging of capillary beds under the tongue of septic and healthy subjects for signatures of microcirculatory dysfunction associated with sepsis. Our published results, in collaboration with Beth Israel Deaconess Medical Center, have shown that machine learning and vision can learn higher-order hierarchical diagnostic and prognostic features for rapid and non-invasive diagnosis of sepsis using these dark field microcirculatory images. A neural network capable of distinguishing between images from non-septic and septic patients with more than 90% accuracy is reported for the first time. This approach can help physicians to rapidly stratify patients, facilitate rational use of antibiotics, and reduce disease burden in hospital emergency rooms.