A Deep Endotracheal Intubation Confirmation System
Keywords: Airway management, Artificial intelligence, Deep learning, Endotracheal intubation, Carina detection
TL;DR: A system for automatic verification of correct endotracheal intubation positioning
Abstract: Misplacement or dislodgement of an endotracheal tube (ETT) can lead to life-threatening complications. We present a novel AI-based system for real-time ETT confirmation using deep learning to detect the carina in video images. A miniature CMOS sensor at the tip of a stylet captures images during intubation. These are analyzed by a ResNet-based algorithm running on a digital signal processor (DSP). The test in 50 clinical videos using case-one leave validation showed 100\% accuracy in carina detection and correct placement of the ETT.
Submission Number: 53
Loading