Prediction Model of Hypoxemia in Gastrointestinal Endoscopy Based on Facial Photography and Clinical Indicators

Gastrointestinal endoscopy is an effective method for diagnosing gastrointestinal diseases. There is a increasing number of patients undergo gastrointestinal examinations year by year. Drug sedation can improve patient comfort and increase the intraoperative detection rate of digestive disease. At present, the most commonly used medication for gastrointestinal endoscopy sedation is propofol, which has a fast onset and short duration of time, making it very suitable for sedation during outpatient short surgeries. However, propofol can cause intraoperative hypoxemia. The hypoxemia is mainly caused by the respiratory inhibition of propofol and upper airway obstruction after anesthesia. Hypoxemia caused by respiratory inhibition can be optimized by adjusting the dose regimen. Hypoxemia caused by upper airway collapse is more urgent and more difficult to deal with, and usually requires suspension of gastroscopy and uses mask to ventilation , even needs undergo tracheal intubation. If patients suffer from hypoxemia for a long time, it can lead to myocardial ischemia, arrhythmia, permanent nerve injury, and even death and other serious complications. Conventional airway assessment methods have poor prediction ability for airway abnormalities, and imaging refined measurement indicators have good prediction ability, but their implementation is difficult and difficult to promote. At present, there is still a lack of accurate and simple prediction model for hypoxemia. It is still difficult to make a complete early warning for intraoperative respiratory complications, and intraoperative hypoxemia still occurs frequently. Therefore, in order to ensure the safety of gastrointestinal endoscopy sedation and make emergency plans in advance, a more complete diagnostic model is urgently needed to help assess the risk of hypoxemia in patients undergoing painless gastrointestinal endoscopy surgery and make early interventions.

Based on the above, the investigators assume that facial photography can provide a comprehensive measurement of risk factors for craniofacial bone and soft tissue (obesity). A prediction model constructed by combining facial photography measurement indicators with other relevant indicators can easily and efficiently predict airway abnormalities. Therefore, this study mainly explores the prediction model of craniofacial phenotype based on facial photos and combined clinical indicators for hypoxemia after gastrointestinal anesthesia.