Reinforcement Learned Automatic Anesthesia System During Painless Gastrointestinal Endoscopy

Automated Anesthesia System-Gastrointestinal Endoscopy (AAS-GE): a Reinforcement Learning Based Control for Anesthesia During Gastrointestinal Endoscopy

The goal of this clinical trial is to compare the performance difference between trained model and clinical professionals in anesthesia control during gastrointestinal endoscopy. The participants would include adults aged between 18-65 years old with ASA scores I-II. The main questions it aims to answer are:

• Can AI model successfully and safely perform anesthesia during gastrointestinal endoscopy?
• Researchers hypothesis that anesthesia performed by professionals is not inferior to those performed by AI model.

The researchers will compare the results of the AI group with the traditional group to see if both groups have similar success rates in performing anesthesia.

Participants will fast for 8 hours before the procedure, and drinking water will be prohibited for 2 hours after taking the intestinal purgatives.

Inclusion Criteria:

• American Society of Anesthesiologists (ASA) physical status class I-II
• Expected operation duration less than 1 hour

Exclusion Criteria:

• History of cardiac, hepatic, renal, or metabolic disorders (e.g., diabetes, hypertension, sleep apnea)
• Known allergy to opioids or ciprofloxacin components
• Cachexia
• Hypothermia; history of narcotic abuse; chronic alcoholism; obesity (BMI >30); patients with apnea
• Current pregnancy or lactation
• History of substance abuse or alcohol dependence
• Participation in any other drug clinical trial within the past 3 months
• Obesity (BMI >30 kg/m²)
• Procedure termination due to gastric retention or inadequate bowel preparation