AI-Based Prediction of Difficult Airway in Bariatric Surgery

Elazıg Fethi Sekin Sehir Hastanesi340 enrolled

Overview

The aim of this prospective study is to evaluate the accuracy of artificial intelligence (AI) and machine learning algorithms in predicting difficult airways in patients undergoing bariatric surgery. Preoperative airway assessments, including the Upper Lip Bite Test (UBLT), Mallampati score, Body Mass Index (BMI), thyromental distance (TMD), and sternomental distance (SMD), will be recorded. The study investigates whether AI models can provide higher sensitivity and specificity in predicting difficult intubation compared to traditional clinical scoring systems in the obese patient population.

Study Type

OBSERVATIONAL

Enrollment

340

Conditions

Obesity Difficult Airway Airway Management

Interventions

Preoperative Airway Assessment and Direct LaryngoscopyDIAGNOSTIC_TEST

Measurement of preoperative airway parameters including Upper Lip Bite Test (UBLT), Mallampati score, Body Mass Index (BMI), thyromental distance, and sternomental distance. Intraoperative airway view is graded using the Cormack-Lehane classification during standard direct laryngoscopy.

Eligibility

Sex: ALLMin age: 18 YearsMax age: 65 YearsHealthy volunteers:

Medical Language ↔ Plain English

Inclusion Criteria: 1. Adult patients aged 18 to 65 years. 2. Scheduled for elective bariatric surgery under general anesthesia. 3. Body Mass Index (BMI) ≥ 35 kg/m². 4. Consenting to participate in the study. Exclusion Criteria: 1. Patients with known upper airway anatomical deformities, head and neck tumors, or a history of head/neck radiotherapy. 2. History of maxillofacial, airway, or cervical spine surgery. 3. Emergency surgeries. 4. Patients requiring planned awake fiberoptic intubation based on obvious preoperative clinical indicators.

Outcomes

Primary Outcomes

Diagnostic Accuracy of the Artificial Intelligence Model in Predicting Difficult Intubation

The predictive performance of the AI model will be evaluated by comparing its preoperative difficult airway prediction against the actual intraoperative direct laryngoscopy view. The intraoperative view is graded using the Cormack-Lehane classification system. Grades 3 and 4 are clinically defined as difficult intubation, while Grades 1 and 2 are defined as easy intubation. The primary metric of diagnostic accuracy will be the Area Under the Receiver Operating Characteristic (AUC-ROC) curve.

Time frame: Intraoperative (assessed during the primary intubation attempt)

Secondary Outcomes

Number of Intubation Attempts

Total number of direct laryngoscopy attempts required to achieve successful tracheal intubation.

Time frame: Intraoperative

Need for Alternative Airway Management Techniques

The frequency of requiring alternative airway devices or strategies (e.g., video laryngoscope, bougie, or fiberoptic bronchoscope) to secure the airway after a primary direct laryngoscopy.