Related Experiment Video
Updated: Jun 7, 2026

Non-Intubated Video-Assisted Thoracoscopic Surgery
Published on: May 26, 2023
AI-Assisted Video Monitoring for Tracheostomy-Dependent Infants: A Proof-Of-Concept Study
Colleen F Cecola1, Christine Settoon1, Lauren S Buck1
1Louisiana State University Health Sciences Center, New Orleans, Louisiana, USA.
Insights
Patient-personalized AI video analysis accurately detects tracheostomy tube status and infant distress. This AI system shows promise for real-world clinical use after further evaluation.
Area of Science:
- Artificial Intelligence in Medicine
- Pediatric Otolaryngology
- Medical Imaging Analysis
Background:
- Tracheostomy-dependent infants require continuous monitoring for tube status and distress.
- Current monitoring methods may be limited in detecting subtle changes.
- AI-assisted video analysis offers a novel approach to objective patient assessment.
Purpose of the Study:
- To evaluate the performance of Eyes-On, an AI system for detecting tracheostomy tube status and visual distress in infants.
- To assess the effectiveness of patient-personalized AI calibration for improved accuracy.
- To determine the system's potential for frame-level classification in clinical settings.
Main Methods:
- Retrospective analysis of video recordings from 25 tracheostomy-dependent infants.
- Development of a YOLOv11 detector for cannulation status and a facial distress classifier.
- Training and evaluation of generalized pretrained models, followed by patient-specific calibration.
- Utilizing decision curve analysis for individualized threshold selection.
Main Results:
- Calibrated AI significantly improved cannulation detection accuracy to 0.940 (sensitivity 0.997) and distress classification accuracy to 0.960 (sensitivity 0.974).
- Pretrained models showed moderate performance, with significant gains after patient-specific calibration.
- The system demonstrated high performance in frame-level classification after individualized adjustments.
Conclusions:
- Patient-personalized AI video analysis shows potential for accurate, frame-level detection of tracheostomy status and infant facial distress.
- The study highlights the importance of within-patient calibration for AI system performance.
- Prospective real-world evaluation is necessary to validate clinical utility and alarm capabilities.
Objective:
To retrospectively evaluate Eyes-On, a patient-personalized AI-assisted video analysis system for frame-level detection of tracheostomy tube status and visual distress in tracheostomy-dependent infants.
Methods:
In an IRB-approved study, 25 tracheostomy-dependent infants aged ≤ 2 years underwent video recording during routine tracheostomy care and tube changes. Data were partitioned at the patient level (22 infants for model development; 3 withheld for staged evaluation). From edited clips, 10,000 frames were extracted and annotated by two blinded pediatric otolaryngologists. A YOLOv11 detector was trained to detect cannulation status, and a facial distress classifier was built using facial features and action unit signals. Generalized pretrained models were tested on held-out infants and then reevaluated after patient-specific calibration. Individualized thresholds were selected using decision curve analysis.
Results:
The pretrained cannulation detector achieved accuracy 0.736, sensitivity 0.806, specificity 0.667, mAP@50 0.645, and a 23.62% Not-Detected rate (n = 1200). After calibration, pooled evaluable-frame cannulation performance improved to accuracy 0.940, sensitivity 0.997, and specificity 0.874 (AUROC 0.962; AUPRC 0.918). The pretrained distress classifier achieved accuracy 0.825, sensitivity 0.877, specificity 0.770, and a 21.8% face-extraction failure rate. After calibration, evaluable-frame distress accuracy increased to 0.960 with sensitivity 0.974 and specificity 0.946 (AUROC 0.993; AUPRC 0.993).
Conclusion:
In this retrospective proof-of-concept study, patient-personalized video-based AI showed promising frame-level classification of tracheostomy-status and facial distress. These results reflect calibrated within-patient deployment and do not validate event-level outcomes, alarm thresholds, or standard monitoring modalities. Prospective real-world evaluation is needed before clinical adoption.
Related Concept Videos
Cardiopulmonary Resuscitation V: Advanced Airway Management Techniques
Oxygen Delivering System III: Tracheostomy and T-piece
Tracheostomy
A tracheostomy is a surgically created opening (stoma) in the anterior part of the trachea. It is used to establish a patient airway, bypass an upper airway obstruction, simplify the removal of secretions, permit long-term...
Tracheostomy Care II: Procedure
Step 1: Perform hand hygiene, and put on personal protective equipment: gown, gloves, mask and...
Tracheostomy: Procedure and Tubes
Tracheostomy tubes can be made of semiflexible plastic (polyurethane or silicone), rigid plastic, or metal, and they come in...
Tracheostomy Care I: Pre-procedural Steps
Required Equipment
The equipment necessary for tracheostomy care includes:
Tracheostomy Suctioning II: Procedure

