Artificial intelligence-based image recognition in bronchoscopy
Software development and randomized controlled trial for training evaluation in intensive care residents
Bibliographic Data
| ID | 22425232 |
|---|---|
| Authors | Beatrice Brunoni (University of Milano-Bicocca), Francesco Zadek (0000-0002-2377-3992, University of Milano-Bicocca), Federica Pampurini (Synlab Czech (Czechia)), Marco Vettorello (Azienda Socio Sanitaria Territoriale Grande Ospedale Metropolitano Niguarda), Francesco Baccoli (University of Milano-Bicocca), Federico Cabitza (0000-0002-4065-3415, Fondazione Bruno Kessler), Roberto Fumagalli (0000-0002-6108-1487, Azienda Socio Sanitaria Territoriale Grande Ospedale Metropolitano Niguarda), Thomas Langer (0000-0002-1599-2052, Azienda Socio Sanitaria Territoriale Grande Ospedale Metropolitano Niguarda, corresponding author) |
| Year | 2026 |
| Volume | 26 |
| Issue | 1 |
| Publication date | 2026-02-24 |
| Peer Reviewed | Yes |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | BMC Medical Education (JOURNAL) |
| Journal identifiers | ISSN: 1472-6920 • E-ISSN: 1472-6920 |
| Publisher | Springer Science and Business Media LLC (PUBLISHER) |
| DOI | 10.1186/s12909-026-08817-4 |
| PMID | 41735925 |
| OpenAlex | W7131210145 |
| Language | EN |
| References cited | 32 |
BACKGROUND: Flexible bronchoscopy is an essential tool for airway management and both diagnostic and therapeutic interventions, particularly in critical care. Accurate identification of tracheobronchial structures is crucial but challenging for less experienced clinicians, often leading to prolonged procedures and increased complication risks. Simulation-based training using virtual reality or manikins has shown promise, and recent studies suggest that artificial intelligence (AI)-based training outperforms self-directed learning. Limited data exist comparing AI-based bronchoscopy training to expert-led instruction. This study aimed to develop and evaluate a custom-made AI-based software for identifying key tracheobronchial structures and assessing its effectiveness as a training tool for anesthesia and intensive care residents. METHODS : An AI-based software using YOLOv8 artificial neural networks was developed to recognize key tracheobronchial structures from bronchoscopy videos of a high-fidelity manikin. In a randomized trial, 22 second-year anesthesia residents with limited bronchoscopy experience were assigned to either AI-based unsupervised training (n=11) or traditional human-led training (n=11). Bronchoscopy skills were assessed using the modified Bronchoscopy Skill and Task Assessment Tool (BSTAT) before and after training. RESULTS : The AI model demonstrated high accuracy, with an average precision-recall AUC of 0.98 and a mean average precision of 0.98. Both groups of residents showed significant improvement in their BSTAT scores (from 30±4 to 53±2, p<0.001) and reduced procedural time (from 217±44 to 101±23 seconds, p<0.001). No significant differences were observed between the AI-based and expert-led training groups. CONCLUSION : We developed an AI-based software capable of real-time guidance during flexible bronchoscopy. The AI-based training demonstrated comparable efficacy to expert-led instruction, suggesting its potential as a viable tool for unsupervised medical training in flexible bronchoscopy.
Airway · Bronchoscopy · Flexible bronchoscopy · Intensive care · Randomized controlled trial · Lung Cancer Diagnosis and Treatment · Surgical Simulation and Training · Tracheal and airway disorders · Software
The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets
Personalized education and Artificial Intelligence in the United States, China, and India
You Only Look Once
Modeling adoption of intelligent agents in medical imaging
“See one, do one, teach one”
| Citation velocity | historical |
|---|---|
| Highly cited | No |