Segmentation of the Upper Airway using Deep learning - nnUNet.
Authors
Affiliations (8)
Affiliations (8)
- Registered Orthodontist. Clinical Assistant Professor, Mike Petric School of Dentistry, University of Alberta. Kaye Edmonton Clinic. 11400 University Ave, 8th floor. Edmonton, AB, Canada. T6G 1Z1. Electronic address: [email protected].
- Edmonton, Alberta, Canada. Electronic address: [email protected].
- Registered Orthodontist. Associate Professor, Université Paris Cité, UFR Odontologie. Service Médecine Bucco-Dentaire, AP-HP, Hôpital Pitié-Salpêtrière, 83 Boulevard de l'Hôpital, 75013 Paris, France. Electronic address: [email protected].
- Maxillo-Facial Surgeon. Professor, Service de Chirurgie maxillo-faciale, Sorbonne Université, AP-HP, Hôpital Pitié Salpêtrière, 83 boulevard de l'Hôpital, Paris 75013, France. Electronic address: [email protected].
- Department of Oral Medicine and Radiology, King George's Medical University, India. Shah Mina Road, Chowk, Lucknow, Uttar Pradesh 226003. Electronic address: [email protected].
- Faculty of Dentistry, University of British Columbia, 2199 Wesbrook Mall, Vancouver, BC V6T 1Z3, Canada. Electronic address: [email protected].
- Mike Petric School of Dentistry, University of Alberta, 11405 87 Ave NW, Edmonton, AB T6G 2V2, Canada. Electronic address: [email protected].
- Department of Radiology and Diagnostic Imaging, University of Alberta. Kaye Edmonton Clinic, 11400 University Ave, Edmonton, AB T6G 1Z1, Canada. Electronic address: [email protected].
Abstract
This study addresses these challenges by developing a robust and efficient deep learning model for automated segmentation using a large and diverse dataset. The objective of this study is to assess the performance of a deep learning-based framework for automated segmentation of the entire upper airway using CBCT and CT datasets. A dataset comprised of 220 multi-source 3D images, including CBCT and CT from institutions in Canada, Chile, and France, covering both adult and pediatric scans, as well as pre- and post-operative scans, was used. A "one-center-out" validation was performed using CBCT scans from an Institution in India. Ground truth was established through manual segmentation using 3D Slicer. Deep learning models were implemented, focusing on a multi-source training approach with nnUNet. The multi-source model achieved a high average Dice score of 0.962. Furthermore, the nnUNet-155 model demonstrated an absolute volume difference of 3.31% compared to manual segmentation, with a prediction time of only 5 minutes per volume. This robust, efficient, and generalizable deep learning model provides a valuable tool for clinicians and researchers, enabling precise and consistent 3D analysis of the upper airway to support clinical decision-making and research across a wide range of patient demographics and imaging modalities. This study introduces a robust and efficient deep learning model for automated 3D upper airway segmentation. This tool is clinically relevant as it provides clinicians with a precise, consistent, and time-saving method for analyzing the upper airway. By overcoming the limitations of manual segmentation-namely, labor intensity and inter-observer variability-this model facilitates more reliable clinical decision-making and research across diverse patient populations and imaging modalities.