Surface Reconstruction of the Pediatric Larynx via Structure From Motion

Sunday, October 1, 2023

5:54 PM - 6:00 PM East Coast USA Time

Location: 207CD

Presenting Author(s)

Shaunak N. Amin, MD

Resident Physician
University of Washington
Seattle, WA, United States

Co-Author(s)

MB

Michael C. Barbour, PhD

University of Washington, Department of Mechanical Engineering, USA
SF

Seth D. Friedman, PhD

Seattle Children's Hospital, USA
Sanjay R. Parikh, MD

Professor
University of Washington / Seattle Children's Hospital
Seattle, WA, United States
Kaalan E. Johnson, MD

Associate Professor, University of Washington
Seattle Children's Hospital
Seattle, WA, United States

Senior Author(s)

Jake P. Dahl, MD, PhD, MBA

Associate Professor of Pediatric Otolaryngology
University of Washington, Seattle Children's Hospital
Seattle, WA, United States
AA

Alberto Aliseda, PhD

Professor/Chair of Mechanical Engineering
University of Washington
Seattle, WA, United States

Disclosure(s):

Shaunak N. Amin, MD: No relevant relationships to disclose.

Jake P. Dahl, MD, PhD, MBA: No relevant relationships to disclose.

Introduction: Endoscopy is currently the gold standard technique for characterizing many pediatric airway diseases, however this method is limited for quantitative analysis due to lack of three-dimensional (3D) vision, depth perception, and ability to measure airway dimensions. We utilize a novel image processing technique, structure from motion (SfM) photogrammetry, for the first time to reconstruct 3D surface renderings of the pediatric larynx using 3D-printed airway models.

Methods: A high-resolution computed tomography scan of an infant with Grade 3 subglottic stenosis was identified and airway segmentation was performed between the supraglottis and carina using 3D Slicer (Slicer, slicer.org). The anatomic subglottis was identified and was manually manipulated in the anterior-posterior axis to simulate four varying degrees of airway constriction. Models were 3D printed using a digital mixture of 95% Agilus and 5% Verowhite. Simulated operative endoscopies were performed on the printed airway models using 2.7mm 0-degree Hopkins telescopes and video data was exported onto a secure drive. 3D airway surface reconstruction was performed using COLMAP (Schönberger, Zurich, Switzerland), an open-source SfM analysis toolkit.

Results: 3D surface renderings were successfully reconstructed from endoscopic videos of all printed models using the COLMAP pipeline, with accurately reflected changes in subglottic lumen geometry and generated distinct, identifiable anatomic landmarks. Average subglottic surface error between SfM reconstructed surfaces and 3D printed models was 0.275mm as measured by Modified Hausdorff Distance. Average volumetric similarity between SfM surfaces and printed models was 0.880 as measured by Jaccard Index.

Conclusions: SfM is a novel technique which can be used to accurately reconstruct 3D surface renderings of the larynx from endoscopy video. This technique has great potential for use in quantitative analysis of airway geometry and virtual surgical planning.