Measuring laryngotracheal stenosis by extracting centerline based on CT 3D reconstruction

Abstract

Objective: To compare the accuracy of the centerline extracted based on CT 3D reconstruction and conventional CT 3D reconstruction in measuring the length and degree of laryngotracheal stenosis. Methods: A retrospective analysis was performed on 35 patients with laryngotracheal stenosis (including 19 cases without tracheotomy and 16 cases with tracheotomy) treated in the Department of Otorhinolaryngology Head and Neck Surgery of the First Affiliated Hospital of Guangxi Medical University from March 2006 to March 2016, including 20 males and 15 females, whose ages ranged from 1 to 73 years, with a median age of 40.5 years. And CT data of 20 normal subjects were included in the same period, including 10 males and 10 females, whose ages ranged from 20 to 63 years, with a median age of 37.0 years. The continuous cross-sectional area of the airway perpendicular to the centerline was obtained by Mimics software. The area was compared with the discontinuous cross-sectional areas reconstructed by conventional CT 3D reconstruction software advantage workstation, also the length of cervical trachea, the length of stenosis, and the minimum airway area were compared. Multi-factor linear stepwise regression method was used to analyze the factors influencing the measuring difference between the two methods. Three patients with laryngotracheal stenosis were selected, and the measured stenosis length was compared with the surgical specimens to evaluate the accuracy of the two methods. SPSS 26.0 software was used for statistical analysis. Results: In normal people, the areas of thyroid cartilage notch, glottis, inferio thyroid cartilage margin, inferio cricoid cartilage margin, and suprasternal notch planes measured by Mimics centerline method were smaller than those measured by conventional CT 3D reconstruction (t thyroid cartilage notch=4.685, tglottis=3.791, tlower thyroid cartilage margin=5.621, tlower cricoid cartilage margin=6.312, tsuprasternal notch plane=6.436, P<0.05). And the airway length measured by Mimics centerline method from the inferior thyroid cartilage to the superior sternal notch was longer (t=9.79, P<0.001). In laryngotracheal stenosis, in the non-tracheotomy group, the minimum airway area measured by Mimics centerline method was smaller and the stenosis length was longer than those measured by the conventional CT 3D reconstruction, and the difference was statistically significant (tminimum airway area=2.562, tstenosis length=5.240, P<0.05). In the tracheotomy group, the stenosis length measured by Mimics centerline method was longer than that measured by conventional CT 3D reconstruction, and the difference was statistically significant (tstenosis length=2.854, P<0.05). Multi-factor linear regression analysis showed that different CT thickness had a statistically significant effect on the difference in the length of stenosis measured by the two methods (b=-5.370, t=-3.306, P=0.004), and different tracheal forward angle had a statistically significant effect on the difference in the minimum airway area measured by the two methods (b=-0.419, t=-2.208, P=0.04). The difference between the measured length of the Mimics centerline method and the intraoperative specimens was less than 0.5 mm. Conclusion: The centerline extracted based on CT 3D reconstruction can precisely reflect the laryngotracheal morphology and measure laryngotracheal stenosis more accurately.