Can lung airway geometry be used to predict autism? A preliminary machine learning-based study

Asef Islam; Anthony Ronco; Stephen M Becker; Jeremiah Blackburn; Johannes C Schittny; Kyoungmi Kim; Rebecca Stein-Wexler; Anthony S Wexler

doi:10.1002/ar.25332

Can lung airway geometry be used to predict autism? A preliminary machine learning-based study

Anat Rec (Hoboken). 2024 Feb;307(2):457-469. doi: 10.1002/ar.25332. Epub 2023 Sep 28.

Authors

Asef Islam¹, Anthony Ronco², Stephen M Becker³, Jeremiah Blackburn³, Johannes C Schittny^{4

5}, Kyoungmi Kim⁶, Rebecca Stein-Wexler², Anthony S Wexler^{3

7

8

9}

Affiliations

¹ Department of Computer Science, Stanford University, Stanford, California, USA.
² Department of Radiology, University of California, Davis, California, USA.
³ Department of Mechanical and Aerospace Engineering, University of California, Davis, California, USA.
⁴ Institute of Anatomy, University of Bern, Bern, Switzerland.
⁵ Center for Health and the Environment, University of California, Davis, California, USA.
⁶ Department of Public Health Science, University of California, Davis, California, USA.
⁷ Department of Civil and Environmental Engineering, University of California, Davis, California, USA.
⁸ Department of Land, Air and Water Resources, University of California, Davis, California, USA.
⁹ Air Quality Research Center, University of California, Davis, California, USA.

PMID: 37771211
DOI: 10.1002/ar.25332

Abstract

The goal of this study is to assess the feasibility of airway geometry as a biomarker for autism spectrum disorder (ASD). Chest computed tomography images of children with a documented diagnosis of ASD as well as healthy controls were identified retrospectively. Fifty-four scans were obtained for analysis, including 31 ASD cases and 23 controls. A feature selection and classification procedure using principal component analysis and support vector machine achieved a peak cross validation accuracy of nearly 89% using a feature set of eight airway branching angles. Sensitivity was 94%, but specificity was only 78%. The results suggest a measurable difference in airway branching angles between children with ASD and the control population.

Keywords: autism spectrum disorder; biomarker; computed tomography; conducting airway geometry; feature selection; machine learning.

MeSH terms

Autism Spectrum Disorder* / diagnostic imaging
Autistic Disorder*
Child
Humans
Lung / diagnostic imaging
Machine Learning
Retrospective Studies