Related Experiment Video
Updated: Jun 9, 2026
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
Airway segmentation on CT - A systematic review of machine learning tools
Nada Lsloum1,2, Ahmed Maiter1,3,4, Turki Alnasser1,5
1School of Medicine and Population Health, The University of Sheffield, Sheffield, United Kingdom.
European Journal of Radiology Open
|June 8, 2026
Summary
Machine learning (ML) methods show high accuracy for automated airway segmentation in chest CT scans. However, topological completeness varies, necessitating standardized evaluation and diverse datasets for clinical use.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in healthcare
- Respiratory disease diagnostics
Background:
- Computed tomography (CT) is crucial for respiratory disease management.
- Manual airway segmentation in CT is labor-intensive and error-prone.
- Automated methods are needed to improve efficiency and reproducibility.
Purpose of the Study:
- To systematically review machine learning (ML) methods for fully automated airway segmentation in chest CT.
- To evaluate the performance and quality of existing ML approaches.
Main Methods:
- Systematic literature search of EMBASE, MEDLINE, and CENTRAL databases.
- Inclusion of studies using fully automated ML for airway segmentation on CT with quantitative metrics.
- Quality assessment using the Must AI Criteria-10 (MAIC-10) checklist.
Main Results:
- Thirty-two studies, mostly using deep learning (DL), were included.
- High voxel-wise accuracy (Dice Similarity Coefficient: 83-96%) was reported.
- Topological metrics showed variability (60-95%), with DL outperforming classical ML.
- External validation was performed in 15 studies; MAIC-10 scores were moderate.
Conclusions:
- ML models demonstrate strong accuracy for airway segmentation but vary in topological completeness.
- Standardized evaluation frameworks and diverse datasets are essential for clinical translation.
- Further research is needed to enhance model generalizability and reliability.
