Related Experiment Video
Updated: Sep 16, 2025
07:53
Author Spotlight: Advancing 3D Modeling for Enhanced Diagnosis and Treatment of Pulmonary Nodules in Early-Stage Lung Cancer
Published on: October 13, 2023
1.6K
A Systematic Review of AI Performance in Lung Cancer Detection on CT Thorax
Hao Min Cheo1, Chern Yue Glen Ong2, Yonghan Ting3
1National University Hospital, Singapore 119074, Singapore.
Healthcare (Basel, Switzerland)
|July 12, 2025
Summary
Artificial intelligence (AI) can aid lung cancer screening by detecting and classifying pulmonary nodules on CT scans. AI models show promise in improving sensitivity and accuracy, potentially easing radiologist workload.
Area of Science:
- Radiology
- Artificial Intelligence
- Oncology
Background:
- Lung cancer screening (LCS) increases imaging demands on radiologists.
- Increased workload risks delayed diagnosis and radiologist burnout.
- Artificial intelligence (AI) offers potential solutions for nodule detection and classification.
Purpose of the Study:
- To systematically review AI performance in lung cancer detection using CT scans.
- To compare AI diagnostic capabilities against radiologists' performance.
Main Methods:
- Systematic review of studies published between January 1, 2010, and December 21, 2022.
- Searched databases: Medline, Embase, PubMed, and Cochrane.
- Included studies focused on AI for pulmonary nodule detection and classification on CT.
Main Results:
- AI models exhibited higher sensitivity for nodule detection (86.0-98.1%) than radiologists (68-76%).
- AI showed higher sensitivity (60.58-93.3%), specificity (64-95.93%), and accuracy (64.96-92.46%) in classifying malignancy.
- AI specificity in detection was lower (77.5-87%) compared to radiologists (87-91.7%).
Conclusions:
- AI models can augment CT thorax interpretation for lung cancer screening.
- AI demonstrates potential in maintaining diagnostic accuracy for pulmonary lesions.
- AI may help overcome implementation challenges in lung cancer screening programs.
