Related Experiment Video
Updated: Jun 25, 2026
10:26
A 3D Digital Model for the Diagnosis and Treatment of Pulmonary Nodules
Published on: May 19, 2023
Benchmarking of AI and Radiologists for Indeterminate Lung Nodule Malignancy Risk Estimation on Screening CT: The
Dré Peeters1, Bogdan Obreja1, Noa Antonissen1
1Diagnostic Image Analysis Group, Department of Medical Imaging, Radboudumc, Radboud University Medical Center, Route 767, Room 2.30, Geert Grooteplein Zuid 10, 6525 GA Nijmegen, the Netherlands.
Radiology. Artificial Intelligence
|June 24, 2026
Summary
An artificial intelligence (AI) system outperformed radiologists in assessing lung nodule malignancy risk on low-dose CT scans. The AI demonstrated superior accuracy in identifying cancerous nodules and reducing false positives.
Area of Science:
- Radiology
- Artificial Intelligence
- Medical Imaging
Background:
- Lung nodules detected on low-dose CT (LDCT) require accurate malignancy risk assessment.
- Indeterminate-sized nodules (5-15 mm) pose a diagnostic challenge.
- Standardized evaluation frameworks are crucial for comparing diagnostic tools.
Purpose of the Study:
- To compare the performance of an AI system against radiologists in estimating malignancy risk of indeterminate lung nodules on LDCT.
- To evaluate diagnostic accuracy within a transparent framework using external datasets.
Main Methods:
- AI systems were developed using a public dataset from the National Lung Screening Trial (NLST).
- The best-performing AI system was selected based on AUC from external European lung cancer screening trial data.
- Radiologists assessed nodules, and their performance was compared to the AI system using AUC.
Main Results:
- The selected AI system showed superior performance (AUC, 0.78) compared to the average radiologist (AUC, 0.70).
- The AI system correctly classified 12% more malignant nodules at matched specificity.
- The AI system yielded 20% fewer false positives at matched sensitivity.
Conclusions:
- The AI system demonstrated superiority over radiologists in malignancy risk estimation for indeterminate lung nodules on LDCT.
- AI holds significant potential to improve the accuracy and efficiency of lung nodule diagnosis.
