Related Experiment Video
Updated: Dec 25, 2025
10:26
Author Spotlight: A 3D Digital Model for the Diagnosis and Treatment of Pulmonary Nodules
Published on: May 19, 2023
2.4K
Pulmonary nodule detection using hybrid two-stage 3D CNNs
1The School of Mathematical Sciences, Zhejiang University, Hangzhou, Zhejiang, 310058, China.
Medical Physics
|April 3, 2020
Summary
This study introduces an efficient computer-aided detection (CAD) system for pulmonary nodules using 3D convolutional neural networks (CNNs). The novel system achieves high detection sensitivity for lung nodules on CT scans.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Radiology
Background:
- Early detection of pulmonary nodules is crucial for improving patient survival rates.
- Computed tomography (CT) scans are a primary tool for identifying lung abnormalities.
Purpose of the Study:
- To develop a novel and efficient computer-aided detection (CAD) system for pulmonary nodules.
- To leverage 3D convolutional neural networks (CNNs) for enhanced nodule detection in CT scans.
Main Methods:
- A two-stage approach involving nodule candidate detection and false positive reduction using 3D CNNs.
- A segmentation-based 3D CNN with hybrid loss for initial nodule segmentation.
- Classification-based 3D CNNs with hybrid inputs for refining detection and reducing false positives.
Main Results:
- The system achieved a high detection sensitivity of 97.5% on the LIDC-IDRI dataset.
- Demonstrated superior performance compared to state-of-the-art methods with an average of only one false positive per scan.
- Validated on both public and local hospital CT scan datasets.
Conclusions:
- The developed CAD system is highly effective for pulmonary nodule detection.
- The proposed 3D CNN approach shows significant promise for clinical application in lung cancer screening.
