Exploration of deep learning-based model's lung cancer screening capabilities.
Although chest radiography has proven inferiority in terms of screening accuracy compared to computed tomography, augmenting human readers with deep learning-based computer-aided detection (CAD) model for pulmonary nodules may provide added benefits in screening. This project will first utilize radiographs from the NLST study as a standardized external validation set and compare the detection accuracies between a deep learning-based model and human readers, then the investigators will develop a new lung cancer prediction model involving the detection results.
- Exploration / validation of deep learning-based nodule detection system and comparison with human readers.
- Development of a novel lung cancer prediction model incorporating deep learning-based nodule detection system's results.
Clinical Research Group, Health & Medical Equipment Business, Samsung Electronics