Study
NLST
(Learn more about this study)
Project ID
NLST-1114
Initial CDAS Request Approval
Aug 17, 2023
Title
Development and validation of a novel deep learning method for diagnosing solitary pulmonary nodules
Summary
Artificial intelligence, especially deep learning-based techniques have emerged as promising decision supporting approaches to automatically analyze medical images for different clinical purposes. However, models based on deep learning algorithms may predict the correct answer for the wrong reasons, resulting in excellent performance on the training set but a large performance drop on the external validation sets. In this project, a novel deep learning model will be developed for classifying benign and malignant solitary pulmonary nodules in CT images. The model will follow the radiologist's reasoning process to make predictions. First, the model will identify nodule signs such as spiculation sign, lobulation sign, etc. Then, the nodule signs, the patient's clinical information, and radiomic features are integrated to predict the malignant risk of the nodule. The model will be trained on a private pulmonary nodule dataset and validated on a multicenter dataset. In order to further illustrate that the model has good generalization ability, additional validation on the NLST dataset is required.
Aims
1. Develop a deep learning-based model to distinguish benign and malignant pulmonary nodule in CT images.
2. Evaluating the generalization ability of the model on a mix of private and NLST data.
3. Compare the diagnostic performance of several AI models with state-of-the-art architectures, traditional risk prediction models, and developed deep learning-based model.
Collaborators
Bo Liu - Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China
Haiping Liu - PET/CT Center, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China
Xiaokui Yang - Yau Mathematical Sciences Center, Tsinghua University, Beijing, China
Ruobing Liang - Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China
Hongqiao Dong - Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China
Yipeng Zhao - Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China
