Lung cancer investigation based on deep learning; detection, diagnosis and prognosis prediction
Principal Investigator
Name
Chin A Yi
Degrees
M.D., Ph.D.
Institution
Sungkyunkwan University School of Medicine
Position Title
Research Professor
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-366
Initial CDAS Request Approval
Nov 14, 2017
Title
Lung cancer investigation based on deep learning; detection, diagnosis and prognosis prediction
Summary
For machine learning, accurate dataset is required in terms of training, test and validation. NLST dataset is composed of big, confident demographic information and CT images. Based on NLST datasets of spiral CT comparison read abnormalities, lung cancer, lung cancer progression, we set aims to develop models of lung segmentation, lung nodule detection, lung cancer diagnosis prediction and prognosis prediction. CT images would be input datum for segmentation by U-Net architecture, the convolutional networks for biomedical image segmentation. After module design and analysis for nodule detection and cancer prediction based on convolutional neural network, we will include lung cancer progression and cause of death data dictionary for the prognosis prediction.
Aims
1.Lung segmentation by U-net based model
2.Lung nodule detection and cancer diagnosis prediction by CNN-based model
3.Prognosis prediction of lung cancer by CNN-based model
Collaborators
Jae-Hun KIm, Samsung medical center
Ehwa Yang, Samsung medical center