Deep learning models to predict lung cancer malignancy
Principal Investigator
Name
Yufeng Deng
Degrees
Ph.D
Institution
Infervision US Inc.
Position Title
President
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-604
Initial CDAS Request Approval
Nov 25, 2019
Title
Deep learning models to predict lung cancer malignancy
Summary
Lung cancer screening by LDCT is a Medicare-covered procedure to eligible patients. However, it is a difficult task to characterize nodules detected from the screening exams.
The National Lung Screening Trial (NLST) contains LDCT images for patients with lung cancers and with benign nodules. We plan to utilize deep learning, radiomics, and clinical data to create an accurate prediction model to estimate lung nodule malignancy from LDCT images.
The National Lung Screening Trial (NLST) contains LDCT images for patients with lung cancers and with benign nodules. We plan to utilize deep learning, radiomics, and clinical data to create an accurate prediction model to estimate lung nodule malignancy from LDCT images.
Aims
We aim to divide the data into training, validation, and testing set. Each set will contain malignant and benign nodules. We plan to train and validate a deep learning model that combines image features, radiomic features, and clinical features to accurately characterize nodule malignancy. We are aiming for sensitivity and specificity over 90%.
Collaborators
NA