Development and evaluation of deep-learning algorithm for the classification of lung nodules detected on CT-Scan
Principal Investigator
Name
Cecile Dupont
Degrees
Msc
Institution
MD Start
Position Title
Associate
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-683
Initial CDAS Request Approval
Jun 26, 2020
Title
Development and evaluation of deep-learning algorithm for the classification of lung nodules detected on CT-Scan
Summary
Lung cancer is the one of the leading cause of death worldwide. Early detection and early treatment are 2 key factors to improve survival outcomes. ). Observational studies in high-risk populations have shown that spiral computed tomography (CT) screening detects more lung cancers than chest radiography screening with 55–85% of CT-detected lung cancers being at a surgically removable stage (stage I).
Due to the increase use of screening programs and the increase of incidentally detected pulmonary nodules by LD-CT, a high number of nodules prompt further, invasive testing but do not result in a lung cancer diagnosis.
This project will entail developing CNN based algorithms, that can help discriminate malignant nodules from benign or indolent lesions to limit the number of unnecessary diagnostic procedures, associated complications and expenditure.
Due to the increase use of screening programs and the increase of incidentally detected pulmonary nodules by LD-CT, a high number of nodules prompt further, invasive testing but do not result in a lung cancer diagnosis.
This project will entail developing CNN based algorithms, that can help discriminate malignant nodules from benign or indolent lesions to limit the number of unnecessary diagnostic procedures, associated complications and expenditure.
Aims
1. Develop computational tools to automatically extract image features from CT data.
2. Develop methods to classify lesions based on the extracted image features.
3. Evaluate the algorithm’s classification accuracy as well as the prognostic value
Collaborators
Anne Osdoit, MD Start
Lukas Guenther, MD, MD start