Quantitative CT Analysis to Predict Future Lung Cancer
Principal Investigator
Name
TIMOTHY CLOUSER
Degrees
B.S., M.B.A.
Institution
Quantitative Imaging Solutions, LLC
Position Title
Managing Director
Email
About this CDAS Project
Study
NLST
(Learn more about this study)
Project ID
NLST-490
Initial CDAS Request Approval
Mar 29, 2019
Title
Quantitative CT Analysis to Predict Future Lung Cancer
Summary
Project Summary: The goal of this work is to develop standard logistic and machine learning learning models to predict the development of future cancer. Specifically, we will apply established quantitative metrics of the lung (e.g. emphysema, vascular density, etc) and body morphology (pectoralis muscle area, visceral adiposity, etc) at the T0 CT scans in patients who do not yet have a diagnosis of lung cancer and use these metrics to predict which patients will have developed lung cancer at T1 or T2.
Aims
Specific Aim 1: To evaluate lung parenchymal metrics as predictors of the development for subsequent cancer from 15000 CT scans taken at T0.
Specific Aim 2: To evaluate body morphology metrics as predictors of the development for subsequent cancer from 15000 CT scans taken at T0.
Specific Aim 3: To develop combined models using information from lung parenchymal metrics, body morphology, unsupervised machine learning, and clinicoepidemiologic variables to predict which patients at an initial CT screen have a higher risk of future lung cancer.
Collaborators
n/a