Skip to Main Content

An official website of the United States government

Principal Investigator
Name
Charles Hatt
Degrees
PhD
Institution
Imbio LLC
Position Title
Development Scientist
Email
About this CDAS Project
Study
NLST (Learn more about this study)
Project ID
NLST-125
Initial CDAS Request Approval
Mar 11, 2015
Title
Screening for COPD with Lung CT Densitometry
Summary
Over 20 million people are diagnosed with COPD in the United States with significantly more undiagnosed. With the growing number high-risk patients undergoing CT-based lung screening due to the published results of the National Lung Screening Trial (NLST) there is an opportunity to screen the same patients for COPD using the images already being acquired. This project aims to implement a fully automated software tool to quantify the extent of emphysema in lung screening CT scans. By using a fully automated quantification tool, COPD screening could take place with minimal burden to the Radiologist and give reliable and repeatable metrics.

This project is a partnership between the University of Michigan and Imbio LLC. Imbio’s commercial lung analysis software, which just received FDA clearance, also includes a tool to quantify emphysema in inspiration CT scans. It is this component of the software that we will use to analyze the NLST data to determine the potential for COPD screening using emphysema quantification. In addition to having an FDA cleared emphysema quantification tool, Imbio has a proprietary cloud-computing platform capable of high throughput image analysis.

Once the emphysema quantification completed, the amount of emphysema will be correlated with pulmonary function tests to determine if the Imbio software can be used to predict the GOLD status of the patient and modify the risk prediction for future cancer development. Other information such as smoking history and demographic information will be included in the statistical model in order to give a more reliable prediction. Emphysema quantification will also be combined with a novel CT-based nodule characterization tool under development at Imbio to investigate the ability of a combined densitometry and nodule characterization technique to predict lung cancer outcomes. If successful, this tool could be used by a primary care physician to determine if a high-risk patient should be referred to a pulmonologist for further tests.

By screening undiagnosed COPD patients, we hope to improve overall health outcomes and lower costs by reducing the number of undiagnosed COPD patients at risk for exacerbation events without being supervised by a physician.
Aims

Aim 1: Perform CT densitometry on the NLST screening CT exams. Imbio’s automated lung segmentation and densitometry algorithm with be run on all of the screening CT exams, giving detailed information about the distribution of HU values throughout the lung
Aim 2: Create a statistical model that combines CT densitometry and patient demographic and medical history in order to predict GOLD status and cancer development.
Aim 3: Perform lung nodule texture characterization on NLST screening CT exams. NLST identified lung nodules will be segmented and analyzed with Imbio’s nodule texture characterization method.
Aim 4: Create a statistical model that combines the model developed in Aim 2 with the nodule texture analysis information to create a model predicting the stage and outcome of the patient presenting with a lung nodule.

Collaborators

MeiLan Han, MD, University of Michigan
Craig Galban, PhD, University of Michigan
Brian Ross, PhD, University of Michigan
Cynthia Maier, PhD, Imbio LLC
Lauren Keith, PhD, Imbio LLC
Isabelle Searcy, PhD, Imbio LLC
Ryan Chamberlain, PhD, Imbio LLC

Related Publications