Skip to Main Content

An official website of the United States government

Government Funding Lapse

Because of a lapse in government funding, the information on this website may not be up to date, transactions submitted via the website may not be processed, and the agency may not be able to respond to inquiries until appropriations are enacted. The NIH Clinical Center (the research hospital of NIH) is open. For more details about its operating status, please visit  cc.nih.gov. Updates regarding government operating status and resumption of normal operations can be found at OPM.gov.

Deep learning and other automated techniques for predicting incident cardiovascular death and cancer from lung cancer screening CT

Principal Investigator
Name
Michael Lu
Degrees
M.D., M.P.H.
Institution
The General Hospital Corporation d/b/a Massachusetts General Hospital
Position Title
Co-Director, Cardiovascular Imaging Research Center
Email
About this CDAS Project
Study
NLST (Learn more about this study)
Project ID
NLST-367
Initial CDAS Request Approval
Oct 25, 2017
Title
Deep learning and other automated techniques for predicting incident cardiovascular death and cancer from lung cancer screening CT
Summary
The Centers for Medicare & Medicaid Services (CMS) recently approved lung cancer screening with chest computed tomography (CT), based on evidence from the National Lung Screening Trial (NLST) that screening reduces lung cancer mortality. CMS defines lung cancer screening eligibility as persons aged 55-77 who are current or recent (≤15 years) former smokers with a ≥30 pack-year smoking history. While these eligibility criteria target those at highest risk of lung cancer, smoking and age are also important risk factors for atherosclerotic cardiovascular disease (ASCVD) and other types of cancer. In fact in the Framingham Heart Study we found that ASCVD is more common than lung cancer in the screening-eligible population.

Non ECG-gated lung screening CT permits measurement of coronary artery calcification (CAC) and ectopic fat depots, which we and others have demonstrated are important predictors of cardiovascular death and cancer. CTs also contain information about the heart, blood vessels, and other thoracic and upper abdominal anatomy. While these data are available on every lung screening CT, they are infrequently used in clinical care due to lack of time, equipment, and/or expertise. We propose to develop deep learning and other automated techniques to quantify these parameters and predict cardiovascular death and cancer.
Aims

1) To develop and assess a deep learning tool to predict incident cardiovascular death and cancer from lung screening CT
2) To compare the tool to manual quantification of coronary artery calcium, other anatomy visible on chest CT, and ectopic fat

Collaborators

Hugo Aerts (Harvard)
Parastou Eslami (Harvard)
Borek Foldyna (Harvard)
Udo Hoffmann (Harvard)
Alexander Ivanov (Harvard)
Chintan Parmar (Harvard)
Roman Zeleznik (Harvard)

Related Publications