Automated assessment of cardiovascular disease risk from CTs using deep learning

Principal Investigator

Name
Shazia Akbar

Degrees
Ph.D.

Institution
Altis Labs, Inc.

Position Title
Machine Learning Engineer

Email
shazia@altislabs.com

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-770

Initial CDAS Request Approval
Mar 10, 2021

Title
Automated assessment of cardiovascular disease risk from CTs using deep learning

Summary
Cardiovascular disease (CVD) is the leading cause of death in the general population and among cancer survivors. CVDs are observed in 23% of newly diagnosed lung cancer patients according to a population-based study of the Comprehensive Cancer Centre South. Cardiovascular event risk scores such as the Framingham Risk Score (FRS) are key prognostic factors for cardiovascular disease (CVD) patients. However, heterogeneity among patients may result in variability of outcomes among patients with the same score. Deep learning-based imaging biomarkers offer additional, individualized prognostic insight derived from rich features present in CTs to facilitate more precise clinical decision making and monitoring.

Aims

- develop state-of-the-art classification algorithms for CVD risk assessment
- test algorithms' utility on lung cancer screening CTs to provide additional clinical insight for at-risk patients, which might help to facilitate lifestyle changes and preventative care

Collaborators

Joint Department of Medical Imaging and Quantitative Imaging for Personalized Care Medicine, UHN Research and Development, Altis Labs