Tuesday, November 28 | 3:00 p.m.-3:10 p.m. | T7-SSCA06-1 | Room E353B
A deep-learning algorithm can predict a patient’s risk of atherosclerotic cardiovascular disease (ASCVD) from analysis of a single chest CT image, according to a group of researchers from Harvard Medical School.
Although a patient’s coronary artery calcium score (CAC) from a cardiac CT exam can be used to refine a patient’s risk of ASCVD, presenter Vineet Kalathur Raghu, PhD, and colleagues wanted to explore the potential utility of chest CT studies for this crucial task. To accomplish this goal, they developed CT-CV-Risk, a deep-learning algorithm trained using chest CT exams from the National Lung Screening Trial to predict a patient’s probability of cardiovascular mortality within 12 years.
In testing on a hold-out set of 7,405 individuals from the NLST with no history of type 2 diabetes, myocardial infarction, or stroke, CT-CV-Risk yielded a statistically significant improvement over a baseline regression model for predicting cardiovascular mortality. What’s more, it achieved similar results for predicting fatal myocardial infarction and stroke and also predicted cardiovascular mortality beyond CAC and baseline risk factors, according to the researchers.
“Deep learning can estimate cardiovascular risk from a chest CT image,” the researchers wrote. “This may enable opportunistic risk assessment to guide decisions for primary prevention of cardiovascular disease.”
Delve further into the data by sitting in on this Tuesday talk.