Advanced AI model predicts risk of CVD death using low-dose CT exams
Researchers have developed an advanced AI algorithm capable of predicting a person’s five-year risk of death from cardiovascular disease (CVD), sharing their findings in Radiology: Cardiothoracic Imaging.
The AI model uses information from low-dose CT exams to make its predictions, helping clinicians gain key information without a significant amount of extra work.
The team’s prediction model works in two stages, using deep learning to identify signs of arterial calcification and then predicting the patient’s risk of five-year mortality based those findings.
To train the model, the authors tracked data from more than 4,400 patients who underwent low-dose CT exams from August 2002 to 2004. Follow-up data for at least five years was available for all participants. The team then tested their model on data from another 1,113 patients from the same cohort.
Overall, the researchers wrote, their model was a success, outperforming other prediction methods based on only patient characteristics.
“The method uses only image information, it is fully automatic, and it is fast,” lead author Bob D. de Vos, PhD, a specialist at Amsterdam University Medical Center and University Medical Center Utrecht, said in a prepared statement. “The method obtains calcium scores in a complete chest CT in less than half a second. This means that the method should be easy to implement in routine patient work ups and screening.”
Read the full analysis from de Vos et al. here.
Michael has more than 18 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.