HealthDay News — Information automatically extracted from low-dose lung computed tomography (CT) imaging can predict 5-year cardiovascular disease (CVD) mortality, according to a study published online April 15 in Radiology: Cardiothoracic Imaging.

Bob D. de Vos, PhD, from Amsterdam University Medical Center, and colleagues conducted a retrospective study involving 5564 participants who underwent low-dose CT from the National Lung Screening Trial. Six types of vascular calcification (thoracic aorta calcification, aortic and mitral valve calcification, and coronary artery calcification of the left main, left anterior descending, and right coronary artery) were quantified after training a deep learning network. Prediction of CVD mortality was performed with multivariable logistic regression; the methods were compared to semiautomatic baseline prediction using self-reported participant characteristics.

The researchers trained the prediction model with data from 4451 participants and tested it on data from 1113 participants. Using calcium scores, the prediction model achieved a C statistic of 0.74, which outperformed the baseline model using participant characteristics (C statistic, 0.69). Combining all variables yielded the best results (C statistic, 0.76).


Continue Reading

“We have shown that 5-year CVD mortality can be predicted for lung screening participants in less than half a second, using only site-specific calcium scores automatically derived from lung screening low-dose CT,” the authors write. “The proposed image-based analysis could aid in identification of lung screening participants at risk for CVD mortality, without relying on self-reported participant data.”

Two authors disclosed financial ties to the medical technology industry.

Abstract/Full Text