Short summary

Early detection of coronary artery calcification (CAC) scores on CT scans is crucial, but routine screening is not performed due to insufficient studies. In Denmark, over 300,000 chest CT scans are done yearly for non-cardiac reasons, with potential for CAC scoring. Manual scoring requires expertise, but AI advancements can automate this process. The DETECT-AI project aims to develop AI algorithms to quantify CAC in chest CT images and identify patients needing early intervention.

Rationale

Coronary artery disease is a global leading cause of disability and death with a high socioeconomic cost. An early marker of coronary artery disease is presence of coronary artery calcification (CAC) which is measured and quantified as a CAC-score on CT scans. Elevated CAC-score is associated with increased risk of cardiovascular disease. Still, cardiovascular screening including a cardiac CT scan to measure CAC is not routinely performed, as benefit, harm and costs have not been sufficiently studied. Yet, more than 300,000 chest CT-scans are performed yearly in Denmark for non-cardiac indications such as lung disease. Although, it is possible to obtain CAC-score from the majority of these scans, they are rarely reported as manual scoring and interpretation requires expertise and time. Recent advances in artificial intelligence have shown its viability in medical applications, including diagnostic imaging. Artificial intelligence has the potential to automate complex image assessments that previously could only be done by radiologists, making largescale CAC scoring feasible, with higher speed and lower cost. By combining artificial intelligence assessed CAC-scores with data from the Danish Health Registries, we will be able to identify healthy patients with early and unprotected coronary artery disease.

Description of the cohort

Patients who had a recent chest CT scan performed for non-cardiac indication.

Collaborating researchers and departments

Department of Cardiology, Odense University Hospital.

• Professor Axel Diederichsen

Applied AI and Data Science Unit The Mærsk Mc-Kinney Møller Institute

• Professor Victoria Blanes-Vidal