A precise image-based signature of patient response to statin therapy

Heart attacks are a leading cause of death in Australia, claiming on average 21 lives every day. The most common cause of a heart attack is coronary heart disease, the build-up of fatty plaques in the coronary arteries which supply blood to the heart muscle. CT scans, a type of 3D X-ray, are used in daily practice to assess for plaques in the coronary arteries. Beyond this, CT images contain a wealth of information which is invisible to the human eye and not captured by current standard measurements. Prior work by our team has shown that converting routine digital CT images into large-scale data, an innovative technique known as “radiomics”, identifies plaques at higher risk of causing heart attacks more accurately than standard measurements. Radiomics creates a unique “signature” of plaque structure based on thousands of data points describing 3D shape and texture. It is unknown whether this signature changes in response to usual medical treatment, especially in patients with higher-risk plaques. Statins are widely prescribed for patients with elevated blood cholesterol and can lead to a decrease in the size of plaques and their fat content, as shown using standard CT measurements. Yet, this type of assessment fails to capture complex changes in the underlying plaque structure. The current study aims to establish a new, precise, image-based signature of patient response to statin treatment by applying radiomics to routine CT scans of the coronary arteries. To achieve this, we will utilise a large database of over 32,500 patients with CT scans at MonashHeart, Victoria. Artificial intelligence techniques will be used to build a unique radiomic signature for each patient, enabling new insights into underlying disease pathways that are typically only explored under the microscope. In future, this image-based signature could potentially be used in clinical trials of new drug therapies.