STAR-Echo: Software to Predict Cardiovascular Disease in Chronic Kidney Disease Patients

Application

A novel software for prognosis of MACE in chronic kidney disease patients using spatiotemporal analysis and transformer-based radiomics models.

Key Benefits

Unique interpretable features – identifies novel features based on longitudinal changes in LVW shape (perimeter & sphericity) and texture (intensity variations) over a heartbeat cycle.
STAR-Echo model shows promise in stratifying MACE risk CKD patient population.

Market Summary

Chronic kidney disease (CKD) is a condition in which the kidneys are damaged and cannot filter blood properly. This condition results in excess fluid and waste remaining within the body that may cause other health problems, such as heart disease and stroke. The Centers for Disease Control and Prevention, reports that more than 1 in 7 (15%) of the U.S. adults or 37 million people have chronic kidney disease. Furthermore, the global prevalence of CKD has been steadily rising due to factors such as rise in aging population, and the increase in prevalence of diabetes and hypertension. Cardiovascular disease (CVD) is the most common cause of mortality among patients with CKD, and patients with CKD are at higher risk for Major Cardiac Events (MACE). A complex relationship exists between CKD and CVD, with each being an established risk factor for the other. Thus, biomarkers for MACE are needed to improve risk stratification and support clinical decision making in patients with CKD.

Technical Summary

Echocardiography evaluates left ventricle (LV) function and heart abnormalities. LV Wall (LVW) pathophysiology and systolic/diastolic dysfunction are linked to MACE outcomes in CKD patients. Traditional LV volume-based measurements like ejection-fraction offer limited predictive value as they rely only on end-phase frames. Analyzing LVW morphology over time, through spatiotemporal analysis, may predict MACE risk in CKD patients. However, accurately delineating and analyzing LVW at every frame is challenging due to noise, poor resolution, and the need for manual intervention. Researchers at Emory developed an automated pipeline for identifying and standardizing heart-beat cycles and segmenting the LVW. Further, researchers introduced a novel computational biomarker—STAR-Echo—which combines spatiotemporal risk from radiomic (MR) and deep learning (MT) models to predict MACE prognosis in CKD patients. This technology demonstrates a superior prognostic performance for characterizing cardiac dysfunction in CKD patients, and potentially outperforms Echonet based LV volume-based approaches.