Coronary Artery Calcium Detected by Ungated CT and CV Outcomes
- Incidental coronary artery calcium (CAC) ≥100 quantified on non–gated CT using a deep-learning (DL) algorithm was associated with worse CVD and mortality outcomes, beyond traditional risk factors and the pooled cohort equation.
- Patients with DL-CAC ≥100 had over two times increased risk of mortality compared to those with DL-CAC = 0.
- These data suggest that an automated DL algorithm used with non–ECG-gated scans can quantify incidental CAC, leading to potential identification of patients at higher risk for CV events.
Does incidental coronary artery calcium (CAC) quantified on routine nonelectrocardiography (ECG)-gated computed tomography (CT) using a deep-learning (DL) algorithm provide cardiovascular (CV) risk stratification beyond traditional risk prediction methods?
Electronic health record data from the Stanford Research Repository Clinical Data Warehouse and CT imaging from the Stanford Picture Archiving and Communication Systems were used for the present analysis. These data were limited to adult patients within the Stanford Healthcare System (includes Stanford Hospital and Stanford Health Care Tri-Valley locations) who had prior and post-CT clinical encounters within the health system. Incidental CAC was quantified using a DL algorithm (DL-CAC) on non–ECG-gated chest CTs performed for routine care in all settings at a large academic medical center from 2014–2019. DL-CAC (0, 1-99, or ≥100) was examined in association with the primary outcome of all-cause death. The two secondary composite outcomes included 1) death, myocardial infarction (MI), and stroke, and 2) death, MI, stroke, and revascularization. Models were adjusted for age, sex, race, ethnicity, comorbidities, systolic blood pressure, lipid levels, smoking status, and antihypertensive use. Ten-year atherosclerotic cardiovascular disease (ASCVD) risk was calculated using the pooled cohort equations (PCEs).
Of 8,040 individuals who underwent non–ECG-gated chest CT imaging between January 1, 2014 and December 31, 2019, 5,678 individuals met inclusion criteria and were free of ASCVD at baseline. Average follow-up was 4.8 ± 2.7 years. The study population was 50.7% women, 18.1% Asian, and 13.0% Hispanic/Latinx, with a mean age of 60.5 ± 16.2 years. Over half (52%) had DL-CAC >0 and 1,899 (33.4%) had DL-CAC ≥100. Those with DL-CAC ≥100 were older, more likely to be male (60.7%), had a higher estimated 10-year ASCVD risk using the PCEs (23.8 ± 15.5%), and were more likely to be on antihypertensive agents (48.1%), statins (26.1%), and aspirin (15.6%), than those with lower DL-CAC. After adjustment, patients with DL-CAC ≥100 had increased risk of death (hazard ratio [HR], 1.51; 95% confidence interval [CI], 1.28-1.79). A similar increased risk was noted for the observed composite endpoint of death, MI, and stroke (HR, 1.57; 95% CI, 1.33-1.84), and the composite endpoint of death, MI, stroke, or revascularization (HR, 1.69; 95% CI, 1.45-1.98) compared to those with a DL-CAC = 0.
The authors concluded that incidental CAC ≥100 was associated with an increased risk of all-cause death and adverse CV outcomes, beyond traditional risk factors. DL-CAC from routine non–ECG-gated CTs identifies patients at increased CV risk and holds promise as a tool for opportunistic screening to facilitate earlier intervention.
This study suggests that use of non–gated CT scans be used to provide clinically relevant information on CV risk, leading to identification of patients at higher risk for clinical events even after adjustment for traditional risk factors. Confirmation using other data would be an important next step.
Keywords: Atherosclerosis, Diagnostic Imaging, Primary Prevention, Tomography, X-Ray Computed, Vascular Diseases
< Back to Listings