Abstract Body (Do not enter title and authors here): Hypothesis/Purpose:
Coronary artery calcium (CAC) can be quantified by artificial intelligence (AI) on routine chest computed tomography (CT) scans and is associated with increased risk of atherosclerotic cardiovascular disease (ASCVD). Our prior work (NOTIFY-1) found that notifying statin-naïve patients free of ASCVD and clinicians of AI-identified incidental CAC (AI-CAC) on remote CTs increased statin initiation.
NOTIFY-PICTURE (Picture of Incidental Calcium To Understand Risk Estimate) tested the hypothesis that real-time notification of AI-CAC detected on CTs would increase statin initiation in patients with and without ASCVD.
Study Design/Methods:
NOTIFY-PICTURE was a pragmatic, randomized controlled trial of individuals with and without ASCVD to either clinician and patient notification of AI-CAC with a personalized image or usual care. Adults who underwent nongated chest CT at Stanford from 1/2024 to 9/2024, not on lipid-lowering therapy, were eligible. Key exclusion criteria included lack of routine outpatient care or acute/life-limiting conditions that precluded appropriateness of the intervention (Figure). CT scans were screened for CAC using an FDA-cleared AI algorithm with radiologist confirmation. Notifications were delivered via the electronic health record (EHR).
Sample Size:
202 individuals.
Population Studied:
Adults ≥18 and <85 years old who underwent non-ECG-gated chest CTs.
Intervention:
EHR notification of clinicians (cardiology or primary care) and patients with personalized images of AI-CAC and guideline recommendations.
Power Calculation:
A sample size of 200 provided 95% power to detect a 20% difference in the rate of statin initiation.
Primary End Point:
Statin initiation at 6 months.
Secondary End Points:
Change in cardiovascular medications, clinical care, and downstream testing.
Outcomes:
Participants had a mean age of 67.9 (SD 9.8), 116 (57.4%) were female, 30 (14.9%) had ASCVD, and 15 (6.0%) preferred a non-English language (Table 1). At 6 months, statins were initiated in 44.0% of the Notification arm vs. 9.8% of the Usual Care arm (p<0.001). In pre-specified subgroups (including baseline ASCVD, sex, age, AI-CAC), only those with baseline ASCVD were less likely to start statin therapy (p-interaction=0.003). There was no difference in antihypertensive therapy or antiplatelet therapy initiation. There was a nonsignificant trend toward increased healthcare utilization, including increased clinical care and cardiovascular testing (Table 2).