Abstract Body (Do not enter title and authors here): Background
Coronary CT Angiography (CCTA) is an effective diagnostic test, which can inform prognosis through the severity of stenosis and the number of vessels with disease. Artificial Intelligence Coronary Plaque Analysis (AI-CPA) provides quantitation of coronary plaque volume offering the opportunity to improve risk assessment.
Methods
Symptomatic patients investigated with CCTA from 2 English centers were included. AI-CPA (Heartflow Inc, Mountain View, CA, USA) was performed using thresholds to categorise plaque burden [total plaque volume (TPV, mm³) Group 1: 0, Group 2: 1-100, Group 3: 101-250, Group 4: 251-750, Group 5: >750]. Clinical outcomes were all-cause death, cardiovascular (CV) death, myocardial infarction (MI) and revascularization. Cox-regression, adjusted for degree coronary stenosis, FFRCT, age, gender, hypertension, hypercholesterolaemia and diabetes, determined risk over time.
Results
Of 7879 patients, average age was 57.4±12.3, 51.4% were male, and 2943 (37.4%) had obstructive coronary artery disease. Average TPV was 31mm³ [0-160], with 2539 (32%) patients in the moderate-extensive (stage 3-5) plaque categories. Over 3.3 years (IQR1.4)) of follow up there were 253 (3.2%) deaths, 95 (1.2%) CV deaths, 107 (1.4%) MI and 467 (5.9%) revascularizations. TPV stages 4-5 independently predicted all-cause (p<0.001) and CV death (p=0.007). TPV stage was a significant independent risk predictor for MI, CV death, or late revascularization (HR5.27(2.89-9.59)), p<0.0001), Figure 1.
Conclusion
TPV measured by AI-CPA predicts all-cause death, CV death and a composite of CV death, MI and late revascularization independently of anatomical coronary stenosis, FFRCT and patient risk factors.