In patients with suspected CAD and a calcium score of zero, AI-QCT detected non-obstructive CAD in >85% of patients, though long-term annual clinical event rates remained very low (0.26%-0.43%).
Cohort (n=606)
No
Does AI-based quantitative analysis of coronary CTA improve the detection of coronary atherosclerosis compared to clinical reading in symptomatic patients with a calcium score of zero?
AI-based quantitative analysis reveals that non-obstructive coronary atherosclerosis is highly prevalent in symptomatic patients with a calcium score of zero, though their long-term clinical prognosis remains excellent.
Abstract Background There is conflicting evidence on the prevalence of obstructive and non-obstructive coronary artery disease (CAD) and outcome in symptomatic patients with coronary calcium score (CACS) of zero. Objectives To characterise coronary atherosclerosis and its sex differences in patients with chest pain and zero CACS. Methods Patients evaluated for suspected CAD at our hospital 2006-2020 with CACS of zero were included. In addition to clinical reading, artificial intelligence-based quantitative analysis (AI-QCT) of coronary computed tomography angiography (CTA) images was performed. Long-term clinical outcome was analysed. Results Out of 2,216 cohort patients with CTA results, 606 had CACS of zero (156 men and 450 women, median age 54 vs. 61). Around 5% of men and 2% of women had obstructive CAD, with comparable findings on clinical CTA and AI-QCT. Non-obstructive disease was found in higher proportion on AI-QCT (88.5% of men and 85.6% of women) than on clinical reading (16.7% of men and 11.6% of women). Non-calcified plaques dominated in both sexes, but small calcifications were not uncommon. Low-density plaques increased with the severity of CAD. After a median follow-up of 7.3 years, only 3 men and 14 women experienced clinical events, translating to an annual event rate of 0.26% in men and 0.43% in women. Conclusions In patients with suspected CAD and CACS of zero obstructive CAD is detected in around 5% of men and 2% of women. Non-obstructive CAD is detected in a vast majority with AI-QCT. Despite this, long-term clinical outcome is generally very good.
Molnar et al. (2026) conducted a cohort in Suspected coronary artery disease with calcium score of zero (n=606). AI-based quantitative analysis (AI-QCT) vs. Clinical reading was evaluated on Clinical events. In patients with suspected CAD and a calcium score of zero, AI-QCT detected non-obstructive CAD in >85% of patients, though long-term annual clinical event rates remained very low (0.26%-0.43%).