Automated Epicardial Adipose Tissue Analysis Identifies Increased Non-Calcified Plaque Burden in Non-Obese Individuals with Low CAC Scores: An AI-CVD Study within the Miami Heart

Can a routine CAC scan reveal hidden coronary plaque risk—even in non-obese individuals with little or no coronary calcium?

A new HeartLung.AI AI-CVD study, to be presented at ESC Congress 2026, explores whether automated measurement of epicardial adipose tissue (EAT) can help identify individuals with increased non-calcified coronary plaque (NCP) despite having low CAC scores and a non-obese BMI.

The study analyzed 1,259 participants from the Miami Heart Study who underwent paired CAC and coronary CT angiography (CCTA). EAT volume was automatically quantified from CAC scans, while non-calcified plaque burden was measured from CCTA.

Key Findings

Among 758 participants with BMI <29 kg/m², non-calcified plaque burden increased by approximately 93% for each standard-deviation increase in EAT.

The relationship became even more striking among the 663 participants with low CAC scores (0–99):

• Those in the highest EAT quartile had a median non-calcified plaque burden of 34.5 mm³

• Compared with only 11.9 mm³ in the lowest EAT quartile

• Representing approximately a 3-fold higher median non-calcified plaque burden

Importantly, the association between EAT and non-calcified plaque remained after adjustment for age and CAC.

Why It Matters

A low or even zero CAC score does not necessarily mean the absence of coronary atherosclerosis. Some individuals may still carry a meaningful burden of non-calcified plaque that conventional calcium scoring does not capture.

These findings suggest that automatically measuring epicardial fat from the same routine CAC scan could provide an additional layer of risk information—particularly in individuals who might otherwise appear lower risk based on CAC and BMI alone.

Low CAC. Normal-to-overweight BMI. But potentially hidden plaque burden.

Adding automated EAT analysis may help uncover this cardiometabolically adverse phenotype without requiring an additional scan.

📍 To be presented at ESC Congress 2026 · Munich

‌By Hamed Zarei, M.D.