Long-term Lung Cancer Risk Prediction Using Sybil AI on Routine Coronary Artery Calcium Scans
- 2 days ago
- 1 min read
ESC Congress 2026 - 28 - 31 Aug 2026
Munich, Germany
Organised by: European Society of Cardiology (ESC) visual
Can a routine cardiac CT predict lung cancer risk years before diagnosis?
A new HeartLung.AI study explores the potential of Sybil AI to predict long-term lung cancer risk using routine coronary artery calcium (CAC) scans.
Using baseline CAC scans from 5,726 participants in the Multi-Ethnic Study of Atherosclerosis (MESA), researchers evaluated incident lung cancer over follow-up extending up to 15 years. Importantly, Sybil analyzed only the baseline CT scan—without smoking history or other clinical risk factors.
The findings showed that:
A single baseline CAC scan contained imaging signals associated with lung cancer risk many years into the future.
Predictive performance remained around 70% AUC through much of the long-term follow-up and was approximately 68% at 15 years.
Even after excluding cancers diagnosed within the first three years, the predictive signal remained around 70%, suggesting the findings were not driven solely by cancers already developing at baseline.
Participants who later developed lung cancer showed consistently higher baseline Sybil scores, including for cancers diagnosed up to 15 years later.
These findings suggest a broader opportunity: routine cardiac CT could potentially become a multidimensional prevention scan, extending its value beyond cardiovascular assessment and helping identify individuals at elevated lung cancer risk who may not be reached by conventional screening pathways.
One cardiac scan. Two major disease pathways. Years of potential insight.
Download the full presentation below to explore the study design, results, and potential implications.
Comments