AI enables tracking of coronary artery calcium growth on routine lung cancer CT scans
Artificial intelligence can improve the longitudinal assessment of coronary artery calcification (CAC) growth on routine lung cancer screening CT scans.
CAC is a common incidental finding in lung cancer screening exams, regularly appearing in 12% or more of patients. Typically, it occurs alongside advanced atherosclerosis, which can be indicative of risk of major cardiovascular events in the future.
In patients who complete annual lung cancer CTs, providers can observe longitudinal changes in CAC growth. However, there is ongoing debate pertaining to how this growth relates to the risk of adverse events, authors of a new paper in Clinical Imaging note.
“CAC growth compared to the baseline CAC status is controversial and insufficiently explored. Although CAC growth is often considered an inevitable aspect of cardiovascular aging, its implications remain insufficiently studied,” Won Gi Jeong, MD, with the department of radiology at the Chonnam National University Hwasun Hospital, in the Republic of Korea, and colleagues explain. “Though studies focus on the predictive role of CAC severity in baseline or initial CT scans, few studies report its association between CAC growth and atherosclerotic cardiovascular disease.”
The use of low-dose CT lung cancer screenings is expected to continue to grow. With this growth provider will have new opportunities to identify patients with concerning changes in CAC growth, the authors suggest. Though the absence of ECG-gating in LDCT exams makes calculating CAC scores difficult, the group believes the emergence of AI can help.
“Recently, artificial intelligence software has been developed to calculate CAC scores using the Agatston method on non-ECG-gated chest CT scans,” the authors note. “This AI software accurately computes CAC scores from non-ECG-gated CT scans, demonstrating a high degree of concordance compared to that of ECG-gated calcium-scoring CT.”
The authors hypothesized that by using AI to track CAC growth, they could identify patients who may be at greater risk of adverse events. To test their theory, they applied the algorithm to the serial scans of 163 patients who underwent LDCT between April 2017 and December 2023. The AI software was used to quantify CAC on each scan and measure the growth progression between exams, while researchers analyzed the patients’ charts for cardiac events during the study period.
Over a mean follow-up period of 4 years, 15.5% of patients experienced adverse cardiac events; of those, 4.1 % were considered major events, while minor events occurred 11.4 %. Patients with greater CAC severity at baseline were more likely to display higher CAC increases in between scans and experience an adverse event at some point. The team determined that older age, higher CAC baseline scores and CAC growth were significantly associated with a greater likelihood of adverse events.
“These findings suggest that longitudinal tracking of CAC scores during annual LDCT follow-up—a core principle of [lung cancer screening]—may help establish more effective follow-up strategies,” the team suggests.
Follow-up research should seek to measure the impact of AI-enabled early prevention in these exams, they added.
Learn more about the findings here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.