RETURN TO ABSTRACT LISTING


Objective:
Coronary artery disease (CAD) remains a leading cause of death worldwide, and early detection and treatment of CAD can significantly improve morbidity and mortality. Coronary artery calcium (CAC) is closely associated with increased risk for CAD. Typically, ECG-gated CT is performed for CAC estimation; however, there is now an FDA-cleared, artificial intelligence (AI)-enabled cardiac imaging algorithm that quantifies CAC burden on nongated CT scans. The purpose of this study was to evaluate the integration of this AI algorithm into an existing radiology workflow to identify patients with high CAC on routine CT chest studies and expedite their referral to cardiology for further clinical evaluation.

Materials and Methods:
The CAC AI algorithm analyzed all nongated chest CT scans performed between January 2022 and August 2023 at an urban academic hospital. The algorithm stratified patients into three risk categories based on the extent of CAC expressed as Agatston scores. A visual report was generated for studies with moderate or severe CAC. After independent review by a board-certified chest radiologist for confirmation, the calcium score information was included in the study report using an encoding phrase that flagged the case for review by a cardiology clinical navigator. Patients who met the eligibility criteria were contacted to schedule a cardiology appointment for clinical evaluation. Longitudinal outcomes were tracked, including continued medical management, further diagnostic testing, or procedural intervention.

Results:
There were 8533 nongated chest CTs analyzed by the AI device, and 2950 cases (34.6%) were classified with a moderate or severe CAC score. So far, 528 cases have been reviewed by a chest radiologist and a cardiology nurse navigator for study eligibility. There were 108 patients (20.5%) eligible for cardiology referral, and as of August 2023, 45 of 108 (41.7%) patients have completed at least their first cardiology appointment, and many have already undergone additional follow-up testing and appointments. Of 45 patients, 39 (86.7%) required further diagnostic workup with an ECG and nuclear stress test. Based on those results, 13 of 39 (33.3%) patients had changes to their medication regimen, and 8/39 (20.5%) patients underwent cardiac catheterization. Following catheterization, two patients received drug-eluting stents (DES), and one patient underwent a coronary artery bypass grafting (CABG). Overall, 23 of 45 (51.1%) of patients who have seen cardiology ultimately had some change in their management.

Conclusion:
These findings illustrate the opportunity to use an AI-enabled method to screen routine nongated chest CT scans to identify patients at high risk for CAD and direct them to receive appropriate and timely clinical workup and treatment. We found that 51.1% of patients identified in this screening process saw some change in their clinical management. This has the potential to significantly decrease the morbidity and mortality of CAD while lowering costs for healthcare systems.