The utility of coronary artery calcium in non-gated high resolution CT thorax scans in predicting cardiac events as compared to the Framingham risk score: a retrospective study

Background There is existing evidence to suggest a correlation between coronary artery calcification (CAC) measured using ECG-triggered chest computed tomography and cardiovascular disease. Further evidence has emerged to suggest a correlation between CAC measured using non-gated CT scans and cardiovascular disease. Herein, we sought to ascertain the utility of incidental findings of CAC on non-triggered high resolution CT (HRCT) thorax used for patients undergoing lung cancer screening or follow-up for interstitial lung disease and Framingham risk score (FRS) in predicting cardiovascular events. Methods The Computerised Radiology Information Service (CRIS) database was manually searched to determine all HRCT scans performed in a single trust from 05/2015 to 05/2016. The reports issued by Radiologists and images of selected studies were reviewed. For patients with CAC, we calculated the calcium score for patients using the Agatston method. Clinical events were determined from the electronic medical record without knowledge of patients' CAC findings. For these patients, the Framingham Risk Score (FRS) was also calculated. The primary end point of the study was composite of all-cause mortality and cardiac events (non-fatal myocardial infarction, coronary revascularization, new atrial fibrillation or heart failure episode requiring hospitalization). Results We selected 300 scans from a total of approximately 2000 scans performed over this time. Data at follow up was available for 100% of the patients, with a median duration of follow up of 1.6 years. Moderate to severe CAC was found in 35% of people. Multivariable analysis showed good concordance between CAC and FRS in predicting composite clinical end point. The Odds Ratio for cardiac events in patients with moderate to severe CAC was 5.3 (p<0.01) and for composite clinical end point was 3.4 (p<0.01). This is similar to the OR predicted by the FRS: 4.8; p<0.01 and 3.1; p<0.01 respectively. Only 6.2% of patients with moderate to severe CAC were currently statin treated. Conclusion In this retrospective study of patients with respiratory disease attending for HRCT scanning, co-incidentally detected CAC predicts cardiac events, with good concordance with the FRS. The incidental finding of CAC on non-gated CT scanning should be reported with Agatston score calculation allowing consideration of intervention to mitigate cardiovascular risk and optimize. Further multi-centre prospective studies of this strategy, with a larger patient cohort should be conducted to clarify the utility of CAC as a prediction tool to modify cardiac risk.