Vender Independent CAC Scoring Improves Risk Assessment

Dec 05, 2023
Font Size
A
A
A
Authors:
van der Werf NR, Dobrolinska MM, Greuter MJ, et al.
Citation:
Vendor Independent Coronary Calcium Scoring Improves Individual Risk Assessment: MESA (Multi-Ethnic Study of Atherosclerosis). JACC Cardiovasc Imaging 2023;16:1552-1564.
Summary By:
Melvyn Rubenfire, MD, FACC

Quick Takes

  • The present generation of multi-detector CTs of different and even the same manufacturer sensitivity for CAC scores may vary by as much as 44%.
  • It would be prudent for guidelines to include use of the calibration tool for current state-of-the-art CT systems so as to have a vendor-neutral Agatston score (vnAS) that improves event prediction for both CHD and ASCVD, and avoid concern when the score is used to decide initiation or deferral of statin therapy.
  • The vnAS calculator will be made publicly available.

Study Questions:

Does a calibration tool for state-of-the-art computed tomography (CT) systems result in vendor-neutral Agatston coronary calcium scores (vnAS), and does a vnAS improve coronary heart disease (CHD) and atherosclerotic cardiovascular disease (ASCVD) event prediction?

Methods:

The vnAS calibration tool was derived by imaging two anthropomorphic calcium containing phantoms on seven different CT and one electron beam tomography (EBT) system, which was used as the reference system. The effect of vnAS on CHD event prediction was analyzed with data from 3,181 participants from the MESA (Multi-Ethnic Study on Atherosclerosis). Chi-square analysis was used to compare CHD event rates between low (vnAS < 100) and high coronary artery calcium (CAC) score (vnAS ≥ 100). Multivariable Cox proportional hazard regression models were used to assess the incremental value of vnAS. The study used CHD event prediction over 16.7 (interquartile range, 4.8 years) years of follow-up.

Results:

For all CT systems, a strong correlation with EBT-AS was found (R2 > 0.932). Of the reclassified individuals, 13 (13/85 = 15.3%) and 20 (20/85 = 23.5%) experienced a CHD or ASCVD event during 16.7 years of follow-up. Of those originally in the low calcium group (n = 781), 85 (11%) were reclassified to a higher risk category based on the recalculated vnAS. The vnAS was always higher with up to 39%. For reclassified participants, the CHD event rate of 15% was significantly higher compared to participants in the low calcium group (7%, p = 0.008) with a CHD hazard ratio of 3.39 (95% confidence interval [CI], 1.82-6.35; p = 0.001). The hazard ratio of ASCVD for reclassified individuals was 1.97 (95% CI, 1.22-3.18). There was no reclassification for persons with a zero CAC score or those in the high CAC group.

Conclusions:

The study group developed a calibration tool, which enables calculation of a vendor-neutral Agatston CAC score. MESA participants who were reclassified to a higher calcium category by means of the vendor-neutral score experienced more CHD events, indicating improved risk categorization.

Perspective:

It is important to consider that using a score of 100 does not separate a low- and high-risk group, particularly when considering low, intermediate, and high by the American College of Cardiology/American Heart Association risk estimate and the risk enhancers.

There is a significant difference in CAC scores comparing the multi-detector CT systems used today from different manufacturers and even between different CTs from the same manufacturer (difference in mean score of up to 44%). Clinically, variations in AS may translate into risk misclassification and improper treatment in up to 6.5% of asymptomatic individuals. It would be prudent for guidelines to include use of the calibration tool for current state-of-the-art CT systems so as to have a vnAS that improves event prediction for both CHD and ASCVD and avoid concern when the score is used to decide initiation or deferral of statin therapy.

The decision of using or deferring statins based on the CACs of 75th percentile for age/sex or >100 is often used so we do not want variability as was found in this sample patient: patient with a body mass index >25 kg/m2 and a multi-detector CT-derived AS of 100, the vnAS varied between 88–148 depending on the specific CT system that was used.

While the vnAS score will be helpful and more accurate, it is only one of the variables that should be used. Surprising and maybe not valid was that in patients with an AS of zero, vnAS was always zero since EBT is much more sensitive to calcium.

Clinical Topics: Noninvasive Imaging, Prevention

Keywords: Computed Tomography, Plaque, Atherosclerotic, Risk Assessment


< Back to Listings