Abstract 15494: Defining Coronary Artery Calcium Concordance and Variability - Implications for Incidence and Progression: The Dallas Heart Study
Background: Incidence and progression of coronary artery calcium (CAC) measured by serial CT are independently associated with future CHD events. Application of this information in clinical practice will depend on better understanding true changes in CAC beyond variability in the test itself.
Methods: We analyzed 2336 Dallas Heart Study (DHS) participants with paired CAC scans (1747 EBCT, 589 MDCT) that were attenuation corrected using a calibration phantom, and obtained minutes apart to determine interscan concordance for scores 0 vs >0. The 95% variability coefficients (95% limits of second scan for each individual CAC score) were determined for the full range of scores, including 0. These data derived cutoffs were then used to determine the frequency of CAC incidence and progression among 1766 subjects with follow up CAC scans 6.9 years later. Using a model with known predictors of CHD, we determined model fitness with these new definitions of CAC incidence and progression as outcome variables compared with prior reported definitions.
Results: Binary concordance (0 vs >0) between scans performed minutes apart was only 88.0%. The 95% variability for a single score of 0 was 3.1 Agatston units (AU) (ie. for a score of 0, 95% of second scan scores <3.1AU). Among those with positive scores, the 95% variability coefficients were determined by the derived formula 9.04√CAC. For a score of 10AU, the expected variability is ±29; a score of 100AU: ±90; a score of 400AU: ± 181. Using 95% variability limits to define binary CAC incidence and progression, respective frequencies were 14% and 44% (Table). Both novel definitions resulted in better measures of model fit compared to prior definitions in a model with known CHD predictors (Table).
Conclusion: There is significant interscan variability in CAC measurement, including around scores of 0. Incorporating variability estimates can help discern true changes in serial CAC scores and may provide superior methods to define CAC progression.
- © 2013 by American Heart Association, Inc.