Automatic quantification of plaque progression dynamics as assessed by serial coronary computed tomography angiography using scan-quality-based vessel specific thresholds

Abstract

Introduction Use of serial coronary computed tomography angiography (CCTA) allows for the early assessment of coronary plaque progression which may aid in the prevention of major adverse cardiac events (MACE). However, assessment of serial CCTA is done by using anatomical landmarks for matching baseline and follow-up scans. Recently, a tool has been developed allowing for automatic quantification of plaque progression dynamics in serial CCTA utilizing plaque delineation. Purpose The aim of this study was to determine the thresholds that define whether there is plaque progression and/or regression. These thresholds depend on the contrast to noise ratio (CNR) which is an objective marker for scan quality as the latter impacts the plaque delineation. Methods Thresholds and CNR ratios were determined on 50 patients referred for a CCTA due to thoracic complaints. Two scan phases were selected from each patient in which maximum and minimum differences in plaque delineation were measured. Also, CNR was calculated separately for all three major epicardial coronary vessels. A total of 100 scans were analyzed in the current analysis accounting for a total of 300 coronary vessels. First, vessel and lumen wall delineation was done semi-automatically for all major epicardial coronary vessels. Secondly, manual drawings of 7 regions of interest (ROI) per scan were used to quantify scan quality which was defined using the CNR and calculated for each vessel separately. As plaque differences of two scans at the same moment in time should always be zero, the minimum and maximum difference in plaque delineation was used in these scans along with the CNR in order to create calibration graphs on which a linear regression analysis was performed (Figure 1, charts A & B). Inter-observer measurements were calculated using Pearson's correlation coefficient. Results A total of 300 coronary vessels were assessed at CCTA. Semi-automatic vessel and lumen wall delineation as well as CNR calculation was successful in all cases. Subsequent linear regression analysis performed on the CNR and maximal and minimal plaque delineation differences and taking into account the standard error of the estimate revealed the following formulas for minimum and maximum cut-off values: Max = [(0.660 − (002 × CNR)] + 0.349 Min = [(−1.028 + (0.012 × CNR)) − 0.61 The average CNR values was 13.4±3.6. Average maximum and minimum difference in plaque delineation was 0.7±0.3mm and −0.9±0.6mm respectively. The inter-observer correlation for CNR values was excellent yielding a correlation coefficient of 0.872 (p<0.001). The importance of using thresholds and subsequent calculation of vessel specific cutoff values is demonstrated in Figure 2. Conclusion Development of vessel-specific quality-based thresholds for the quantification and visualization of plaque progression dynamics as assessed by serial CCTA seems feasible and may aid in the early detection of atherosclerosis progression. Funding Acknowledgement Type of funding sources: None.