Noninvasive definition of anatomic coronary artery disease by ultrafast computed tomographic scanning: a quantitative pathologic comparison study

Abstract

Objectives: The aim of this study was to determine the relation between coronary artery calcification detected by ultrafast computed tomographic scanning and histopathologic coronary artery disease.

Background: Recent studies suggest that discrete coronary artery calcification as visualized by ultrafast computed tomographic scanning may facilitate the noninvasive detection or estimation, or both, of the in situ extent of coronary disease. Such quantitative relations have not been established.

Methods: Thirteen consecutive perfusion-fixed autopsy hearts (from eight male and five female patients aged 17 to 83 years) were scanned by ultrafast computed tomographic scanning in contiguous 3-mm tomographic sections. The major epicardial arteries were dissected free, positioned longitudinally and scanned again in cross section. Coronary artery calcification in a coronary segment was defined as the presence of one or more voxels with a computed tomographic density > 130 Hounsfield units. Each epicardial artery was sectioned longitudinally, stained and measured with a planimeter for quantification of cross-sectional and atherosclerotic plaque areas at 3-mm intervals, corresponding to the computed tomographic scans. A total of 522 paired coronary computed tomographic and histologic sections were studied.

Results: Direct relations were found between ultrafast computed tomographic scanning coronary artery calcium burden and atherosclerotic plaque area and percent lumen area stenosis. However, the range for plaque area or percent lumen stenosis, or both, associated with a given calcium burden was broad. Three hundred thirty-one coronary segments showed no calcification by computed tomography. Although atherosclerotic disease was found in several corresponding pathologic specimens, > 97% of these noncalcified segments were associated with nonobstructive disease (< 75% area stenosis); if no calcification was determined in an entire coronary vessel, all corresponding coronary disease was found to be nonobstructive. To determine the relation between arterial calcification and any atheromatous disease, computed tomographic calcium burden for each segment was paired with the histologic absence or presence of disease. Ultrafast computed tomographic scanning had a sensitivity and specificity of 59% and 90% and a negative and positive predictive value of 65% and 87%, respectively. A direct correlation was found (r = 0.99) between total calcium burden calculated from tomographic scans of the heart as a whole and scans of the arteries obtained in cross section.

Conclusions: The detection of coronary calcification by ultrafast computed tomographic scanning is highly predictive of the presence of histopathologic coronary disease, but the use of this technique to define the extent of coronary disease may be limited. However, the absence of coronary calcification at any site is highly specific for the absence of obstructive disease.