Accuracy of density measurements within plaques located in artificial coronary arteries by X-ray multislice CT: results of a phantom study.

PURPOSE: Clinical studies indicate that coronary plaque morphology might be differentiated noninvasively using multislice CT by determining tissue density within the lesions. The aim of the present experimental study was to evaluate factors that influence density measurements within small vessels.
METHOD: A coronary phantom model was developed, consisting of silicon tubes (lumen diameter 4 mm) with two plaques of known density inside, simulating soft and intermediate lesions (Plaque 1: -39 HU; Plaque 2: 72 HU). Density measurement were conducted in three different contrast medium concentrations (1:30, 1:40, 1:50) and two different slice widths (4 x 2.5 mm, 4 x 1 mm). All scans were performed on a Somatom Volume Zoom (Siemens, Forchheim, Germany). Experimental results were compared with calculated data based on computer simulation.
RESULTS: The two plaques could be clearly differentiated from each other on both collimations (4 x 2.5 mm: Plaque 1, 85 +/- 61 HU vs. Plaque 2, 119 +/- 26 HU, p < 0.0001; 4 x 1 mm: Plaque 1, 50 +/- 54 HU vs. Plaque 2, 91 +/- 17 HU, p < 0.0001). Significantly lower and more accurate results were achieved with 1.0 mm collimation (p < 0.0001). Contrast medium concentration contributed significantly to the measurements (p < 0.001). The experimental findings were confirmed by computer simulation, which revealed even more accurate results when using a 0.5 mm collimation (Plaque 1, 0.5 mm: -9 HU vs. 4 x 1 mm: 14 HU, Plaque 2, 4 x 0.5 mm: 83 HU vs. 4 x 1 mm: 93 HU).
CONCLUSION: Density measurements were found to be highly dependent on slice width and surrounding contrast enhancement. Our results indicate that standardization of methodology is required before the noninvasive differentiation of human plaque morphology by multislice CT can be applied in the clinical setting as a screening test for coronary soft plaques.