Fig 1.
Method of quantitative stent wall thickness measurement.
Coronary stent reformation in short axis perpendicular to centerline. Upper row (A–E), images after reconstruction with a medium-smooth kernel (XCB, Philips Healthcare, Cleveland, OH, USA); lower row, (F–J), images after reconstruction with an edge-enhancing kernel (XCD, Philips Healthcare, Cleveland, OH, USA). Internal (B, G) and external (C, H) stent areas measured using a semi-automatic segmentation software, with manual correction when needed (circumference thickness method). Internal (D, I) and external (E, J) stent diameters measured using manual placement of an electronic caliper for distance measurements (orthogonal thickness method).
Table 1.
Patient characteristics and scan-related parameters.
Table 2.
Coronary artery stent segmental distribution* (N = 24 patients).
Table 3.
Types of stents*.
Table 4.
Stent CT measurements (71 stented segments).
Fig 2.
Stent wall thickness overestimation with smooth and sharp kernels.
Box plots: Stent wall thickness overestimation using the XCD kernel was significantly less than with the XCB kernel (p < 0.005). In this figure, merged data is used for both diameter and circumference methods, as well as for both observers no 1 and 2.
Fig 3.
Qualitative stent image quality with smooth and sharp kernels.
Average scores of visual assessment of image quality in 71 stented coronary artery segments. Distribution of average scores of visual assessment of image quality in 71 stented segments, as evaluated with 256-slice MDCT and prospective ECG-gating, after image reconstruction with medium-smooth (XCB) and edge-enhancing (XCD) reconstruction kernels. In this figure, each score is the average of the visual score of the two independent observers 1 and 2 for a given stent.
Fig 4.
Reduced stent blooming artifacts and improved strut definition with sharp (XCD) in comparison to smooth (XCB) kernel.
77-year-old female, right proximal coronary artery patent bare-metal stent (Abbott Vision, length 15 mm, diameter 4 mm, nominal thickness 0.08 mm). 256-slice CT acquisition with prospective ECG-gating, and image reconstuction with a medium-soft (XCB, left) and edge-enhancing (XCD, right) reconstruction kernels, multiplanar reformat. Window width, 1500 HU; window centre: 300 HU, for both kernels. For observer 1, stent wall thickness for the XCB and XCD kernels with the orthogonal thickness method was 1.44 mm and 1.24 mm, and 1.66 mm and 1.58 mm with the circumference method, respectively. Image quality scores for observer 1 were 3 and 4, respectively. For observer 2, stent wall thickness for the XCB and XCD kernels with the orthogonal thickness method was 0.89 mm and 1.15 mm, and 1.20 mm and 0.93 mm with the circumference method, respectively. Image quality scores for observer 2 were 3 and 4, respectively.
Fig 5.
Reduced stent blooming artifacts and improved strut definition with sharp (XCD) in comparison to smooth (XCB) kernel.
69-year-old female, first obtuse marginal artery stent. 256-slice CT acquisition with prospective ECG-gating, and image reconstuction with a medium-soft (XCB, left) and edge-enhancing (XCD, right) reconstruction kernels, multiplanar reformat. Window width, 1500 HU; window centre: 300 HU, for both kernels. For observer 1, stent wall thickness for the XCB and XCD kernels with the orthogonal thickness method was 1.29 mm and 1.05 mm, and 1.26 mm and 1.26 mm with the circumference method, respectively. Image quality scores for observer 1 were 3 and 4, respectively. For observer 2, stent wall thickness for the XCB and XCD kernels with the orthogonal thickness method was 0.97 mm and 0.71 mm, and 0.82 mm and 0.83 mm with the circumference method, respectively. Image quality scores for observer 2 were 2 and 3, respectively.