Fig 1.
Three-dimensional ultrasound (3DUS) vessel reconstruction.
(A) Two-dimensional power mode images were grabbed from the video port of the ultrasound system and concatenated with spatial and temporal information to be stored in a virtual 3D-volume. Within this virtual 3D-volume navigation in all 3 orthogonal spatial planes is possible ad libitum. Plane orientation: green = transversal, red = coronal, blue = sagittal. (B) Based on the transversal planes (green), the lumen of the common and internal carotid artery was traced manually at variable distances of 1 to 4 mm with smaller (1 mm) intersection intervals at level of the stenosis. This manual segmentation (all yellow lines) was followed by an automatic vessel reconstruction (C). The reconstructed carotid artery can be rotated freely and cross-sectional area can be measured at every point perpendicular to the vessel’s course. Asterisk indicates origin of external carotid artery (not shown). Scale bar each 1 cm.
Fig 2.
Stenotic value of internal carotid artery stenosis (ICAS) depends on projection view.
(A and B) Anterior-posterior projection of power-mode 3D reconstructed ICAS; original lumen is also reconstructed and visualised in yellow (A). (B—D) Stenotic value of ICAS assessed as distal cross-sectional area reduction percentage (93%) does not change since it is not affected by rotation or tilting. By contrast, clockwise rotation of ICAS results in a decrease of stenotic value assessed as distal diameter reduction percentage from 82% (B) respectively 83% (C) to 72% (D). Smallest luminal diameter is marked with filled arrowhead while distal luminal diameter is marked with open arrowhead. Asterisk indicates origin of external carotid artery (not shown). CCA common carotid artery. Scale bar: 1 cm.
Fig 3.
Inter-method agreement between power-mode 3DUS and 2D-CDS.
(A and B) Inter-method agreements between power-mode 3DUS and 2D-CDS for grading ICAS are visualised by Bland and Altman analyses with the differences in stenotic values—assessed as distal diameter (A = examiner 1 and A’ = examiner 2) respectively CSA (B = examiner 1 and B’ examiner 2) reduction percentage—plotted against the mean stenotic value of both modalities. For the comparison of stenotic values assessed as distal diameter reduction percentage with 2D-CDS the Bland and Altman analyses (A and A’) showed no evidence of bias between methods but moderate limits of agreement (A bias 0.4%, limits of agreement 20.0 and -19.2; A’ bias 0.4, limits of agreement 25.0 and -24.2). Assessment of stenotic value with power-mode 3DUS as distal CSA reduction percentage accounted for a permanent overestimation of ICAS in comparison to 2D-CDS and wide limits of agreement (B bias -9.3, limits of agreement 17.1 and -35.7; B’ bias -11.4, limits of agreement 15.2 and -37.6). Note that ICAS number 33 (marked in red) was assumed to be an outlier when assessing stenotic value via CSA reduction percentage by both examiners as shown in the box-and-whisker plots (C and C’). Hence, Bland and Altman analyses with stenotic value assessed as distal CSA reduction percentage (B and B’) were performed without ICAS number 33. 3DUS three-dimensional ultrasound, 2D-CDS 2D colour-coded duplexsonography, ICAS internal carotid artery stenosis, SD standard deviation.
Table 1.
Intra-rater, inter-rater and inter-method agreement of 3D ultrasound (3DUS) for quantification of internal carotid artery stenosis (ICAS).