Impact of coronary lumen reconstruction on the estimation of endothelial shear stress: in vivo comparison of three-dimensional quantitative coronary angiography and three-dimensional fusion combining optical coherent tomography

Impact of coronary lumen reconstruction on the estimation of endothelial shear stress: in vivo... Abstract Aims It is not clearly elucidated how the fusion technique improves the accuracy of endothelial shear stress (ESS) prediction, in comparison with that of three-dimensional (3D) quantitative coronary angiography (QCA) alone. We aimed to evaluate the difference in geometric measurements and haemodynamic estimation between 3D QCA and a 3D fusion model combining 3D QCA and optical coherence tomography (OCT). Methods and results Computational fluid dynamics was assessed in the coronary models of 20 patients. In the plane-per-plane comparison, the difference and agreement were assessed using a generalized linear mixed model and concordance correlation coefficient (CCC), respectively. The haemodynamic feature around minimum-lumen-diameter (MLD) was characterized using CCC values calculated for 1-mm segments. In comparison with the 3D fusion model, 3D QCA showed a shorter maximum lumen diameter (2.54 ± 0.67 mm vs. 2.78 ± 0.73 mm, P < 0.001) and smaller lumen area (4.81 ± 2.56 mm2 vs. 5.66 ± 2.97 mm2, P < 0.001), resulting in a significantly higher ESS (4.64 Pa vs. 3.78 Pa, p = 0.029). A more asymmetric lumen shape of the 3D fusion model was more likely associated with under- and over-estimation of the maximum and minimum lumen diameters in the 3D QCA model, respectively. The circumferential ESS variations, which were blunted by 3D QCA, showed the worst concordance near the MLD site (CCC = 0.370) on segment-based comparison. Conclusion The 3D fusion technique may be a more relevant tool for the haemodynamic simulation of coronary arteries through providing more accurate lumen characterization than 3D QCA. atherosclerotic plaque, coronary lumen reconstruction, endothelial shear stress, optical coherent tomography, invasive coronary angiography Introduction Endothelial shear stress (ESS) plays a key role in early atheroma formation and the evolution of vulnerable plaque rupture1,2 as main components of the mechanism underlying acute coronary syndrome.3,4 A computational fluid dynamics (CFD) study using three-dimensional (3D) reconstruction of coronary trees facilitates the in vivo evaluation of intra-luminal ESS patterns in order to predict the risk of plaque rupture and future events. Although modelling of high-fidelity coronary geometry is a prerequisite for the precise assessment of fluid dynamics,5–8 the intravascular imaging modalities used in daily practice rarely support an appropriate 3D model for ESS analysis. At present, the novel fusion model that combined 3D reconstructed quantitative coronary angiography (QCA) with cross-sectional imaging [optical coherent tomography (OCT) or intravascular ultrasound (IVUS)-based structural and compositional information]7,8 provides more reliable regional mapping of the ESS, which may enable better linkage between the baseline haemodynamic features and the natural history of coronary atherosclerosis.2,5,9,10 Nevertheless, despite the complexity of the pre-processing steps, including lumen segmentation, geometric matching, orientation adjustment, and surface regeneration,6 the manner in which the fusion technique improves the accuracy of ESS prediction, in comparison with 3D QCA alone, remains unclear. In the present study, we aimed to compare the geometric measurements and CFD data between the 3D fusion model (combining 3D QCA and OCT) and 3D QCA alone. Methods Study population A total of 20 patients (20 coronary artery lesions) who underwent pre-procedural invasive coronary angiography (two planes with a viewing angle difference >40°) and OCT between October 2011 and December 2013 were enrolled. All the patients had a lesion with a diameter stenosis (DS) >30% (visual estimation) within a native coronary artery. The exclusion criteria were the presence of acute myocardial infarction, haemodynamic instability, inaccessibility of the OCT catheter across the lesion, left main or ostial lesions, side branch lesions, in-stent restenosis, diffuse or tandem lesions with a total lesion length >30 mm, vessel size >4.0 mm, and poor OCT quality. All patients signed written informed consent prior to the study. The study was approved by the Institutional Review Board at Asan Medical Center. Invasive imaging: coronary angiography and optical coherent tomography Catheterization was performed through the femoral or radial routes with standard catheters and coronary angiograms were digitally recorded. OCT images were acquired using a non-occlusive technique with the C7XRTM system and DragonFlyTM catheters (LightLab Imaging, Westford, MA, USA). The artery was cleared of blood by continuous flushing with iodixanol 370 (Visipaque; GE Health Care, Cork, Ireland) at a flow rate of 3.0 mL/s.11 3D coronary model reconstruction and morphological analysis For the construction of the ‘3D QCA’ model, two angiograms were selected at the end-diastolic phase of the cardiac cycle (Figure 1A). The corresponding images were imported into CAAS Workstation 5.11 (Pie Medical Imaging, Maastricht, the Netherlands), and the luminal area was semi-automatically segmented by experienced personnel according to the standard clinical definitions.12 The 3D QCA model was automatically constructed by combining segmented lumen boundaries on CAAS Workstation 5.11 (Figure 1B). The ‘3D fusion’ model was constructed by replacing the cross-section of the 3D QCA model with the luminal morphology segmented on OCT images (Supplementary data online). The lumen boundary of the OCT image in the region of interest (ROI in Figure 1D) was equidistantly positioned along the centreline of the 3D QCA model (Figure 1C), and the position and orientation of the OCT imaging plane were subsequently adjusted to match with the locations of anatomical landmarks such as the stenosis site and side branch.6 By integrating the OCT data (ROI) and the rest of the 3D QCA model (except for the ROI), the 3D fusion model was generated (Figure 1F). Figure 1 View largeDownload slide Model reconstruction techniques and analysis method. (A) Coronary angiograms. (B) Reconstruction of 3D QCA model. (C) Centreline extraction of 3D QCA model. (D) Geometrical matching of OCT imaging planes in ROI. (E) OCT lumen segmentation. (F) Reconstruction of 3D fusion model. (G) Geometries of 3D fusion models for 20 patients. (H) Mesh generation and boundary conditions for CFD analysis. (I) Comparison method for morphological and haemodynamic variables. QCA, coronary angiography; OCT, optical coherent tomography; ROI, region of interest; CFD, computational fluid dynamics. Figure 1 View largeDownload slide Model reconstruction techniques and analysis method. (A) Coronary angiograms. (B) Reconstruction of 3D QCA model. (C) Centreline extraction of 3D QCA model. (D) Geometrical matching of OCT imaging planes in ROI. (E) OCT lumen segmentation. (F) Reconstruction of 3D fusion model. (G) Geometries of 3D fusion models for 20 patients. (H) Mesh generation and boundary conditions for CFD analysis. (I) Comparison method for morphological and haemodynamic variables. QCA, coronary angiography; OCT, optical coherent tomography; ROI, region of interest; CFD, computational fluid dynamics. CFD simulation and ESS analysis For ESS analysis, CFD simulations were conducted using the coronary models of 20 vessels (Figure 1H). The flow area in the coronary models that were assumed as rigid was discretized using Ansys ICEM CFD 15.0 (Ansys Inc., Canonsburg, PA, USA). The blood was modelled as incompressible (density = 1060 kg/m3) and Newtonian (0.0035 kg/m·s). A pulsatile flow was imposed, with the inflow rate estimated by the echocardiographic left-ventricular mass and resting perfusion (0.9 mL/g·min).13 Outflow boundary conditions were employed, and the flow fraction of each outlet was calculated by applying allometric laws.14 The flow simulations were performed using Ansys Fluent 15.0 (Ansys Inc.) by setting the same conditions to the 3D QCA and 3D fusion models. Morphological and haemodynamic comparison between coronary models The 3D QCA and 3D fusion models were compared in the OCT imaging plane to identify the direct correlation with the morphological and haemodynamic metrics. For morphological assessment, the luminal area, and the maximum and minimum diameters were calculated and compared (Figure 1I). The ratio of the maximum to minimum diameters was also computed to estimate lumen asymmetry. To evaluate the haemodynamic impact of the morphological differences, the ESS and oscillatory shear index (OSI) were evaluated, and in-plane asymmetry (ESS max 30°/ESS min 30°) was assessed by measuring their maximum and minimum values in the 30° lumen section with respect to the lumen centre (Supplementary data online). To characterize the haemodynamic feature around the narrowed lesion, haemodynamic comparison was performed for the segment that was divided into 1-mm intervals with respect to the position of the angiographic minimum lumen diameter (MLD). The domains for the 1-mm-segment analysis included the entire ROI, MLD site (MLD −5 mm, MLD +5 mm), and distal-MLD site (MLD, MLD +5 mm), respectively. Statistical analysis Continuous values are presented as mean ± standard deviation or median and interquartile range, as appropriate. Categorical variables are presented as numbers and percentages. Generalized linear mixed models were applied to remove the clustering effect of the vessel when evaluating the differences in the morphological and haemodynamic variables. Bland–Altman analysis was used to calculate the bias and limits of agreement in morphological assessment. Linear regression analysis was used to evaluate the correlation of lumen asymmetry with the morphological differences. The concordance correlation coefficient (CCC), which evaluates the agreement between variables on a scale of 0 (no agreement) to 1 (perfect agreement), was applied for all the variables.15 Values of P < 0.05 were considered statistically significant. Statistical analyses were performed using R package and SPSS 17.0 for Windows (SPSS, Inc., Chicago, IL, USA). Results Patients The baseline clinical and angiographic data are summarized in Table 1. A total of 20 vessels from 20 patients (11 left anterior descending arteries and 9 right coronary arteries) were included in the present study. The mean length of the ROI was 30.8 ± 8.8 mm. After excluding the branching site, a total of 2884 planes within 612 segments were finally included. The DS of the ROI was 41% ± 11%. Table 1 Baseline clinical, angiographic, and OCT characteristics of the study patients (n = 20) Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 QCA, quantitative coronary angiography; OCT, optical coherent tomography. Table 1 Baseline clinical, angiographic, and OCT characteristics of the study patients (n = 20) Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 QCA, quantitative coronary angiography; OCT, optical coherent tomography. Morphological comparison Compared to the 3D fusion model, 3D QCA showed a shorter maximum lumen diameter and a smaller lumen area (Table 2). The underestimation of the lumen diameter with the 3D QCA model was observed in most OCT imaging planes (81.1%, 62.0%, and 73.7% for the maximum and minimum lumen diameters, and lumen area, respectively, Figure 2). A more asymmetric lumen shape of the 3D fusion model was more likely associated with under- and over-estimation of the maximum and minimum lumen diameters in the 3D QCA model, respectively. Although the luminal area showed a high concordance (CCC = 0.802), lumen asymmetry demonstrated a weak agreement in the luminal shape (CCC = 0.118). In case examples (Figure 3), detailed changes in the luminal area were neglected in the 3D QCA model, and the discrepancy was most prominent in the ruptured area (Figure 3B). Table 2 Plane-per-plane comparison of morphological and haemodynamic variables between the 3D QCA and 3D fusion models (n = 2884) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; D, lumen diameter; Alumen, luminal area; ESS, endothelial shear stress; OSI, oscillatory shear index; CCC, concordance correlation coefficient; CI, confidence interval; max, maximum; min, minimum; *, statistically significant. Table 2 Plane-per-plane comparison of morphological and haemodynamic variables between the 3D QCA and 3D fusion models (n = 2884) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; D, lumen diameter; Alumen, luminal area; ESS, endothelial shear stress; OSI, oscillatory shear index; CCC, concordance correlation coefficient; CI, confidence interval; max, maximum; min, minimum; *, statistically significant. Figure 2 View largeDownload slide Comparison of morphological variables between reconstruction methods. (A) Maximum lumen diameter (Dmax). (B) Minimum lumen diameter (Dmin). (C) Luminal area (Alumen). Bland–Altman plot and effect of lumen asymmetry, which estimated by Dmax/Dmin calculated from the 3D fusion model, are presented in top and bottom rows, respectively. QCA, coronary angiography; β, standardized beta coefficient; *, P< 0.001. Figure 2 View largeDownload slide Comparison of morphological variables between reconstruction methods. (A) Maximum lumen diameter (Dmax). (B) Minimum lumen diameter (Dmin). (C) Luminal area (Alumen). Bland–Altman plot and effect of lumen asymmetry, which estimated by Dmax/Dmin calculated from the 3D fusion model, are presented in top and bottom rows, respectively. QCA, coronary angiography; β, standardized beta coefficient; *, P< 0.001. Figure 3 View largeDownload slide View largeDownload slide Case examples. (A) With multiple luminal narrowing, the morphological variations along the blood flow direction were not demonstrated by 3D QCA model. Thus, the low oscillatory ESS regions were indicated only by the 3D fusion model (but not by the 3D QCA model). (B) A large plaque rupture, detected by OCT, was not visualized by coronary angiography. At the rupture site, the haemodynamic features (low ESS and high OSI) shown in the 3D fusion model were not demonstrated by the 3D QCA model, wherein morphology reconstruction was hindered by the overlapping of coronary angiography images. QCA, coronary angiography; ESS, endothelial shear stress; OCT, optical coherent tomography; RCA, right coronary artery; LM, left main coronary artery; LCX, left circumflex artery; LAD, left anterior descending artery; ROI, region of interest. Figure 3 View largeDownload slide View largeDownload slide Case examples. (A) With multiple luminal narrowing, the morphological variations along the blood flow direction were not demonstrated by 3D QCA model. Thus, the low oscillatory ESS regions were indicated only by the 3D fusion model (but not by the 3D QCA model). (B) A large plaque rupture, detected by OCT, was not visualized by coronary angiography. At the rupture site, the haemodynamic features (low ESS and high OSI) shown in the 3D fusion model were not demonstrated by the 3D QCA model, wherein morphology reconstruction was hindered by the overlapping of coronary angiography images. QCA, coronary angiography; ESS, endothelial shear stress; OCT, optical coherent tomography; RCA, right coronary artery; LM, left main coronary artery; LCX, left circumflex artery; LAD, left anterior descending artery; ROI, region of interest. Haemodynamic comparison The 3D QCA model showed a significantly higher ESS as compared to the 3D fusion model (Table 2). ESS overestimation >0.5 Pa was found in 50.5% of the overall planes (Supplementary data online, Figure S2). The circumferential variations of the ESS (ESS max 30°/ESS min 30°) that were blunted by the 3D QCA (vs. 3D fusion) model were lower in 1331 planes (Table 2). The ESS max 30°/ESS min 30° and the maximum OSI section (OSI max 30°) showed low agreement in plane-per-plane comparison (CCC < 0.5), and in the 1-mm-segment comparison on the MLD or distal-MLD site (Table 3), the agreement between reconstruction methods was more impaired for those variables. For the lesion with multiple narrowing, only 3D fusion model detected the low and oscillatory patterns of ESS (Figure 3A), and in the ruptured lesion identified by OCT imaging, the haemodynamic variations were absent in 3D QCA model (Figure 3B). Table 3 Concordance correlation coefficient (CCC) of haemodynamic variables between the 3D QCA and 3D fusion models in 1-mm-segment analysis Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) With respect to the angiographically minimum-lumen-diameter (MLD) location, the MLD site includes the segments within 5 mm in both the proximal and distal sides. The distal-MLD site indicates the distal part of MLD site. 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; ROI, region of interest; ESS, endothelial shear stress; OSI, oscillatory shear index; CI, confidence interval; max, maximum; min, minimum. Table 3 Concordance correlation coefficient (CCC) of haemodynamic variables between the 3D QCA and 3D fusion models in 1-mm-segment analysis Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) With respect to the angiographically minimum-lumen-diameter (MLD) location, the MLD site includes the segments within 5 mm in both the proximal and distal sides. The distal-MLD site indicates the distal part of MLD site. 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; ROI, region of interest; ESS, endothelial shear stress; OSI, oscillatory shear index; CI, confidence interval; max, maximum; min, minimum. Discussion The major findings of the present study are as follows: (i) the 3D QCA (compared to 3D fusion) model generally underestimated the lumen size, which resulted in significant overestimation of the ESS; (ii) the discrepancies between the two models were particularly accentuated in the lesions with an asymmetric lumen; and (iii) as 3D QCA could not identify the delicate lumen contour, the circumferential variations and the oscillatory behaviours of the ESS were blunted by the 3D QCA (vs. 3D fusion) model. In the present study, the fusion technique combining 3D QCA and OCT showed significant differences in the assessment of both coronary geometry and haemodynamic forces with 3D QCA alone. The 3D QCA underestimated the luminal area and the lumen asymmetry compared with 3D fusion, as reported in the comparison studies of QCA and intravascular imaging.8,16,17 In phantom experiments and histological comparison, OCT imaging has shown that it provides reproducible quantitative measurements of the coronary lumen with high accuracy and precision, including stented lesions.17–19 On the other hand, although 3D QCA reconstructed by integrating multiple angiographic planes provides better information regarding luminal geometry than the 2D projected profile,20,21 inaccurate edge detection and oversimplification of the lumen geometry remain as fundamental limitations.20,21 Thus, the accuracy of 3D QCA reconstruction is susceptible to luminal shape, angiographic angle, cardiac cycles, and vessel curvature.16,22 Combining imprecise lumen boundaries for 3D QCA model might exclude the longitudinal changes in the flow area and therefore hinder more accurate ESS estimation. Smaller lumen of 3D QCA model yielded a higher ESS, since the ESS in a circular conduit is inversely proportional to the cube of the lumen diameter (as per the Hagen–Poiseuille theory). Also, the 3D fusion model demonstrated the local flow patterns, including oscillating or circumferentially varying ESS, which were not captured by the 3D QCA model. This result was supported by the finding of Papafaklis et al.23 that 3D fusion with OCT could detect detailed variations on ESS mapping in good agreement with IVUS-based 3D models. Meanwhile, in vitro experiment study reported that in 3D QCA model of concentric stenosis phantom, the lumen contour of the stenotic segment was smoothed, and the dissimilarity in ESS and OSI distributions was more pronounced around the lesions with significant stenosis.24 Although the 3D QCA model was useful for assessing the ESS profile of normal-looking vessels,25 it may lead to significant errors in the quantitative assessment of ESS metrics in the stenotic lesions. In particular, coronary angiography was not sufficient to depict the complicated morphology of plaque rupture, due to the presence of image overlapping and poor resolution (Figure 3B). In the 3D fusion model, however, low ESS regions in the cavity containing recirculating or oscillatory flow were clearly demonstrated. At the site of plaque rupture, coronary flow is disturbed by the haemodynamic interplay with the deformed lumen, which might also affect the functional significance.26 Thus, the fusion technique combining 3D QCA and high-resolution OCT may facilitate the determination of the link between ESS distribution and the biological responses at the rupture site. In the technical and practical points of view, the longitudinal positioning of intravascular imaging planes could be equivocal without appropriate landmarks.27 For vessels with multiple bifurcations, the incorrect identification of side branches led to the disorientation of imaging planes in the ROI.6 Also, 3D fusion model may suffer from the motion artifact28 and geometric errors inherited from 3D QCA. In particular, oblique imaging, which imposed unrealistic roughness at the vessel wall, created locally low and oscillating ESS, unless it was systemically removed.29 Nonetheless, the introduction of the novel technique and use of dedicated software enabled the reconstruction of coronary models and estimation of ESS with a high reproducibility within a feasible time on a desktop PC.6,8,30,31 In the present study, our high-fidelity simulations using eight CPUs required an average of 202 ± 46 min, and for practical use, time requirement can be further reduced by adjusting mesh size and optimizing computational domain. The application of 3D reconstruction of coronary arteries to CFD analyses has helped elucidate the pathophysiologic mechanism underlying the dynamic changes over time. The regional ESS patterns and circumferential ESS distribution were influenced by vascular remodelling2,5,9,10 and plaque eccentricity.23 Moreover, serial intravascular imaging studies demonstrated that low oscillatory ESS played a role in the development and progression of coronary atherosclerosis32,33 in line with histological analysis of animal models.34 Furthermore, low ESS led to plaque development and progression, as well as the formation of a rupture-prone thin-cap fibroatheroma.35–37 In the Prediction of Progression of Coronary Artery Disease and Clinical Outcome Using Vascular Profiling of Shear Stress and Wall Morphology (PREDICTION) trial, low local ESS was an independent predictor of plaque enlargement and clinically relevant luminal obstruction.10 Along with introduction of novel methods for ESS assessment,38 haemodynamic analysis using elaborated fusion techniques will provide a better insight into the detection of high-risk plaques and event-prone lesions.28 There are several limitations. First, the small sample size and potential selection bias may have affected the results. With limited view angles, the left circumflex coronary artery was not included. Second, although the flow rate calculated from the echocardiographically-measured left-ventricular mass was theoretically reasonable, the direct measurement of blood flow might provide better interpretation of the haemodynamic metrics. Third, although the major side branches in the ROI were considered, the exclusion of side branches that could not be reconstructed might lead to the overestimation of the flow rate in the distal main branch.6 Forth, ECG-gating was not considered in the 3D fusion reconstruction. Nonetheless, because OCT imaging with higher pullback speed and shorter acquisition length is less affected by cardiac phase than IVUS,39,40 non-ECG-gated OCT showed a high agreement with ECG-gated IVUS in geometric reconstruction, suggesting that the consideration of ECG-gating was not likely necessary for the 3 D fusion with OCT.31 Finally, the prognostic implication of different flow haemodynamics between two models was not investigated. The 3D QCA model (compared to the 3D fusion) underestimated the lumen size, thus resulting in the significant overestimation of the ESS. As 3D QCA could not identify the lumen contour in detail, the circumferential variations and oscillatory behaviours of the ESS were blunted by 3D QCA. In contrast, the fusion technique may facilitate the determination of the link between ESS distribution and biological responses of coronary lesions. Funding This study was supported by the Bio & Medical Technology Development Program of the NRF funded by the Korean government (Grant Number: NRF-2017R1A2B3009800) and by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Republic of Korea (Grant Number: HI14C0517 and HI15C1790). Conflict of interest: None declared. References 1 Chatzizisis YS , Coskun AU , Jonas M , Edelman ER , Feldman CL , Stone PH. Role of endothelial shear stress in the natural history of coronary atherosclerosis and vascular remodeling: molecular, cellular, and vascular behavior . J Am Coll Cardiol 2007 ; 49 : 2379 – 93 . Google Scholar Crossref Search ADS PubMed 2 Stone PH , Coskun AU , Kinlay S , Popma JJ , Sonka M , Wahle A et al. Regions of low endothelial shear stress are the sites where coronary plaque progresses and vascular remodelling occurs in humans: an in vivo serial study . Eur Heart J 2007 ;28:705–710. 3 Virmani R , Kolodgie FD , Burke AP , Farb A , Schwartz SM. Lessons from sudden coronary death a comprehensive morphological classification scheme for atherosclerotic lesions . Arterioscler Thromb Vasc Biol 2000 ; 20 : 1262 – 75 . Google Scholar Crossref Search ADS PubMed 4 Falk E , Shah PK , Fuster V. Coronary plaque disruption . Circulation 1995 ; 92 : 657 – 71 . Google Scholar Crossref Search ADS PubMed 5 Samady H , Eshtehardi P , McDaniel MC , Suo J , Dhawan SS , Maynard C et al. Coronary artery wall shear stress is associated with progression and transformation of atherosclerotic plaque and arterial remodeling in patients with coronary artery disease . Circulation 2011 ; 124 : 779 – 88 . Google Scholar Crossref Search ADS PubMed 6 Li Y , Gutiérrez-Chico JL , Holm NR , Yang W , Hebsgaard L , Christiansen EH et al. Impact of side branch modeling on computation of endothelial shear stress in coronary artery disease coronary tree reconstruction by fusion of 3D angiography and OCT . J Am Coll Cardiol 2015 ; 66 : 125 – 35 . Google Scholar Crossref Search ADS PubMed 7 Kousera C , Nijjer S , Torii R , Petraco R , Sen S , Foin N et al. Patient-specific coronary stenoses can be modeled using a combination of OCT and flow velocities to accurately predict hyperemic pressure gradients . IEEE Trans Biomed Eng 2014 ; 61 : 1902 – 13 . Google Scholar Crossref Search ADS PubMed 8 Papafaklis MI , Bourantas CV , Yonetsu T , Vergallo R , Kotsia A , Nakatani S et al. Anatomically correct three-dimensional coronary artery reconstruction using frequency domain optical coherence tomographic and angiographic data: head-to-head comparison with intravascular ultrasound for endothelial shear stress assessment in humans . EuroIntervention 2015;11:407–15. 9 Wentzel JJ , Janssen E , Vos J , Schuurbiers JC , Krams R , Serruys PW et al. Extension of increased atherosclerotic wall thickness into high shear stress regions is associated with loss of compensatory remodeling . Circulation 2003 ; 108 : 17 – 23 . Google Scholar Crossref Search ADS PubMed 10 Stone PH , Saito S , Takahashi S , Makita Y , Nakamura S , Kawasaki T et al. Prediction of progression of coronary artery disease and clinical outcomes using vascular profiling of endothelial shear stress and arterial plaque characteristics: the PREDICTION Study . Circulation 2012 ; 126 : 172 – 81 . Google Scholar Crossref Search ADS PubMed 11 Gonzalo N , Garcia-Garcia HM , Regar E , Barlis P , Wentzel J , Onuma Y et al. In vivo assessment of high-risk coronary plaques at bifurcations with combined intravascular ultrasound and optical coherence tomography . JACC Cardiovasc Imaging 2009 ; 2 : 473 – 82 . Google Scholar Crossref Search ADS PubMed 12 Popma JJ , Leon MB , Moses JW , Holmes DR , Cox N , Fitzpatrick M et al. Quantitative assessment of angiographic restenosis after sirolimus-eluting stent implantation in native coronary arteries . Circulation 2004 ; 110 : 3773 – 80 . Google Scholar Crossref Search ADS PubMed 13 Westerhof N , Stergiopulos N , Noble MI. Snapshots of Hemodynamics: An Aid for Clinical Research and Graduate Education . Springer Science & Business Media , 2010 . 14 Huo Y , Finet G , Lefevre T , Louvard Y , Moussa I , Kassab GS. Which diameter and angle rule provides optimal flow patterns in a coronary bifurcation? J Biomech 2012 ; 45 : 1273 – 9 . Google Scholar Crossref Search ADS PubMed 15 King TS , Chinchilli VM , Carrasco JL. A repeated measures concordance correlation coefficient . Statist Med 2007 ; 26 : 3095 – 113 . Google Scholar Crossref Search ADS 16 Tu S , Xu L , Ligthart J , Xu B , Witberg K , Sun Z et al. In vivo comparison of arterial lumen dimensions assessed by co-registered three-dimensional (3D) quantitative coronary angiography, intravascular ultrasound and optical coherence tomography . Int J Cardiovasc Imaging 2012 ; 28 : 1315 – 27 . Google Scholar Crossref Search ADS PubMed 17 Kubo T , Akasaka T , Shite J , Suzuki T , Uemura S , Yu B et al. OCT compared with IVUS in a coronary lesion assessment: the OPUS-CLASS study . JACC Cardiovasc Imaging 2013 ; 6 : 1095 – 104 . Google Scholar Crossref Search ADS PubMed 18 Suzuki Y , Ikeno F , Koizumi T , Tio F , Yeung AC , Yock PG et al. In vivo comparison between optical coherence tomography and intravascular ultrasound for detecting small degrees of in-stent neointima after stent implantation . JACC Cardiovasc Interv 2008 ; 1 : 168 – 73 . Google Scholar Crossref Search ADS PubMed 19 Gutiérrez-Chico JL , Serruys PW , Girasis C , Garg S , Onuma Y , Brugaletta S et al. Quantitative multi-modality imaging analysis of a fully bioresorbable stent: a head-to-head comparison between QCA, IVUS and OCT . Int J Cardiovasc Imaging 2012 ; 28 : 467 – 78 . Google Scholar Crossref Search ADS PubMed 20 Koning G , Hekking E , Kemppainen J , Richardson G , Rothman M , Reiber J. Suitability of the cordis stabilizer™ marker guide wire for quantitative coronary angiography calibration: an in vitro and in vivo study . Catheter Cardiovasc Interv 2001 ; 52 : 334 – 41 . Google Scholar Crossref Search ADS PubMed 21 Rittger H , Schertel B , Schmidt M , Justiz J , Brachmann J , Sinha A-M. Three-dimensional reconstruction allows accurate quantification and length measurements of coronary artery stenoses . EuroIntervention 2009 ; 5 : 127 – 32 . Google Scholar Crossref Search ADS PubMed 22 Çimen S , Gooya A , Grass M , Frangi AF. Reconstruction of coronary arteries from X-ray angiography: a review . Med Image Anal 2016 ; 32 : 46 – 68 . Google Scholar Crossref Search ADS PubMed 23 Papafaklis MI , Takahashi S , Antoniadis AP , Coskun AU , Tsuda M , Mizuno S et al. Effect of the local hemodynamic environment on the de novo development and progression of eccentric coronary atherosclerosis in humans: insights from PREDICTION . Atherosclerosis 2015 ; 240 : 205 – 11 . Google Scholar Crossref Search ADS PubMed 24 Schrauwen JT , Karanasos A , van Ditzhuijzen NS , Aben J-P , van der Steen AF , Wentzel JJ et al. Influence of the accuracy of angiography-based reconstructions on velocity and wall shear stress computations in coronary bifurcations: a phantom study . PLoS One 2015 ; 10 : e0145114 . Google Scholar Crossref Search ADS PubMed 25 Goubergrits L , Wellnhofer E , Kertzscher U , Affeld K , Petz C , Hege H-C. Coronary artery WSS profiling using a geometry reconstruction based on biplane angiography . Ann Biomed Eng 2009 ; 37 : 682 – 91 . Google Scholar Crossref Search ADS PubMed 26 Park S-J , Kang S-J , Ahn J-M , Shim EB , Kim Y-T , Yun S-C et al. Visual-functional mismatch between coronary angiography and fractional flow reserve . JACC Cardiovasc Interv 2012 ; 5 : 1029 – 36 . Google Scholar Crossref Search ADS PubMed 27 Tu S , Holm NR , Koning G , Huang Z , Reiber JH. Fusion of 3D QCA and IVUS/OCT . Int J Cardiovasc Imaging 2011 ; 27 : 197 – 207 . Google Scholar Crossref Search ADS PubMed 28 Bourantas CV , Jaffer FA , Gijsen FJ , van Soest G , Madden SP , Courtney BK et al. Hybrid intravascular imaging: recent advances, technical considerations, and current applications in the study of plaque pathophysiology . Eur Heart J 2017 ; 38 : 400 – 12 . Google Scholar Crossref Search ADS PubMed 29 Rybicki FJ , Melchionna S , Mitsouras D , Coskun AU , Whitmore AG , Steigner M et al. Prediction of coronary artery plaque progression and potential rupture from 320-detector row prospectively ECG-gated single heart beat CT angiography: Lattice Boltzmann evaluation of endothelial shear stress . Int J Cardiovasc Imaging 2009 ; 25 : 289 – 99 . Google Scholar Crossref Search ADS PubMed 30 Bourantas CV , Papafaklis MI , Athanasiou L , Kalatzis FG , Naka KK , Siogkas PK et al. A new methodology for accurate 3-dimensional coronary artery reconstruction using routine intravascular ultrasound and angiographic data: implications for widespread assessment of endothelial shear stress in humans . EuroIntervention 2013 ; 9 : 582 – 93 . Google Scholar Crossref Search ADS PubMed 31 Toutouzas K , Chatzizisis YS , Riga M , Giannopoulos A , Antoniadis AP , Tu S et al. Accurate and reproducible reconstruction of coronary arteries and endothelial shear stress calculation using 3D OCT: comparative study to 3D IVUS and 3D QCA . Atherosclerosis 2015 ; 240 : 510 – 9 . Google Scholar Crossref Search ADS PubMed 32 Timmins LH , Molony DS , Eshtehardi P , McDaniel MC , Oshinski JN , Giddens DP et al. Oscillatory wall shear stress is a dominant flow characteristic affecting lesion progression patterns and plaque vulnerability in patients with coronary artery disease . J R Soc Interface 2017 ; 14 : 20160972 . Google Scholar Crossref Search ADS PubMed 33 Vergallo R , Papafaklis MI , Yonetsu T , Bourantas CV , Andreou I , Wang Z et al. Endothelial shear stress and coronary plaque characteristics in humans: combined frequency-domain optical coherence tomography and computational fluid dynamics study . Circ Cardiovasc Imaging 2014 ; 7 : 905 – 11 . Google Scholar Crossref Search ADS PubMed 34 Pedrigi R , Poulsen C , Mehta V , Ramsing HN , Pareek N , Post A et al. Inducing persistent flow disturbances accelerates atherogenesis and promotes thin cap fibroatheroma development in D374Y-PCSK9 hypercholesterolemic minipigs . Circulation 2015 ; 132 : 1003 – 12 . Google Scholar Crossref Search ADS PubMed 35 Chatzizisis YS , Jonas M , Coskun AU , Beigel R , Stone BV , Maynard C et al. Prediction of the localization of high-risk coronary atherosclerotic plaques on the basis of low endothelial shear stress an intravascular ultrasound and histopathology natural history study . Circulation 2008 ; 117 : 993 – 1002 . Google Scholar Crossref Search ADS PubMed 36 Koskinas KC , Feldman CL , Chatzizisis YS , Coskun AU , Jonas M , Maynard C et al. Natural history of experimental coronary atherosclerosis and vascular remodeling in relation to endothelial shear stress a serial, in vivo intravascular ultrasound study . Circulation 2010 ; 121 : 2092 – 101 . Google Scholar Crossref Search ADS PubMed 37 Yamamoto E , Siasos G , Zaromytidou M , Coskun AU , Xing L , Bryniarski K et al. Low endothelial shear stress predicts evolution to high-risk coronary plaque phenotype in the future: a serial optical coherence tomography and computational fluid dynamics study . Circ Cardiovasc Interv 2017 ; 10 : e005455 . Google Scholar Crossref Search ADS PubMed 38 Chatzizisis YS , Toutouzas K , Giannopoulos AA , Riga M , Antoniadis AP , Fujinom Y et al. Association of global and local low endothelial shear stress with high-risk plaque using intracoronary 3D optical coherence tomography: introduction of ‘shear stress score’ . Eur Heart J Cardiovasc Imaging 2017;18:888–897. 39 Prati F , Guagliumi G , Mintz GS , Costa M , Regar E , Akasaka T et al. Expert review document part 2: methodology, terminology and clinical applications of optical coherence tomography for the assessment of interventional procedures . Eur Heart J 2012 ; 33 : 2513 – 20 . Google Scholar Crossref Search ADS PubMed 40 Fedele S , Biondi-Zoccai G , Kwiatkowski P , Di Vito L , Occhipinti M , Cremonesi A et al. Reproducibility of coronary optical coherence tomography for lumen and length measurements in humans (The CLI-VAR [Centro per la Lotta contro l'Infarto-VARiability] study) . Am J Cardiol 2012 ; 110 : 1106 – 12 . Google Scholar Crossref Search ADS PubMed Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Heart Journal – Cardiovascular Imaging Oxford University Press

Impact of coronary lumen reconstruction on the estimation of endothelial shear stress: in vivo comparison of three-dimensional quantitative coronary angiography and three-dimensional fusion combining optical coherent tomography

Loading next page...
 
/lp/ou_press/impact-of-coronary-lumen-reconstruction-on-the-estimation-of-MaGSTG7UOM
Publisher
Oxford University Press
Copyright
Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com.
ISSN
2047-2404
D.O.I.
10.1093/ehjci/jex222
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims It is not clearly elucidated how the fusion technique improves the accuracy of endothelial shear stress (ESS) prediction, in comparison with that of three-dimensional (3D) quantitative coronary angiography (QCA) alone. We aimed to evaluate the difference in geometric measurements and haemodynamic estimation between 3D QCA and a 3D fusion model combining 3D QCA and optical coherence tomography (OCT). Methods and results Computational fluid dynamics was assessed in the coronary models of 20 patients. In the plane-per-plane comparison, the difference and agreement were assessed using a generalized linear mixed model and concordance correlation coefficient (CCC), respectively. The haemodynamic feature around minimum-lumen-diameter (MLD) was characterized using CCC values calculated for 1-mm segments. In comparison with the 3D fusion model, 3D QCA showed a shorter maximum lumen diameter (2.54 ± 0.67 mm vs. 2.78 ± 0.73 mm, P < 0.001) and smaller lumen area (4.81 ± 2.56 mm2 vs. 5.66 ± 2.97 mm2, P < 0.001), resulting in a significantly higher ESS (4.64 Pa vs. 3.78 Pa, p = 0.029). A more asymmetric lumen shape of the 3D fusion model was more likely associated with under- and over-estimation of the maximum and minimum lumen diameters in the 3D QCA model, respectively. The circumferential ESS variations, which were blunted by 3D QCA, showed the worst concordance near the MLD site (CCC = 0.370) on segment-based comparison. Conclusion The 3D fusion technique may be a more relevant tool for the haemodynamic simulation of coronary arteries through providing more accurate lumen characterization than 3D QCA. atherosclerotic plaque, coronary lumen reconstruction, endothelial shear stress, optical coherent tomography, invasive coronary angiography Introduction Endothelial shear stress (ESS) plays a key role in early atheroma formation and the evolution of vulnerable plaque rupture1,2 as main components of the mechanism underlying acute coronary syndrome.3,4 A computational fluid dynamics (CFD) study using three-dimensional (3D) reconstruction of coronary trees facilitates the in vivo evaluation of intra-luminal ESS patterns in order to predict the risk of plaque rupture and future events. Although modelling of high-fidelity coronary geometry is a prerequisite for the precise assessment of fluid dynamics,5–8 the intravascular imaging modalities used in daily practice rarely support an appropriate 3D model for ESS analysis. At present, the novel fusion model that combined 3D reconstructed quantitative coronary angiography (QCA) with cross-sectional imaging [optical coherent tomography (OCT) or intravascular ultrasound (IVUS)-based structural and compositional information]7,8 provides more reliable regional mapping of the ESS, which may enable better linkage between the baseline haemodynamic features and the natural history of coronary atherosclerosis.2,5,9,10 Nevertheless, despite the complexity of the pre-processing steps, including lumen segmentation, geometric matching, orientation adjustment, and surface regeneration,6 the manner in which the fusion technique improves the accuracy of ESS prediction, in comparison with 3D QCA alone, remains unclear. In the present study, we aimed to compare the geometric measurements and CFD data between the 3D fusion model (combining 3D QCA and OCT) and 3D QCA alone. Methods Study population A total of 20 patients (20 coronary artery lesions) who underwent pre-procedural invasive coronary angiography (two planes with a viewing angle difference >40°) and OCT between October 2011 and December 2013 were enrolled. All the patients had a lesion with a diameter stenosis (DS) >30% (visual estimation) within a native coronary artery. The exclusion criteria were the presence of acute myocardial infarction, haemodynamic instability, inaccessibility of the OCT catheter across the lesion, left main or ostial lesions, side branch lesions, in-stent restenosis, diffuse or tandem lesions with a total lesion length >30 mm, vessel size >4.0 mm, and poor OCT quality. All patients signed written informed consent prior to the study. The study was approved by the Institutional Review Board at Asan Medical Center. Invasive imaging: coronary angiography and optical coherent tomography Catheterization was performed through the femoral or radial routes with standard catheters and coronary angiograms were digitally recorded. OCT images were acquired using a non-occlusive technique with the C7XRTM system and DragonFlyTM catheters (LightLab Imaging, Westford, MA, USA). The artery was cleared of blood by continuous flushing with iodixanol 370 (Visipaque; GE Health Care, Cork, Ireland) at a flow rate of 3.0 mL/s.11 3D coronary model reconstruction and morphological analysis For the construction of the ‘3D QCA’ model, two angiograms were selected at the end-diastolic phase of the cardiac cycle (Figure 1A). The corresponding images were imported into CAAS Workstation 5.11 (Pie Medical Imaging, Maastricht, the Netherlands), and the luminal area was semi-automatically segmented by experienced personnel according to the standard clinical definitions.12 The 3D QCA model was automatically constructed by combining segmented lumen boundaries on CAAS Workstation 5.11 (Figure 1B). The ‘3D fusion’ model was constructed by replacing the cross-section of the 3D QCA model with the luminal morphology segmented on OCT images (Supplementary data online). The lumen boundary of the OCT image in the region of interest (ROI in Figure 1D) was equidistantly positioned along the centreline of the 3D QCA model (Figure 1C), and the position and orientation of the OCT imaging plane were subsequently adjusted to match with the locations of anatomical landmarks such as the stenosis site and side branch.6 By integrating the OCT data (ROI) and the rest of the 3D QCA model (except for the ROI), the 3D fusion model was generated (Figure 1F). Figure 1 View largeDownload slide Model reconstruction techniques and analysis method. (A) Coronary angiograms. (B) Reconstruction of 3D QCA model. (C) Centreline extraction of 3D QCA model. (D) Geometrical matching of OCT imaging planes in ROI. (E) OCT lumen segmentation. (F) Reconstruction of 3D fusion model. (G) Geometries of 3D fusion models for 20 patients. (H) Mesh generation and boundary conditions for CFD analysis. (I) Comparison method for morphological and haemodynamic variables. QCA, coronary angiography; OCT, optical coherent tomography; ROI, region of interest; CFD, computational fluid dynamics. Figure 1 View largeDownload slide Model reconstruction techniques and analysis method. (A) Coronary angiograms. (B) Reconstruction of 3D QCA model. (C) Centreline extraction of 3D QCA model. (D) Geometrical matching of OCT imaging planes in ROI. (E) OCT lumen segmentation. (F) Reconstruction of 3D fusion model. (G) Geometries of 3D fusion models for 20 patients. (H) Mesh generation and boundary conditions for CFD analysis. (I) Comparison method for morphological and haemodynamic variables. QCA, coronary angiography; OCT, optical coherent tomography; ROI, region of interest; CFD, computational fluid dynamics. CFD simulation and ESS analysis For ESS analysis, CFD simulations were conducted using the coronary models of 20 vessels (Figure 1H). The flow area in the coronary models that were assumed as rigid was discretized using Ansys ICEM CFD 15.0 (Ansys Inc., Canonsburg, PA, USA). The blood was modelled as incompressible (density = 1060 kg/m3) and Newtonian (0.0035 kg/m·s). A pulsatile flow was imposed, with the inflow rate estimated by the echocardiographic left-ventricular mass and resting perfusion (0.9 mL/g·min).13 Outflow boundary conditions were employed, and the flow fraction of each outlet was calculated by applying allometric laws.14 The flow simulations were performed using Ansys Fluent 15.0 (Ansys Inc.) by setting the same conditions to the 3D QCA and 3D fusion models. Morphological and haemodynamic comparison between coronary models The 3D QCA and 3D fusion models were compared in the OCT imaging plane to identify the direct correlation with the morphological and haemodynamic metrics. For morphological assessment, the luminal area, and the maximum and minimum diameters were calculated and compared (Figure 1I). The ratio of the maximum to minimum diameters was also computed to estimate lumen asymmetry. To evaluate the haemodynamic impact of the morphological differences, the ESS and oscillatory shear index (OSI) were evaluated, and in-plane asymmetry (ESS max 30°/ESS min 30°) was assessed by measuring their maximum and minimum values in the 30° lumen section with respect to the lumen centre (Supplementary data online). To characterize the haemodynamic feature around the narrowed lesion, haemodynamic comparison was performed for the segment that was divided into 1-mm intervals with respect to the position of the angiographic minimum lumen diameter (MLD). The domains for the 1-mm-segment analysis included the entire ROI, MLD site (MLD −5 mm, MLD +5 mm), and distal-MLD site (MLD, MLD +5 mm), respectively. Statistical analysis Continuous values are presented as mean ± standard deviation or median and interquartile range, as appropriate. Categorical variables are presented as numbers and percentages. Generalized linear mixed models were applied to remove the clustering effect of the vessel when evaluating the differences in the morphological and haemodynamic variables. Bland–Altman analysis was used to calculate the bias and limits of agreement in morphological assessment. Linear regression analysis was used to evaluate the correlation of lumen asymmetry with the morphological differences. The concordance correlation coefficient (CCC), which evaluates the agreement between variables on a scale of 0 (no agreement) to 1 (perfect agreement), was applied for all the variables.15 Values of P < 0.05 were considered statistically significant. Statistical analyses were performed using R package and SPSS 17.0 for Windows (SPSS, Inc., Chicago, IL, USA). Results Patients The baseline clinical and angiographic data are summarized in Table 1. A total of 20 vessels from 20 patients (11 left anterior descending arteries and 9 right coronary arteries) were included in the present study. The mean length of the ROI was 30.8 ± 8.8 mm. After excluding the branching site, a total of 2884 planes within 612 segments were finally included. The DS of the ROI was 41% ± 11%. Table 1 Baseline clinical, angiographic, and OCT characteristics of the study patients (n = 20) Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 QCA, quantitative coronary angiography; OCT, optical coherent tomography. Table 1 Baseline clinical, angiographic, and OCT characteristics of the study patients (n = 20) Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 Variable Value Age (years) 62.9 ± 7.6 Male, N (%) 11 (55%) Diabetes mellitus, N (%) 8 (40%) Hypertension, N (%) 12 (60%) Smoker, N (%) 9 (45%) Hyperlipidaemia, N (%) 6 (30%) Chronic renal failure, N (%) 0 (0%) Acute coronary syndrome, N (%) 2 (10%) Diseased coronary artery, N (%)  Left anterior descending 11 (55%)  Right coronary 9 (45%) % Diameter stenosis (QCA) 41 ± 11 Lesion length (mm) 30.8 ± 8.8 Number of side branches 1.50 ± 0.74 Total OCT images 3054 OCT images excluded 170 (5.57%) Average OCT images analysed per patient 144 ± 43 QCA, quantitative coronary angiography; OCT, optical coherent tomography. Morphological comparison Compared to the 3D fusion model, 3D QCA showed a shorter maximum lumen diameter and a smaller lumen area (Table 2). The underestimation of the lumen diameter with the 3D QCA model was observed in most OCT imaging planes (81.1%, 62.0%, and 73.7% for the maximum and minimum lumen diameters, and lumen area, respectively, Figure 2). A more asymmetric lumen shape of the 3D fusion model was more likely associated with under- and over-estimation of the maximum and minimum lumen diameters in the 3D QCA model, respectively. Although the luminal area showed a high concordance (CCC = 0.802), lumen asymmetry demonstrated a weak agreement in the luminal shape (CCC = 0.118). In case examples (Figure 3), detailed changes in the luminal area were neglected in the 3D QCA model, and the discrepancy was most prominent in the ruptured area (Figure 3B). Table 2 Plane-per-plane comparison of morphological and haemodynamic variables between the 3D QCA and 3D fusion models (n = 2884) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; D, lumen diameter; Alumen, luminal area; ESS, endothelial shear stress; OSI, oscillatory shear index; CCC, concordance correlation coefficient; CI, confidence interval; max, maximum; min, minimum; *, statistically significant. Table 2 Plane-per-plane comparison of morphological and haemodynamic variables between the 3D QCA and 3D fusion models (n = 2884) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 3D QCA model 3D fusion model P-value CCC (95% CI) Morphological variables Dmax (mm) 2.54 ± 0.67 2.78 ± 0.73 <0.001* 0.804 (0.709, 0.870) Dmin (mm) 2.30 ± 0.63 2.40 ± 0.68 <0.001* 0.850 (0.770, 0.904) Dmax/Dmin 1.11 ± 0.10 1.17 ± 0.12 0.749 0.118 (-0.017, 0.249) Alumen (mm2) 4.81 ± 2.56 5.66 ± 2.97 <0.001* 0.802 (0.707, 0.868) Haemodynamic variables ESS (Pa) 4.64 (2.46–9.87) 3.78 (1.78–7.94) 0.029* 0.712 (0.628, 0.779) ESS min 30° (Pa) 2.06 (0.81–5.09) 1.41 (0.56–3.50) 0.378 0.608 (0.547, 0.663) ESS max 30° (Pa) 7.58 (3.78–15.65) 6.62 (3.04–13.61) <0.001* 0.767 (0.686, 0.829) ESS max 30°/ESS min 30° 2.79 (1.89–5.56) 3.66 (2.26–7.76) <0.001* 0.499 (0.456, 0.539) OSI 0.001 (0–0.002) 0.001 (0–0.006) 0.055 0.608 (0.560, 0.652) OSI max 30˚ 0.002 (0–0.007) 0.003 (0.001–0.026) 0.167 0.475 (0.172, 0.625) 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; D, lumen diameter; Alumen, luminal area; ESS, endothelial shear stress; OSI, oscillatory shear index; CCC, concordance correlation coefficient; CI, confidence interval; max, maximum; min, minimum; *, statistically significant. Figure 2 View largeDownload slide Comparison of morphological variables between reconstruction methods. (A) Maximum lumen diameter (Dmax). (B) Minimum lumen diameter (Dmin). (C) Luminal area (Alumen). Bland–Altman plot and effect of lumen asymmetry, which estimated by Dmax/Dmin calculated from the 3D fusion model, are presented in top and bottom rows, respectively. QCA, coronary angiography; β, standardized beta coefficient; *, P< 0.001. Figure 2 View largeDownload slide Comparison of morphological variables between reconstruction methods. (A) Maximum lumen diameter (Dmax). (B) Minimum lumen diameter (Dmin). (C) Luminal area (Alumen). Bland–Altman plot and effect of lumen asymmetry, which estimated by Dmax/Dmin calculated from the 3D fusion model, are presented in top and bottom rows, respectively. QCA, coronary angiography; β, standardized beta coefficient; *, P< 0.001. Figure 3 View largeDownload slide View largeDownload slide Case examples. (A) With multiple luminal narrowing, the morphological variations along the blood flow direction were not demonstrated by 3D QCA model. Thus, the low oscillatory ESS regions were indicated only by the 3D fusion model (but not by the 3D QCA model). (B) A large plaque rupture, detected by OCT, was not visualized by coronary angiography. At the rupture site, the haemodynamic features (low ESS and high OSI) shown in the 3D fusion model were not demonstrated by the 3D QCA model, wherein morphology reconstruction was hindered by the overlapping of coronary angiography images. QCA, coronary angiography; ESS, endothelial shear stress; OCT, optical coherent tomography; RCA, right coronary artery; LM, left main coronary artery; LCX, left circumflex artery; LAD, left anterior descending artery; ROI, region of interest. Figure 3 View largeDownload slide View largeDownload slide Case examples. (A) With multiple luminal narrowing, the morphological variations along the blood flow direction were not demonstrated by 3D QCA model. Thus, the low oscillatory ESS regions were indicated only by the 3D fusion model (but not by the 3D QCA model). (B) A large plaque rupture, detected by OCT, was not visualized by coronary angiography. At the rupture site, the haemodynamic features (low ESS and high OSI) shown in the 3D fusion model were not demonstrated by the 3D QCA model, wherein morphology reconstruction was hindered by the overlapping of coronary angiography images. QCA, coronary angiography; ESS, endothelial shear stress; OCT, optical coherent tomography; RCA, right coronary artery; LM, left main coronary artery; LCX, left circumflex artery; LAD, left anterior descending artery; ROI, region of interest. Haemodynamic comparison The 3D QCA model showed a significantly higher ESS as compared to the 3D fusion model (Table 2). ESS overestimation >0.5 Pa was found in 50.5% of the overall planes (Supplementary data online, Figure S2). The circumferential variations of the ESS (ESS max 30°/ESS min 30°) that were blunted by the 3D QCA (vs. 3D fusion) model were lower in 1331 planes (Table 2). The ESS max 30°/ESS min 30° and the maximum OSI section (OSI max 30°) showed low agreement in plane-per-plane comparison (CCC < 0.5), and in the 1-mm-segment comparison on the MLD or distal-MLD site (Table 3), the agreement between reconstruction methods was more impaired for those variables. For the lesion with multiple narrowing, only 3D fusion model detected the low and oscillatory patterns of ESS (Figure 3A), and in the ruptured lesion identified by OCT imaging, the haemodynamic variations were absent in 3D QCA model (Figure 3B). Table 3 Concordance correlation coefficient (CCC) of haemodynamic variables between the 3D QCA and 3D fusion models in 1-mm-segment analysis Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) With respect to the angiographically minimum-lumen-diameter (MLD) location, the MLD site includes the segments within 5 mm in both the proximal and distal sides. The distal-MLD site indicates the distal part of MLD site. 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; ROI, region of interest; ESS, endothelial shear stress; OSI, oscillatory shear index; CI, confidence interval; max, maximum; min, minimum. Table 3 Concordance correlation coefficient (CCC) of haemodynamic variables between the 3D QCA and 3D fusion models in 1-mm-segment analysis Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) Entire ROI (n = 612) MLD site (n = 194) Distal-MLD site (n = 98) ESS (Pa) 0.687 (0.589, 0.765) 0.655 (0.501, 0.768) 0.757 (0.620, 0.849) ESS min 30° (Pa) 0.583 (0.490, 0.664) 0.573 (0.439, 0.681) 0.709 (0.576, 0.806) ESS max 30° (Pa) 0.701 (0.608, 0.775) 0.698 (0.546, 0.806) 0.698 (0.525, 0.816) ESS max 30°/ESS min 30° 0.625 (0.561, 0.681) 0.370 (0.224, 0.501) 0.321 (0.089, 0.519) OSI 0.649 (0.564, 0.721) 0.605 (0.465, 0.715) 0.661 (0.462, 0.796) OSI max 30° 0.573 (0.497, 0.640) 0.371 (0.209, 0.512) 0.425 (0.172, 0.625) With respect to the angiographically minimum-lumen-diameter (MLD) location, the MLD site includes the segments within 5 mm in both the proximal and distal sides. The distal-MLD site indicates the distal part of MLD site. 30° indicates the haemodynamic metric measured in the 30° lumen section with respect to the lumen centre. QCA, quantitative coronary angiography; ROI, region of interest; ESS, endothelial shear stress; OSI, oscillatory shear index; CI, confidence interval; max, maximum; min, minimum. Discussion The major findings of the present study are as follows: (i) the 3D QCA (compared to 3D fusion) model generally underestimated the lumen size, which resulted in significant overestimation of the ESS; (ii) the discrepancies between the two models were particularly accentuated in the lesions with an asymmetric lumen; and (iii) as 3D QCA could not identify the delicate lumen contour, the circumferential variations and the oscillatory behaviours of the ESS were blunted by the 3D QCA (vs. 3D fusion) model. In the present study, the fusion technique combining 3D QCA and OCT showed significant differences in the assessment of both coronary geometry and haemodynamic forces with 3D QCA alone. The 3D QCA underestimated the luminal area and the lumen asymmetry compared with 3D fusion, as reported in the comparison studies of QCA and intravascular imaging.8,16,17 In phantom experiments and histological comparison, OCT imaging has shown that it provides reproducible quantitative measurements of the coronary lumen with high accuracy and precision, including stented lesions.17–19 On the other hand, although 3D QCA reconstructed by integrating multiple angiographic planes provides better information regarding luminal geometry than the 2D projected profile,20,21 inaccurate edge detection and oversimplification of the lumen geometry remain as fundamental limitations.20,21 Thus, the accuracy of 3D QCA reconstruction is susceptible to luminal shape, angiographic angle, cardiac cycles, and vessel curvature.16,22 Combining imprecise lumen boundaries for 3D QCA model might exclude the longitudinal changes in the flow area and therefore hinder more accurate ESS estimation. Smaller lumen of 3D QCA model yielded a higher ESS, since the ESS in a circular conduit is inversely proportional to the cube of the lumen diameter (as per the Hagen–Poiseuille theory). Also, the 3D fusion model demonstrated the local flow patterns, including oscillating or circumferentially varying ESS, which were not captured by the 3D QCA model. This result was supported by the finding of Papafaklis et al.23 that 3D fusion with OCT could detect detailed variations on ESS mapping in good agreement with IVUS-based 3D models. Meanwhile, in vitro experiment study reported that in 3D QCA model of concentric stenosis phantom, the lumen contour of the stenotic segment was smoothed, and the dissimilarity in ESS and OSI distributions was more pronounced around the lesions with significant stenosis.24 Although the 3D QCA model was useful for assessing the ESS profile of normal-looking vessels,25 it may lead to significant errors in the quantitative assessment of ESS metrics in the stenotic lesions. In particular, coronary angiography was not sufficient to depict the complicated morphology of plaque rupture, due to the presence of image overlapping and poor resolution (Figure 3B). In the 3D fusion model, however, low ESS regions in the cavity containing recirculating or oscillatory flow were clearly demonstrated. At the site of plaque rupture, coronary flow is disturbed by the haemodynamic interplay with the deformed lumen, which might also affect the functional significance.26 Thus, the fusion technique combining 3D QCA and high-resolution OCT may facilitate the determination of the link between ESS distribution and the biological responses at the rupture site. In the technical and practical points of view, the longitudinal positioning of intravascular imaging planes could be equivocal without appropriate landmarks.27 For vessels with multiple bifurcations, the incorrect identification of side branches led to the disorientation of imaging planes in the ROI.6 Also, 3D fusion model may suffer from the motion artifact28 and geometric errors inherited from 3D QCA. In particular, oblique imaging, which imposed unrealistic roughness at the vessel wall, created locally low and oscillating ESS, unless it was systemically removed.29 Nonetheless, the introduction of the novel technique and use of dedicated software enabled the reconstruction of coronary models and estimation of ESS with a high reproducibility within a feasible time on a desktop PC.6,8,30,31 In the present study, our high-fidelity simulations using eight CPUs required an average of 202 ± 46 min, and for practical use, time requirement can be further reduced by adjusting mesh size and optimizing computational domain. The application of 3D reconstruction of coronary arteries to CFD analyses has helped elucidate the pathophysiologic mechanism underlying the dynamic changes over time. The regional ESS patterns and circumferential ESS distribution were influenced by vascular remodelling2,5,9,10 and plaque eccentricity.23 Moreover, serial intravascular imaging studies demonstrated that low oscillatory ESS played a role in the development and progression of coronary atherosclerosis32,33 in line with histological analysis of animal models.34 Furthermore, low ESS led to plaque development and progression, as well as the formation of a rupture-prone thin-cap fibroatheroma.35–37 In the Prediction of Progression of Coronary Artery Disease and Clinical Outcome Using Vascular Profiling of Shear Stress and Wall Morphology (PREDICTION) trial, low local ESS was an independent predictor of plaque enlargement and clinically relevant luminal obstruction.10 Along with introduction of novel methods for ESS assessment,38 haemodynamic analysis using elaborated fusion techniques will provide a better insight into the detection of high-risk plaques and event-prone lesions.28 There are several limitations. First, the small sample size and potential selection bias may have affected the results. With limited view angles, the left circumflex coronary artery was not included. Second, although the flow rate calculated from the echocardiographically-measured left-ventricular mass was theoretically reasonable, the direct measurement of blood flow might provide better interpretation of the haemodynamic metrics. Third, although the major side branches in the ROI were considered, the exclusion of side branches that could not be reconstructed might lead to the overestimation of the flow rate in the distal main branch.6 Forth, ECG-gating was not considered in the 3D fusion reconstruction. Nonetheless, because OCT imaging with higher pullback speed and shorter acquisition length is less affected by cardiac phase than IVUS,39,40 non-ECG-gated OCT showed a high agreement with ECG-gated IVUS in geometric reconstruction, suggesting that the consideration of ECG-gating was not likely necessary for the 3 D fusion with OCT.31 Finally, the prognostic implication of different flow haemodynamics between two models was not investigated. The 3D QCA model (compared to the 3D fusion) underestimated the lumen size, thus resulting in the significant overestimation of the ESS. As 3D QCA could not identify the lumen contour in detail, the circumferential variations and oscillatory behaviours of the ESS were blunted by 3D QCA. In contrast, the fusion technique may facilitate the determination of the link between ESS distribution and biological responses of coronary lesions. Funding This study was supported by the Bio & Medical Technology Development Program of the NRF funded by the Korean government (Grant Number: NRF-2017R1A2B3009800) and by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI) funded by the Ministry of Health & Welfare, Republic of Korea (Grant Number: HI14C0517 and HI15C1790). Conflict of interest: None declared. References 1 Chatzizisis YS , Coskun AU , Jonas M , Edelman ER , Feldman CL , Stone PH. Role of endothelial shear stress in the natural history of coronary atherosclerosis and vascular remodeling: molecular, cellular, and vascular behavior . J Am Coll Cardiol 2007 ; 49 : 2379 – 93 . Google Scholar Crossref Search ADS PubMed 2 Stone PH , Coskun AU , Kinlay S , Popma JJ , Sonka M , Wahle A et al. Regions of low endothelial shear stress are the sites where coronary plaque progresses and vascular remodelling occurs in humans: an in vivo serial study . Eur Heart J 2007 ;28:705–710. 3 Virmani R , Kolodgie FD , Burke AP , Farb A , Schwartz SM. Lessons from sudden coronary death a comprehensive morphological classification scheme for atherosclerotic lesions . Arterioscler Thromb Vasc Biol 2000 ; 20 : 1262 – 75 . Google Scholar Crossref Search ADS PubMed 4 Falk E , Shah PK , Fuster V. Coronary plaque disruption . Circulation 1995 ; 92 : 657 – 71 . Google Scholar Crossref Search ADS PubMed 5 Samady H , Eshtehardi P , McDaniel MC , Suo J , Dhawan SS , Maynard C et al. Coronary artery wall shear stress is associated with progression and transformation of atherosclerotic plaque and arterial remodeling in patients with coronary artery disease . Circulation 2011 ; 124 : 779 – 88 . Google Scholar Crossref Search ADS PubMed 6 Li Y , Gutiérrez-Chico JL , Holm NR , Yang W , Hebsgaard L , Christiansen EH et al. Impact of side branch modeling on computation of endothelial shear stress in coronary artery disease coronary tree reconstruction by fusion of 3D angiography and OCT . J Am Coll Cardiol 2015 ; 66 : 125 – 35 . Google Scholar Crossref Search ADS PubMed 7 Kousera C , Nijjer S , Torii R , Petraco R , Sen S , Foin N et al. Patient-specific coronary stenoses can be modeled using a combination of OCT and flow velocities to accurately predict hyperemic pressure gradients . IEEE Trans Biomed Eng 2014 ; 61 : 1902 – 13 . Google Scholar Crossref Search ADS PubMed 8 Papafaklis MI , Bourantas CV , Yonetsu T , Vergallo R , Kotsia A , Nakatani S et al. Anatomically correct three-dimensional coronary artery reconstruction using frequency domain optical coherence tomographic and angiographic data: head-to-head comparison with intravascular ultrasound for endothelial shear stress assessment in humans . EuroIntervention 2015;11:407–15. 9 Wentzel JJ , Janssen E , Vos J , Schuurbiers JC , Krams R , Serruys PW et al. Extension of increased atherosclerotic wall thickness into high shear stress regions is associated with loss of compensatory remodeling . Circulation 2003 ; 108 : 17 – 23 . Google Scholar Crossref Search ADS PubMed 10 Stone PH , Saito S , Takahashi S , Makita Y , Nakamura S , Kawasaki T et al. Prediction of progression of coronary artery disease and clinical outcomes using vascular profiling of endothelial shear stress and arterial plaque characteristics: the PREDICTION Study . Circulation 2012 ; 126 : 172 – 81 . Google Scholar Crossref Search ADS PubMed 11 Gonzalo N , Garcia-Garcia HM , Regar E , Barlis P , Wentzel J , Onuma Y et al. In vivo assessment of high-risk coronary plaques at bifurcations with combined intravascular ultrasound and optical coherence tomography . JACC Cardiovasc Imaging 2009 ; 2 : 473 – 82 . Google Scholar Crossref Search ADS PubMed 12 Popma JJ , Leon MB , Moses JW , Holmes DR , Cox N , Fitzpatrick M et al. Quantitative assessment of angiographic restenosis after sirolimus-eluting stent implantation in native coronary arteries . Circulation 2004 ; 110 : 3773 – 80 . Google Scholar Crossref Search ADS PubMed 13 Westerhof N , Stergiopulos N , Noble MI. Snapshots of Hemodynamics: An Aid for Clinical Research and Graduate Education . Springer Science & Business Media , 2010 . 14 Huo Y , Finet G , Lefevre T , Louvard Y , Moussa I , Kassab GS. Which diameter and angle rule provides optimal flow patterns in a coronary bifurcation? J Biomech 2012 ; 45 : 1273 – 9 . Google Scholar Crossref Search ADS PubMed 15 King TS , Chinchilli VM , Carrasco JL. A repeated measures concordance correlation coefficient . Statist Med 2007 ; 26 : 3095 – 113 . Google Scholar Crossref Search ADS 16 Tu S , Xu L , Ligthart J , Xu B , Witberg K , Sun Z et al. In vivo comparison of arterial lumen dimensions assessed by co-registered three-dimensional (3D) quantitative coronary angiography, intravascular ultrasound and optical coherence tomography . Int J Cardiovasc Imaging 2012 ; 28 : 1315 – 27 . Google Scholar Crossref Search ADS PubMed 17 Kubo T , Akasaka T , Shite J , Suzuki T , Uemura S , Yu B et al. OCT compared with IVUS in a coronary lesion assessment: the OPUS-CLASS study . JACC Cardiovasc Imaging 2013 ; 6 : 1095 – 104 . Google Scholar Crossref Search ADS PubMed 18 Suzuki Y , Ikeno F , Koizumi T , Tio F , Yeung AC , Yock PG et al. In vivo comparison between optical coherence tomography and intravascular ultrasound for detecting small degrees of in-stent neointima after stent implantation . JACC Cardiovasc Interv 2008 ; 1 : 168 – 73 . Google Scholar Crossref Search ADS PubMed 19 Gutiérrez-Chico JL , Serruys PW , Girasis C , Garg S , Onuma Y , Brugaletta S et al. Quantitative multi-modality imaging analysis of a fully bioresorbable stent: a head-to-head comparison between QCA, IVUS and OCT . Int J Cardiovasc Imaging 2012 ; 28 : 467 – 78 . Google Scholar Crossref Search ADS PubMed 20 Koning G , Hekking E , Kemppainen J , Richardson G , Rothman M , Reiber J. Suitability of the cordis stabilizer™ marker guide wire for quantitative coronary angiography calibration: an in vitro and in vivo study . Catheter Cardiovasc Interv 2001 ; 52 : 334 – 41 . Google Scholar Crossref Search ADS PubMed 21 Rittger H , Schertel B , Schmidt M , Justiz J , Brachmann J , Sinha A-M. Three-dimensional reconstruction allows accurate quantification and length measurements of coronary artery stenoses . EuroIntervention 2009 ; 5 : 127 – 32 . Google Scholar Crossref Search ADS PubMed 22 Çimen S , Gooya A , Grass M , Frangi AF. Reconstruction of coronary arteries from X-ray angiography: a review . Med Image Anal 2016 ; 32 : 46 – 68 . Google Scholar Crossref Search ADS PubMed 23 Papafaklis MI , Takahashi S , Antoniadis AP , Coskun AU , Tsuda M , Mizuno S et al. Effect of the local hemodynamic environment on the de novo development and progression of eccentric coronary atherosclerosis in humans: insights from PREDICTION . Atherosclerosis 2015 ; 240 : 205 – 11 . Google Scholar Crossref Search ADS PubMed 24 Schrauwen JT , Karanasos A , van Ditzhuijzen NS , Aben J-P , van der Steen AF , Wentzel JJ et al. Influence of the accuracy of angiography-based reconstructions on velocity and wall shear stress computations in coronary bifurcations: a phantom study . PLoS One 2015 ; 10 : e0145114 . Google Scholar Crossref Search ADS PubMed 25 Goubergrits L , Wellnhofer E , Kertzscher U , Affeld K , Petz C , Hege H-C. Coronary artery WSS profiling using a geometry reconstruction based on biplane angiography . Ann Biomed Eng 2009 ; 37 : 682 – 91 . Google Scholar Crossref Search ADS PubMed 26 Park S-J , Kang S-J , Ahn J-M , Shim EB , Kim Y-T , Yun S-C et al. Visual-functional mismatch between coronary angiography and fractional flow reserve . JACC Cardiovasc Interv 2012 ; 5 : 1029 – 36 . Google Scholar Crossref Search ADS PubMed 27 Tu S , Holm NR , Koning G , Huang Z , Reiber JH. Fusion of 3D QCA and IVUS/OCT . Int J Cardiovasc Imaging 2011 ; 27 : 197 – 207 . Google Scholar Crossref Search ADS PubMed 28 Bourantas CV , Jaffer FA , Gijsen FJ , van Soest G , Madden SP , Courtney BK et al. Hybrid intravascular imaging: recent advances, technical considerations, and current applications in the study of plaque pathophysiology . Eur Heart J 2017 ; 38 : 400 – 12 . Google Scholar Crossref Search ADS PubMed 29 Rybicki FJ , Melchionna S , Mitsouras D , Coskun AU , Whitmore AG , Steigner M et al. Prediction of coronary artery plaque progression and potential rupture from 320-detector row prospectively ECG-gated single heart beat CT angiography: Lattice Boltzmann evaluation of endothelial shear stress . Int J Cardiovasc Imaging 2009 ; 25 : 289 – 99 . Google Scholar Crossref Search ADS PubMed 30 Bourantas CV , Papafaklis MI , Athanasiou L , Kalatzis FG , Naka KK , Siogkas PK et al. A new methodology for accurate 3-dimensional coronary artery reconstruction using routine intravascular ultrasound and angiographic data: implications for widespread assessment of endothelial shear stress in humans . EuroIntervention 2013 ; 9 : 582 – 93 . Google Scholar Crossref Search ADS PubMed 31 Toutouzas K , Chatzizisis YS , Riga M , Giannopoulos A , Antoniadis AP , Tu S et al. Accurate and reproducible reconstruction of coronary arteries and endothelial shear stress calculation using 3D OCT: comparative study to 3D IVUS and 3D QCA . Atherosclerosis 2015 ; 240 : 510 – 9 . Google Scholar Crossref Search ADS PubMed 32 Timmins LH , Molony DS , Eshtehardi P , McDaniel MC , Oshinski JN , Giddens DP et al. Oscillatory wall shear stress is a dominant flow characteristic affecting lesion progression patterns and plaque vulnerability in patients with coronary artery disease . J R Soc Interface 2017 ; 14 : 20160972 . Google Scholar Crossref Search ADS PubMed 33 Vergallo R , Papafaklis MI , Yonetsu T , Bourantas CV , Andreou I , Wang Z et al. Endothelial shear stress and coronary plaque characteristics in humans: combined frequency-domain optical coherence tomography and computational fluid dynamics study . Circ Cardiovasc Imaging 2014 ; 7 : 905 – 11 . Google Scholar Crossref Search ADS PubMed 34 Pedrigi R , Poulsen C , Mehta V , Ramsing HN , Pareek N , Post A et al. Inducing persistent flow disturbances accelerates atherogenesis and promotes thin cap fibroatheroma development in D374Y-PCSK9 hypercholesterolemic minipigs . Circulation 2015 ; 132 : 1003 – 12 . Google Scholar Crossref Search ADS PubMed 35 Chatzizisis YS , Jonas M , Coskun AU , Beigel R , Stone BV , Maynard C et al. Prediction of the localization of high-risk coronary atherosclerotic plaques on the basis of low endothelial shear stress an intravascular ultrasound and histopathology natural history study . Circulation 2008 ; 117 : 993 – 1002 . Google Scholar Crossref Search ADS PubMed 36 Koskinas KC , Feldman CL , Chatzizisis YS , Coskun AU , Jonas M , Maynard C et al. Natural history of experimental coronary atherosclerosis and vascular remodeling in relation to endothelial shear stress a serial, in vivo intravascular ultrasound study . Circulation 2010 ; 121 : 2092 – 101 . Google Scholar Crossref Search ADS PubMed 37 Yamamoto E , Siasos G , Zaromytidou M , Coskun AU , Xing L , Bryniarski K et al. Low endothelial shear stress predicts evolution to high-risk coronary plaque phenotype in the future: a serial optical coherence tomography and computational fluid dynamics study . Circ Cardiovasc Interv 2017 ; 10 : e005455 . Google Scholar Crossref Search ADS PubMed 38 Chatzizisis YS , Toutouzas K , Giannopoulos AA , Riga M , Antoniadis AP , Fujinom Y et al. Association of global and local low endothelial shear stress with high-risk plaque using intracoronary 3D optical coherence tomography: introduction of ‘shear stress score’ . Eur Heart J Cardiovasc Imaging 2017;18:888–897. 39 Prati F , Guagliumi G , Mintz GS , Costa M , Regar E , Akasaka T et al. Expert review document part 2: methodology, terminology and clinical applications of optical coherence tomography for the assessment of interventional procedures . Eur Heart J 2012 ; 33 : 2513 – 20 . Google Scholar Crossref Search ADS PubMed 40 Fedele S , Biondi-Zoccai G , Kwiatkowski P , Di Vito L , Occhipinti M , Cremonesi A et al. Reproducibility of coronary optical coherence tomography for lumen and length measurements in humans (The CLI-VAR [Centro per la Lotta contro l'Infarto-VARiability] study) . Am J Cardiol 2012 ; 110 : 1106 – 12 . Google Scholar Crossref Search ADS PubMed Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Journal

European Heart Journal – Cardiovascular ImagingOxford University Press

Published: Oct 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off