Visualization of Swirling Flow in a 3-dimensional Helical Stent Model using Color-coded Circulation and 4-dimensional Flow Magnetic Resonance Imaging Authors: Mitsunari Maruyama, Takaya Kawasumi, Yuki Kimura, Hiroya Aso, Kei Kaneko, Shintaro Sasaki, Makoto Ohta Published: 2026 ABSTRACT AbstractPurpose: To clarify how intraluminal flow in 3-dimensional helical and straight stent models is displayed differently using color-coded circulation and 4-dimensional flow magnetic resonance imaging.Material and Methods: Three-dimensional-printed stent models were created to evaluate intraluminal flow by 4-dimensional flow magnetic resonance imaging without stent-related artifacts. Three different helical shapes with different curvatures were created by deploying a 3-dimensional helical stent into poly(vinyl alcohol) hydrogel tubes with different stiffness. Stented poly(vinyl alcohol) hydrogel tube geometries were acquired by micro-computed tomography and replicated using a 3-dimensional printer. Three helical models with different curvatures (high, moderate, and mild) and one straight model were tested under constant flow conditions (200 mL/min, 40% glycerin solution). Flow dynamics were visualized using color-coded circulation based on time-to-arrival digital subtraction angiography and by 4-dimensional flow magnetic resonance imaging.Results: A swirling flow was observed in the helical models, whereas laminar flow was seen in the straight model. The swirling flow decreased as the curvature decreased. Color-coded circulation imaging revealed eccentric color distribution in the helical models and central-to-peripheral laminar stratification in the straight model. Similarly, 4-dimensional flow magnetic resonance imaging showed eccentric high- and low-velocity regions in the helical models and a central high-velocity region in the straight model.Conclusions: 4-dimensional flow magnetic resonance imaging revealed swirling flow in the helical models and laminar flow in the straight model. Color-coded circulation demonstrated eccentric color distribution in the helical models and central-to-peripheral laminar distribution in the straight model, which might reflect the high-and-low-velocity regions according to the streamline image.Keywords: swirling flow, laminar flow, color-coded circulation, 4D flow MRI Introduction The BioMimics 3D stent (Veryan Medical Ltd., Horsham, UK) is a 3-dimensional (3D) helical device designed to generate swirling flow. Inducing wall shear stress prevents vascular endothelial cell activation and the secretion of arteriosclerosis-promoting substances, thereby suppressing intimal thickening and restenosis [1]. Previously, it has been reported that swirling flow was visualized as layered colors on color-coded circulation after placement of a 3D helical stent in the superficial femoral artery [2]. Intraluminal flow assessment with 4-dimensional (4D) flow MRI was attempted but was limited by stent-related artifacts in that case report. To circumvent these artifacts, 3D-printed stent models were developed for 4D flow MRI evaluation. Both color-coded circulation (time-to-arrival digital subtraction angiography) and 4D flow MRI can visualize flow dynamics. 4D flow MRI requires time for both imaging and analysis, which limits its use in routine clinical practice. On the other hand, color-coded circulation obtained from DSA can be quickly confirmed during angiography. There have been no reports of a direct, one-to-one comparison of color-coded circulation and 4D flow MRI in vascular models, using both to visualize swirling and laminar flow. If the flow patterns displayed by both techniques are visually consistent, color-coded circulation could be useful for easily evaluating hemodynamics. This study aimed to clarify differences in intraluminal flow appearance between 3D helical and straight stent models using color-coded circulation and 4D flow MRI. Material and Methods Stent model preparation Three different helical shapes were created by developing 3D helical stents within poly(vinyl alcohol) hydrogel (PVA-H) tubes with different mechanical properties [3, 4]. Table 1 lists model names and their corresponding Young's modulus. PVA-H tubes with the different Young's moduli were used to mimic different stiffnesses of the vascular wall. The model in which a stent was developed inside a PVA-H tube with the lowest Young's modulus was defined as the high-helical model, the model in which a stent was developed inside a PVA-H tube with the highest Young's modulus was defined as the mild-helical model, and the model in which a stent was deployed inside a PVA-H tube with the intermediate Young's modulus was defined as the moderate-helical model. The unstented PVA-H tubes had an inner diameter of 4 mm, a wall thickness of 1 mm, and a length of 150 mm, and were straight. The stents had a centerline length of 100 mm and an expanded diameter of 6 mm. Figure 1a shows a stent developed in a PVA-H tube with an intermediate Young's modulus. The stented PVA-H tubes were scanned by micro-CT (ScanXmate-D225RSS270, Comscantecno Co., Ltd., Kanagawa, Japan). Figure 1b-d show the MIP images by micro-CT scans of stents developed within three different PVA-H tubes. The stented PVA-H tube centerline coordinates and lumen diameter were extracted in MATLAB R2023a (MathWorks, Natick, MA, USA) and reconstructed with vascularmd [5]. Cylindrical segments (inner diameter 4 mm, length 27 mm) were appended to both ends. A straight model with the same lumen diameter as the moderate-helical model was also prepared. All stent models (Figure 2) were 3D-printed on a Form 4 printer (Formlabs Inc., Somerville, MA, USA) with Clear V5 Resin. Table 1 shows the centerline curvature and lumen diameter of each stent model. Table 1.Curvature and Lumen Diameter of the Stent Models.Stent modelCurvature (mm−1)Lumen diameter (mm)Young’s modulus of the corresponding PVA-H tube (kPa)High-helical0.0422 ± 0.00456.23 ± 0.15166Moderate-helical0.0263 ± 0.00485.91 ± 0.08179Mild-helical0.0140 ± 0.00485.02 ± 0.08342Straight-5.91 ± 0.08-Three helical stent models and one straight stent model were 3D -printed. The straight model has the same lumen diameter as the moderate-helical model.3D: 3-dimensional; PVA-H: poly(vinyl alcohol) hydrogel Curvature and Lumen Diameter of the Stent Models. Three helical stent models and one straight stent model were 3D -printed. The straight model has the same lumen diameter as the moderate-helical model. 3D: 3-dimensional; PVA-H: poly(vinyl alcohol) hydrogel Stented PVA-H tube and MIP images obtained by micro-CT. (a) Stented PVA-H tube; (b-d) MIP images of stents developed within three different PVA-H tubes. The photograph shows a stent developed within a PVA-H tube with an intermediate Young’s modulus (a). The stent morphology within the PVA-H tube changes depending on its respective Young’s modulus (b-d). The corresponding PVA-H tubes are those with the lowest Young’s modulus (b), intermediate Young’s modulus (c), and the highest Young’s modulus (d). PVA-H: poly(vinyl alcohol) hydrogel Helical and straight stent models. (a) High-helical model; (b) Moderate-helical model; (c) Mild-helical model; (d) Straight model. Each model includes cylindrical end segments (inner diameter 4 mm, length 27 mm). Curvature and lumen diameter are listed in Table 1. Flow-path Settings for color-coded circulation A pump (ALPHA FLOW EC−2, Fuyo Corporation, Tokyo, Japan) delivered a constant flow rate of 200 mL/min of 40% glycerin (working fluid). Figure 3 depicts the setup; the flow direction is from bottom to top (dotted arrow). The pump inflow (black arrow) and one end of a Y-connecter were connected to a 1 m urethane tube (outer diameter 8 mm, inner diameter 5 mm). A straight acrylic tube (outer diameter 6 mm, inner diameter 4 mm, length 20 cm; white arrowhead) was attached before and after the stent model (white arrow) to ensure fully developed inflow and stabilize outflow conditions. The other Y-connector port held a 4.2-Fr straight catheter (100 cm, Hanako Medical, Saitama, Japan, black arrowhead) for contrast injection. A 1 m urethane tube connected the acrylic tube downstream, leaving the flow path open. Flow path settings for color-coded circulation. Black arrow, inflow from the pump; black arrowhead, 4-Fr straight catheter; *, pump pressure sensor; white arrow, stent models; white arrowhead, straight section. The flow direction is from bottom to top within the image (dotted arrow). Color-coded circulation Angiography was performed on an X-ray angiography system (Alphenix Sky, Canon Medical System, Tochigi, Japan). Contrast (300 mgI/mL, Omnipaque, GE HealthCare Japan, Tokyo, Japan) was injected at 2 mL/sec for 10 mL using an injector (PRESS DUO elite, NEMOTO & Co., Ltd., Tokyo, Japan). DSA parameters were as follows: six frames per second, 15-second exposure, 80 kVp, 20 mA, automatic dose rate control, tabletop height 95 cm, source-image distance 100 cm, field 12 × 12 inches. Color-coded circulation displays static color maps of contrast arrival times via parametric imaging and produces videos on a workstation (Alphenix WorkStation Pro, Canon Medical System, Tochigi, Japan). Flow-path settings for 4D flow MRI The same pump delivered 200 mL/min of 40% glycerin through a closed-circuit urethane tubing (inflow and outflow 10 m each) and an interposed 20 cm straight acrylic tube before and after the stent model. 4D flow MRI Streamline images were obtained on a 3-T scanner (Ingenia Elition X, Philips Medical Systems, Best, The Netherlands) with SYNAPSE VINCENT (Fujifilm Medical, Tokyo, Japan). Four-dimensional flow MRI employed an electrocardiogram-gated, time-resolved, 3D phase-contrast MR sequence with 3-directional velocity encoding (parameters: TR/TE/flip angle 6.4 ms/4.6 ms/12°, k-space segmentation/temporal resolution 9.3 ms/51.5 ms, field of view 380 × 380 mm, slice thickness 3.0 mm, voxel size 1.83 × 3.26 × 3.00 mm3). Acquisition time was approximately 3.08 min, producing 10 time frames per R-R interval; electrocardiogram synchronization was 60 bpm. Assessment Comparison of intraluminal flow in stent models We examined whether swirling flow appeared on streamline images of the helical models and how the color-coded circulation differed between helical and straight lumens. DSA versus color-coded circulation For both moderate-helical and straight models, early (2 seconds) and delayed (9 seconds) DSA images were compared with color-coded circulation images. Referring to the DSA image of the straight model, the time when the contrast reached the distal end of the straight model was defined as the early phase, and the time when the center of the distal end of the straight model was almost filled with contrast was defined as the late phase. We hypothesized that early DSA reflects high-flow-velocity regions, and delayed DSA reflects low-flow-velocity regions. Differences in the distribution of assumed high- and low-flow-velocity components were assessed on color-coded circulation images. Mean flow velocity distribution Uneven mean flow velocity distribution was evaluated in axial sections at the center of both moderate-helical and straight models using 4D flow MRI mean-velocity images. Results Swirling flow was evident in the helical models on streamline images (Figure 4a-c), but it was absent in the straight model, which showed laminar flow (Figure 4d). As curvature decreased, swirling flow decreased. On color-coded circulation, the helical models displayed eccentric color patterns (Figure 4e-g), whereas the straight models showed colors distributed centrally and peripherally (Figure 4h). Supplementary material shows the 4D flow MRI streamline images and color-coded circulation in the high-helical model. In the moderate-helical model, the assumed high- and low-velocity components were eccentrically located (Figure 5a-c), while in the straight model, the assumed high-velocity component occupied the center and the assumed low-velocity component was located at the periphery (Figure 5d-f). Assumed high and low-velocity components on DSA were consistent with streamline findings (Figure 4b and d). Mean flow-velocity axial images from 4D flow MRI likewise showed an eccentric high-velocity component in the moderate-helical model (Figure 6a) and a centrally located high-velocity component in the straight model (Figure 6b). Comparison of intraluminal flow in stent models. (a-d) Streamline images; (e-h) color-coded circulation images. Images were acquired for the high-helical (a, e), moderate-helical (b, f), mild-helical (c, g), and straight (d, h) models. In the streamline images, swirling flow is seen in the helical models (a-c) and laminar flow in the straight model (d). In color-coded circulation images, eccentric colors appear in the helical models (e-g), whereas colors are observed in the center and periphery in the straight model (h). The flow direction is from bottom to top within the image (dotted arrow). Electronic supplementary material shows the 4D flow MRI streamline images and color-coded circulation in the high-helical model. Color-coded circulation does not have a color bar, but displays colors in a cyclical manner, making it easier to visually understand hemodynamics. 4D: 4-dimensional Comparison of DSA and color-coded circulation images in moderate-helical and straight models. (a) DSA early phase, (b) DSA delayed phase, and (c) color-coded circulation images of the moderate-helical model; (d) DSA early phase, (e) DSA delayed phase, and (f) color-coded circulation images of the straight model. In the moderate-helical model (a-c), assumed high- and low-velocity components are distributed eccentrically, whereas in the straight model (d-f), the assumed high-velocity region is located in the center, and the assumed low-velocity component region is located in the periphery. The flow direction is from bottom to top within the image (dotted arrow). Mean flow velocity axial images from 4D flow MRI. (a) Moderate-helical model; (b) straight model. In the moderate-helical model (a), the high-velocity component is eccentric, whereas in the straight model (b) it is centrally located. The flow direction is from front to back within the image. 4D: 4-dimensional Discussion Color-coded circulation based on time-to-arrival might reflect high- and low-velocity components. However, color-coded circulation is a parametric map based on contrast arrival time, not a direct velocity map. It is thought to depend on contrast injection conditions and concentration. In the present study, the areas showing early contrast arrival in DSA images appeared to reflect relatively high flow-velocity regions compared to 4D flow MRI streamline images, but this might hold true under certain conditions, such as flow rate and contrast injection speed. Four-dimensional flow MRI confirmed that swirling flow exhibited eccentric high- and low-velocity regions. Separation of red and blue on color Doppler has been reported to indicate swirling flow [6], and our findings were similar. The visual correspondence between the streamline images obtained by 4D flow MRI and color-coded circulation images suggests the possibility of easily extracting hemodynamic information, such as relative flow velocity distribution, from DSA, which is widely used in daily clinical practice. From a clinical perspective, color-coded circulation images can be used to rapidly evaluate hemodynamic characteristics during angiography. Furthermore, by comparing hemodynamics before and after stent placement under the same conditions, it may be possible to evaluate the relative effectiveness of treatment and confirm the appropriateness of device placement. Although color-coded circulation images cannot be positioned as a direct substitute for 4D flow MRI, they have the potential to be used in clinical practice as a complementary tool, emphasizing immediacy and versatility. In the present study, quantitative evaluations or comparisons of flow velocity and other parameters were not performed between models. For 4D flow MRI imaging, the MRI bed position had to be moved due to changes in the placement of the models. As a result, uniformity correction using a reference scan was applied for each scan, making it difficult to compare the models under completely identical conditions. Furthermore, the version of 4D flow analysis software available in the present study required manual operations for shape extraction, making it difficult to consistently calculate quantitative indices such as velocity. Furthermore, the analysis software was unable to calculate quantitative values for wall shear stress. Given these technical limitations, the present study focused only on qualitative and comparative evaluations of swirling and laminar flow visualized by color-coded circulation and 4D flow MRI. In 4D flow MRI, the combination of metal artifact reduction techniques and optimized imaging protocols is likely to increase the feasibility of evaluating hemodynamics within nitinol stent lumens. However, due to the residual influence of magnetic susceptibility artifacts, it is currently difficult to quantitatively evaluate stent lumen flow with high accuracy using 4D flow MRI. Furthermore, because the target vessel diameter of the femoropopliteal artery is small, high spatial resolution data must be acquired, resulting in extended imaging times. These limitations restrict the clinical applicability of 4D flow MRI. However, with advances in metal artifact reduction and the use of technology to reduce imaging times, it is expected that it will be widely applied in clinical settings in the future. Swirling flow produces a high-velocity component eccentric to the wall, generating high wall shear stress. By contrast, laminar flow in the straight model has a peripheral low velocity, resulting in low wall shear stress. The BioMimics 3D stent's helical centerline geometry enhances wall shear stress by maintaining or imposing natural arterial curvature [1], thereby preventing atherosclerosis and restenosis [4]. Straight stents reduce the artery's natural curvature, obstructing blood flow and creating regions of low wall shear stress [7]. Animal studies have shown that helical stents reduce intimal hyperplasia, underscoring the superiority of helical geometry [8]. MIMICS-3D, a European real-world registry, demonstrated that patients receiving 3D helical stents achieved 3-year outcomes comparable to those with drug-eluting stents [9]. After propensity-score matching, clinically driven target lesion vascularization and primary patency with 3D helical stents were unaffected by unfavorable lesion characteristics such as chronic total occlusion, chronic limb ischemia, and severe calcification [10]. There were several limitations. First, optimal flow-path settings are unknown; we used a flow rate of 200 mL/min and clinically applicable imaging parameters. Second, only steady-state flow was evaluated; further work should test pulsatile flow. Reproducing pulsatile flow was attempted, but air mixing occurred during DSA imaging performed in an open circuit, preventing proper DSA imaging. Future circuit design modifications are necessary and remain a topic for reliable pulsatile flow investigation. In the present study, considering these limitations, evaluating steady-state flow conditions was the focus. Both color-coded circulation derived from DSA and 4D flow MRI streamline images were used to relatively assess the hemodynamics of swirling flow and laminar flow. Third, resin stent models do not replicate in vivo vessel stiffness, and we assumed no luminal change. Nonetheless, evaluating three different helical shapes with 3D-printed models mimicking vascular stiffness enabled reliable 4D flow MRI visualization. Such models may support image-based evaluation of intravascular hemodynamics and optimization of stent design and angiographic parameters. Conclusion A swirling flow was observed in the helical models and a laminar flow in the straight model. Color-coded circulation might reflect high- and low-velocity components consistent with streamline images, showing eccentric color in the helical models and central-to-peripheral color in the straight model. Author Contributions All authors contributed to the study conception and design. Stent model preparation was performed: KK, SS, and MO. Color-coded circulation and 4-dimensional flow MRI data collection and analysis were performed: MM, TK, YK, and HA. The first draft of the manuscript was written: MM and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Conflicts of Interest There are no conflicts of interest. Supplementary Material Acknowledgments We would like to thank Enago (www.enago.com) for English language editing. Funding Statement This study was partly supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number JP24K18798. References Associated Data Supplementary Materials