Simultaneous quantitative evaluation of both iliac arteries and veins via accelerated four-dimensional flow in healthy controls and patients with deep vein thrombosis: a pilot study

Introduction

Deep vein thrombosis (DVT) is a common multifactorial disease and a major cause of morbidity and mortality (1). The estimated annual incidence of DVT ranges from 50 to 160 cases per 100,000 person-years (2). Although risk factors of DVT including hypercoagulation, Virchow’s triad, and iliac vein compression syndrome (IVCS) (3) have been identified, the exact pathogenesis of DVT remains incompletely understood. Moreover, recent evidence suggests a potential bidirectional relationship between DVT and arterial diseases (4), underscoring the need for a greater understanding of DVT and the importance of accurate hemodynamic assessment in this field.

Four-dimensional (4D) flow magnetic resonance imaging (MRI) has been proven to be effective for the evaluation of various vascular disorders, although its current applications are primarily related to the assessment of cardiac, cerebral, and hepatic vasculature, including valvular diseases (5), aortopathies (6), congenital heart diseases (7), cerebrovascular stenosis (8), and hepatic diseases (9). However, its use in the pelvis remains limited due to its long acquisition time and a lack of population-based references. An abundance of research has demonstrated compressed sensing (CS), a widely used acceleration method, to be an effective 4D flow imaging approach. Garreau et al. reported that a CS factor of 7.6 yielded good agreement with the conventional generalized autocalibrating partially parallel acquisition (GRAPPA) method in a phantom aortic model (10). Aono et al. compared sensitivity encoding (SENSE), echo-planar imaging (EPI), and a CS factor of 12 for 4D flow acquisition, and their results showed that EPI and CS 12 could save 71–73% of time while obtaining accurate flow parameters (11). Ma et al. further tested a CS factor of 14.1 in a pulsatile phantom and found excellent agreement in net flow assessment (12). This evidence indicates the reliability of the 4D flow acceleration technique; however, no study has examined the application of accelerated 4D flow techniques for the evaluation of iliac vasculature.

Moreover, the majority of current 4D flow acquisition is performed with a single-velocity encoding (VENC) approach, which limits the simultaneous estimation of flow in vessels with different blood flow velocities. Therefore, dual- or multi-VENC 4D techniques have been developed to facilitate the evaluation of complex flow environments within a limited scan time. In their study, Kroeger et al. used multi-VENC 4D flow MRI acquisition to evaluate both the ascending aorta and main pulmonary artery, demonstrating the feasibility of multi-VENC 4D flow MRI for complex cardiac flow assessment within a clinically acceptable scan duration (13). Kroeger et al.’s subsequent study further confirmed that multi-VENC 4D flow MRI enhances image quality and improves interreader agreement for vortex evaluation in the pulmonary artery (14). Collectively, these findings highlight the clinical potential of 4D flow with multiple VENCs. The imaging of iliac vasculature, which includes both iliac arteries and veins and a comprehensive flow environment, will likely benefit from this technique, but further investigation is needed to clarify its value.

To our knowledge, no prior study has evaluated the application of CS 4D flow MRI in examining the iliac vasculature, nor has any multi-VENC 4D flow MRI sequence been tested for complex flow assessment in this region. Therefore, this study aimed to evaluate the performance of CS-accelerated dual-VENC 4D flow in assessing both iliac arteries and veins in healthy controls (HCs) and patients with DVT. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1886/rc).

Methods

The visual summary of this study is presented in Figure 1.

Figure 1 Schematic diagram of 4D flow MRI analysis. (A) Image acquisition for both conventional and CS-accelerated 4D flow at the same position to include the iliac vasculature. (B) Segmentation and processing of both veins and arteries for flow parameter analysis. (C) Phase synchronization and analysis, the peak phase was defined as the phase with the highest mean velocity among three consecutive phases at the AA level and was designated as new phase 10. 4D, four-dimensional; AA, abdominal aorta; CS, compressed sensing; GRAPPA, generalized autocalibrating partial parallel acquisition; MRI, magnetic resonance imaging; VENC, velocity encoding.

Study population

This prospective study enrolled HCs and patients with DVT who underwent 4D flow MRI of the iliac vasculature at Beijing Chaoyang Hospital between December 2024 and April 2025. The inclusion criteria for HCs were (I) age >18 years and (II) no known cardiovascular, cerebrovascular, systemic, or malignant diseases; the exclusion criterion for HCs was claustrophobia. Meanwhile, the inclusion criteria for patients with DVT were as follows: (I) age >18 years; (II) a confirmed DVT diagnosis via ultrasound, computed tomography (CT), or MRI within 1 week before the 4D flow MRI examination; and (III) no surgical thrombectomy before MRI. The exclusion criteria for patients were (I) accompanying malignancy, (II) pregnancy, and (III) contraindications to magnetic resonance (MR) scans.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and was approved by the Ethics Committee of Beijing Chaoyang Hospital (No.: 2024-11-29-4). The registered clinical trial ID is ChiCTR2500099260. All participants provided written informed consent.

MRI acquisition

All participants underwent MR scans with a 3T system (MAGNETOM Vida, Siemens Healthineers, Erlangen, Germany). Three sequential acquisitions, including two conventional single-VENC 4D flow MRI sequences with both high and low VENCs, together with a CS-accelerated dual-VENC research 4D flow MRI sequence, were performed without contrast injection, with random repeat acquisitions performed for the test-retest reliability assessment. Participants were scanned in the supine position with a 32-channel spine coil and an 18-channel body phased-array coil. All 4D flow MRI data were reconstructed to 25 cardiac frames via retrospective pulse oximeter gating. Initial VENC scout images were acquired at the abdominal aorta (AA) and inferior vena cava (IVC) levels with VENC settings of 50 and 120 cm/s. If aliasing was observed, the VENC was increased by 20 cm/s until a radiologist confirmed its absence. In one participant, all 4D flow MRI acquisitions maintained identical resolution, field of view (FOV), and oversampling parameters. The acquisition parameters were the following: acquisition resolution =1.50–1.76 × 1.50 × 1.50–1.76 mm³, FOV =112 (foot-head) × 144 (right-left) × 54–90 (anterior-posterior) mm³ (adjusted for vascular curvature), and phase oversampling = 25–50% (based on body size). For conventional 4D flow, the GRAPPA factor was 3, and for CS 4D flow MRI, the CS factor was 8.9 (Table S1).

Image processing

One experienced radiologist (Qiming Liu, with 5 years of experience in cardiac imaging) initially excluded images with significant artifacts and performed postprocessing using the freely available 4D-Flow-Matlab-Toolbox (MathWorks, Natick, MA, USA). This study also adhered to the recommendations of the most recent 4D flow consensus statement (15). The data processing included background phase offset correction (16), the semiautomatic segmentation of iliac arteries and veins, the tetrahedral finite element mesh generation (17), the velocity vector interpolation at each node of the mesh using linear interpolation, and the quantification of mean velocity and wall shear stress (WSS) (18,19). The quantification of hemodynamic parameters was carried out with a finite element approach, which was based on the least square stress projection method and Laplacian method (20).

For each participant, five arterial segments [including the AA, left common iliac artery (LCIA), right common iliac artery (RCIA), left external iliac artery (LEIA), and right external iliac artery (REIA)] and five venous segments [including the IVC, left common iliac vein (LCIV), right common iliac vein (RCIV), left external iliac vein (LEIV), and right external iliac vein (REIV)] were manually selected for calculation of the hemodynamic parameters.

For each participant, the mean values of each parameter were calculated as follows:

m¯k=∑i=1Npi×hi∑i=1Nhi[1]

where m¯k is the weighted average (mean) of each parameter in a segment k, and pi, and hi are the parameter and the Voronoi volume at node i in the segment k, respectively.

To obtain the mean values of WSS, only the wall nodes were considered, and the relative mean error was calculated as follows:

Δm¯k=m¯k,CO−m¯k,CSm¯k,CO×100%[2]

where m¯k,CO is the mean value of the parameter k under the conventional single-VENC acquisition, and m¯k,CS is the mean value of the parameter k under the CS dual-VENC acquisition. Each segment k corresponds to one of the vascular segments.

Phase synchronization

In this study, retrospective pulse oximeter gating was employed, resulting in various peak velocity phases due to participants’ pulse wave velocity, height, vessel curvature, heart rate, and other conditions. To standardize the analysis, the peak phase was defined as the phase with the highest mean velocity across three consecutive phases at the AA level and was designated as the new phase 10 (out of a total of 25 phases), which was defined as the “systole”, while phase 20 was defined as the “diastole” (21).

Quality control of quantitative 4D flow parameters

Random repeat acquisitions were performed 18, 19, and 16 times for single high-VENC, single low-VENC, and CS-accelerated dual-VENC acquisition, and these data were used for test-retest reliability examination. Moreover, to improve the reliability of this study, approximately 25% of HCs were randomly selected for interobserver and intra-observer correlation analyses (conducted by Qifan Lu, a cardiologist with 5 years of experience in cardiac imaging). The re-evaluation for intraobserver correlation analysis was performed after an interval of at least 30 days.

Clinical data

The left ventricular ejection fraction (LVEF) and pulmonary hypertension status were obtained from echocardiography data, and the information on pulmonary embolism was obtained through CT. For the description of DVT, the lower extremity deep veins on each side were divided into five segments: the common iliac, external iliac, femoral, popliteal, and calf veins. The IVC was assessed as a separate, non-laterality-specific segment. All images were independently reviewed by an experienced radiologist (X.W., with 7 years of experience) to identify thrombus in each segment. The location of the most proximal segment of each patient was recorded.

Statistical analysis

Statistical analysis was performed with SPSS 23.0 (IBM Corp., Armonk, NY, USA) and Python 3.7.10 (Python Software Foundation, Wilmington, DE, USA) and its associated packages (including pingouin, sklearn, and scipy). The Kolmogorov-Smirnov test was used to evaluate the normality of continuous data. Normally distributed continuous variables are expressed as the mean ± standard deviation (SD), while nonnormally distributed variables are presented as the median and interquartile range (IQR). Meanwhile, categorical variables are reported as frequencies (%). Flow parameters were compared via paired-samples t tests or Wilcoxon rank-sum tests, as appropriate. Intraclass correlation coefficients (ICCs) were used to assess test-retest reliability, interreader agreement, and intrareader agreement under the following scheme: 0.0–0.3, no agreement; 0.31–0.5, weak agreement; 0.51–0.7, moderate agreement; 0.71–0.9, strong agreement; and 0.91–1.00, very strong agreement. The Pearson correlation coefficient (R) was used to evaluate the correlations between measurements obtained with different techniques and the association between key flow parameters and age, with R≥0.7 indicating strong correlation, R>0.3 but <0.7 indicating moderate correlation, and R<0.3 indicating weak correlation. A P value <0.05 was considered statistically significant.

Results

Baseline characteristics

A total of 62 HCs and 20 patients with DVT were initially enrolled. After the exclusion criteria were applied, 47 HCs (median age 49 years; 66% male) and 15 patients with DVT (median age 67 years; 60% male) were included in the final analysis (Figure 2). The age of patients with DVT (median age 67.0 years, IQR 50.0–75.0 years) was significantly higher compared with the HCs (median age 49.0 years, IQR 28.5–57.0 years) (P<0.001), while other baseline characteristics were similar between the two groups. The LVEF results indicated that all patients with DVT had normal cardiac function (all >55.0%; median 68.0%, IQR 66.5–71.5%), with three patients with pulmonary hypertension and four patients with pulmonary embolism (Table 1).

Figure 2 Flow diagram of the study datasets. Twenty patients with DVT and 62 health controls were initially screened; after inclusion and exclusion, 15 patients with DVT and 47 healthy controls were included in the final analysis. 4D, four-dimensional; DVT, deep vein thrombosis; MR, magnetic resonance.

The detailed analysis of DVT showed that 1, 9, and 6 thrombi were located at the IVC, the left side deep veins, and the right side deep veins, respectively. Most thrombi were present at the femoral vein (n=6, 40.0%), while 3, 3, 2, and 1 thrombi were present at the common iliac vein, popliteal vein, calf vein, and IVC, respectively (Table 1).

The average acquisition times for conventional high-VENC, conventional low-VENC, and CS dual-VENC 4D flow MRI were 10.6, 10.7, and 8.8 min, respectively, demonstrating a significant improvement in acquisition efficiency with an average reduction of 59%. Representative 4D flow visualization of the iliac arteries and veins with conventional and CS-accelerated techniques in an HC and a patient with DVT is shown in Figure 3, which also provides DVT visualized via the MR black-blood thrombus imaging technique.

Figure 3 Visualization of velocity mapping of the iliac vasculature in an HC and a patient with DVT. (A) Conventional flow mapping of an HC. (B) CS-accelerated flow mapping of an HC. (C) Conventional flow mapping of a patient with DVT. (D) CS-accelerated flow mapping of a patient with DVT. (E) DVT visualized with the MRBTI technique (yellow triangle). CS, compressed sensing; DVT, deep vein thrombosis; HC, healthy control; MRBTI, magnetic resonance black-blood thrombus imaging.

Quality control

In the test-retest reliability analysis, arterial velocity with conventional high-VENC and CS high-VENC imaging yielded ICCs ranging from 0.832 to 0.952 (all P values <0.001), while venous parameter quantification with low-VENC imaging yielded ICCs of 0.898 to 0.981 (all P values <0.001). Systolic WSS quantification produced similar results: except for RCIA WSS assessed with the conventional method (ICC =0.808; P=0.001), all WSS parameters exhibited ICCs >0.8 and P<0.001, indicating strong technical reproducibility for both conventional and CS methods. Intra- and inter-observer correlation analyses revealed comparable findings, with arterial and venous velocity and WSS measurements demonstrating ICCs >0.85 and P<0.001, confirming robust flow parameter assessment across experienced readers (Table S2).

The correlation between measurements obtained from interobserver, intraobserver, and test-retest experiments was strong. The R values all exceeded 0.9 for intra- and inter-observer velocity parameters and were greater than 0.85 for test-retest velocity parameters (all P values <0.001).

Comparison of flow parameters obtained with conventional and CS methods

The flow measurements of arterial segments at systole are presented in Figure 4A,4B and Figure S1A,S1B. Specifically, in HCs, the velocity at the AA measured with the conventional method (median 0.46 m/s, IQR 0.4–0.53 m/s) was higher than that with the CS method (median 0.44 m/s, IQR 0.39–0.51 m/s) (P<0.001). Similarly, measured WSS was also significantly higher under conventional methods (median 1.45, IQR 1.3–1.75) than with the CS method (median 1.44, IQR 1.19–1.64) (P<0.001) (Table 2). In patients with DVT, no significant differences were observed in velocity (conventional: median 0.41 m/s, IQR 0.39–0.46 m/s; CS: median 0.4 m/s, IQR 0.38–0.46 m/s; P=0.85) or WSS (conventional: median 1.31, IQR 1.17–1.49; CS: median 1.3, IQR 1.17–1.43; P=0.622) at the AA due to limited sample size, although the tendency to underestimation remained (2.1% for velocity and 5.1% for WSS).

Figure 4 Comparison of velocity and WSS evaluated with CS and conventional methods. (A) AA velocity. (B) AA WSS. (C) IVC velocity. (D) IVC WSS. The red lines indicate the parameters obtained with the conventional method, while the green lines indicate the parameters obtained with the CS-accelerated method. AA, abdominal aorta; CS, compressed sensing; IVC, inferior vena cava; WSS, wall shear stress.

The flow measurements of veins at systole are presented in Figure 4C,4D and Figure S1C,S1D. Similar to arterial flow parameters, all parameters in the HCs exhibited underestimation under the CS method (0.0% to 13.2%). In the DVT group at the IVC level, conventional measurements yielded a significantly higher peak velocity (conventional: median 0.12 m/s, IQR 0.06–0.15 m/s; CS: median 0.09 m/s, IQR 0.07–0.14 m/s; P=0.009) and WSS (conventional: median 0.37, IQR 0.21–0.54; CS: median 0.31, IQR 0.23–0.51; P=0.016). For HCs, conventional measurements exhibited a higher peak velocity (median 0.1 m/s, IQR 0.07–0.13 m/s; CS: median 0.09 m/s, IQR 0.07–0.12 m/s; P=0.078) and WSS (median 0.3, IQR 0.24–0.43; CS: median 0.3, IQR 0.21–0.39; P=0.109) compared to CS methods, but not significantly so (Table 3).

Bland-Altman plots were also generated to compare the differences in flow parameters between HCs and patients with DVTs. The results indicated that for velocity, the mean difference for the AA and IVC were 0.02 m/s [95% confidence interval (CI): 0.00–0.03 m/s] and 0.01 m/s (95% CI: 0.00–0.01 m/s), respectively; for WSS, the mean differences for AA and IVC were 0.08 N/m² (95% CI: 0.03–0.12 N/m²) and 0.02 N/m² (95% CI: −0.01 to 0.04 N/m²), respectively (Figure S2).

Correlation between age and flow parameters

In the correlation analysis, there were significant negative correlations were observed between the age of HCs and velocity (R=−0.33; P=0.024) and WSS (R=−0.34; P=0.019) of the AA, while for patients with DVT, age demonstrated nonsignificant negative correlations with velocity (R=−0.25; P=0.366) and WSS (R=−0.23; P=0.408) (Figure 5A,5B).

Figure 5 Correlation between age and flow parameters. (A) AA velocity. (B) AA WSS. (C) IVC velocity. (D) IVC WSS. The green color represents the measurements and correlation in HCs, and the red color represents the measurement and correlation in patients with DVT. AA, abdominal aorta; DVT, deep vein thrombosis; HCs, healthy controls; IVC, inferior vena cava; WSS, wall shear stress.

Regarding the IVC in HCs, there was a significant negative correlation between velocity and age (R=−0.29; P=0.047); however, the correlation between WSS and age was not significant (R=−0.21; P=0.158). Meanwhile, in patients with DVT, although no statistical significance was observed, there were positive correlations between velocity/WSS and age (Figure 5C,5D).

Discussion

This study systematically evaluated and quantified flow parameters in iliac arteries and veins using conventional and dual-VENC 4D flow MRI techniques in both HCs and patients with DVT. The results demonstrated that both the conventional and CS techniques had robust performance when applied to the iliac vasculature; moreover, negative correlations between velocity/WSS and age were confirmed in HCs. In contrast, in patients with DVT, a nonsignificant positive correlation was found between venous velocity/WSS and age, a finding that needs to be scrutinized further.

High spatial and temporal resolution is key to accurate flow estimation, and according to the expert consensus on 4D flow MRI (15), the optimal spatial resolution for large vessels should be 2 mm³, and the temporal resolution should not exceed 50 ms. For iliac vasculature, an even higher spatial resolution is required due to the smaller vessel diameters. However, without acceleration, this strict acquisition protocol often results in prolonged scan times in clinical practice which limits its application. Ghibes et al. reported a median acquisition time of 12 min and 12 s for venous imaging with a spatial resolution of 1.5×1.5×1.6 mm³ and a temporal resolution (22) of 74.6 ms; meanwhile, You et al. used a coarser resolution of 1.7×2.1×3.0 mm³ for flow quantification in the LCIA and LCIV (23). In our study, we adopted a protocol with high temporal (~40 ms) and spatial (1.5 mm isotropic) resolution to evaluate the feasibility of 4D flow in the pelvis. We found that dual-VENC CS 4D flow MRI could reduce the acquisition time by 59% time and produce highly consistent results for visualization and velocity/WSS assessment. This parameter setting also improved region-of-interest segmentation accuracy. Moreover, in clinic, patients with DVT often experience limb pain, swelling, or reduced mobility. A shorter examination time could improve patient comfort, reduce motion artifacts, and enhance the feasibility of follow-up imaging acquisition in symptomatic or acute settings. Another practical challenge was the choice of triggering technique. Electrocardiogram gating was initially tested; however, low VENC increased signal interference, leading to corrupted triggering and low-quality data (22,24). Consequently, pulse oximeter gating was selected, and combining the phase synchronization processing proposed in this study, we developed a complete acquisition-processing-analysis pipeline for 4D flow imaging of the iliac vasculature. This study also found that the CS 4D technique flow underestimates velocity and WSS as compared to GRAPPA methods, with the percent error ranging from 1% to 14%, a finding consistent with previous studies. For instance, Ikoma et al. found that high-VENC 4D flow for pulmonary artery evaluation may result in underestimation of flow parameters (25), and Varga-Szemes et al. reported a 10% underestimation of flow measurements with a CS factor of 7.7 (26). These findings suggest that the comparison of 4D flow measurements acquired with different techniques should be treated with caution, especially in clinical scenarios. Moreover, we believe that the underestimation observed in the CS techniques should be carefully considered, especially in future multicenter studies in which data can be collected with different MR acquisition protocols. This underestimation also supports the development of a novel algorithm that can convert the CS-based results into GRAPPA-based ones, thus facilitating the comparison of CS values and GRAPPA values.

Vascular aging is also an important factor in the evaluation of flow parameters. Roberts et al. found that with aging, cerebral blood flow reduces while pulsatility increases (24). Ramaekers et al. reported similar findings in the aorta, with both velocity and WSS decreasing with age (27). We also analyzed the correlation of the flow parameters of the AA and IVC with age, and results in HCs were highly consistent with those of previous studies (24,27), confirming the presence of iliac vascular aging. Interestingly, a nonsignificant positive correlation of WSS at the IVC level with age was observed in patients with DVT, indicating that DVT formation may play an important role in vascular remodeling, especially the venous system; however, this remains speculative and further validation is required.

Functional change of vessels is a key factor in the development of DVT, and WSS is a sensitive parameter for the evaluation of vascular change (28). Wang et al. demonstrated that iliac veins with higher time-averaged WSS and lower relative residence time exhibit narrowed luminal areas, which could impede blood flow and promote the accumulation of blood components, thereby increasing DVT risk (29). Meanwhile, Lyu et al. found that positive WSS divergence may promote thrombus formation in patients with intracranial aneurysms (30). Although our study did not directly compare flow parameters between HCs and patients with DVT due to age differences and the limited sample size, the positive correlation between age and flow parameters observed in patients with DVT suggests that future research should focus more on patients with DVT. Furthermore, different thrombus locations and extent may alter venous hemodynamics, although we did not quantify the effect of thrombus location on hemodynamic change, our example case presented in Figure 3C,3D, suggests that hemodynamic alterations may also serve as valuable indicators for thrombus detection.

Besides MR, the evaluation of the iliac vasculature can also be performed using other imaging modalities. For instance, Assi et al. found that ultrasound and computational fluid dynamics (CFD) methods had excellent performance in assessing the iliac veins in patients with IVCS (31). Moreover, Guo et al. reported that a positive iliac vein ultrasound finding was associated with a higher risk of concurrent IVCS (32). Computed tomography venography (CTV) is another widely used modality for quantifying the iliac vasculature. Zhang et al. identified female and advanced age as risk factors for IVCS (33), while Jiang et al. found that hemodynamic changes derived from CTV and CFD could aid in predicting iliac vein stenosis (34). In contrast to CTV, 4D flow MR is contrast-free and can record dynamic flow information, and compared with ultrasound, it provides higher objectivity and a larger FOV for minimizing the operator dependence. Additionally, flow patterns can be directly calculated via 4D imaging, avoiding potential CFD simulation errors.

Certain limitations to this study should be acknowledged. First, only one CS factor and the conventional GRAPPA acceleration technique were examined in this study due to time constraints, and future studies should include and compare other acceleration methods. Second, we conducted a pilot study focused on technical validation; therefore, certain baseline characteristics were not measured. In future large-scale, prospective studies, cardiac function status and vascular indices (e.g., cardio-ankle vascular index, ankle-brachial index, and pulse wave velocity) should be obtained along with 4D flow MRI to better characterize the association between hemodynamic parameters and the underlying DVT mechanisms. Finally, the comparison of flow dynamics should be compared in a large, age- and sex-matched cohort, and the hemodynamic alteration under DVT conditions should be better quantified.

Conclusions

This study evaluated and quantified the flow parameters of iliac arteries and veins in HCs and patients with DVT using conventional and dual-VENC 4D flow techniques. Although underestimation was observed, dual-VENC 4D flow enabled comprehensive velocity and WSS assessment. However, the WSS measurement exhibited insufficient stability, indicating a need for further sequence optimization.

Acknowledgments

The authors would like to thank all volunteers for their participation in this study. J.S. thanks to Millennium Science Initiative Program—ICN2021_004 and Department of Medical Imaging and Radiation Sciences at Monash University.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1886/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1886/dss

Funding: This research received funding from National Science Fund of National Natural Science Foundation of China (No. 82025018) and Beijing Hospitals Authority Clinical Medicine Development of Special Funding Support (No. ZLRK202306).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1886/coif). C.Z. is from Siemens Healthineers and G.D. is from Siemens Healthineers AG. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by Ethics Committee of Beijing Chaoyang Hospital (No.: 2024-11-29-4). All participants provided written informed consent.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Colling ME, Tourdot BE, Kanthi Y. Inflammation, Infection and Venous Thromboembolism. Circ Res 2021;128:2017-36. [Crossref] [PubMed]
2. Raya-Benítez J, Heredia-Ciuró A, Calvache-Mateo A, Martín-Núñez J, Valenza-Peña G, López-López L, Valenza MC. Effectiveness of non-instrumental early mobilization to reduce the incidence of deep vein thrombosis in hospitalized patients: A systematic review and meta-analysis. Int J Nurs Stud 2025;161:104917. [Crossref] [PubMed]
3. Lüscher TF, Davies A, Beer JH, Valgimigli M, Nienaber CA, Camm JA, Baumgartner I, Diener HC, Konstantinides SV. Towards personalized antithrombotic management with drugs and devices across the cardiovascular spectrum. Eur Heart J 2022;43:940-58. [Crossref] [PubMed]
4. Donadini MP, Calcaterra F, Romualdi E, Ciceri R, Cancellara A, Lodigiani C, Bacci M, Della Bella S, Ageno W, Mavilio D. The Link Between Venous and Arterial Thrombosis: Is There a Role for Endothelial Dysfunction? Cells 2025;14:144. [Crossref] [PubMed]
5. Spampinato RA, Jahnke C, Crelier G, Lindemann F, Fahr F, Czaja-Ziolkowska M, Sieg F, Strotdrees E, Hindricks G, Borger MA, Paetsch I. Quantification of regurgitation in mitral valve prolapse with four-dimensional flow cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2021;23:87. [Crossref] [PubMed]
6. Franco P, Sotelo J, Guala A, Dux-Santoy L, Evangelista A, Rodríguez-Palomares J, Mery D, Salas R, Uribe S. Identification of hemodynamic biomarkers for bicuspid aortic valve induced aortic dilation using machine learning. Comput Biol Med 2022;141:105147. [Crossref] [PubMed]
7. Bai W, Luo S, Xu H. Combining 4D flow magnetic resonance imaging with cardiac catheterization for pre-operative blood flow assessment in an older child with complex congenital heart disease. Eur Heart J 2024;45:4868. [Crossref] [PubMed]
8. Zarrinkoob L, Myrnäs S, Wåhlin A, Eklund A, Malm J. Cerebral Blood Flow Patterns in Patients With Low-Flow Carotid Artery Stenosis, a 4D-PCMRI Assessment. J Magn Reson Imaging 2024;60:1521-9. [Crossref] [PubMed]
9. Hyodo R, Takehara Y, Naganawa S. 4D Flow MRI in the portal venous system: imaging and analysis methods, and clinical applications. Radiol Med 2022;127:1181-98. [Crossref] [PubMed]
10. Garreau M, Puiseux T, Toupin S, Giese D, Mendez S, Nicoud F, Moreno R. Accelerated sequences of 4D flow MRI using GRAPPA and compressed sensing: A comparison against conventional MRI and computational fluid dynamics. Magn Reson Med 2022;88:2432-46. [Crossref] [PubMed]
11. Aono S, Tsuneta S, Nishioka N, Aoike T, Hirayama H, Ishizaka K, Kwon J, Yoneyama M, Fujima N, Kudo K. Comparison of Echo-Planar Imaging and Compressed Sensing in the Estimation of Flow Metrics from Aortic 4D Flow MR Imaging: A Healthy Volunteer Study. Magn Reson Med Sci 2025;24:241-52. [Crossref] [PubMed]
12. Ma LE, Markl M, Chow K, Huh H, Forman C, Vali A, Greiser A, Carr J, Schnell S, Barker AJ, Jin N. Aortic 4D flow MRI in 2 minutes using compressed sensing, respiratory controlled adaptive k-space reordering, and inline reconstruction. Magn Reson Med 2019;81:3675-90. [Crossref] [PubMed]
13. Kroeger JR, Pavesio FC, Mörsdorf R, Weiss K, Bunck AC, Baeßler B, Maintz D, Giese D. Velocity quantification in 44 healthy volunteers using accelerated multi-VENC 4D flow CMR. Eur J Radiol 2021;137:109570. [Crossref] [PubMed]
14. Kroeger JR, Stackl M, Weiss K, Baeßler B, Gerhardt F, Rosenkranz S, Maintz D, Giese D, Bunck AC. k-t accelerated multi-VENC 4D flow MRI improves vortex assessment in pulmonary hypertension. Eur J Radiol 2021;145:110035. [Crossref] [PubMed]
15. Bissell MM, Raimondi F, Ait Ali L, Allen BD, Barker AJ, Bolger A, et al. 4D Flow cardiovascular magnetic resonance consensus statement: 2023 update. J Cardiovasc Magn Reson 2023;25:40. [Crossref] [PubMed]
16. Chernobelsky A, Shubayev O, Comeau CR, Wolff SD. Baseline correction of phase contrast images improves quantification of blood flow in the great vessels. J Cardiovasc Magn Reson 2007;9:681-5. [Crossref] [PubMed]
17. Fang Q, Boas DA. Tetrahedral mesh generation from volumetric binary and grayscale images. 2009 IEEE International Symposium on Biomedical Imaging: From Nano to Macro; 28 June 2009 - 01 July 2009; Boston, MA, USA. IEEE; 2009:1142-5.
18. Sotelo J, Urbina J, Valverde I, Tejos C, Irarrazaval P, Andia ME, Uribe S, Hurtado DE. 3D Quantification of Wall Shear Stress and Oscillatory Shear Index Using a Finite-Element Method in 3D CINE PC-MRI Data of the Thoracic Aorta. IEEE Trans Med Imaging 2016;35:1475-87. [Crossref] [PubMed]
19. Sotelo J, Dux-Santoy L, Guala A, Rodríguez-Palomares J, Evangelista A, Sing-Long C, Urbina J, Mura J, Hurtado DE, Uribe S. 3D axial and circumferential wall shear stress from 4D flow MRI data using a finite element method and a laplacian approach. Magn Reson Med 2018;79:2816-23. [Crossref] [PubMed]
20. Sotelo J, Franco P, Guala A, Dux-Santoy L, Ruiz-Muñoz A, Evangelista A, Mella H, Mura J, Hurtado DE, Rodríguez-Palomares JF, Uribe S. Fully Three-Dimensional Hemodynamic Characterization of Altered Blood Flow in Bicuspid Aortic Valve Patients With Respect to Aortic Dilatation: A Finite Element Approach. Front Cardiovasc Med 2022;9:885338. [Crossref] [PubMed]
21. Bachani N, Vijay S, Vyas A, Jadwani J, Panicker G, Lokhandwala Y. The diastolic duration as a percentage of the cardiac cycle in healthy adults: A pilot study. Indian Heart J 2025;77:73-7. [Crossref] [PubMed]
22. Ghibes P, Martirosian P, Grözinger G, Plajer D, Estler A, Partovi S. Quantitative Flow Measurements of Pelvic Venous Vasculature Using 4D Flow MRI. Acad Radiol 2024;31:929-38. [Crossref] [PubMed]
23. You S, Masutani EM, Alley MT, Vasanawala SS, Taub PR, Liau J, Roberts AC, Hsiao A. Deep Learning Automated Background Phase Error Correction for Abdominopelvic 4D Flow MRI. Radiology 2022;302:584-92. [Crossref] [PubMed]
24. Roberts GS, Peret A, Jonaitis EM, Koscik RL, Hoffman CA, Rivera-Rivera LA, Cody KA, Rowley HA, Johnson SC, Wieben O, Johnson KM, Eisenmenger LB. Normative Cerebral Hemodynamics in Middle-aged and Older Adults Using 4D Flow MRI: Initial Analysis of Vascular Aging. Radiology 2023;307:e222685. [Crossref] [PubMed]
25. Ikoma T, Suwa K, Sano M, Ushio T, Saotome M, Ogawa N, Satoh H, Maekawa Y. Early changes of pulmonary arterial hemodynamics in patients with systemic sclerosis: flow pattern, WSS, and OSI analysis with 4D flow MRI. Eur Radiol 2021;31:4253-63. [Crossref] [PubMed]
26. Varga-Szemes A, Halfmann M, Schoepf UJ, Jin N, Kilburg A, Dargis DM, Düber C, Ese A, Aquino G, Xiong F, Kreitner KF, Markl M, Emrich T. Highly Accelerated Compressed-Sensing 4D Flow for Intracardiac Flow Assessment. J Magn Reson Imaging 2023;58:496-507. [Crossref] [PubMed]
27. Ramaekers MJFG, Te Kiefte BJC, Adriaans BP, Juffermans JF, van Assen HC, Winkens B, Wildberger JE, Lamb HJ, Schalla S, Westenberg JJM. Comprehensive sex-specific and age-dependent analysis of four-dimensional flow cardiovascular magnetic resonance assessed aortic blood flow-related parameters in normal subjects using single-vendor magnetic resonance systems and single-vendor software. J Cardiovasc Magn Reson 2024;26:101083. [Crossref] [PubMed]
28. El Sayed R, Park CC, Shah Z, Nahab FB, Haussen DC, Allen JW, Oshinski JN. Assessment of Complex Flow Patterns in Patients With Carotid Webs, Patients With Carotid Atherosclerosis, and Healthy Subjects Using 4D Flow MRI. J Magn Reson Imaging 2024;59:2001-10. [Crossref] [PubMed]
29. Wang H, Jia W, Xi Y, Li Y, Fan Y, Deng X, Chen Z. Morphometric and Hemodynamic Analysis of the Compressed Iliac Vein. J Endovasc Ther 2024;31:744-55. [Crossref] [PubMed]
30. Lyu Z, Gudino A, Dier C, Sagues E, Salinas I, Chiriboga G, Setia S, Shenoy N, Samaniego EA, Jiang J. Hemodynamic analysis of thrombosed intracranial aneurysms: a comparative correlation study. Neurosurg Rev 2025;48:417. [Crossref] [PubMed]
31. Assi IZ, Lynch SR, Samulak K, Williams DM, Wakefield TW, Obi AT, Figueroa CA. An ultrasound imaging and computational fluid dynamics protocol to assess hemodynamics in iliac vein compression syndrome. J Vasc Surg Venous Lymphat Disord 2023;11:1023-1033.e5. [Crossref] [PubMed]
32. Guo C, Gao S, Hu L, Shang D, Li Y. Predictive factors for iliac vein compression syndrome in patients with varicose veins. Vascular 2025;33:999-1006. [Crossref] [PubMed]
33. Zhang F, Song HX, He ZP, Zheng LH, Han YR, Wang BY, Liu P. Analysis of computed tomography venography for the diagnosis and endovascular treatment of iliac venous compression syndrome with venous leg ulcers: a retrospective study. Sci Rep 2024;14:22314. [Crossref] [PubMed]
34. Jiang XD, Ye SL, Zhang M, Li XQ, Sun LL. Clinical implications of hemodynamic analysis for the three-dimension iliac vein model with different stenosis. Heliyon 2023;9:e13681. [Crossref] [PubMed]