4D Flow MRI Research Articles

3D Vortex-Energetics in the Left Pulmonary Artery for Differentiating Pulmonary Arterial Hypertension and Pulmonary Venous Hypertension Groups Using 4D Flow MRI.

Pulmonary hypertension (PH) is a life-threatening. Differentiation pulmonary arterial hypertension (PAH) from pulmonary venous hypertension (PVH) is important due to distinct treatment protocols. Invasive right heart catheterization (RHC) remains the reference standard but noninvasive alternatives are needed. To evaluate 4D Flow MRI-derived 3D vortex energetics in the left pulmonary artery (LPA) for distinguishing PAH from PVH. Prospective case-control. Fourteen PAH patients (11 female) and 18 PVH patients (9 female) diagnosed from RHC, 23 healthy controls (9 female). 1.5 T; gradient recalled echo 4D flow and balanced steady-state free precession (bSSFP) cardiac cine sequences. LPA 3D vortex cores were identified using the lambda2 method. Peak vortex-contained kinetic energy (vortex-KE) and viscous energy loss (vortex-EL) were computed from 4D flow MRI. Left and right ventricular (LV, RV) stroke volume (LVSV, RVSV) and ejection fraction (LVEF, RVEF) were computed from bSSFP. In PH patients, mean pulmonary artery pressure (mPAP), pulmonary capillary wedge pressure (PCWR) and pulmonary vascular resistance (PVR) were determined from RHC. Mann-Whitney U test for group comparisons, Spearman's rho for correlations, logistic regression for identifying predictors of PAH vs. PVH and develop models, area under the receiver operating characteristic curve (AUC) for model performance. Significance was set at P < 0.05. PAH patients showed significantly lower vortex-KE (37.14 [14.68-78.52] vs. 76.48 [51.07-120.51]) and vortex-EL (9.93 [5.69-25.70] vs. 24.22 [12.20-32.01]) than PVH patients. The combined vortex-KE and LVEF model achieved an AUC of 0.89 for differentiating PAH from PVH. Vortex-EL showed significant negative correlations with mPAP (rho = -0.43), PCWP (rho = 0.37), PVR (rho = -0.64). In the PAH group, PVR was significantly negatively correlated with LPA vortex-KE (rho = -0.73) and vortex-EL (rho = -0.71), and vortex-KE significantly correlated with RVEF (rho = 0.69), RVSV, (rho = 0.70). In the PVH group, vortex-KE (rho = 0.52), vortex-EL significantly correlated with RVSV (rho = 0.58). These preliminary findings suggest that 4D flow MRI-derived LPA vortex energetics have potential to noninvasively differentiate PAH from PVH and correlate with invasive hemodynamic parameters. 1 TECHNICAL EFFICACY: Stage 3.

Development of a Porcine Model of Arteriovenous Fistula Venous Stenosis Treated with Percutaneous Transluminal Angioplasty

To develop a porcine model for arteriovenous fistula (AVF) venous stenosis treated with percutaneous transluminal angioplasty (PTA) and compare outcomes of plain balloon angioplasty (POBA) to IN.PACT paclitaxel drug-coated balloons (DCB). Twelve castrated male Yorkshire pigs (4-5 months, 35-45kg) underwent renal artery embolization to induce chronic kidney disease (CKD). Twenty-eight days later, AVF was created by anastomosing the left external jugular vein to left common carotid artery. The pigs were divided into a pilot group (n=6) for optimizing the AVF technique (euthanized at day 4) and a definitive group (n=6) for validating PTA outcomes (euthanized at day 42). Stenosis developed at juxta-anastomosis 28 days later and was treated with POBA [pilot group (n=6), definitive group (n=3)] or DCB [definitive group only (n=3)]. Definitive group underwent biweekly 4D Flow MRI scans. All animals developed CKD, with significant increases in BUN (median increase: 2.6 to 3.2 mmol/L, P<0.001) and creatinine (median increase: 100 to 187 μmol/L, P<0.001). In the pilot group, one animal had an infected fistula, and AVF patency was 1/5. In the definitive group, the patency was 5/6 because the AVF technique was modified by resecting the sternomastoid muscle and increasing the spatulation. At day 42 post PTA, DCB treated AVF outflow vein showed increasing blood flow compared to POBA (DCB 209.8 ± 64.42 mm2 vs POBA 170.9 ± 95.52 mm2 p = 0.934). A porcine model of AVF venous stenosis treated with PTA was developed, with blood flow trends favoring DCB over POBA.

4D-Flow MRI and Vector Ultrasound in the In-Vitro Evaluation of Surgical Aortic Heart Valves - a Pilot Study.

The aim of this study was the initial investigation of 4D-Flow MRI and Vector Ultrasound as novel imaging techniques in the in-vitro analysis of hemodynamics in anatomical models. Specifically, by looking at the hemodynamic performance of state-of-the-art surgical heart valves in a 3D-printed aortic arch. The mock circulatory loop simulated physiological, pulsatile flow. Two mechanical and three biological aortic valves prostheses were compared in a 3D-printed aortic arch. 4D magnetic resonance imaging and vector flow Doppler ultrasound served as imaging methods. Hemodynamic parameters such as wall shear stress, flow velocities and pressure gradients were analyzed. The flow analysis revealed characteristic flow-patterns in the 3D-printed aortic arch. The blood-flow in the arch presented complex patterns, including the formation of helixes and vortices. Higher proximal peak velocities and lower flow volumes were found for biological valves. The mock circulatory loop in combination with modern radiological imaging provides a sufficient basis for the hemodynamic comparison of aortic valves.

Four-dimensional Flow MRI-based Computational Fluid Dynamics Simulation for Noninvasive Portosystemic Pressure Gradient Assessment in Patients with Cirrhosis and Transjugular Intrahepatic Portosystemic Shunt.

Background Transjugular intrahepatic portosystemic shunt (TIPS) dysfunction in patients with liver cirrhosis and recurrent symptoms of portal hypertension is primarily assessed with US and confirmed with invasive catheter venography, which can be used to measure the portosystemic pressure gradient (PSPG) to identify TIPS-refractory portal hypertension. To avoid the risks and costs of invasive catheter venography, noninvasive PSPG evaluation strategies are needed. Purpose To demonstrate the feasibility of the combination of four-dimensional (4D) flow MRI with computational fluid dynamics (CFD) for noninvasive PSPG assessment in participants with cirrhosis and TIPS. Materials and Methods Abdominal 4D flow MRI was performed prospectively in participants with cirrhosis and TIPS between January 2019 and September 2020. Flow rates were measured within the TIPS and inferior vena cava (IVC). The portal vein (PV), TIPS, right hepatic vein, and IVC were segmented on MRI scans to create a CFD mesh. The PV and infrahepatic IVC were defined as inflows for 4D flow MRI-derived flow rates. The suprahepatic IVC was defined as the outflow. CFD simulations were used to noninvasively estimate PSPG as the difference between the simulated pressures in the PV and suprahepatic IVC. Invasive venographic measurements of the PSPG served as the reference standard, and Pearson correlation analysis was conducted to evaluate the relationship between noninvasive estimates and invasive measurements. Results In all 20 participants with cirrhosis (mean age, 58 years ± 9 [SD]; 11 men), 4D flow MRI-based CFD simulations enabled visualization of flow velocities and pressure distributions within the segmented vasculature and TIPS. Noninvasive estimates and invasive measures of PSPG were strongly correlated (r = 0.77; P < .001). The 4D flow MRI-based CFD simulations correctly classified the presence or absence of a post-TIPS PSPG greater than 12 mm Hg in 16 of 20 participants (80%). Conclusion The combination of 4D flow MRI and CFD was feasible for noninvasive PSPG assessment in participants with cirrhosis, portal hypertension, and TIPS. © RSNA, 2024 See also the editorial by Motosugi and Watanabe in this issue.

137 - WSS and velocity changes of common iliac arteries in patients with abdominal aortic aneurysm: assessment using 4D flow MRI

137 - WSS and velocity changes of common iliac arteries in patients with abdominal aortic aneurysm: assessment using 4D flow MRI

Evaluation of the presence and severity of spontaneous splenorenal or gastrorenal shunts via four-dimensional flow magnetic resonance imaging: a preliminary study.

Four-dimensional phase-contrast magnetic resonance imaging (4D flow MRI) is a relatively new type of MRI acquisition technique that provides a unique and comprehensive set of information within a single acquisition, including hemodynamic and anatomical information. This study was designed to noninvasively evaluate the correlation between the presence and severity of spontaneous splenorenal shunt (SRS) or gastrorenal shunt (GRS) and 4D flow MRI-derived parameters. This retrospective case-control study enrolled 70 patients who were diagnosed with hepatocirrhosis portal hypertension and admitted to the Second Affiliated Hospital of Chongqing Medical University. Patients were divided into three groups according to the diameter of the SRS and GRS. 4D flow MRI-derived parameters, including the turbulent kinetic energy, total volume (TV), flow velocity, blood flow volume (BFV), maximum flow (MF), wall shear stress, and relative pressure, were obtained for eight cut planes: proximal to the splenomesenteric confluence and liver hilum of the portal vein (PV1/PV2); the left/right branch of the bifurcation of the PV (LPV/RPV), at the mesosplenic confluence of the splenic vein (SV1), at the splenic hilum of the SV (SV2); at the proximal to the splenomesenteric confluence of the superior mesenteric vein (SMV1), and 5 cm from the splenomesenteric confluence of the SMV (SMV2). Comparisons among the three groups were based on one-way analysis of variance (ANOVA). Logistic regression was used to identify the risk factors for small SRS/GRS (S-SRS/GRS) and for large SRS/GRS (L-SRS/GRS). Receiver operating characteristic curves were used to evaluate the diagnostic performance of the independent risk factors for SRS and GRS. The associations between the clinical data and the 4D flow MRI-derived parameters of GRS and SRS were assessed via Spearman correlation coefficient analysis. The presence of SRS or GRS was correlated with TVLPV (r=-0.302; P=0.035), TVPV1 (r=-0.385; P=0.001), TVPV2 (r=-0.301; P=0.013), BFVPV1 (r=-0.360; P=0.010), BFVSMV2 (r=0.371; P=0.008), MFPV1 (r=-0.341; P=0.004), and MFPV2 (r=-0.291; P=0.017). Meanwhile, the severity of the SRS or GRS was correlated with alanine aminotransferase level (r=-0.535; P<0.001), BFVLPV (r=-0.560; P=0.008), aspartate aminotransferase level (r=-0.321; P=0.038), and model for end-stage liver disease score (r=0.323; P=0.039). TVPV1, TVPV2, BFVPV1, BFVPV2, and MFSMV2 were found to be independent risk factors for L-SRS/GRS, with intermediate diagnostic efficacy, with the area under the curve (AUC)TV PV1=0.706 [95% confidence interval (CI): 0.519-0.853; sensitivity, 61.54%; specificity, 80.77%; P=0.018], AUCBFV PV1 =0.694 (95% CI: 0.507-0.844; sensitivity, 95.00%; specificity, 63.16%; P=0.035), AUCTV PV2 =0.729 (95% CI: 0.544-0.870; sensitivity, 77.78%; specificity, 66.67%; P=0.016), AUCBFV PV2 =0.718 (95% CI: 0.531-0.862; sensitivity, 60.00%; specificity, 82.35%; P=0.017), and AUCMF SMV2 =0.788 (95% CI: 0.608-0.912; sensitivity, 44.00%; specificity, 84.46%; P=0.005), respectively. As the TV of PV1 and PV2 and the BFV of PV1 and PV2 decreased, the risk of L-SRS/GRS increased. As the MF of SMV2 increased, the risk of the presence of L-SRS/GRS increased. 4D flow MRI-derived parameters correlated with the presence and severity of SRS or GRS. Meanwhile, the independent risk factors for the presence of L-SRS/GRS were the TV of LPV, PV1, and PV2; the BFV of PV1 and SMV2; and the MF of PV1 and PV2.

Porto-mesenteric four-dimensional flow MRI: a novel non-invasive technique for assessment of gastro-oesophageal varices

ObjectivesTo assess the role of 4D flow MRI in the assessment of gastro-oesophageal varices and in the prediction of high-risk varices in patients with chronic liver disease.MethodsThirty-eight patients diagnosed with either oesophageal or gastric varices were included in this single-centre prospective study. 4D flow MRI was used to calculate peak flow, average flow and peak velocity at the portal vein confluence (PV1) and hilum (PV2), splenic vein hilum (SV1) and confluence (SV2), and superior mesenteric vein (SMV). PV and SV fractional flow changes were also measured.ResultsROC analysis revealed that both PV2 average flow and PV fractional average flow change had 100% sensitivity to predict high-risk patients with the PV fractional peak flow change having the widest area under the curve (AUC) and the highest specificity (92.3%). SV1 average flow, SV2 average flow, SV2 peak flow, and SV2 peak velocity increased significantly in patients with oesophageal compared to gastric varices included (p = 0.022, < 0.001, < 0.001 and 0.001, respectively).ConclusionBased on certain porto-mesenteric blood flow, velocity, and fractional flow change parameters, 4D flow MRI showed excellent performance in identifying high-risk patients and giving an idea about the grade and location of varices.Critical relevance statementVariceal bleeding is a major consequence of unidentified risky upper GI varices. Thus, by identifying and locating high-risk varices early, either oesophageal or gastric, using a non-invasive method like MRI, adverse events might be avoided.Key Points4D flow MRI can be used as a potential alternative for endoscopy to predict patients with high-risk varices.Based on portal vein fractional flow change, splenic flow and velocity, 4D MRI can predict and locate high-risk varices.Earlier identification of high-risk varices can allow for interventions to prevent adverse events.Graphical

Role of Non-Invasive Hemodynamic Forces through Four-Dimensional-Flow Magnetic Resonance Imaging (4D-Flow MRI) in Evaluating Mitral Regurgitation with Preserved Ejection Fraction: Seeking Novel Biomarkers

Mitral regurgitation (MR) is the systolic retrograde flow from the left ventricle (LV) to the left atrium. Despite the recognized importance of hemodynamic force (HDF) in cardiology, its exploration in MR has been limited. Therefore, we aimed to explore non-invasively assessed HDF as a novel biomarker for evaluating MR utilizing 4D-flow MRI. The study cohort comprised 15 healthy controls (19–61 years, 53% men) and 26 MR patients with preserved ejection fraction (EF) (33–75 years, trivial–severe, 54% men). The HDF analysis involved the semi-automatic calculation of systolic–diastolic root mean square (RMS), average, and transverse/longitudinal ratio across three directions (S-L: septal–lateral, I-A: inferior–anterior, and B-A: basal–apical) using Segment, v2.2 R6410 (Lund, Sweden, Medviso). A noticeable trend shift emerged in HDF as the MR severity increased (p-value &lt; 0.05). The MR severity demonstrated a noteworthy correlation with systolic RMS B-A, average B-A, diastolic average B-A, systolic average S-L, B-A, and systolic–diastolic ratio (rho = 0.621, 0.457, 0.317, 0.318, 0.555, −0.543, −0.35, respectively; p-value &lt; 0.05). HDF significantly correlated with LV function (end-diastolic volume, end-systolic volume, EF, and mass; p-value &lt; 0.05). Systolic RMS B-A and diastolic RMS S-L emerged as significant predictors of MR (Beta, 95% CI [3.253, 1.204–5.301], [5.413, 0.227–10.6], p-value &lt; 0.05). This study emphasizes HDF as a potential hemodynamic biomarker for evaluating MR.

Intracranial aneurysm stiffness assessment using 4D Flow MRI

BackgroundAlthough arterial stiffness is known as a biomarker for cardiovascular events and stroke, there is limited information in the literature regarding the stiffness of intracranial aneurysms. In this study, we aim to assess the stiffness of intracranial aneurysms using 4D Flow MRI. MethodsA total of 27 aneurysms in 25 patients with internal carotid artery aneurysms were included in this study. Using 4D Flow MRI, we measured the arterial pulse wave form during a cardiac cycle at planes proximal and distal to the target aneurysm. The damping of these waveforms through the aneurysm was defined as the aneurysm damping index (ADI) and compared to the contralateral side. We also investigated the clinical factors related to the ADI. ResultsADI assessment was successful in all cases. The average ADI was 1.18±0.28, which was significantly larger than 1.0 (P = 0.0027 [t-test]). The ADI on the aneurysm side was larger than on the contralateral side (1.19±0.30 vs 1.05±0.17, P = 0.029 [t-test]). On multivariate analysis, the use of beta-blockers (β=0.46, P = 0.015) and smoking history (β=-0.22, P = 0.024) showed a significant correlation with ADI. ConclusionWe have proposed a novel method to observe arterial pulse wave dumping through intracranial aneurysm using 4D Flow MRI. The damping can be quantitatively observed, and the ADI has correlations with clinical factors such as antihypertensive drugs and smoking. Further studies should focus more on evaluating aneurysm stiffness and its clinical applications.

Cluster analysis of 100 Marfan patients based on aortic 4D flow MRI and Z-score: insights into disease heterogeneity and stratification of subgroups.

4D flow MRI-derived variables from Marfan patients are highly heterogeneous. Our aim was to identify distinct Marfan patient subgroups based on aortic 4D flow MRI and Z-score for stratification of distinct hemodynamic profiles and clinical features by means of hierarchical cluster analysis. One hundred Marfan patients underwent baseline aortic 4D flow MRI at 3 T. Z-scores, degree of helical and vortical flow, wall shear stress, flow displacement, and peak velocity were determined in the ascending aorta. Sex, age, BMI, antihypertensive medication, and dural ectasia were recorded. Hierarchical cluster analysis was performed using 4D flow MRI variables and Z-scores as input. Cluster analysis resulted in three distinct clusters characterized by different Z-scores (mean ± SD); cluster 1: 0.4 ± 1.1 vs. cluster 2: 3.1 ± 1.1 vs. cluster 3: 3.6 ± 1.9. The three clusters delivered differences in helical and vortical flow patterns (global p = 0.003 and p < 0.001, respectively), wall shear stress (0.49 ± 0.11 vs. 0.44 ± 0.12 vs. 0.37 ± 0.09 N/m2, global p < 0.001), flow displacement (0.11 ± 0.05 vs. 0.16 ± 0.08 vs. 0.15 ± 0.07, global p = 0.006), and peak velocity (76.3 ± 9.0 vs. 60.1 ± 7.3 vs. 56.0 ± 7.8 cm/s, global p < 0.001). Patients in cluster 1 and 2 were relevantly younger than in cluster 3 (32.3 ± 13.8 vs. 32.8 ± 12.6 vs. 40.2 ± 15.0 years, all pairwise ∆p < 0.0297). Hierarchical cluster analysis based on aortic 4D flow MRI and Z-score revealed three distinct subgroups of Marfan patients, each characterized by specific hemodynamic profiles and clinical features. Follow-up of our patients is warranted to assess if 4D flow MRI- and Z-score-based stratification can predict future aortic diameter growth and ultimately improve outcomes. A combination of Z-score and 4D flow MRI-derived parameters may help identify subgroups of Marfan patients representing different stages or phenotypes of aortic disease, which require specific management strategies. Four-dimensional (4D) flow MRI-derived variables of Marfan patients are highly heterogeneous across varying Z-scores. Cluster analysis based on 4D flow MRI and Z-score revealed three distinct subgroups of Marfan patients. A combination of Z-score and 4D flow MRI-derived parameters may identify different stages of aortic disease in Marfan patients.

4D flow MRI velocity uncertainty quantification.

An automatic method is presented for estimating 4D flow MRI velocity measurement uncertainty in each voxel. The velocity distance (VD) metric, a statistical distance between the measured velocity and local error distribution, is introduced as a novel measure of 4D flow MRI velocity measurement quality. The method uses mass conservation to assess the local velocity error variance and the standardized difference of means (SDM) velocity to estimate the velocity error correlations. VD is evaluated as the Mahalanobis distance between the local velocity measurement and the local error distribution. The uncertainty model is validated synthetically and tested in vitro under different flow resolutions and noise levels. The VD's application is demonstrated on two in vivo thoracic vasculature 4D flow datasets. Synthetic results show the proposed uncertainty quantification method is sensitive to aliased regions across various velocity-to-noise ratios and assesses velocity error correlations in four- and six-point acquisitions with correlation errors at or under 3.2%. In vitro results demonstrate the method's sensitivity to spatial resolution, venc settings, partial volume effects, and phase wrapping error sources. Applying VD to assess in vivo 4D flow MRI in the aorta demonstrates the expected increase in measured velocity quality with contrast administration and systolic flow. The proposed 4D flow MRI uncertainty quantification method assesses velocity measurement error owing to sources including noise, intravoxel phase dispersion, and velocity aliasing. This method enables rigorous comparison of 4D flow MRI datasets obtained in longitudinal studies, across patient populations, and with different MRI systems.

Transverse sinus blood flow characteristics of pulsatile tinnitus with dehiscent sigmoid plate based on 4D flow MRI

Objective: To investigate the hemodynamic characteristics of transverse sinus with sigmoid sinus wall dehiscence (SSWD) of pulsatile tinnitus (PT) based on 4D flow MRI. Methods: Retrospective analysis was performed on all patients admitted to Beijing Friendship Hospital, Capital Medical University from January 2019 to January 2021 for dehiscent sigmoid plate pulsatile tinnitus. A total of 26 patients (sides) who met the criteria and underwent 4D flow MRI were included. A total of 26 subjects (46 sides), matched 1∶1 according to gender and age, were included in the normal healthy control group. Nonparametric rank sum test, Student's t test, and ANOVA were performed by SPSS 19.0 software. Binary Logistic regression was applied to the data with statistical significance. Results: There were more patients with dominant drainage on the affected side in PT group than in control group (73.1% vs. 42.3%). The incidence of transverse with a focal intraluminal filling defect and tapered stenosis was higher than that in control group (21.7% vs. 69.2%; 17.4% vs. 42.3%). Average through-plane velocity and maximum through-plane velocity in PT group were higher than those in control group [(33.75±13.88) cm/s vs. (15.84±7.21) cm/s; (93.19±33.55) cm/s vs. (40.40±14.40) cm/s]. The middle part and proximal end of Flowavg (ml/s) in PT group were larger than those in control group [4.69 (2.87; 5.62) ml/s vs. 2.76 (1.67; 4.99) ml/s; 3.41 (2.16; 5.47) ml/s vs. 2.67 (1.68; 4.41) ml/s]. In control group, the velocity of transverse sinus changed relatively gently, while in PT group, the velocity of proximal sinus increased significantly. Binary Logistic regression showed that SSWD PT was independently correlated with proximal maximum flow velocity [OR=1.086(1.029-1.146),P=0.003]. Conclusion: 4D flow MRI showed that the dominant drainage and higher velocity at the proximal end of the transverse sinus might be an important hemodynamic characteristics of dehiscent sigmoid plate pulsatile tinnitus.

Assessing Cerebral Microvascular Volumetric Pulsatility with High-Resolution 4D CBV MRI at 7T.

Arterial pulsation is crucial for promoting fluid circulation and for influencing neuronal activity. Previous studies assessed the pulsatility index based on blood flow velocity pulsatility in relatively large cerebral arteries of human. Here, we introduce a novel method to quantify the volumetric pulsatility of cerebral microvasculature across cortical layers and in white matter (WM), using high-resolution 4D vascular space occupancy (VASO) MRI with simultaneous recording of pulse signals at 7T. Microvascular volumetric pulsatility index (mvPI) and cerebral blood volume (CBV) changes across cardiac cycles are assessed through retrospective sorting of VASO signals into cardiac phases and estimating mean CBV in resting state (CBV0) by arterial spin labeling (ASL) MRI at 7T. Using data from 11 young (28.4±5.8 years) and 7 older (61.3±6.2 years) healthy participants, we investigated the aging effect on mvPI and compared microvascular pulsatility with large arterial pulsatility assessed by 4D-flow MRI. We observed the highest mvPI in the cerebrospinal fluid (CSF) on the cortical surface (0.19±0.06), which decreased towards the cortical layers as well as in larger arteries. In the deep WM, a significantly increased mvPI (p = 0.029) was observed in the older participants compared to younger ones. Additionally, mvPI in deep WM is significantly associated with the velocity pulsatility index (vePI) of large arteries (r = 0.5997, p = 0.0181). We further performed test-retest scans, non-parametric reliability test and simulations to demonstrate the reproducibility and accuracy of our method. To the best of our knowledge, our method offers the first in vivo measurement of microvascular volumetric pulsatility in human brain which has implications for cerebral microvascular health and its relationship research with glymphatic system, aging and neurodegenerative diseases.

Super-Resolving and Denoising 4D flow MRI of Neurofluids Using Physics-Guided Neural Networks.

To obtain high-resolution velocity fields of cerebrospinal fluid (CSF) and cerebral blood flow by applying a physics-guided neural network (div-mDCSRN-Flow) to 4D flow MRI. The div-mDCSRN-Flow network was developed to improve spatial resolution and denoise 4D flow MRI. The network was trained with patches of paired high-resolution and low-resolution synthetic 4D flow MRI data derived from computational fluid dynamic simulations of CSF flow within the cerebral ventricles of five healthy cases and five Alzheimer's disease cases. The loss function combined mean squared error with a binary cross-entropy term for segmentation and a divergence-based regularization term for the conservation of mass. Performance was assessed using synthetic 4D flow MRI in one healthy and one Alzheimer' disease cases, an in vitro study of healthy cerebral ventricles, and in vivo 4D flow imaging of CSF as well as flow in arterial and venous blood vessels. Comparison was performed to trilinear interpolation, divergence-free radial basis functions, divergence-free wavelets, 4DFlowNet, and our network without divergence constraints. The proposed network div-mDCSRN-Flow outperformed other methods in reconstructing high-resolution velocity fields from synthetic 4D flow MRI in healthy and AD cases. The div-mDCSRN-Flow network reduced error by 22.5% relative to linear interpolation for in vitro core voxels and by 49.5% in edge voxels. The results demonstrate generalizability of our 4D flow MRI super-resolution and denoising approach due to network training using flow patches and physics-based constraints. The mDCSRN-Flow network can facilitate MRI studies involving CSF flow measurements in cerebral ventricles and association of MRI-based flow metrics with cerebrovascular health.

The calculation method of blood flow pressure based on four-dimensional flow magnetic resonance imaging and deep learning

Utilizing artificial intelligence methods for blood flow pressure estimation can significantly enhance the computational speed of blood flow pressure. However, current related research can only calculate the blood flow pressure parameters of vessels with different geometric shapes under fixed boundary conditions, thus fail to achieve transient flow field calculation and consider the hemodynamic differences formed by patients' varying physiological and pathological conditions. In view of this, this study proposes a method for relative pressure estimation based on four-dimensional flow magnetic resonance imaging (4D flow MRI) of patient blood flow and deep learning. 4D flow MRI was used to obtain the patient's blood flow velocity gradient data, and feature engineering processing is performed on the sampled data. Then, a novel neural network was proposed to acquire the characteristic relationship between velocity gradient and pressure gradient in the vicinity of the point to be measured and within adjacent sampling time periods, thereby achieving the calculation of the relative pressure in the vicinity of the point to be measured. Statistical analysis was performed to evaluate the efficacy of the method, comparing it with computational fluid dynamics methods and catheter pressure measurement techniques. The accuracy of the proposed method exceeded 96%, while computational efficiency was improved by several tens of times, and no manual setting of physiological parameters was required. Furthermore, the results were compared with clinical catheter-measured pressure results, r2 = 0.9053, indicating a significant consistency between the two methods. Compared to previous research, the method proposed in this study can take the blood flow velocity conditions of different patients at different times as input features via 4D flow MRI, thus enabling the calculation of pressure in transient flow fields, which significantly improved computational efficiency and reduced costs while maintaining a high level of calculation accuracy. This provides new direction for future research on machine learning prediction of blood flow pressure.

CSF dynamics throughout the ventricular system using 4D flow MRI: associations to arterial pulsatility, ventricular volumes, and age

BackgroundCerebrospinal fluid (CSF) dynamics are increasingly studied in aging and neurological disorders. Models of CSF-mediated waste clearance suggest that altered CSF dynamics could play a role in the accumulation of toxic waste in the CNS, with implications for Alzheimer’s disease and other proteinopathies. Therefore, approaches that enable quantitative and volumetric assessment of CSF flow velocities could be of value. In this study we demonstrate the feasibility of 4D flow MRI for simultaneous assessment of CSF dynamics throughout the ventricular system, and evaluate associations to arterial pulsatility, ventricular volumes, and age.MethodsIn a cognitively unimpaired cohort (N = 43; age 41–83 years), cardiac-resolved 4D flow MRI CSF velocities were obtained in the lateral ventricles (LV), foramens of Monro, third and fourth ventricles (V3 and V4), the cerebral aqueduct (CA) and the spinal canal (SC), using a velocity encoding (venc) of 5 cm/s. Cerebral blood flow pulsatility was also assessed with 4D flow (venc = 80 cm/s), and CSF volumes were obtained from T1- and T2-weighted MRI. Multiple linear regression was used to assess effects of age, ventricular volumes, and arterial pulsatility on CSF velocities.ResultsCardiac-driven CSF dynamics were observed in all CSF spaces, with region-averaged velocity range and root-mean-square (RMS) velocity encompassing from very low in the LVs (RMS 0.25 ± 0.08; range 0.85 ± 0.28 mm/s) to relatively high in the CA (RMS 6.29 ± 2.87; range 18.6 ± 15.2 mm/s). In the regression models, CSF velocity was significantly related to age in 5/6 regions, to CSF space volume in 2/3 regions, and to arterial pulsatility in 3/6 regions. Group-averaged waveforms indicated distinct CSF flow propagation delays throughout CSF spaces, particularly between the SC and LVs.ConclusionsOur findings show that 4D flow MRI enables assessment of CSF dynamics throughout the ventricular system, and captures independent effects of age, CSF space morphology, and arterial pulsatility on CSF motion.

A comparison of machine learning methods for recovering noisy and missing 4D flow MRI data.

Experimental blood flow measurement techniques are invaluable for a better understanding of cardiovascular disease formation, progression, and treatment. One of the emerging methods is time-resolved three-dimensional phase-contrast magnetic resonance imaging (4D flow MRI), which enables noninvasive time-dependent velocity measurements within large vessels. However, several limitations hinder the usability of 4D flow MRI and other experimental methods for quantitative hemodynamics analysis. These mainly include measurement noise, corrupt or missing data, low spatiotemporal resolution, and other artifacts. Traditional filtering is routinely applied for denoising experimental blood flow data without any detailed discussion on why it is preferred over other methods. In this study, filtering is compared to different singular value decomposition (SVD)-based machine learning and autoencoder-type deep learning methods for denoising and filling in missing data (imputation). An artificially corrupted and voxelized computational fluid dynamics (CFD) simulation as well as in vitro 4D flow MRI data are used to test the methods. SVD-based algorithms achieve excellent results for the idealized case but severely struggle when applied to in vitro data. The autoencoders are shown to be versatile and applicable to all investigated cases. For denoising, the in vitro 4D flow MRI data, the denoising autoencoder (DAE), and the Noise2Noise (N2N) autoencoder produced better reconstructions than filtering both qualitatively and quantitatively. Deep learning methods such as N2N can result in noise-free velocity fields even though they did not use clean data during training. This work presents one of the first comprehensive assessments and comparisons of various classical and modern machine-learning methods for enhancing corrupt cardiovascular flow data in diseased arteries for both synthetic and experimental test cases.

Optimizing encoding strategies for 4D Flow MRI of mean and turbulent flow

For 4D Flow MRI of mean and turbulent flow a compromise between spatiotemporal undersampling and velocity encodings needs to be found. Assuming a fixed scan time budget, the impact of trading off spatiotemporal undersampling versus velocity encodings on quantification of velocity and turbulence for aortic 4D Flow MRI was investigated. For this purpose, patient-specific mean and turbulent aortic flow data were generated using computational fluid dynamics which were embedded into the patient-specific background image data to generate synthetic MRI data with corresponding ground truth flow. Cardiac and respiratory motion were included. Using the synthetic MRI data as input, 4D Flow MRI was subsequently simulated with undersampling along pseudo-spiral Golden angle Cartesian trajectories for various velocity encoding schemes. Data were reconstructed using a locally low rank approach to obtain mean and turbulent flow fields to be compared to ground truth. Results show that, for a 15-min scan, velocity magnitudes can be reconstructed with good accuracy relatively independent of the velocity encoding scheme (SSIMU=0.938±0.003)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$SSI{M}_{U}=0.938\\pm 0.003)$$\\end{document}, good accuracy (SSIMU≥0.933\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$SSI{M}_{U}\\ge 0.933$$\\end{document}) and with peak velocity errors limited to 10%. Turbulence maps on the other hand suffer from both lower reconstruction quality (SSIMTKE≥0.323\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$SSI{M}_{TKE}\\ge 0.323$$\\end{document}) and larger sensitivity to undersampling, motion and velocity encoding strengths (SSIMTKE=0.570±0.110)\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$SSI{M}_{TKE}=0.570\\pm 0.110)$$\\end{document} when compared to velocity maps. The best compromise to measure unwrapped velocity maps and turbulent kinetic energy given a fixed 15-min scan budget was found to be a 7-point multi-Venc\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${V}_{enc}$$\\end{document} acquisition with a low Venc\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${V}_{enc}$$\\end{document} tuned for best sensitivity to the range of expected intra-voxel standard deviations and a high Venc\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$${V}_{enc}$$\\end{document} larger than the expected peak velocity.

Patient-specific compliant simulation framework informed by 4DMRI-extracted pulse wave Velocity: Application post-TEVAR

We introduce a new computational framework that utilises Pulse Wave Velocity (PWV) extracted directly from 4D flow MRI (4DMRI) to inform patient-specific compliant computational fluid dynamics (CFD) simulations of a Type-B aortic dissection (TBAD), post-thoracic endovascular aortic repair (TEVAR). The thoracic aortic geometry, a 3D inlet velocity profile (IVP) and dynamic outlet boundary conditions are derived from 4DMRI and brachial pressure patient data. A moving boundary method (MBM) is applied to simulate aortic wall displacement. The aortic wall stiffness is estimated through two methods: one relying on area-based distensibility and the other utilising regional pulse wave velocity (RPWV) distensibility, further fine-tuned to align with in vivo values. Predicted pressures and outlet flow rates were within 2.3 % of target values. RPWV-based simulations were more accurate in replicating in vivo hemodynamics than the area-based ones. RPWVs were closely predicted in most regions, except the endograft. Systolic flow reversal ratios (SFRR) were accurately captured, while differences above 60 % in in-plane rotational flow (IRF) between the simulations were observed. Significant disparities in predicted wall shear stress (WSS)-based indices were observed between the two approaches, especially the endothelial cell activation potential (ECAP). At the isthmus, the RPWV-driven simulation indicated a mean ECAP>1.4 Pa-1 (critical threshold), indicating areas potentially prone to thrombosis, not captured by the area-based simulation.RPWV-driven simulation results agree well with 4DMRI measurements, validating the proposed pipeline and facilitating a comprehensive assessment of surgical decision-making scenarios and potential complications, such as thrombosis and aortic growth.

Prediction Modelling for Gastroesophageal Variceal Bleeding in Patients With Chronic Hepatitis B Using Four-dimensional Flow MRI

Introduction and objectivesIn this study, we aim to develop a model for predicting Gastroesophageal varices (GEV) bleeding in patients with chronic hepatitis B (CHB) by utilizing hemodynamic parameters obtained through four-dimensional flow MRI (4D flow MRI). Materials and methodsThis study conducted a prospective enrollment of CHB patients suspected of GEV from October 2021 to May 2022. The severity of varices and bleeding risk were evaluated using clinical findings and upper gastrointestinal endoscopy, and patients were classified into high-risk and non-high-risk groups. The study utilized serological examination, ultrasonographic examination, and 4D flow MRI. Relevant parameters were selected through univariate and multivariate analyses, and a prediction model was established using binary logistic regression analysis. The model was combined with the Baveno Ⅵ/Ⅶ and Expanded Baveno Ⅵ/Ⅶ criteria to evaluate diagnostic efficacy and the risk of avoiding endoscopic examination. ResultsA total of 40 CHB patients were enrolled and categorized into the high-risk group (n=15) and the non-high-risk group (n=25). The spleen diameter and regurgitant fraction (R%) were independent predictors of variceal bleeding and a predictive model was established. The combination of this prediction model and the Baveno Ⅵ/Ⅶ criteria achieved high diagnostic efficiency, enabling 45.00% (18/40) of patients to be exempted from the unnecessary endoscopic procedure and the high-risk misclassification rate (0%) was less than 5%. ConclusionThe prediction model generated by 4D flow MRI has the potential to assess the likelihood of varices and can be supplemented by the Baveno VI/VII criteria to improve diagnostic accuracy in CHB patients.