Creating hemodynamic atlases of cardiac 4D flow MRI

Coupling synthetic and real-world data for a deep learning-based segmentation process of 4D flow MRI

BACKGROUND AND OBJECTIVE

Phase contrast magnetic resonance imaging (4D flow MRI) is an imaging technique able to provide blood velocity in vivo and morphological information. This capability has been used to study mainly the hemodynamics of large vessels, such as the thoracic aorta. However, the segmentation of 4D flow MRI data is a complex and time-consuming task. In recent years, neural networks have shown great accuracy in segmentation tasks if large datasets are provided. Unfortunately, in the context of 4D flow MRI, the availability of these data is limited due to its recent adoption in clinical settings. In this study, we propose a pipeline for generating synthetic thoracic aorta phase contrast magnetic resonance angiography (PCMRA) to expand the limited dataset of patient-specific PCMRA images, ultimately improving the accuracy of the neural network segmentation even with a small real dataset.

METHODS

The pipeline involves several steps. First, a statistical shape model is used to synthesize new artificial geometries to improve data numerosity and variability. Secondly, computational fluid dynamics simulations are employed to simulate the velocity fields and, finally, after a downsampling and a signal-to-noise and velocity limit adjustment in both frequency and spatial domains, volumes are obtained using the PCMRA formula. These synthesized volumes are used in combination with real-world data to train a 3D U-Net neural network. Different settings of real and synthetic data are tested.

RESULTS

Incorporating synthetic data into the training set significantly improved the segmentation performance compared to using only real data. The experiments with synthetic data achieved a DICE score (DS) value of 0.83 and a better target reconstruction with respect to the case with only real data (DS = 0.65).

CONCLUSION

The proposed pipeline demonstrated the ability to increase the dataset in terms of numerosity and variability and to improve the segmentation accuracy for the thoracic aorta using PCMRA.

Four-Dimensional Flow MR Imaging: Technique and Advances

4D Flow MRI is an advanced imaging technique for comprehensive non-invasive assessment of the cardiovascular system. The capture of the blood velocity vector field throughout the cardiac cycle enables measures of flow, pulse wave velocity, kinetic energy, wall shear stress, and more. Advances in hardware, MRI data acquisition and reconstruction methodology allow for clinically feasible scan times. The availability of 4D Flow analysis packages allows for more widespread use in research and the clinic and will facilitate much needed multi-center, multi-vendor studies in order to establish consistency across scanner platforms and to enable larger scale studies to demonstrate clinical value.

Abdominal applications of quantitative 4D flow MRI

4D flow MRI is a quantitative MRI technique that allows the comprehensive assessment of time-resolved hemodynamics and vascular anatomy over a 3-dimensional imaging volume. It effectively combines several advantages of invasive and non-invasive imaging modalities like ultrasound, angiography, and computed tomography in a single MRI acquisition and provides an unprecedented characterization of velocity fields acquired non-invasively in vivo. Functional and morphological imaging of the abdominal vasculature is especially challenging due to its complex and variable anatomy with a wide range of vessel calibers and flow velocities and the need for large volumetric coverage. Despite these challenges, 4D flow MRI is a promising diagnostic and prognostic tool as many pathologies in the abdomen are associated with changes of either hemodynamics or morphology of arteries, veins, or the portal venous system. In this review article, we will discuss technical aspects of the implementation of abdominal 4D flow MRI ranging from patient preparation and acquisition protocol over post-processing and quality control to final data analysis. In recent years, the range of applications for 4D flow in the abdomen has increased profoundly. Therefore, we will review potential clinical applications and address their clinical importance, relevant quantitative and qualitative parameters, and unmet challenges.

SRflow: Deep learning based super-resolution of 4D-flow MRI data

Exploiting 4D-flow magnetic resonance imaging (MRI) data to quantify hemodynamics requires an adequate spatio-temporal vector field resolution at a low noise level. To address this challenge, we provide a learned solution to super-resolve in vivo 4D-flow MRI data at a post-processing level. We propose a deep convolutional neural network (CNN) that learns the inter-scale relationship of the velocity vector map and leverages an efficient residual learning scheme to make it computationally feasible. A novel, direction-sensitive, and robust loss function is crucial to learning vector-field data. We present a detailed comparative study between the proposed super-resolution and the conventional cubic B-spline based vector-field super-resolution. Our method improves the peak-velocity to noise ratio of the flow field by 10 and 30% for in vivo cardiovascular and cerebrovascular data, respectively, for 4 × super-resolution over the state-of-the-art cubic B-spline. Significantly, our method offers 10x faster inference over the cubic B-spline. The proposed approach for super-resolution of 4D-flow data would potentially improve the subsequent calculation of hemodynamic quantities.

Fully automated intracardiac 4D flow MRI post-processing using deep learning for biventricular segmentation

OBJECTIVES

4D flow MRI allows for a comprehensive assessment of intracardiac blood flow, useful for assessing cardiovascular diseases, but post-processing requires time-consuming ventricular segmentation throughout the cardiac cycle and is prone to subjective errors. Here, we evaluate the use of automatic left and right ventricular (LV and RV) segmentation based on deep learning (DL) network that operates on short-axis cine bSSFP images.

METHODS

A previously published DL network was fine-tuned via retraining on a local database of 106 subjects scanned at our institution. In 26 test subjects, the ventricles were segmented automatically by the network and manually by 3 human observers on bSSFP MRI. The bSSFP images were then registered to the corresponding 4D flow images to apply the segmentation to 4D flow velocity data. Dice coefficients and the relative deviation between measurements (automatic vs. manual and interobserver manual) of various hemodynamic parameters were assessed.

RESULTS

The automated segmentation resulted in similar Dice scores (LV: 0.92, RV: 0.86) and lower relative deviations from manual segmentation in left ventricular (LV) average kinetic energy (KE) (8%) and RV KE (15%) than the Dice scores (LV: 0.91, RV: 0.87) and relative deviations between manual segmentation observers (LV KE: 11%, p = 0.01; RV KE: 19%, p = 0.03).

CONCLUSIONS

The automated post-processing method using deep learning resulted in hemodynamic measurements that differ from a manual observer's measurements equally or less than the variation between manual observers. This approach can be used to decrease post-processing time on intraventricular 4D flow data and mitigate interobserver variability.

KEY POINTS

• Our proposed method allows for fully automated post-processing of intraventricular 4D flow MRI data. • Our method resulted in hemodynamic measurements that matched those derived from manual segmentation equally as well as interobserver variability. • Our method can be used to greatly accelerate intraventricular 4D flow post-processing and improve interobserver repeatability.

Hemodynamic Parameters for Cardiovascular System in 4D Flow MRI: Mathematical Definition and Clinical Applications

Blood flow imaging becomes an emerging trend in cardiology with the recent progress in computer technology. It not only visualizes colorful flow velocity streamlines but also quantifies the mechanical stress on cardiovascular structures; thus, it can provide the detailed inspections of the pathophysiology of diseases and predict the prognosis of cardiovascular functions. Clinical applications include the comprehensive assessment of hemodynamics and cardiac functions in echocardiography vector flow mapping (VFM), 4D flow MRI, and surgical planning as a simulation medicine in computational fluid dynamics (CFD).For evaluation of the hemodynamics, novel mathematically derived parameters obtained using measured velocity distributions are essential. Among them, the traditional and typical parameters are wall shear stress (WSS) and its related parameters. These parameters indicate the mechanical damages to endothelial cells, resulting in degenerative intimal change in vascular diseases. Apart from WSS, there are abundant parameters that describe the strength of the vortical and/or helical flow patterns. For instance, vorticity, enstrophy, and circulation indicate the rotating flow strength or power of 2D vortical flows. In addition, helicity, which is defined as the cross-linking number of the vortex filaments, indicates the 3D helical flow strength and adequately describes the turbulent flow in the aortic root in cases with complicated anatomies. For the description of turbulence caused by the diseased flow, there exist two types of parameters based on completely different concepts, namely: energy loss (EL) and turbulent kinetic energy (TKE). EL is the dissipated energy with blood viscosity and evaluates the cardiac workload related to the prognosis of heart failure. TKE describes the fluctuation in kinetic energy during turbulence, which describes the severity of the diseases that cause jet flow. These parameters are based on intuitive and clear physiological concepts, and are suitable for in vivo flow measurements using inner velocity profiles.

4D Flow with MRI

Magnetic resonance imaging (MRI) has become an important tool for the clinical evaluation of patients with cardiac and vascular diseases. Since its introduction in the late 1980s, quantitative flow imaging with MRI has become a routine part of standard-of-care cardiothoracic and vascular MRI for the assessment of pathological changes in blood flow in patients with cardiovascular disease. More recently, time-resolved flow imaging with velocity encoding along all three flow directions and three-dimensional (3D) anatomic coverage (4D flow MRI) has been developed and applied to enable comprehensive 3D visualization and quantification of hemodynamics throughout the human circulatory system. This article provides an overview of the use of 4D flow applications in different cardiac and vascular regions in the human circulatory system, with a focus on using 4D flow MRI in cardiothoracic and cerebrovascular diseases.

Hypertrophic Cardiomyopathy Is Associated with Altered Left Ventricular 3D Blood Flow Dynamics

Purpose

To employ four-dimensional (4D) flow MRI to investigate associations between hemodynamic parameters with systolic anterior motion (SAM), mitral regurgitation (MR), stroke volume, and cardiac mass in patients with hypertrophic cardiomyopathy (HCM).

Materials and Methods

A total of 13 patients with HCM (51 years ± 16 [standard deviation]; 10 men) and 11 age-matched healthy control subjects (54 years ± 15; eight men) underwent cardiac 4D flow MRI data analysis including calculation of peak systolic and diastolic control-averaged left ventricular (LV) velocity maps to quantify volumes of elevated velocity (EVV) in the left ventricle. Standard-of-care cine imaging was performed in short-axis, LV outflow tract (LVOT), and two-, three-, and four-chamber views on which the presence of SAM, presence of MR, total stroke volume, and cardiac mass were assessed.

Results

Systolic EVV in patients with HCM was 7 mL ± 5, which was significantly associated with elevated aortic peak velocity (R = 0.87; P < .001), decreased LVOT diameter (R = 0.68; P = .01), and increased cardiac mass (R = 0.62; P = .02). In addition, EVV differed significantly between patients with and those without SAM (10 mL ± 4.7 vs 3 mL ± 2.3; P = .03) and those with and those without MR (9.9 mL ± 4.8 vs 4.0 mL ± 3.2; P < .05). In the atrial systolic phase, peak diastolic velocity in the LV correlated with septal thickness (R = 0.66; P = .01).

Conclusion

Quantification and visualization of EVV in the LV is feasible and may provide further insight into the clinical manifestations of altered hemodynamics in HCM.© RSNA, 2020.

Ventricular Flow Field Visualization During Mechanical Circulatory Support in the Assisted Isolated Beating Heart

Investigations of ventricular flow patterns during mechanical circulatory support are limited to in vitro flow models or in silico simulations, which cannot fully replicate the complex anatomy and contraction of the heart. Therefore, the feasibility of using echocardiographic particle image velocimetry (Echo-PIV) was evaluated in an isolated working heart setup. Porcine hearts were connected to an isolated, working heart setup and a left ventricular assist device (LVAD) was implanted. During different levels of LVAD support (unsupported, partial support, full support), microbubbles were injected and echocardiographic images were acquired. Iterative PIV algorithms were applied to calculate flow fields. The isolated heart setup allowed different hemodynamic situations. In the unsupported heart, diastolic intra-ventricular blood flow was redirected at the heart’s apex towards the left ventricular outflow tract (LVOT). With increasing pump speed, large vortex formation was suppressed, and blood flow from the mitral valve directly entered the pump cannula. The maximum velocities in the LVOT were significantly reduced with increasing support. For the first time, cardiac blood flow patterns during LVAD support were visualized and quantified in an ex vivo model using Echo-PIV. The results reveal potential regions of stagnation in the LVOT and, in future the methods might be also used in clinical routine to evaluate intraventricular flow fields during LVAD support.

Parametric Hemodynamic 4D Flow MRI Maps for the Characterization of Chronic Thoracic Descending Aortic Dissection

BACKGROUND

Systematic evaluation of complex flow in the true lumen and false lumen (TL, FL) is needed to better understand which patients with chronic descending aortic dissection (DAD) are predisposed to complications.

PURPOSE

To develop quantitative hemodynamic maps from 4D flow MRI for evaluating TL and FL flow characteristics.

STUDY TYPE

Retrospective.

POPULATION

In all, 20 DAD patients (age = 60 ± 11 years; 12 male) (six medically managed type B AD [TBAD], 14 repaired type A AD [rTAAD] now with ascending aortic graft [AAo] or elephant trunk [ET1] repair) and 21 age-matched controls (age = 59 ± 10 years; 13 male) were included.

FIELD STRENGTH/SEQUENCE

1.5T, 3T, 4D flow MRI.

ASSESSMENT

4D flow MRI was acquired in all subjects. Data analysis included 3D segmentation of TL and FL and voxelwise calculation of forward flow, reverse flow, flow stasis, and kinetic energy as quantitative hemodynamics maps.

STATISTICAL TESTS

Analysis of variance (ANOVA) or Kruskal-Wallis tests were performed for comparing subject groups. Correlation and Bland-Altman analysis was performed for the interobserver study.

RESULTS

Patients with rTAAD presented with elevated TL reverse flow (AAo repair: P = 0.004, ET1: P = 0.018) and increased TL kinetic energy (AAo repair: P = 0.0002, ET1: P = 0.011) compared to controls. In addition, TL kinetic energy was increased vs. patients with TBAD (AAo repair: P = 0.021, ET1: P = 0.048). rTAAD was associated with higher FL kinetic energy and lower FL stasis compared to patients with TBAD (AAo repair: P = 0.002, ET1: P = 0.024 and AAo repair: P = 0.003, ET1: P = 0.048, respectively).

DATA CONCLUSION

Quantitative maps from 4D flow MRI demonstrated global and regional hemodynamic differences between DAD patients and controls. Patients with rTAAD vs. TBAD had significantly altered regional TL and FL hemodynamics. These findings indicate the potential of 4D flow MRI-derived hemodynamic maps to help better evaluate patients with DAD.

LEVEL OF EVIDENCE

3 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:1357-1368.

In silico Optimization of Left Atrial Appendage Occluder Implantation Using Interactive and Modeling Tools

According to clinical studies, around one third of patients with atrial fibrillation (AF) will suffer a stroke during their lifetime. Between 70 and 90% of these strokes are caused by thrombus formed in the left atrial appendage. In patients with contraindications to oral anticoagulants, a left atrial appendage occluder (LAAO) is often implanted to prevent blood flow entering in the LAA. A limited range of LAAO devices is available, with different designs and sizes. Together with the heterogeneity of LAA morphology, these factors make LAAO success dependent on clinician's experience. A sub-optimal LAAO implantation can generate thrombi outside the device, eventually leading to stroke if not treated. The aim of this study was to develop clinician-friendly tools based on biophysical models to optimize LAAO device therapies. A web-based 3D interactive virtual implantation platform, so-called VIDAA, was created to select the most appropriate LAAO configurations (type of device, size, landing zone) for a given patient-specific LAA morphology. An initial LAAO configuration is proposed in VIDAA, automatically computed from LAA shape features (centreline, diameters). The most promising LAAO settings and LAA geometries were exported from VIDAA to build volumetric meshes and run Computational Fluid Dynamics (CFD) simulations to assess blood flow patterns after implantation. Risk of thrombus formation was estimated from the simulated hemodynamics with an index combining information from blood flow velocity and complexity. The combination of the VIDAA platform with in silico indices allowed to identify the LAAO configurations associated to a lower risk of thrombus formation; device positioning was key to the creation of regions with turbulent flows after implantation. Our results demonstrate the potential for optimizing LAAO therapy settings during pre-implant planning based on modeling tools and contribute to reduce the risk of thrombus formation after treatment.

Wall shear stress estimation in the aorta: Impact of wall motion, spatiotemporal resolution, and phase noise

BACKGROUND

Wall shear stress (WSS) presents an important parameter for assessing blood flow characteristics and evaluating flow-mediated lesions in the aorta.

PURPOSE

To investigate the robustness of WSS and oscillatory shear index (OSI) estimation based on 4D flow MRI against vessel wall motion, spatiotemporal resolution, and velocity encoding (VENC).

STUDY TYPE

Simulated and prospective.

POPULATION

Synthetic 4D flow MRI data of the aorta, simulated using the Lattice-Boltzmann method; in vivo 4D flow MRI data of the aorta from healthy volunteers (n = 11) and patients with congenital heart defects (n = 17).

FIELD STRENGTH/SEQUENCE

1.5T; 4D flow MRI with PEAK-GRAPPA acceleration and prospective electrocardiogram triggering.

ASSESSMENT

Predicated upon 3D cubic B-splines interpolation of the image velocity field, WSS was estimated in mid-systole, early-diastole, and late-diastole and OSI was derived. We assessed the impact of spatiotemporal resolution and phase noise, and compared results based on tracked-using deformable registration-and static vessel wall location.

STATISTICAL TESTS

Bland-Altman analysis to assess WSS/OSI differences; Hausdorff distance (HD) to assess wall motion; and Pearson's correlation coefficient (PCC) to assess correlation of HD with WSS.

RESULTS

Synthetic data results show systematic over-/underestimation of WSS when different spatial resolution (mean ± 1.96 SD up to -0.24 ± 0.40 N/m2 and 0.5 ± 1.38 N/m2 for 8-fold and 27-fold voxel size, respectively) and VENC-depending phase noise (mean ± 1.96 SD up to 0.31 ± 0.12 N/m2 and 0.94 ± 0.28 N/m2 for 2-fold and 4-fold VENC increase, respectively) are given. Neglecting wall motion when defining the vessel wall perturbs WSS estimates to a considerable extent (1.96 SD up to 1.21 N/m2 ) without systematic over-/underestimation (Bland-Altman mean range -0.06 to 0.05).

DATA CONCLUSION

In addition to sufficient spatial resolution and velocity to noise ratio, accurate tracking of the vessel wall is essential for reliable image-based WSS estimation and should not be neglected if wall motion is present.

LEVEL OF EVIDENCE

2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.