Automatic Time-Resolved Cardiovascular Segmentation of 4D Flow MRI Using Deep Learning

Diastolic function assessment with four-dimensional flow cardiovascular magnetic resonance using automatic deep learning E/A ratio analysis

BACKGROUND

Diastolic left ventricular (LV) dysfunction is a powerful contributor to the symptoms and prognosis of patients with heart failure. In patients with depressed LV systolic function, the E/A ratio, the ratio between the peak early (E) and the peak late (A) transmitral flow velocity, is the first step to defining the grade of diastolic dysfunction. Doppler echocardiography (echo) is the preferred imaging technique for diastolic function assessment, while cardiovascular magnetic resonance (CMR) is less established as a method. Previous four-dimensional (4D) Flow-based studies have looked at the E/A ratio proximal to the mitral valve, requiring manual interaction. In this study, we compare an automated, deep learning-based and two semi-automated approaches for 4D Flow CMR-based E/A ratio assessment to conventional, gold-standard echo-based methods.

METHODS

Ninety-seven subjects with chronic ischemic heart disease underwent a cardiac echo followed by CMR investigation. 4D Flow-based E/A ratio values were computed using three different approaches; two semi-automated, assessing the E/A ratio by measuring the inflow velocity (MVvel) and the inflow volume (MVflow) at the mitral valve plane, and one fully automated, creating a full LV segmentation using a deep learning-based method with which the E/A ratio could be assessed without constraint to the mitral plane (LVvel).

RESULTS

MVvel, MVflow, and LVvel E/A ratios were strongly associated with echocardiographically derived E/A ratio (R2 = 0.60, 0.58, 0.72). LVvel peak E and A showed moderate association to Echo peak E and A, while MVvel values were weakly associated. MVvel and MVflow EA ratios were very strongly associated with LVvel (R2 = 0.84, 0.86). MVvel peak E was moderately associated with LVvel, while peak A showed a strong association (R2 = 0.26, 0.57).

CONCLUSION

Peak E, peak A, and E/A ratio are integral to the assessment of diastolic dysfunction and may expand the utility of CMR studies in patients with cardiovascular disease. While underestimation of absolute peak E and A velocities was noted, the E/A ratio measured with all three 4D Flow methods was strongly associated with the gold standard Doppler echocardiography. The automatic, deep learning-based method performed best, with the most favorable runtime of ∼40 seconds. As both semi-automatic methods associated very strongly to LVvel, they could be employed as an alternative for estimation of E/A ratio.

Evaluation of 4D Flow MRI-Derived Relative Residence Time as a Marker for Cirrhosis Associated Portal Vein Thrombosis

BACKGROUND

Portal vein thrombosis (PVT) is thought to arise from stagnant blood flow, yet conclusive evidence is lacking. Relative residence time (RRT) assessed using 4D Flow MRI may offer insight into portal flow stagnation.

PURPOSE

To explore the relationship between RRT values and the presence of PVT in cirrhotic participants.

STUDY TYPE

Prospective.

POPULATION

Forty-eight participants with liver cirrhosis (27 males, median age 67 years [IQR: 57-73]) and 20 healthy control participants (12 males, median age 45 years [IQR: 40-54]).

FIELD STRENGTH/SEQUENCE

3 T/4D Flow MRI.

ASSESSMENT

Laboratory (liver and kidney function test results and platelet count) and clinical data (presence of tumors and other imaging findings), and portal hemodynamics derived from 4D Flow MRI (spatiotemporally averaged RRT [RRT-mean], flow velocity, and flow rate) were analyzed.

STATISTICAL TESTS

We used multivariable logistic regression, adjusted by selected covariates through the Lasso method, to explore whether RRT-mean is an independent risk factor for PVT. The area under the ROC curve (AUC) was also calculated to assess the model's discriminative ability. P < 0.05 indicated statistical significance.

RESULTS

The liver cirrhosis group consisted of 16 participants with PVT and 32 without PVT. Higher RRT-mean values (odds ratio [OR] 11.4 [95% CI: 2.19, 118]) and lower platelet count (OR 0.98 per 1000 μL [95% CI: 0.96, 0.99]) were independent risk factors for PVT. The incorporation of RRT-mean (AUC, 0.77) alongside platelet count (AUC, 0.75) resulted in an AUC of 0.84. When including healthy control participants, RRT-mean had an adjusted OR of 12.4 and the AUC of the combined model (RRT-mean and platelet count) was 0.90.

DATA CONCLUSION

Prolonged RRT values and low platelet count were significantly associated with the presence of PVT in cirrhotic participants. RRT values derived from 4D Flow MRI may have potential clinical relevance in the management of PVT.

EVIDENCE LEVEL

2 TECHNICAL EFFICACY: Stage 2.

Deep Learning For Automatic Gross Tumor Volumes Contouring in Esophageal Cancer Based on Contrast-Enhanced Computed Tomography Images: A Multi-Institutional Study

PURPOSE

To develop and externally validate an automatic artificial intelligence (AI) tool for delineating gross tumor volume (GTV) in patients with esophageal squamous cell carcinoma (ESCC), which can assist in neo-adjuvant or radical radiation therapy treatment planning.

METHODS AND MATERIALS

In this multi-institutional study, contrast-enhanced CT images from 580 eligible ESCC patients were retrospectively collected. The GTV contours delineated by 2 experts via consensus were used as ground truth. A 3-dimensional deep learning model was developed for GTV contouring in the training cohort and internally and externally validated in 3 validation cohorts. The AI tool was compared against 12 board-certified experts in 25 patients randomly selected from the external validation cohort to evaluate its assistance in improving contouring performance and reducing variation. Contouring performance was measured using dice similarity coefficient (DSC) and average surface distance. Additionally, our previously established radiomics model for predicting pathologic complete response was used to compare AI-generated and ground truth contours, to assess the potential of the AI contouring tool in radiomics analysis.

RESULTS

The AI tool demonstrated good GTV contouring performance in multicenter validation cohorts, with median DSC values of 0.865, 0.876, and 0.866 and median average surface distance values of 0.939, 0.789, and 0.875 mm, respectively. Furthermore, the AI tool significantly improved contouring performance for half of 12 board-certified experts (DSC values, 0.794-0.835 vs 0.856-0.881, P = .003-0.048), reduced the intra- and interobserver variations by 37.4% and 55.2%, respectively, and saved contouring time by 77.6%. In the radiomics analysis, 88.7% of radiomic features from ground truth and AI-generated contours demonstrated stable reproducibility, and similar pathologic complete response prediction performance for these contours (P = .430) was observed.

CONCLUSIONS

Our AI contouring tool can improve GTV contouring performance and facilitate radiomics analysis in ESCC patients, which indicates its potential for GTV contouring during radiation therapy treatment planning and radiomics studies.

Reproducibility of 3D thoracic aortic displacement from 3D cine balanced SSFP at 3 T without contrast enhancement

PURPOSE

Aortic motion has direct impact on the mechanical stresses acting on the aorta. In aortic disease, increased stiffness of the aorta may lead to decreased aortic motion over time, which could be a predictor for aortic dissection or rupture. This study investigates the reproducibility of obtaining 3D displacement and diameter maps quantified using accelerated 3D cine MRI at 3 T.

METHODS

A noncontrast-enhanced, free-breathing 3D cine sequence based on balanced SSFP and pseudo-spiral undersampling with high spatial isotropic resolution was developed (spatial/temporal resolution [1.6 mm]3 /67 ms). The thoracic aorta of 14 healthy volunteers was prospectively scanned three times at 3 T: twice on the same day and a third time 2 weeks later. Aortic displacement was calculated using iterative closest point nonrigid registration of manual segmentations of the 3D aorta at end-systole and mid-diastole. Interexamination and interobserver regional analysis of mean displacement for five regions of interest was performed using Bland-Altman analysis. Additionally, a complementary voxel-by-voxel analysis was done, allowing a more local inspection of the method.

RESULTS

No significant differences were found in mean and maximum displacement for any of the regions of interest for the interexamination and interobserver analysis. The maximum displacement measured in the lower half of the ascending aorta was 11.0 ± 3.4 mm (range: 3.0-17.5 mm) for the first scan. The smallest detectable change in mean displacement in the lower half of the ascending aorta was 3 mm.

CONCLUSION

Detailed 3D cine balanced SSFP at 3 T allows for reproducible quantification of systolic-diastolic mean aortic displacement within acceptable limits.

Deep learning based automated left ventricle segmentation and flow quantification in 4D flow cardiac MRI

BACKGROUND

4D flow MRI enables assessment of cardiac function and intra-cardiac blood flow dynamics from a single acquisition. However, due to the poor contrast between the chambers and surrounding tissue, quantitative analysis relies on the segmentation derived from a registered cine MRI acquisition. This requires an additional acquisition and is prone to imperfect spatial and temporal inter-scan alignment. Therefore, in this work we developed and evaluated deep learning-based methods to segment the left ventricle (LV) from 4D flow MRI directly.

METHODS

We compared five deep learning-based approaches with different network structures, data pre-processing and feature fusion methods. For the data pre-processing, the 4D flow MRI data was reformatted into a stack of short-axis view slices. Two feature fusion approaches were proposed to integrate the features from magnitude and velocity images. The networks were trained and evaluated on an in-house dataset of 101 subjects with 67,567 2D images and 3030 3D volumes. The performance was evaluated using various metrics including Dice, average surface distance (ASD), end-diastolic volume (EDV), end-systolic volume (ESV), LV ejection fraction (LVEF), LV blood flow kinetic energy (KE) and LV flow components. The Monte Carlo dropout method was used to assess the confidence and to describe the uncertainty area in the segmentation results.

RESULTS

Among the five models, the model combining 2D U-Net with late fusion method operating on short-axis reformatted 4D flow volumes achieved the best results with Dice of 84.52% and ASD of 3.14 mm. The best averaged absolute and relative error between manual and automated segmentation for EDV, ESV, LVEF and KE was 19.93 ml (10.39%), 17.38 ml (22.22%), 7.37% (13.93%) and 0.07 mJ (5.61%), respectively. Flow component results derived from automated segmentation showed high correlation and small average error compared to results derived from manual segmentation.

CONCLUSIONS

Deep learning-based methods can achieve accurate automated LV segmentation and subsequent quantification of volumetric and hemodynamic LV parameters from 4D flow MRI without requiring an additional cine MRI acquisition.

Present and Future Innovations in AI and Cardiac MRI

Cardiac MRI is used to diagnose and treat patients with a multitude of cardiovascular diseases. Despite the growth of clinical cardiac MRI, complicated image prescriptions and long acquisition protocols limit the specialty and restrain its impact on the practice of medicine. Artificial intelligence (AI)-the ability to mimic human intelligence in learning and performing tasks-will impact nearly all aspects of MRI. Deep learning (DL) primarily uses an artificial neural network to learn a specific task from example data sets. Self-driving scanners are increasingly available, where AI automatically controls cardiac image prescriptions. These scanners offer faster image collection with higher spatial and temporal resolution, eliminating the need for cardiac triggering or breath holding. In the future, fully automated inline image analysis will most likely provide all contour drawings and initial measurements to the reader. Advanced analysis using radiomic or DL features may provide new insights and information not typically extracted in the current analysis workflow. AI may further help integrate these features with clinical, genetic, wearable-device, and "omics" data to improve patient outcomes. This article presents an overview of AI and its application in cardiac MRI, including in image acquisition, reconstruction, and processing, and opportunities for more personalized cardiovascular care through extraction of novel imaging markers.

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.

4Dflow-VP-Net: A deep convolutional neural network for noninvasive estimation of relative pressures in stenotic flows from 4D flow MRI

PURPOSE

To estimate relative transvalvular pressure gradient (TVPG) noninvasively from 4D flow MRI.

METHODS

A novel deep learning-based approach is proposed to estimate pressure gradient across stenosis from four-dimensional flow MRI (4D flow MRI) velocities. A deep neural network 4D flow Velocity-to-Presure Network (4Dflow-VP-Net) was trained to learn the spatiotemporal relationship between velocities and pressure in stenotic vessels. Training data were simulated by computational fluid dynamics (CFD) for different pulsatile flow conditions under an aortic flow waveform. The network was tested to predict pressure from CFD-simulated velocity data, in vitro 4D flow MRI data, and in vivo 4D flow MRI data of patients with both moderate and severe aortic stenosis. TVPG derived from 4Dflow-VP-Net was compared to catheter-based pressure measurements for available flow rates, in vitro and Doppler echocardiography-based pressure measurement, in vivo.

RESULTS

Relative pressures calculated by 4Dflow-VP-Net and in vitro pressure catheterization revealed strong correlation (r2  = 0.91). Correlations analysis of TVPG from reference CFD and 4Dflow-VP-Net for 450 simulated flow conditions showed strong correlation (r2  = 0.99). TVPG from in vitro MRI had a correlation coefficient of r2  = 0.98 with reference CFD. 4Dflow-VP-Net, applied to 4D flow MRI in 16 patients, showed comparable TVPG measurement with Doppler echocardiography (r2  = 0.85). Bland-Altman analysis of TVPG measurements showed mean bias and limits of agreement of -0.20 ± 2.07 mmHg and 0.19 ± 0.45 mmHg for CFD-simulated velocities and in vitro 4D flow velocities. In patients, overestimation of Doppler echocardiography relative to TVPG from 4Dflow-VP-Net (10.99 ± 6.77 mmHg) was observed.

CONCLUSION

The proposed approach can predict relative pressure in both in vitro and in vivo 4D flow MRI of aortic stenotic patients with high fidelity.

Segmentation of the aorta in systolic phase from 4D flow MRI: multi-atlas vs. deep learning

OBJECTIVE

In the management of the aortic aneurysm, 4D flow magnetic resonance Imaging provides valuable information for the computation of new biomarkers using computational fluid dynamics (CFD). However, accurate segmentation of the aorta is required. Thus, our objective is to evaluate the performance of two automatic segmentation methods on the calculation of aortic wall pressure.

METHODS

Automatic segmentation of the aorta was performed with methods based on deep learning and multi-atlas using the systolic phase in the 4D flow MRI magnitude image of 36 patients. Using mesh morphing, isotopological meshes were generated, and CFD was performed to calculate the aortic wall pressure. Node-to-node comparisons of the pressure results were made to identify the most robust automatic method respect to the pressures obtained with a manually segmented model.

RESULTS

Deep learning approach presented the best segmentation performance with a mean Dice similarity coefficient and a mean Hausdorff distance (HD) equal to 0.92+/- 0.02 and 21.02+/- 24.20 mm, respectively. At the global level HD is affected by the performance in the abdominal aorta. Locally, this distance decreases to 9.41+/- 3.45 and 5.82+/- 6.23 for the ascending and descending thoracic aorta, respectively. Moreover, with respect to the pressures from the manual segmentations, the differences in the pressures computed from deep learning were lower than those computed from multi-atlas method.

CONCLUSION

To reduce biases in the calculation of aortic wall pressure, accurate segmentation is needed, particularly in regions with high blood flow velocities. Thus, the deep learning segmen-tation method should be preferred.

Segmentation of 4D Flow MRI: Comparison between 3D Deep Learning and Velocity-Based Level Sets

A thoracic aortic aneurysm is an abnormal dilatation of the aorta that can progress and lead to rupture. The decision to conduct surgery is made by considering the maximum diameter, but it is now well known that this metric alone is not completely reliable. The advent of 4D flow magnetic resonance imaging has allowed for the calculation of new biomarkers for the study of aortic diseases, such as wall shear stress. However, the calculation of these biomarkers requires the precise segmentation of the aorta during all phases of the cardiac cycle. The objective of this work was to compare two different methods for automatically segmenting the thoracic aorta in the systolic phase using 4D flow MRI. The first method is based on a level set framework and uses the velocity field in addition to 3D phase contrast magnetic resonance imaging. The second method is a U-Net-like approach that is only applied to magnitude images from 4D flow MRI. The used dataset was composed of 36 exams from different patients, with ground truth data for the systolic phase of the cardiac cycle. The comparison was performed based on selected metrics, such as the Dice similarity coefficient (DSC) and Hausdorf distance (HD), for the whole aorta and also three aortic regions. Wall shear stress was also assessed and the maximum wall shear stress values were used for comparison. The U-Net-based approach provided statistically better results for the 3D segmentation of the aorta, with a DSC of 0.92 ± 0.02 vs. 0.86 ± 0.5 and an HD of 21.49 ± 24.8 mm vs. 35.79 ± 31.33 mm for the whole aorta. The absolute difference between the wall shear stress and ground truth slightly favored the level set method, but not significantly (0.754 ± 1.07 Pa vs. 0.737 ± 0.79 Pa). The results showed that the deep learning-based method should be considered for the segmentation of all time steps in order to evaluate biomarkers based on 4D flow MRI.

Advances in machine learning applications for cardiovascular 4D flow MRI

Four-dimensional flow magnetic resonance imaging (MRI) has evolved as a non-invasive imaging technique to visualize and quantify blood flow in the heart and vessels. Hemodynamic parameters derived from 4D flow MRI, such as net flow and peak velocities, but also kinetic energy, turbulent kinetic energy, viscous energy loss, and wall shear stress have shown to be of diagnostic relevance for cardiovascular diseases. 4D flow MRI, however, has several limitations. Its long acquisition times and its limited spatio-temporal resolutions lead to inaccuracies in velocity measurements in small and low-flow vessels and near the vessel wall. Additionally, 4D flow MRI requires long post-processing times, since inaccuracies due to the measurement process need to be corrected for and parameter quantification requires 2D and 3D contour drawing. Several machine learning (ML) techniques have been proposed to overcome these limitations. Existing scan acceleration methods have been extended using ML for image reconstruction and ML based super-resolution methods have been used to assimilate high-resolution computational fluid dynamic simulations and 4D flow MRI, which leads to more realistic velocity results. ML efforts have also focused on the automation of other post-processing steps, by learning phase corrections and anti-aliasing. To automate contour drawing and 3D segmentation, networks such as the U-Net have been widely applied. This review summarizes the latest ML advances in 4D flow MRI with a focus on technical aspects and applications. It is divided into the current status of fast and accurate 4D flow MRI data generation, ML based post-processing tools for phase correction and vessel delineation and the statistical evaluation of blood flow.