Super-resolution and denoising of 4D-Flow MRI using physics-Informed deep neural nets

Fathi, Mojtaba F.; Perez-Raya, Isaac; Baghaie, Ahmadreza; Berg, Philipp; Janiga, Gábor; Arzani, Amirhossein; D’Souza, Roshan

doi:10.1016/j.cmpb.2020.105729

Cited by 83 publications

(59 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Under the same pulse pressure, the compliant aorta produced a higher peak flow rate and a more negative flow rate at end-systole than the rigid simulation, resulting in favourable agreement with 4D-Flow MRI data in the compliant case. This effect is expected due to the accumulation and subse- The boundary layer thickness of ≈ 1mm (see 4) is much smaller than the highest achievable resolution of 4D-Flow MRI in the large vessels, which typically exceeds 2mm (Fathi et al, 2020). As discussed, the compliant simulation achieved a higher peak flow rate.…”

Section: Discussionmentioning

confidence: 89%

See 1 more Smart Citation

A novel MRI-based data fusion methodology for efficient, personalised, compliant simulations of aortic haemodynamics

Stokes

Bonfanti

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

We present a novel, cost-efficient methodology to simulate aortic haemodynamics in a patient-specific, compliant aorta using an MRI data fusion process. Based on a previously-developed Moving Boundary Method, this technique circumvents the high computational cost and numerous structural modelling assumptions required by traditional Fluid-Structure Interaction techniques. Without the need for Computed Tomography (CT) data, the MRI images required to construct the simulation can be obtained during a single imaging session. Black Blood MR Angiography and 2D Cine-MRI data were used to reconstruct the luminal geometry and calibrate wall movement specifically to each region of the aorta. 4D-Flow MRI and non-invasive pressure measurements informed patient-specific inlet and outlet boundary conditions. Simulated wall movement closely matched 2D Cine-MRI measurements throughout the aorta, and physiological pressure and flow distributions in CFD were achieved within 3.3% of patient-specific targets. Excellent agreement with 4D-Flow MRI velocity data was observed. Conversely, a rigid-wall simulation under-predicted peak flow rate and systolic maximum velocities whilst predicting a mean Time-Averaged Wall Shear Stress (TAWSS) 13% higher than the compliant simulation. The excellent agreement observed between compliant simulation results and MRI is testament to the accuracy and efficiency of this MRI-based technique.

show abstract

Section: Discussionmentioning

confidence: 89%

“…The boundary layer thickness of ≈ 1mm (see 4) is much smaller than the highest achievable resolution of 4D-Flow MRI in the large vessels, which typically exceeds 2mm (Fathi et al, 2020). As discussed, the compliant simulation achieved a higher peak flow rate.…”

Section: Discussionmentioning

confidence: 90%

A novel MRI-based data fusion methodology for efficient, personalised, compliant simulations of aortic haemodynamics

Stokes

Bonfanti

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Continuing into the cerebrovascular space, a few very recent works have shown how merging physics-informed analysis, machine learning, and imaging can have particular promise for improving non-invasive cerebrovascular assessment. Fathi et al [18] used a patient-specific PINN to recover regional flow and pressure from input 4D Flow MRI, promising virtually unrestricted spatiotemporal refinements on recovered velocity fields. Similarly, Rutkowski et al [12] recently presented a CNN-based network to reconstruct super-resolution 4D Flow MRI in a cerebrovascular setting, using patient-specific in-vitro models for both training and testing.…”

Section: Discussionmentioning

confidence: 99%

Cerebrovascular super-resolution 4D Flow MRI – using deep learning to non-invasively quantify velocity, flow, and relative pressure

Ferdian

Marlevi

Schollenberger

et al. 2021

Preprint

View full text Add to dashboard Cite

The development of cerebrovascular disease is tightly coupled to changes in cerebrovascular hemodynamics, with altered flow and relative pressure indicative of the onset, development, and acute manifestation of pathology. Image-based monitoring of cerebrovascular hemodynamics is, however, complicated by the narrow and tortuous vasculature, where accurate output directly depends on sufficient spatial resolution. To address this, we present a method combining dedicated deep learning and state-of-the-art 4D Flow MRI to generate super-resolution full-field images with coupled quantification of relative pressure using a physics-driven image processing approach. The method is trained and validated in a patient-specific in-silico cohort, showing good accuracy in estimating velocity (relative error: 12.0 ± 0.1%, mean absolute error (MAE): 0.07 ± 0.06 m/s at peak velocity), flow (relative error: 6.6 ± 4.7%, root mean square error (RMSE): 0.5 ± 0.1 mL/s at peak flow), and with maintained recovery of relative pressure through the circle of Willis (relative error: 11.0 ± 7.3%, RMSE: 0.3 ± 0.2 mmHg). Furthermore, the method is applied to an in-vivo volunteer cohort, effectively generating data at <0.5mm resolution and showing potential in reducing low-resolution bias in relative pressure estimation. Our approach presents a promising method to non-invasively quantify cerebrovascular hemodynamics, applicable to dedicated clinical cohorts in the future.

show abstract

“…dynamic contrast-enhanced imaging). Finally, PINNs have been recently used for improving 4D flow MRI data fidelity (super-resolution and denoising) [127].…”

Section: Machine Learning and Neural Networkmentioning

confidence: 99%

Data-driven cardiovascular flow modelling: examples and opportunities

Arzani

Dawson

2021

J. R. Soc. Interface.

Self Cite

View full text Add to dashboard Cite

High-fidelity blood flow modelling is crucial for enhancing our understanding of cardiovascular disease. Despite significant advances in computational and experimental characterization of blood flow, the knowledge that we can acquire from such investigations remains limited by the presence of uncertainty in parameters, low resolution, and measurement noise. Additionally, extracting useful information from these datasets is challenging. Data-driven modelling techniques have the potential to overcome these challenges and transform cardiovascular flow modelling. Here, we review several data-driven modelling techniques, highlight the common ideas and principles that emerge across numerous such techniques, and provide illustrative examples of how they could be used in the context of cardiovascular fluid mechanics. In particular, we discuss principal component analysis (PCA), robust PCA, compressed sensing, the Kalman filter for data assimilation, low-rank data recovery, and several additional methods for reduced-order modelling of cardiovascular flows, including the dynamic mode decomposition and the sparse identification of nonlinear dynamics. All techniques are presented in the context of cardiovascular flows with simple examples. These data-driven modelling techniques have the potential to transform computational and experimental cardiovascular research, and we discuss challenges and opportunities in applying these techniques in the field, looking ultimately towards data-driven patient-specific blood flow modelling.

show abstract

Super-resolution and denoising of 4D-Flow MRI using physics-Informed deep neural nets

Cited by 83 publications

References 57 publications

A novel MRI-based data fusion methodology for efficient, personalised, compliant simulations of aortic haemodynamics

A novel MRI-based data fusion methodology for efficient, personalised, compliant simulations of aortic haemodynamics

Cerebrovascular super-resolution 4D Flow MRI – using deep learning to non-invasively quantify velocity, flow, and relative pressure

Data-driven cardiovascular flow modelling: examples and opportunities

Contact Info

Product

Resources

About