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.

[1]  A. Huber,et al.  Advances in machine learning applications for cardiovascular 4D flow MRI , 2022, Frontiers in Cardiovascular Medicine.

[2]  A. Amini,et al.  FlowRAU-Net: Accelerated 4D Flow MRI of Aortic Valvular Flows With a Deep 2D Residual Attention Network , 2022, IEEE Transactions on Biomedical Engineering.

[3]  T. Ebbers,et al.  Automatic Time‐Resolved Cardiovascular Segmentation of 4D Flow MRI Using Deep Learning , 2022, Journal of magnetic resonance imaging : JMRI.

[4]  Bradley D. Allen,et al.  Deep learning–based velocity antialiasing of 4D‐flow MRI , 2022, Magnetic resonance in medicine.

[5]  A. Amini,et al.  Relative pressure estimation from 4D flow MRI using generalized Bernoulli equation in a phantom model of arterial stenosis , 2022, Magnetic Resonance Materials in Physics, Biology and Medicine.

[6]  A. Amini,et al.  4D flow MRI and CFD simulations of pulsatile flow in a phantom model of arterial stenosis: visualizing the vortex dynamics , 2021 .

[7]  M. Mack,et al.  2020 ACC/AHA Guideline for the Management of Patients With Valvular Heart Disease: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. , 2020, Journal of the American College of Cardiology.

[8]  David Nolte,et al.  Validation of 4D Flow based relative pressure maps in aortic flows , 2020, Medical Image Anal..

[9]  Avan Suinesiaputra,et al.  4DFlowNet: Super-Resolution 4D Flow MRI Using Deep Learning and Computational Fluid Dynamics , 2020, Frontiers in Physics.

[10]  Michael B. Scott,et al.  Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning , 2020, Magnetic resonance in medicine.

[11]  Chenguang Huang,et al.  A flow feature detection method for modeling pressure distribution around a cylinder in non-uniform flows by using a convolutional neural network , 2020, Scientific Reports.

[12]  A. Amini,et al.  Dual‐Venc acquisition for 4D flow MRI in aortic stenosis with spiral readouts , 2019, Journal of magnetic resonance imaging : JMRI.

[13]  Hojin Ha,et al.  Non-invasive estimation of relative pressure in turbulent flow using virtual work-energy , 2019, Medical Image Anal..

[14]  Yiping Lu,et al.  PDE-Net 2.0: Learning PDEs from Data with A Numeric-Symbolic Hybrid Deep Network , 2018, J. Comput. Phys..

[15]  Hui Li,et al.  Prediction model of velocity field around circular cylinder over various Reynolds numbers by fusion convolutional neural networks based on pressure on the cylinder , 2018 .

[16]  A. Kheradvar,et al.  Simplified Bernoulli's method significantly underestimates pulmonary transvalvular pressure drop , 2016, Journal of magnetic resonance imaging : JMRI.

[17]  T. Ebbers,et al.  4D Flow MRI‐based pressure loss estimation in stenotic flows: Evaluation using numerical simulations , 2016, Magnetic resonance in medicine.

[18]  G. Kim,et al.  Turbulent Kinetic Energy Measurement Using Phase Contrast MRI for Estimating the Post-Stenotic Pressure Drop: In Vitro Validation and Clinical Application , 2016, PloS one.

[19]  Marcus F Stoddard,et al.  4D spiral imaging of flows in stenotic phantoms and subjects with aortic stenosis , 2016, Magnetic resonance in medicine.

[20]  Pablo Lamata,et al.  Non-invasive pressure difference estimation from PC-MRI using the work-energy equation , 2015, Medical Image Anal..

[21]  M. Markl,et al.  4D flow cardiovascular magnetic resonance consensus statement , 2015, Journal of Cardiovascular Magnetic Resonance.

[22]  Mo Kadbi,et al.  Noninvasive 3D pressure calculation from PC-MRI via non-iterative harmonics-based orthogonal projection: Constant flow experiment , 2013, 2013 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[23]  Amir A Amini,et al.  Factors affecting the accuracy of pressure measurements in vascular stenoses from phase‐contrast MRI , 2004, Magnetic resonance in medicine.

[24]  D N Firmin,et al.  Computation of flow pressure fields from magnetic resonance velocity mapping , 1996, Magnetic resonance in medicine.

[25]  H. Gudbjartsson,et al.  The rician distribution of noisy mri data , 1995, Magnetic resonance in medicine.

[26]  L Hatle,et al.  Non-invasive assessment of aortic stenosis by Doppler ultrasound. , 1980, British heart journal.

[27]  E. Tseng,et al.  Magnetic resonance measurement of turbulent kinetic energy for the estimation of irreversible pressure loss in aortic stenosis. , 2013, JACC. Cardiovascular imaging.

[28]  J. Chambers,et al.  Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice. , 2009, European journal of echocardiography : the journal of the Working Group on Echocardiography of the European Society of Cardiology.