Fully automated 3D aortic segmentation of 4D flow MRI for hemodynamic analysis using deep learning

Abstract

Purpose To generate fully automated and fast 4D‐flow MRI‐based 3D segmentations of the aorta using deep learning for reproducible quantification of aortic flow, peak velocity, and dimensions. Methods A total of 1018 subjects with aortic 4D‐flow MRI (528 with bicuspid aortic valve, 376 with tricuspid aortic valve and aortic dilation, 114 healthy controls) comprised the data set. A convolutional neural network was trained to generate 3D aortic segmentations from 4D‐flow data. Manual segmentations served as the ground truth (N = 499 training, N = 101 validation, N = 418 testing). Dice scores, Hausdorff distance, and average symmetrical surface distance were calculated to assess performance. Aortic flow, peak velocity, and lumen dimensions were quantified at the ascending, arch, and descending aorta and compared using Bland‐Altman analysis. Interobserver variability of manual analysis was assessed on a subset of 40. Results Convolutional neural network segmentation required 0.438 ± 0.355 seconds versus 630 ± 254 seconds for manual analysis and demonstrated excellent performance with a median Dice score of 0.951 (0.930‐0.966), Hausdorff distance of 2.80 (2.13‐4.35), and average symmetrical surface distance of 0.176 (0.119‐0.290). Excellent agreement was found for flow, peak velocity, and dimensions with low bias and limits of agreement less than 10% difference versus manual analysis. For aortic volume, limits of agreement were moderate within 16.3%. Interobserver variability (median Dice score: 0.950; Hausdorff distance: 2.45; and average symmetrical surface distance: 0.145) and convolutional neural network–based analysis (median Dice score: 0.953‐0.959; Hausdorff distance: 2.24‐2.91; and average symmetrical surface distance: 0.145‐1.98 to observers) demonstrated similar reproducibility. Conclusions Deep learning enabled fast and automated 3D aortic segmentation from 4D‐flow MRI, demonstrating its potential for efficient clinical workflows. Future studies should investigate its utility for other vasculature and multivendor applications.

Highly accelerated aortic 4D flow MRI using compressed sensing: Performance at different acceleration factors in patients with aortic disease

“…A 3D phase‐contrast MR angiogram (PC‐MRA) was calculated from 4D flow data as previously described 39 . The 3D segmentations of the thoracic aorta were generated to mask the blood flow velocities in the thoracic aorta either manually (Mimics; Materialise, Leuven, Belgium) or automatically using an in‐house deep learning‐based method 40 . These segmentations were used to create peak‐systolic velocity maximum intensity projections (MIPs) 41 to quantify V max in three contiguous regions of interest (ROIs) for each scan: the ascending aorta (AAo), aortic arch (arch), and descending aorta (DAo).…”

“…39 The 3D segmentations of the thoracic aorta were generated to mask the blood flow velocities in the thoracic aorta either manually (Mimics; Materialise, Leuven, Belgium) or automatically using an in-house deep learning-based method. 40 These segmentations were used to create peak-systolic velocity maximum intensity projections (MIPs) 41 to quantify V max in three contiguous regions of interest (ROIs) for each scan: the ascending aorta (AAo), aortic arch (arch), and descending aorta (DAo). For time-resolved flow evaluation and Q max and Q net quantification, three 2D planes were placed orthogonal to the midline at the AAo, arch, and DAo on the segmented volume of the aorta derived from the conventional 4D flow scan (EnSight, version 10.0.3; CEI, Apex, NC, USA).…”

Automatic Vertebral Body Segmentation Based on Deep Learning of Dixon Images for Bone Marrow Fat Fraction Quantification

“…Although quantitative methods for BMF measurement are promising, BMF analysis from the resulting images requires manual identification of vertebral bodies and segmentation of regions of interest which, in MR images, is laborious and time consuming. One potential solution is deep learning, which has demonstrated robust automated segmentation performance in various biomedical imaging problems ( 13 – 16 ). Given large amounts of data, deep learning algorithms, mostly in the form of convolutional neural networks (CNNs), can automatically learn and thus predict representative features for a given medical imaging problem.…”

Complicated Double-Orifice Mitral Regurgitation: Combined Hemodynamic Assessment Using Echocardiography and Four-Dimensional Flow Magnetic Resonance Imaging

“…Efforts are under way to reduce scan time (∼2 min) 11 and to use deep learning to streamline data processing. 12 Also, commercial products are now equipped with 4D flow MRI analysis function (e.g., Circle CVI, Pie Medial Imaging, and Arterys), which should extend availability of the technique. However, it remains too early to demonstrate the clinical benefit of adding 4D flow MRI to a clinical MR assessment routine.…”