FReSCO: Flow Reconstruction and Segmentation for low‐latency Cardiac Output monitoring using deep artifact suppression and segmentation

Abstract

Purpose Real‐time monitoring of cardiac output (CO) requires low‐latency reconstruction and segmentation of real‐time phase‐contrast MR, which has previously been difficult to perform. Here we propose a deep learning framework for “FReSCO” (Flow Reconstruction and Segmentation for low latency Cardiac Output monitoring). Methods Deep artifact suppression and segmentation U‐Nets were independently trained. Breath‐hold spiral phase‐contrast MR data (N = 516) were synthetically undersampled using a variable‐density spiral sampling pattern and gridded to create aliased data for training of the artifact suppression U‐net. A subset of the data (N = 96) was segmented and used to train the segmentation U‐net. Real‐time spiral phase‐contrast MR was prospectively acquired and then reconstructed and segmented using the trained models (FReSCO) at low latency at the scanner in 10 healthy subjects during rest, exercise, and recovery periods. Cardiac output obtained via FReSCO was compared with a reference rest CO and rest and exercise compressed‐sensing CO. Results The FReSCO framework was demonstrated prospectively at the scanner. Beat‐to‐beat heartrate, stroke volume, and CO could be visualized with a mean latency of 622 ms. No significant differences were noted when compared with reference at rest (bias = −0.21 ± 0.50 L/min, p = 0.246) or compressed sensing at peak exercise (bias = 0.12 ± 0.48 L/min, p = 0.458). Conclusions The FReSCO framework was successfully demonstrated for real‐time monitoring of CO during exercise and could provide a convenient tool for assessment of the hemodynamic response to a range of stressors.

Movienet: Deep space–time‐coil reconstruction network without k‐space data consistency for fast motion‐resolved 4D MRI

“…Movienet would enable rapid motion dictionary calculation and adaptation to increase the motion‐tracking performance of MRSIGMA. Subsecond reconstructions were also accomplished in the MR‐MOTUS 9 technique by exploiting the low‐rank structure of respiratory motion to prelearn a motion model and frequent updates based on sparse k‐space data, and in the FReSCO 32 technique that uses deep learning to perform simultaneous artifact suppression and segmentation for low‐latency cardiac imaging.…”

Deep Learning-Based Reconstruction for Cardiac MRI: A Review

“…Compared to the ECG-gated 2D-Flow with R = 2 PI, their prospectively deployed real-time radial 2D-Flow sequence with DL-based reconstruction provided comparable measurements of the stroke volume (−0.8 [−10.5, 8.9] mL) and peak mean velocity (−2.5 [−13.7, 8.7] cm/s) in the aAo while also demonstrating a reduced bias and tighter LOA for the stroke volume but not the peak mean velocity, compared to the CS [ 107 ]. They later expanded their original work to demonstrate real-time 2D-Flow measurements, including visualization on the scanner of the beat-to-beat heart rate, stroke volume, and cardiac output with a mean latency of 622 ms [ 101 ]. Importantly, no significant differences were observed between the reference R = 2 PI methods at rest (−0.21 ± 0.50 L/min, p = 0.25) or the CS at peak exercise (0.12 ± 0.48 L/min, p = 0.46).…”

The future of CMR: All-in-one vs. real-time CMR (Part 2)

“…Another such study used EPI to generate 5x5mm 2 resolution, 40 ms temporal frames, and studied large artery and vein flow while undergoing Valsalva maneuver [85] . Many acceleration methods have since been applied to phase-contrast imaging [86] , including k-t view sharing [87] , parallel imaging [88] , compressed sensing, low-rank reconstructions [58] , and deep-learning [89] , [90] , with testing primarily focused on aortic flow, and demonstration of good temporal and spatial resolution. Temporal resolution is critical in phase-contrast [91] , and flow wave-forms, especially flow peaks, are highly sensitive to actual temporal resolution.…”