Rapid reconstruction of highly undersampled, non‐Cartesian real‐time cine k‐space data using a perceptual complex neural network (PCNN). (1st September 2020)

• Record Type: Journal Article
• Authors: Shen, Daming; Ghosh, Sushobhan; Haji‐Valizadeh, Hassan; Pathrose, Ashitha; Schiffers, Florian; Lee, Daniel C.; Freed, Benjamin H.; Markl, Michael; Cossairt, Oliver S.; Katsaggelos, Aggelos K.; Kim, Daniel
• Is Part Of: NMR in biomedicine. Volume 34:Number 1(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-09-01
• Subjects: compressed sensing (CS) -- convolutional neural network (CNN) -- deep learning (DL) -- perceptual complex neural network (PCNN) -- perceptual loss -- real‐time cine MRI; Nuclear magnetic resonance -- Periodicals; Magnetic Resonance Spectroscopy -- Periodicals; 574
• DOI: 10.1002/nbm.4405
• Languages: English
• ISSNs: 0952-3480

Abstract

Abstract : Highly accelerated real‐time cine MRI using compressed sensing (CS) is a promising approach to achieve high spatio‐temporal resolution and clinically acceptable image quality in patients with arrhythmia and/or dyspnea. However, its lengthy image reconstruction time may hinder its clinical translation. The purpose of this study was to develop a neural network for reconstruction of non‐Cartesian real‐time cine MRI k ‐space data faster (<1 min per slice with 80 frames) than graphics processing unit (GPU)‐accelerated CS reconstruction, without significant loss in image quality or accuracy in left ventricular (LV) functional parameters. We introduce a perceptual complex neural network (PCNN) that trains on complex‐valued MRI signal and incorporates a perceptual loss term to suppress incoherent image details. This PCNN was trained and tested with multi‐slice, multi‐phase, cine images from 40 patients (20 for training, 20 for testing), where the zero‐filled images were used as input and the corresponding CS reconstructed images were used as practical ground truth. The resulting images were compared using quantitative metrics (structural similarity index (SSIM) and normalized root mean square error (NRMSE)) and visual scores (conspicuity, temporal fidelity, artifacts, and noise scores), individually graded on a five‐point scale (1, worst; 3, acceptable; 5, best), and LV ejection fraction (LVEF). The mean processing time per slice with 80 frames for PCNN was 23.7 ± 1.9 s for pre‐processing (Step 1, same as CS) and 0.822 ± 0.004 s for dealiasing (Step 2, 166 times faster than CS). Our PCNN produced higher data fidelity metrics (SSIM = 0.88 ± 0.02, NRMSE = 0.014 ± 0.004) compared with CS. While all the visual scores were significantly different ( P < 0.05), the median scores were all 4.0 or higher for both CS and PCNN. LVEFs measured from CS and PCNN were strongly correlated ( R 2 = 0.92) and in good agreement (mean difference = −1.4% [2.3% of mean]; limit of agreement = 10.6% [17.6% of mean]). The proposed PCNN is capable of rapid reconstruction (25 s per slice with 80 frames) of non‐Cartesian real‐time cine MRI k ‐space data, without significant loss in image quality or accuracy in LV functional parameters. Abstract : This study describes implementation of a neural network for reconstruction of non‐Cartesian real‐time cine MRI k ‐space data faster (<30 s per slice with 80 frames) than graphics processing unit‐accelerated compressed sensing (CS) reconstruction, without significant loss in image quality or accuracy in left ventricular (LV) functional parameters. We implemented a perceptual complex neural network (PCNN) that trains on complex MRI signal and incorporates a perceptual loss term to retain high‐level features better than a pixel‐wise loss term. This PCNN was trained and tested with multi‐slice, multi‐phase, cine images from 40 patients (20 for training and 20 for testing), where the zero‐filled images were used as input and the corresponding CS reconstructed images were used as practical ground truth. The mean processing time per slice with 80 frames for PCNN was 23.7 ± 1.9 s for pre‐processing (Step 1, same as CS) and 0.822 ± 0.004 s for dealiasing (Step 2, 166 times faster than CS). Our PCNN produced higher data fidelity metrics (structural similarity index = 0.88 ± 0.02, normalized root mean square error = 0.014 ± 0.004) compared with CS. LV ejection fraction measurements were strongly correlated ( R 2 = 0.82) and in good agreement (mean difference = −3.1% [5.2% of mean]; limit of agreement = 15.6% [25.7% of mean]) between CS and PCNN. The proposed PCNN is capable of rapid reconstruction (25 s per slice with 80 frames) of non‐Cartesian real‐time cine MRI k ‐space data, without significant loss in image quality or accuracy in LV functional parameters.