Abstract:Background:Cardiac T2 mapping is a valuable tool for diagnosing myocardial edema, inflammation, and infiltration, yet its spatial resolution is limited by the single-shot balanced steady-state free precession acquisition and duration of the cardiac quiescent period, which may reduce sensitivity in detecting focal lesions in the myocardium. To improve spatial resolution without extending the acquisition window, this study examined a novel accelerated Cartesian cardiac T2 mapping technique. Methods:We introduce a novel improved-resolution cardiac T2 mapping approach leveraging a calibrationless space-contrast-coil locally low-rank tensor (SCC-LLRT)-constrained reconstruction algorithm in conjunction with Cartesian undersampling trajectory. The method was validated with phantom imaging and in vivo imaging that involved 13 healthy participants and 20 patients. The SCC-LLRT algorithm was compared with a conventional locally low-rank (LLR)-constrained algorithm and a nonlinear inversion (NLINV) reconstruction algorithm. The improved-resolution T2 mapping (1.4 mm × 1.4 mm) was compared globally and regionally with the regular-resolution T2 mapping (2.3 mm × 1.9 mm) according to the 16-segment model of the American Heart Association. The agreement between the improved-resolution and regular-resolution T2 mappings was evaluated by linear regression and Bland-Altman analyses. Image quality was scored by two experienced reviewers on a five-point scale (1, worst; 5, best). Results:In healthy participants, SCC-LLRT significantly reduced artifacts (4.50±0.39) compared with LLR (2.31±0.60; P<0.001) and NLINV (3.65±0.56; P<0.01), suppressed noise (4.12±0.35) compared with NLINV (2.65±0.50; P<0.001), and improved the overall image quality (4.38±0.40) compared with LLR (2.54±0.41; P<0.001) and NLINV (3.04±0.50; P<0.001). Compared with the regular-resolution T2 mapping, the proposed method significantly improved the sharpness of myocardial boundaries (4.46±0.60 vs. 3.04±0.50; P<0.001) and the conspicuity of papillary muscles and fine structures (4.46±0.63 vs. 2.65±0.30; P<0.001). Myocardial T2 values obtained with the proposed method correlated significantly with those from regular-resolution T2 mapping in both healthy participants (r=0.79; P<0.01) and patients (r=0.94; P<0.001). Conclusions:The proposed SCC-LLRT-constrained reconstruction algorithm in conjunction with Cartesian undersampling pattern achieved improved-resolution cardiac T2 mapping of comparable accuracy, precision, and scan-rescan reproducibility compared with the regular-resolution T2 mapping. The higher resolution improved the sharpness of myocardial borders and the conspicuity of image fine details, which may increase diagnostic confidence in cardiac T2 mapping for detecting small lesions.

Automated quantification of myocardial tissue characteristics from native T1 mapping using neural networks with uncertainty-based quality-control

Automated quantification of myocardial tissue characteristics from native T1 mapping using neural networks with Bayesian inference for uncertainty-based quality-control

Free-breathing, Non-Ecg, Simultaneous Myocardial T1 , T2 , T2 *, and Fat-Fraction Mapping with Motion-Resolved Cardiovascular MR Multitasking.

High Spatial-Resolution and Acquisition-Efficiency Cardiac MR T1 Mapping Based on Radial bSSFP and a Low-Rank Tensor Constraint

Automatic pipeline for segmentation of LV myocardium on quantitative MR T1 maps using deep learning model and computation of radial T1 and ECV values

Development and performance evaluation of fully automated deep learning-based models for myocardial segmentation on T1 mapping MRI data

Quality assurance of quantitative cardiac T1-mapping in multicenter clinical trials – A T1 phantom program from the hypertrophic cardiomyopathy registry (HCMR) study

Deep neural network ensemble for on-the-fly quality control-driven segmentation of cardiac MRI T1 mapping

Highly automatic quantification of myocardial oedema in patients with acute myocardial infarction using bright blood T2-weighted CMR

PCMC-T1: Free-breathing myocardial T1 mapping with Physically-Constrained Motion Correction

Automated T1 and T2 mapping segmentation on cardiovascular magnetic resonance imaging using deep learning

Fully Automated Regional Analysis of Myocardial T2* Values for Iron Quantification Using Deep Learning

$\textbf{MyoMapNet}$: Accelerated Modified Look-Locker Inversion Recovery Myocardial T1 Mapping via Neural Networks

Accelerated Cartesian Cardiac T2 Mapping Based on a Calibrationless Locally Low-Rank Tensor Constraint

Automated quantitative analysis of first-pass myocardial perfusion magnetic resonance imaging data

Fully Automatic initialization and segmentation of left and right ventricles for large-scale cardiac MRI using a deeply supervised network and 3D-ASM

MBSS-T1: Model-Based Self-Supervised Motion Correction for Robust Cardiac T1 Mapping

A Robust Segmentation Method with Triple-Factor Non-Negative Matrix Factorization for Myocardial Blood Flow Quantification from Dynamic Rb-82 Positron Emission Tomography

MOCOnet: Robust Motion Correction of Cardiovascular Magnetic Resonance T1 Mapping Using Convolutional Neural Networks

Motion Correction for Native Myocardial T1 Mapping Using Self-Supervised Deep Learning Registration with Contrast Separation.

Deep learning with attention supervision for automated motion artefact detection in quality control of cardiac T1-mapping