Imaging sequence for joint myocardial T1 mapping and fat/water separation.

PURPOSE: To develop and evaluate an imaging sequence to simultaneously quantify the epicardial fat volume and myocardial T1 relaxation time. METHODS: We introduced a novel simultaneous myocardial T1 mapping and fat/water separation sequence (joint T1 -fat/water separation). Dixon reconstruction is performed on a dual-echo data set to generate water/fat images. T1 maps are computed using the water images, whereas the epicardial fat volume is calculated from the fat images. A phantom experiment using vials with different T1 /T2 values and a bottle of oil was performed. Additional phantom experiment using vials of mixed fat/water was performed to show the potential of this sequence to mitigate the effect of intravoxel fat on estimated T1 maps. In vivo evaluation was performed in 17 subjects. Epicardial fat volume, native myocardial T1 measurements and precision were compared among slice-interleaved T1 mapping, Dixon, and the proposed sequence. RESULTS: In the first phantom, the proposed sequence separated oil from water vials and there were no differences in T1 of the fat-free vials (P = .1). In the second phantom, the T1 error decreased from 22%, 36%, 57%, and 73% to 8%, 9%, 16%, and 26%, respectively. In vivo there was no difference between myocardial T1 values (1067 ± 17 ms versus 1077 ± 24 ms, P = .6). The epicardial fat volume was similar for both sequences (54.3 ± 33 cm3 versus 52.4 ± 32 cm3 , P = .8). CONCLUSION: The proposed sequence provides simultaneous quantification of native myocardial T1 and epicardial fat volume. This will eliminate the need for an additional sequence in the cardiac imaging protocol if both measurements are clinically indicated.

Integrated motion correction and dictionary learning for free-breathing myocardial T1 mapping.

PURPOSE: To develop and evaluate an integrated motion correction and dictionary learning (MoDic) technique to accelerate data acquisition for myocardial T1 mapping with improved accuracy. METHODS: MoDic integrates motion correction with dictionary learning-based reconstruction. A random undersampling scheme was implemented for slice-interleaved T1 mapping sequence to allow prospective undersampled data acquisition. Phantom experiments were performed to evaluate the effect of reconstruction on T1 measurement. In vivo T1 mappings were acquired in 8 healthy subjects using 6 different acceleration approaches: uniform or randomly undersampled k-space data with reduction factors (R) of 2, 3, and 4. Uniform undersampled data were reconstructed with SENSE, and randomly undersampled k-space data were reconstructed using dictionary learning, compressed sensing SENSE, and MoDic methods. Three expert readers subjectively evaluated the quality of T1 maps using a 4-point scoring system. The agreement between T1 values was assessed by Bland-Altman analysis. RESULTS: In the phantom study, the accuracy of T1 measurements improved with increasing reduction factors ( - 31 ± 35 ms, - 13 ± 18 ms, and - 5 ± 11 ms for reduction factor (R) = 2 to 4, respectively). The image quality of in vivo T1 maps assessed by subjective scoring using MoDic was similar to that of SENSE at R = 2 (P = .61) but improved at R = 3 and 4 (P < .01). The scores of dictionary learning (2.98 ± 0.71, 2.91 ± 0.60, and 2.67 ± 0.71 for R = 2 to 4) and CS-SENSE (3.32 ± 0.42, 3.05 ± 0.43, and 2.53 ± 0.43) were lower than those of MoDic (3.48 ± 0.46, 3.38 ± 0.52, and 2.9 ± 0.60) for all reduction factors (P < .05 for all). CONCLUSION: The MoDic method accelerates data acquisition for myocardial T1 mapping with improved T1 measurement accuracy.

Black blood myocardial T2 mapping.

PURPOSE: To develop a black blood heart-rate adaptive T2 -prepared balanced steady-state free-precession (BEATS) sequence for myocardial T2 mapping. METHODS: In BEATS, blood suppression is achieved by using a combination of preexcitation and double inversion recovery pulses. The timing and flip angles of the preexcitation pulse are auto-calculated in each patient based on heart rate. Numerical simulations, phantom studies, and in vivo studies were conducted to evaluate the performance of BEATS. BEATS T2 maps were acquired in 36 patients referred for clinical cardiac MRI and in 1 swine with recent myocardial infarction. Two readers assessed all images acquired in patients to identify the presence of artifacts associated with slow blood flow. RESULTS: Phantom experiments showed that the BEATS sequence provided accurate T2 values over a wide range of simulated heart rates. Black blood myocardial T2 maps were successfully obtained in all subjects. No significant difference was found between the average T2 measurements obtained from the BEATS and conventional bright-blood T2 ; however, there was a decrease in precision using the BEATS sequence. A suppression of the blood pool resulted in sharper definition of the blood-myocardium border and reduced partial voluming effect. The subjective assessment showed that 16% (18 out of 108) of short-axis slices have residual blood artifacts (12 in the apical slice, 4 in the midventricular slice, and 2 in the basal slice). CONCLUSION: The BEATS sequence yields dark blood myocardial T2 maps with better definition of the blood-myocardium border. Further studies are warranted to evaluate diagnostic accuracy of black blood T2 mapping.

Automated analysis of cardiovascular magnetic resonance myocardial native T1 mapping images using fully convolutional neural networks.

BACKGROUND: Cardiovascular magnetic resonance (CMR) myocardial native T1 mapping allows assessment of interstitial diffuse fibrosis. In this technique, the global and regional T1 are measured manually by drawing region of interest in motion-corrected T1 maps. The manual analysis contributes to an already lengthy CMR analysis workflow and impacts measurements reproducibility. In this study, we propose an automated method for combined myocardium segmentation, alignment, and T1 calculation for myocardial T1 mapping. METHODS: A deep fully convolutional neural network (FCN) was used for myocardium segmentation in T1 weighted images. The segmented myocardium was then resampled on a polar grid, whose origin is located at the center-of-mass of the segmented myocardium. Myocardium T1 maps were reconstructed from the resampled T1 weighted images using curve fitting. The FCN was trained and tested using manually segmented images for 210 patients (5 slices, 11 inversion times per patient). An additional image dataset for 455 patients (5 slices and 11 inversion times per patient), analyzed by an expert reader using a semi-automatic tool, was used to validate the automatically calculated global and regional T1 values. Bland-Altman analysis, Pearson correlation coefficient, r, and the Dice similarity coefficient (DSC) were used to evaluate the performance of the FCN-based analysis on per-patient and per-slice basis. Inter-observer variability was assessed using intraclass correlation coefficient (ICC) of the T1 values calculated by the FCN-based automatic method and two readers. RESULTS: The FCN achieved fast segmentation (< 0.3 s/image) with high DSC (0.85 ± 0.07). The automatically and manually calculated T1 values (1091 ± 59 ms and 1089 ± 59 ms, respectively) were highly correlated in per-patient (r = 0.82; slope = 1.01; p < 0.0001) and per-slice (r = 0.72; slope = 1.01; p < 0.0001) analyses. Bland-Altman analysis showed good agreement between the automated and manual measurements with 95% of measurements within the limits-of-agreement in both per-patient and per-slice analyses. The intraclass correllation of the T1 calculations by the automatic method vs reader 1 and reader 2 was respectively 0.86/0.56 and 0.74/0.49 in the per-patient/per-slice analyses, which were comparable to that between two expert readers (=0.72/0.58 in per-patient/per-slice analyses). CONCLUSION: The proposed FCN-based image processing platform allows fast and automatic analysis of myocardial native T1 mapping images mitigating the burden and observer-related variability of manual analysis.

Nonrigid active shape model-based registration framework for motion correction of cardiac T1 mapping.

PURPOSE: Accurate reconstruction of myocardial T1 maps from a series of T1 -weighted images consists of cardiac motions induced from breathing and diaphragmatic drifts. We propose and evaluate a new framework based on active shape models to correct for motion in myocardial T1 maps. METHODS: Multiple appearance models were built at different inversion time intervals to model the blood-myocardium contrast and brightness changes during the longitudinal relaxation. Myocardial inner and outer borders were automatically segmented using the built models, and the extracted contours were used to register the T1 -weighted images. Data acquired from 210 patients using a free-breathing acquisition protocol were used to train and evaluate the proposed framework. Two independent readers evaluated the quality of the T1 maps before and after correction using a four-point score. The mean absolute distance and Dice index were used to validate the registration process. RESULTS: The testing data set from 180 patients at 5 short axial slices showed a significant decrease of mean absolute distance (from 3.3 ± 1.6 to 2.3 ± 0.8 mm, P < 0.001) and increase of Dice (from 0.89 ± 0.08 to 0.94 ± 0.4%, P < 0.001) before and after correction, respectively. The T1 map quality improved in 70 ± 0.3% of the motion-affected maps after correction. Motion-corrupted segments of the myocardium reduced from 21.8 to 8.5% (P < 0.001) after correction. CONCLUSION: The proposed method for nonrigid registration of T1 -weighted images allows T1 measurements in more myocardial segments by reducing motion-induced T1 estimation errors in myocardial segments. Magn Reson Med 80:780-791, 2018. © 2018 International Society for Magnetic Resonance in Medicine.

Comparison of spoiled gradient echo and steady-state free-precession imaging for native myocardial T1 mapping using the slice-interleaved T1 mapping (STONE) sequence.

Cardiac T1 mapping allows non-invasive imaging of interstitial diffuse fibrosis. Myocardial T1 is commonly calculated by voxel-wise fitting of the images acquired using balanced steady-state free precession (SSFP) after an inversion pulse. However, SSFP imaging is sensitive to B1 and B0 imperfection, which may result in additional artifacts. A gradient echo (GRE) imaging sequence has been used for myocardial T1 mapping; however, its use has been limited to higher magnetic field to compensate for the lower signal-to-noise ratio (SNR) of GRE versus SSFP imaging. A slice-interleaved T1 mapping (STONE) sequence with SSFP readout (STONE-SSFP) has been recently proposed for native myocardial T1 mapping, which allows longer recovery of magnetization (>8 R-R) after each inversion pulse. In this study, we hypothesize that a longer recovery allows higher SNR and enables native myocardial T1 mapping using STONE with GRE imaging readout (STONE-GRE) at 1.5T. Numerical simulations and phantom and in vivo imaging were performed to compare the performance of STONE-GRE and STONE-SSFP for native myocardial T1 mapping at 1.5T. In numerical simulations, STONE-SSFP shows sensitivity to both T2 and off resonance. Despite the insensitivity of GRE imaging to T2 , STONE-GRE remains sensitive to T2 due to the dependence of the inversion pulse performance on T2 . In the phantom study, STONE-GRE had inferior accuracy and precision and similar repeatability as compared with STONE-SSFP. In in vivo studies, STONE-GRE and STONE-SSFP had similar myocardial native T1 times, precisions, repeatabilities and subjective T1 map qualities. Despite the lower SNR of the GRE imaging readout compared with SSFP, STONE-GRE provides similar native myocardial T1 measurements, precision, repeatability, and subjective image quality when compared with STONE-SSFP at 1.5T.

Coronary MR angiography at 3T: fat suppression versus water-fat separation.

OBJECTIVES: To compare Dixon water-fat suppression with spectral pre-saturation with inversion recovery (SPIR) at 3T for coronary magnetic resonance angiography (MRA) and to demonstrate the feasibility of fat suppressed coronary MRA at 3T without administration of a contrast agent. MATERIALS AND METHODS: Coronary MRA with Dixon water-fat separation or with SPIR fat suppression was compared on a 3T scanner equipped with a 32-channel cardiac receiver coil. Eight healthy volunteers were examined. Contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), right coronary artery (RCA), and left anterior descending (LAD) coronary artery sharpness and length were measured and statistically compared. Two experienced cardiologists graded the visual image quality of reformatted Dixon and SPIR images (1: poor quality to 5: excellent quality). RESULTS: Coronary MRA images in healthy volunteers showed improved contrast with the Dixon technique compared to SPIR (CNR blood-fat: Dixon = 14.9 ± 2.9 and SPIR = 13.9 ± 2.1; p = 0.08, CNR blood-myocardium: Dixon = 10.2 ± 2.7 and SPIR = 9.11 ± 2.6; p = 0.1). The Dixon method led to similar fat suppression (fat SNR with Dixon: 2.1 ± 0.5 vs. SPIR: 2.4 ± 1.2, p = 0.3), but resulted in significantly increased SNR of blood (blood SNR with Dixon: 19.9 ± 4.5 vs. SPIR: 15.5 ± 3.1, p < 0.05). This means the residual fat signal is slightly lower with the Dixon compared to the SIPR technique (although not significant), while the SNR of blood is significantly higher with the Dixon technique. Vessel sharpness of the RCA was similar for Dixon and SPIR (57 ± 7 % vs. 56 ± 9 %, p = 0.2), while the RCA visualized vessel length was increased compared to SPIR fat suppression (107 ± 21 vs. 101 ± 21 mm, p < 0.001). For the LAD, vessel sharpness (50 ± 13 % vs. 50 ± 7 %, p = 0.4) and vessel length (92 ± 46 vs. 90 ± 47 mm, p = 0.4) were similar with both techniques. Consequently, the Dixon technique resulted in an improved visual score of the coronary arteries in the water fat separated images of healthy subjects (RCA: 4.6 ± 0.5 vs. 4.1 ± 0.7, p = 0.01, LAD: 4.1 ± 0.7 vs. 3.5 ± 0.8, p = 0.007). CONCLUSIONS: Dixon water-fat separation can significantly improve coronary artery image quality without the use of a contrast agent at 3T.

Deep convolution neural networks based artifact suppression in under-sampled radial acquisitions of myocardial T 1 mapping images.

We developed a deep convolutional neural network (CNN) based method to remove streaking artefact from accelerated radial acquisitions of myocardial T 1-mapping images. A deep CNN based on a modified U-Net architecture was developed and trained to remove the streaking artefacts from under-sampled T 1 mapping images. A total of 2090 T 1-weighted images for 33 patients (55 ± 15 years, 19 males) and five healthy subjects (30 ± 14 years, 2 males) were used for training and testing the network. The images were acquired using radial slice interleaved T 1 mapping sequence (STONE) and retrospectively under-sampled to achieve acceleration rate of 4 (corresponding to 48 spokes). The dataset was split into training and testing subsets with 23 subjects (60%) and 15 subjects (40%), respectively. For generating voxel-wise T 1 maps, a two-parameter fitting model was used. Network performance was evaluated using normalized mean square error (NMSE), structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR) metrics. The proposed network allowed fast (<0.3 s/image) removal of the artefact from all T 1-weighted testing images and the corresponding T 1 maps with PSNR = 64.3 ± 1.02, NMSE = 0.2 ± 0.09 and SSIM = 0.9 ± 0.3 × 10-4. There was no statistically significant difference between the measured T 1 maps for both per-subject (reference: 1085 ± 37 ms, CNN: 1088 ± 37 ms, p = 0.4) and per-segment (reference: 1084 ± 48 ms, CNN: 1083 ± 58 ms, p = 0.9) analyses. In summary, deep CNN allows fast and reliable removal of streaking artefact from under-sampled radial T 1 mapping images. Our results show that the highly non-linear operations of deep CNN processing of T 1 mapping images do not impact accurate reconstruction of myocardial T 1 maps.

Left Atrial Epicardial Fat Volume Is Associated With Atrial Fibrillation: A Prospective Cardiovascular Magnetic Resonance 3D Dixon Study.

BACKGROUND: Recent studies demonstrated a strong association between atrial fibrillation (AF) and epicardial fat around the left atrium (LA). We sought to assess whether epicardial fat volume around the LA is associated with AF, and to determine the additive value of LA-epicardial fat measurements to LA structural remodeling for identifying patients with AF using 3-dimensional multi-echo Dixon fat-water separated cardiovascular magnetic resonance. METHODS AND RESULTS: A total of 105 subjects were studied: 53 patients with a history of AF and 52 age-matched patients with other cardiovascular diseases but no history of AF. The 3-dimensional multi-echo Dixon fat-water separated sequence was performed for LA-epicardial fat measurements. AF patients had significantly greater LA-epicardial fat (28.9±12.3 and 14.2±7.3 mL for AF and non-AF, respectively; P<0.001) and LA volume (110.8±38.2 and 89.7±30.3 mL for AF and non-AF, respectively; P=0.002). LA-epicardial fat adjusted for LA volume was still higher in patients with AF compared with those without AF (P<0.001). LA-epicardial fat and hypertension were independently associated with the risk of AF (odds ratio, 1.17; 95% confidence interval, 1.10%-1.25%, P<0.001, and odds ratio, 3.29; 95% confidence interval, 1.17%-9.27%, P=0.03, respectively). In multivariable logistic regression analysis adjusted for body surface area, LA-epicardial fat remained significant and an increase per mL was associated with a 42% increase in the odds of AF presence (odds ratio, 1.42; 95% confidence interval, 1.23%-1.62%, P<0.001). Combined assessment of LA-epicardial fat and LA volume provided greater discriminatory performance for detecting AF than LA volume alone (c-statistic=0.88 and 0.74, respectively, DeLong test; P<0.001). CONCLUSIONS: Cardiovascular magnetic resonance 3-dimensional Dixon-based LA-epicardial fat volume is significantly increased in AF patients. LA-epicardial fat measured by 3-dimensional Dixon provides greater performance for detecting AF beyond LA structural remodeling.

Simultaneous Assessment of Cardiac Inflammation and Extracellular Matrix Remodeling after Myocardial Infarction.

Background: Optimal healing of the myocardium following myocardial infarction (MI) requires a suitable degree of inflammation and its timely resolution, together with a well-orchestrated deposition and degradation of extracellular matrix (ECM) proteins. Methods and Results: MI and SHAM-operated animals were imaged at 3,7,14 and 21 days with 3T magnetic resonance imaging (MRI) using a 19F/1H surface coil. Mice were injected with 19F-perfluorocarbon (PFC) nanoparticles to study inflammatory cell recruitment, and with a gadolinium-based elastin-binding contrast agent (Gd-ESMA) to evaluate elastin content. 19F MRI signal co-localized with infarction areas, as confirmed by late-gadolinium enhancement, and was highest 7days post-MI, correlating with macrophage content (MAC-3 immunohistochemistry) (ρ=0.89,P<0.0001). 19F quantification with in vivo (MRI) and ex vivo nuclear magnetic resonance (NMR) spectroscopy correlated linearly (ρ=0.58,P=0.020). T1 mapping after Gd-ESMA injection showed increased relaxation rate (R1) in the infarcted regions and was significantly higher at 21days compared with 7days post-MI (R1[s-1]:21days=2.8 [IQR,2.69-3.30] vs 7days=2.3 [IQR,2.12-2.5], P<0.05), which agreed with an increased tropoelastin content (ρ=0.89, P<0.0001). The predictive value of each contrast agent for beneficial remodeling was evaluated in a longitudinal proof-of-principle study. Neither R1 nor 19F at day 7 were significant predictors for beneficial remodeling (P=0.68;P=0.062). However, the combination of both measurements (R1<2.34Hz and 0.55≤19F≤1.85) resulted in an odds ratio of 30.0 (CI95%:1.41-638.15;P=0.029) for favorable post-MI remodeling. Conclusions: Multinuclear 1H/19F MRI allows the simultaneous assessment of inflammation and elastin remodeling in a murine MI model. The interplay of these biological processes affects cardiac outcome and may have potential for improved diagnosis and personalized treatment.

Improved segmented modified Look-Locker inversion recovery T1 mapping sequence in mice.

OBJECT: To develop and evaluate a 2D modified Look-Locker (MOLLI) for high-resolution T1 mapping in mice using a 3T MRI scanner. MATERIALS AND METHODS: To allow high-resolution T1 mapping in mice at high heart rates a multi-shot ECG-triggered 2D MOLLI sequence was developed. In the proposed T1 mapping sequence the optimal number of sampling points and pause cardiac cycles following an initial adiabatic inversion pulse was investigated in a phantom. Seven native control and eight mice, 3 days post myocardial infarction (MI) after administration of gadolinium were scanned. Two experienced readers graded the visual T1 map quality. RESULTS: In T1 phantoms, there were no significant differences (<0.4% error) between 12, 15 and 20 pause cardiac cycles (p = 0.1, 0.2 and 0.6 respectively) for 8 acquisition cardiac cycles for 600bpm in comparison to the conventional inversion recovery spin echo T1 mapping sequence for short T1's (<600 ms). Subsequently, all in-vivo scans were performed with 8 data acquisitions and 12 pause cardiac cycles to minimize scan time. The mean native T1 value of myocardium in control animal was 820.5±52 ms. The post-contrast T1 measured 3 days after MI in scar was 264±59 ms and in healthy myocardium was 512±62 ms. The Bland-Altman analysis revealed mean difference of only -1.06% of infarct size percentage between T1 maps and LGE. CONCLUSIONS: A multi-shot 2D MOLLI sequence has been presented that allows reliable measurement of high spatial resolution T1 maps in mice for heart rates up to 600bpm.