Figure 1: Post-processing steps used to combine DENSE and cDTI data and calculate myofiber strain E_ff. (A) Myofiber orientations and Cardiac displacement fields were represented using nodes after reconstruction. (B) 2D Displacement fields from long-axis (LA) and short-axis (SA) are combined to obtain a SA 3D displacement field. (C) The SA 3D displacement field is used to generate a deformed cine of myofiber orientation(D) Finally, cardiac strains are calculated from the SA 3D displacement field and the myofiber orientations and using beginning of systole as the initial config.

Figure 5: Computed strain (N=30) calculated after combining cDTI and DENSE. (A) Computed E_ff and (B) E_cc at endo, mid, epi layers, and across the LV wall. The black dashed and dotted lines represent the median [Q1, Q3] across volunteers; the red dashed line provides a -0.15 strain ref. the colored solid lines are individual medians per volunteers. (C) Transmural distribution of E_ff, E_cc, and their difference at peak-systole. Blue solid lines are the median per volunteer and the box plots are across volunteers.p-values<0.01 are represented by (**) and p-values<0.05 by (*).

Figure 2. (a) Cr CEST map overlaid on a CEST-weighted image of a representative animal. Negative Cr CEST contrast was largely consistent with the scar region shown on the T₁w-LGE image (b). Segmental based circumferential strain, (e) radial strain (f), wall thickening (g) and wall motion (h) were measured from the respective cine images at end-diastolic (c) and end-systolic (d) cardiac phases.

Figure 4. Correlation of segmental based Cr CEST with respective myocardium contractile function indices of (a) circumferential strain, (b) radial strain, (c) wall thickening and (d) wall motion.

Figure 4. A) Temporal evolution of the injected [2-¹³C]pyruvate and downstream metabolites [2-¹³C]lactate, [5-¹³C]glutamate and [1-¹³C]acetylcarnitine(ALCAR) in Volunteer 1 before and 30-minutes after oral glucose challenge. Time was calculated from the start of injection. B) A table showing the metabolites ratios normalized to pyruvate. The levels of lactate, glutamate and ALCAR were elevated after oral glucose compared to baseline. C) Rate constants of pyruvate-to-product conversion were calculated to a first-order approximation.

Figure 2. Representative summed spectra (Volunteer 1) reflecting baseline Krebs cycle energetics in human heart probed with hyperpolarized C2 pyruvate. The substrates and products in the spectra were assigned [5-¹³C]glutamate(182.5ppm), [1-¹³C]acetylcarnitine(ALCAR, 173.6ppm), [2-¹³C]lactate(70.8, 66.2ppm), [1,2-¹³C]pyruvate(171.8, 169.7ppm), and [2-¹³C]hydrate(94.2ppm), respectively.

Figure 3: Examples of 4 patients with evidence of myocardial injury on LGE and motion-corrected T1ρ mapping. (A) 59-year-old male patient with sub-epicardial LGE in the latero-apical segment. (B) 53-year-old male patient with ischemic cardiomyopathy and transmural LGE in the inferior and infero-septal mid segments. (C) 51-year-old male patient with acute myocarditis and extensive patchy intramural and subepicardial LGE in the left ventricular free wall. (D) 35-year-old male patient with myocarditis and intramural LGE in the antero-septo-basal segment.

Figure 1: Schematic of the proposed single breath-hold 2D myocardial T1ρ mapping technique (A) with joint T1ρ fitting and model-based motion correction (B). T1ρ mapping is performed using a single-shot electrocardiogram-triggered bSSFP acquisition where five images with different spin lock times are acquired within a single breath-hold. Motion correction is performed by iterating between a T1ρ fitting (step 1), the simulation of T1ρ-weighted images (step 2) and a pair-wise non-rigid motion correction (step 3).

Figure 1. Example T₁ and T₂ maps obtained from the cMRF sequence for a healthy control subject and a patient with light-chain cardiac amyloidosis (AL). Myocardial T₁ and T₂ appear to be elevated in the AL patient as compared to the healthy control.

Figure 5. Linear discriminant analysis (LDA) of relaxometric data and signal data. (a) Bar plots of LDA score means and standard deviations obtained for patients and controls for relaxometric data and signal evolution data. LDA scores for relaxometric data and signal data were normalized by a constant factor for each group for visualization. The * indicates a statistically significant difference (p<0.05). (b) Fisher coefficient indicating separability of amyloid and control groups (greater value indicates greater separation).

Fig.5 (animated): Cardiac motion-resolved images of two subjects of proposed 3D Stack-of-Stars Multitasking in short axis orientation with 3D (1.4x1.4x8.0 mm³ resolution) whole-ventricle coverage in TA=9:14min. Comparative multi-slice and multi breath-hold 2D TrueFisp CINE in short axis (non-fat suppressed, 1.3x1.3x8.0 mm³ resolution, TA=11s per slice).

Fig.3: The proposed 3D Multitasking T1,T2, B1+ cine mapping results (basal, mid and apical) and short-axis 2D MOLLI T1 maps, T2-prep FLASH T2 maps for one subject (F, 27).

Fig 4: Short axis view of T₁, T₂ and T_1ρ quantitative maps in two representative healthy subjects. Reference maps (top row) of T₁-MOLLI, T₂-GRaSE and T_1⍴-Reference are compared against the regularized T₁, T₂ and T_1⍴ cardiac MRF maps (bottom row) obtained from the same ~16s cardiac MRF acquisition.

Fig 5: a) Scatter plots and T₁, T₂ and T_1ρ cardiac MRF quantification compared against their reference values obtained from T₁-MOLLI, T₂-GRaSE and T_1⍴-Reference. Green dashed lines depict the identity line. b) Bland-Altman plots of T₁, T₂ and T_1ρ cardiac MRF against their respective references. Black lines denote the mean bias and dashed-dotted red lines show the limits of agreement (95% confidence interval).

Sharp source images and precise T2 maps in healthy volunteers. A-E) The source images contain sharp features (papillary muscles, edge definition) and are well aligned to one another. F-H) The resulting short-axis T2 map and its two perpendicular views are similarly sharp.

T2 mapping in a patient with myocardial infarction. A) LGE can be observed in the septum. B) The corresponding T2 map obtained with 2D T2-prepared bSSFP shows a small T2 elevation in the infarcted area. C) The proposed technique accurately confirms the T2 elevation.

Figure 4. Representative 3D T₂ map from one healthy human subject by BH3DT2 sequence. A: 3D whole left ventricle T₂ maps from apex to base. B: mean and standard deviation of T₂ values within every slice. C: Histogram of myocardium T₂ over the whole left ventricle. The mean (μ) and standard deviation (σ) across the whole left ventricle are shown on the graph. D: AHA 16-segment bull's-eye plot for showing T₂ of each region.

Figure 2. Representative slice from a swine with acute MI, nRMSE was increased at higher acceleration factor (R). Image artefacts were subtle at R=2 and 4. At R=6, obvious artefacts were observed at the blood pool in the T₂ weighted image (red arrow), and septal T₂ value was overestimated (white arrow).

Figure 1. Schematic of the acquisition. Data in the first two seconds acquired with off-resonance Fermi pulses were used to reconstruct the Bloch-Siegert B1 map. Data acquired with repetitive inversion pulses were used to reconstruct the T1* map.

Figure 4. Human results. Short-axis basal (a) and middle (b) slices from one human subject are shown. As indicated by the bracket, 3° and 15° T1* maps are used to generate the 2FAs T1 map, while 3° T1* and BS B1 maps are used to generate the 1FA+B1 T1 map. MOLLI and SASHA T1 maps are also shown as comparison.

Figure 4 - 12 patients (30%) out of the total 40 patients showed T1 value below the normal range (<966ms), despite having T2* value in the normal range (>20ms)

Figure 1 – T2* colour map in a normal patient (A) and a patient with severe cardiac iron overload (B). T1 colour map in a normal patient (C) and a patient with severe cardiac iron overload (D). T2 colour map in a normal patient (E) and a patient with severe cardiac iron overload (F)

A representative cardiac motion-resolved T1 mapping at the end-expiration state for subject 1.

A. MOLLI (mid-diastolic) and the representative motion-resolved myocardial T1 maps (mid-diastolic and mid-systolic) for two repetitive scans and two subjects. The line profiles along the motion dimension are presented in the bottom row for both subjects. B. Quantitative myocardial T1 values (ms, mean $$$\pm$$$ standard deviation) in left-ventricular septum for T1 maps in Figure 4A.

Figure 4 - T1 maps (in ms) estimated from ProMyoT1 and MOLLI with: a. Linear filling order; and b. Centric filling order. Vial segmentation is presented on the left.

Figure 3 - T1 weighted image series of NIST/ISMRM phantom obtained with ProMyoT1 sequence with: a. Linear filling order; and b. Centric filling order. The time of inversion (in ms) for each image is presented according to the acquisition order and before the reordering required for T1 estimation. Simulated heart rate of 60 bpm.

Figure: 2. T1 map obtained using the data-matching analysis method

Figure: 3. Comparison of the measured T1 values obtained using the conventional and data-matching analysis methods

Images descriptors for the same slice in pre- (first line) and post-injected (second line) MOLLI. The original images (first column) correspond to the higher TI images.

Global evaluation of the performances of the registration algorithms in term of Dice (1st line), HD (2nd line) and NMI (3rd line). The second column presents the difference between the algorithm’s performance before and after motion correction. The table below presents the average results (+/-sd) for each parameter (rows) and each descriptor (columns).

Human Multiecho-GRASP magnitude images and T₂* maps for 20 cardiac cycles. T₂* values with low magnitude data (< 5% of mean signal intensity) was omitted.

Reference multi-echo gradient echo acquisition compared to multiecho GRASP (5th echo not shown). The T₂* map shows the color-coded values in the myocardium together with a bullseye plot of the mean T₂* values in each segment.

(Fig:3)A short-axis image of color T2map. Comparison between conventional method and C-SENSE combined color T2map

(Fig:2)A short-axis image of T2map. Comparison between conventional method and C-SENSE combined T2map

A man in his 60s man who underwent heart transplantation 16 years ago due to dilated cardiomyopathy. NH3-PET can quantify the absolute MBF and MFR by coronary territory (upper row). MRI T1 and T2 values in four territories were measured using the short-axis of mid left ventricular T1 and T2 mappings (lower row).

Comparison of T1 (left) and T2 (right) between territories with MFR >2.0 and <2.0.

Fig. 4. In comparison to the reference T1 values for the T1 phantom, the actual T1 values measured by the MOLLI techniques are shown in (a) and (b). The top row is for HR=60 bpm, while the bottom is for HR=80 bpm. The comparisons were made three times for the flip angle (FA) 20°, 35° and 50°. The apparent T1 comparisons for the three FAs are shown in (c) and (d), and the correction factors are in (e) and (f).

Fig. 2. If the relaxation under the influence of the B1 field is non-negligible, the longitudinal relaxation of the magnetization can be described as Eq. 4, where the T2rho relaxation is added. If the apparent T1 relaxation, T1’ in Eq. 5, can be expressed as Eq. 5, the signal model can still be the same way as Eq. 1 in Fig.1, and the correction factor can be expressed as Eq. 6. The actual T1 can still be obtained by multiplying the apparent T1 (T1’) and the correction factor, as in Eq. 7.

Figure 4:Full field of view example STEAM-SASHA images after averaging two sequence repeats (top line) with the associated STEAM-SASHA T1 map (bottom left) and the equivalent systolic product MOLLI T1 map acquired in one typical volunteer (T1 maps cropped). The STEAM-SASHA images show the reduction in signal intensity at the top and bottom of the image due to the in-plane slice select gradient used. The STEAM-SASHA T1 map provides excellent blood and fat suppression in the left and right ventricles.

Figure 2:Example STEAM-SASHA images (full field of view) after averaging two sequence repeats (top line) with the associated STEAM-SASHA map (cropped) and the equivalent product MOLLI T1 map acquired in the T1MES phantom⁷.

Figure 4. T1 mapping result. T1 maps obtained by simple fitting of low-rank reconstruction (top row), the proposed direct estimation method (middle row) and MOLLI (bottom row) are shown over multiple slices. Green arrows indicate regions in the myocardium where the Fit (top row) and Direct (middle row) differ. The T1 maps from the proposed method better match those from MOLLI, with correction of artifacts resulting in overestimation of T1 values within the myocardium.

Figure 2. MR images at different time frames. Top-row shows images from low-rank reconstruction (LR). The central row shows synthesized images obtained from a T1 model fit of the low-rank reconstruction (Fit). The bottom row shows temporal images obtained using the proposed direct method (Direct). Low-rank reconstructions are degraded by aliasing artifacts, which are reduced in synthesized images obtained from T1 model fitting. The proposed method shows improvement by further reducing the artifacts.

Figure 1: A) Overlay of manually drawn regions on T2-weighted cross-section. B & C) Corresponding T1 and T2 relaxation time maps. MG = medial gastrocnemius, LG = lateral gastrocnemius, and SOL = soleus muscle regions.

Figure 2: Ankle Brachial Index (ABI) is positively correlated with T1 relaxation time in the medial gastrocnemius and soleus (upper panel). Decreasing ABI reflects worsening arterial blockages. Ischemic window is negatively correlated with T2 relaxation time in the soleus and lateral gastrocnemius (lower panel). Increasing ischemic window reflects worsening arterial blockages.

Figure 1. The flowchart of in-vivo data analysis for the T1 and B1 mapping calculation.

Figure 2. The T1 values of left, right and both carotid vessel walls on the T1 mapping without or with B1 correction.

Figure 2. The first row corresponds to a sample of blood pool extraction result, the rest to whole-heart identification results. The first column shows the contours of the user-defined masks used as target to train the model. The second column corresponds to the contours of the inferred mask. The third column shows the overlay of both segmentations: in green, true positives; in red, false negatives and in yellow, false positives.

Figure 1. U-net architecture implemented for both segmentation tasks.

Figure 2: 3D imaging of acetic acid chemoablations (green arrows). Acutely, acetic acid produces strong T1 enhancement, which in combination with SPGR (inherently T1-weighted) results in signal contamination of T2W imaging. LGE appears hyperenhanced though heterogeneous. At later time points the T1W highlights lesion extent with good correspondence with ex-vivo MRI but with potentially confounding microvascular obstruction (orange arrows). Comparison with ex-vivo LGE indicates good correspondence to in-vivo imaging with both LGE and T1W imaging.

Figure 1: 3D imaging of ethanol chemoablation lesions (red arrows). Acutely, edema is mildly visible on T2-W imaging, T1-W imaging reflects hypointense signal, and lesions appear hypointense on LGE, indicating no contrast agent was present within lesions. At later time points the inherent T1 weighting of SPGR resulted in lesion visualization on T2-W imaging. 3D T1-W imaging resulted in hyperintense lesion cores, surrounded by thin layer of hypointensity depicting onset of encapsulating scar (yellow arrows). Comparison with ex-vivo imaging demonstrates good correspondence.

Fig.2: Cardiac and respiratory motion-resolved T1 images of one subject (F,27) of proposed 2D SMS Multitasking protocol in base, mid, apex slices (1.7x1.7x8.0 mm³ resolution, TA=3min). Comparative ECG-gated, breath-held MOLLI T1 maps (base, mid, apex, 1.4x1.4x8.0 mm³ resolution, TA=11s) in short-axis orientation.

Fig.3: Cardiac and respiratory motion-resolved T2 images of one subject (F,27) of proposed 2D SMS Multitasking protocol in base, mid, apex slices (1.7x1.7x8.0 mm³ resolution, TA=3min). Comparative ECG-gated, breath-held T2-prep FLASH T2 maps (base, mid, apex, 1.4x1.4x8.0 mm³ resolution, TA=11s) in short-axis orientation.

Figure 2

Figure 1

Figure 2. Typical volunteer multitasking images. (a) 30.7s imaging data were acquired and used for spiral, navigator-less multitasking reconstruction. (b) Both imaging data and navigator data were acquired from a total of 61.4s scan and used for spiral multitasking reconstruction.

Figure 1. Spiral sampling trajectory used in self-navigated multitasking.

Figure 2.- Scatter plot of the improvement in the anatomical alignment metric, obtained with each of the motion correction methods, with respect to uncorrected data. Each data sample corresponds to one 2D MOLLI series.

Figure 3.- Box & whisker plot illustrating the quantitative metric values for the three myocardial motion classes used in the qualitative analysis.

Registration result. The green contours and bars refer to the post-injection T1 map mask and the yellow one to the pre-T1 map mask. The second row presents the overlap (in red) of the masks before (1st column) and after (2nd column) Moco. The histogram plot shows the distribution of the pre-T1 values in the mask defined on the pre-T1 map before registration (yellow) and the distribution of the pre-T1 values in the mask defined on the post-T1 map after registration (green). HD: Hausdorff Distance.

Global results. The 1st row presents the results in term of ECV for the remote (left) and infarct part (right), and for manual computed ECV (blue) and registration-based computation of ECV (orange). The divergence plot shows the difference between the distribution of pre-T1 value of the myocardium before and after registration in term of L2 distance and Jeffrey’s divergence. The final plot represents the difference between manual and Moco ECV regarding the obtained Dice after registration.

Figure 1. The demonstration of CFD analysis parameters, to multi-contrast MRI images and histological images.

Figure 2. ROC curves of TAWSS combined with max wall thickness (MWT) and normal wall index (NWI) in discriminating IPH.

(The black line represents the ROC curves for predicting IPH with TAWSS. The orange and green lines represent the ROC curves for predicting IPH with TAWSS of combined MWT and of combined NWI.)

Figure 3. Parameter maps in control and isoproterenol hearts generated by fitting the (top-bottom) DTI, covariance and gamma models scaled to [0 1]. DTI and covariance methods yield mean diffusivity (MD in µm²/ms) and fractional anisotropy (FA). Additional maps of microscopic fractional anisotropy (µFA), mean anisotropic kurtosis (MKa), mean isotropic kurtosis (MKi), mean total kurtosis (MKt), microscopic orientation coherence (Cc) and normalised size variance (CMD) are shown, along with ROI masks (bottom right).

Figure 2. (Left) Normalised signal attenuation curves for spherical and linear tensor encoding in myocardium (myo) of the control heart and gel. Non-monoexponential behaviour of gel at high b was governed by noise. (Right) Zoomed section showing signal kurtosis attributable to isotropic and anisotropic components (MKi; MKa). (Top) Example LTE images averaged across diffusion directions.

Figure 1. Example of a typical free-hand region of interest drawn on a short axis bSSFP image, encompassing the left ventricular myocardium.

Figure 2. Box-Whisker plots illustrating the differences for the significant texture features between patients with myocarditis and controls on bSSFP images. The median is represented by the centerline of the boxplot with upper and lower limits of 25th and 75th percentiles, respectively. The Whiskers extending from the boxes indicates the most extreme values within 25th and 75th percentiles ±1.5*interquartile range; data points beyond the whiskers are displayed as +. Texture features are dimensionless. bSSFP = balanced steady-state-free-precession.

Figure 3. Relationship between CNR and T_RAFF2 pulse duration. Mean and SEM of CNR, CNR, contrast to noise ratio. CNR is calculated as follow CNR=[T(fibrous skeleton)-T(myocardium)]/σ_o(myocardium)×100%; σ_o, standard deviation of the noise; SEM, standard error of the mean.*p<0.05, **p<0.01, Paired two-tail Student’s t-test.

Figure 2. Relationship between relaxation times in fibrous skeleton and myocardium areas and T_RAFF2 pulse duration.

Figure 1: Overview of study design. (Left) Flow chart of the main study steps from imaging, through segmentation and analysis to comparison of results. (Top Right) Histology image produced from the extended volume confocal microscopy, presented using a ‘Glow’ look up table. (Bottom right) The same image after segmentation of the following compartments: intracellular (blue), extracellular (red) and blood vessel/cleavage space (green).

Figure 3: Results of the diffusion tensor simulation. Vector representations of the primary eigenvector of the diffusion tensor analysis from the four rows of voxel-blocks (top row and middle row). Color maps of the helix angle (bottom). The vector plots show a transition from longitudinal myofibers at the epicardium, through circumferential fibers at the mid-wall, to longitudinal fibers at the endocardium.

Figure 1. Gradient echo (GRE), QSM and R2* in representative cases of A) moderate MAC, B) mild MAC, and C) non-calcification. The corresponding computed tomography of each case is shown in the first column as the reference for presence and severity of calcification. The red arrows indicate the location of mitral annular calcification.

Figure 3. Correlation between CT value and susceptibility in calcification regions from MAC patients. Susceptibility detected by thresholding aligned well with CT reference as linear relationship. H.U., Hounsfield unit; ppm, parts per million. Red points: moderate calcification; blue points: mild calcification.

Figure 1: 1 Example images over the time course of the experiment. The systolic timeframe shows hypocontraction in the inferior lateral wall coinciding with late gadolinium enhancement (LGE), native T1, extra cellular volume fraction (ECV), T2 mapping, mean diffusivity (MD) and fractional anisotropy (FA).

Figure 3 relative change in native T1, extra cellular volume (ECV), T2, mean diffusivity (MD) and fractional anisotropy (FA) in the infarcted area compared to the remote area. Error bars indicate one standard deviation across cases. The asterisk indicates statistically significanct (p<0.05) difference compared to baseline contrast.

Figure 2: PLS-DA plot showing separation of patients with CAD (Red) from controls (Green).

Figure 1: 1D 1H NMR spectrum of blood plasma of a patient with CAD acquired at 700 MHz.

Figure 2: (A) A sample segmentation map of one slice of the heart, showing (blue) chest wall, (orange) heart and (green) other compartments, and (black) the midseptal voxel (64% threshold of voxel PSF) used in the short AW CSI reconstruction. (B) Sample spectra for each compartment in (A) reconstructed using the SLAM algorithm and AW acquisition. (C) Fit, using the OXSA toolbox of the heart compartment spectra in (B). (D) SLIM and SLAM reconstruction of heart compartment in (A) with AW data and the mid-septal voxel of the short AW CSI reconstruction. Spectra in B and D are normalized by noise.

Figure 4: Box-plot showing PCr SNR values for each acquisition, median and IQR indicated by box. * indicates significant difference to midseptal CSI reconstruction (Wilcoxon signed-rank paired, α=0.05/6, P-values above box-plot). SLAM/SLIM reconstructions which use the same data acquisition as the midseptal reconstruction are highlighted.

Figure 4 Typical example of IVIM-derived parameters, strain and LGE for (A) a healthy control and (B) a LGE(+) patient with cardiac amyloidosis. Other abbreviations as in Figure 2&3.

Figure 2 Box plot indicating the distribution of the ADC_slow, ADC_fast, and F values among the healthy controls and LGE(+) or LGE(-) patients. LGE= late gadolinium enhancement.

(a.) Fully sampled reconstruction compared to a 2x2 CAIPI scheme and a VD Poisson 4-fold retrospectively accelerated. The red arrow indicates the infarct. (b.) Relative error between the fully sampled and the accelerated data.

Sequence diagrams. Zoom in on the Trajectory Calibration Sequence (b.) and on the Wave sequence (c.).

a.) The sequence is triggered in diastole and respiratory gating was performed using an echo navigator played before data acquisition.

b.) The trajectories calibration module takes place during the first 25 seconds (4 averages are performed).

c.) The wave encoding gradients are applied during the readout, in both phase and slice directions. The slice encoding wave gradient starts ¼ of a cycle before the phase wave gradient.

Table 1. Analysis for CMR of Baseline and Flow-up

Figure 2. Comparison of left ventricular global strain parameters between baseline and follow-up based on per-patient subgroup analysis. In the subgroup per-patient analysis, the global peak systolic radial strain (GRS), global circumferential strain (GCS), and global longitudinal strain (GLS) of the viable (SIE < 50%) and nonviable (SIE ≥ 50%) groups were not significantly improved after successful CTO-PCI in 1-year follow-up.

3D grey-blood PSIR image and fat volume in coronal and short axis views for two representative cases with scar. Excellent scar to myocardium SNR can be observed in both cases.

Comparison between the 2D and 3D grey-blood PSIR images for 3 patients. Corresponding slice positions of the 2D short-axis images were reformatted for the 3D grey-blood PSIR images. Image quality is comparable across the slices for all the cases with a good depiction of scar observed across the whole 3D volume compared to the 2D acquisition.

Figure 2: Water and fat phantom images obtained using complex-difference (CD) and MICSR for a trigger-delay-time of 45 ms (the delays are different between the images due to differences in the duration of the tagging prepulse). The images were acquired using (a) CSPAMM, (b) ORI-CSPAMM, (c) O-CSPAMM, and (d) ORI-O-CSPAMM sequences. Images acquired without ORI prepulse showed a shift in the tagging pattern at the water-fat interface, while in ORI versions ((b) and (d)) the shift was corrected.

Course of late gadolinium enhancement (LGE), Native myocardial T1 and ECV map images in PM and DM patients, respectively. LGE detected mid-wall enhancement in the interventricular septum and inferior in PM and subepicardial enhancement in the interventricular septum and inferior in DM. LGE volume of 16% in both the PM and DM patients. Mean global native T1 and ECV values were 1281 ms and 32%, 1283 ms and 30% in PM and DM patients, respectively.

Global native T1 (A) in control (green), Polymyositis (PM) (red), and Dermatomyositis (DM) (blue) and global ECV (B) in Polymyositis (PM) (red), and Dermatomyositis (DM) (blue) showed by box plots.

Figure 1. Diffusion weighted images, MD maps, FA maps, and direction-encoded color FA maps at b values of 300, 500, 800 s/mm² of one volunteer.

Figure 3. SNR, CNR of images acquired with different myocardium quiescent duration of all subjects.

Figure 1. Receiver operating characteristic analysis of model1, model2 and model3 for differentiation of patients with IIM from the controls. Area under curve for model 3:0.876 (95% CI: 0.787-0.965; P<0.001); model 2:0.851(95% CI: 0.758-0.944; P<0.001); model 1:0.745 (95% CI: 0.623-0.868; P=0.001).

Table 3.LA volumetric and deformation parameters assessed by CMR-FT

The statistical analysis of the myofiber angle for (A) the swine subject (A) and (B) the swine subject (B).

wall thickness and helix angle distribution are shown on a 2D flattened polar plot of swine subject (A). (A) Swine Subject (A) 3D endocardial surface with wall-thickness contour, (B) 2D-flattened wall thickness contour, (C) 2D-flattened helix angle at the subendocardial surface, and (D) 2D-flattened helix angle at the subepicardial surface. White and black curves represent the marking of right ventricle anterior and interior insertion points, respectively.

Figure 2. Comparison of simulated B₀ fields in the heart between the FFT-based method and dipole method. A) Exemplary field distribution in the heart was calculated using dipole method based on the susceptibility distribution of Ella’s entire body at 3 mm isotropic. The standard deviation of the field difference was shown B) from fine to coarse resolution and C) a range of zero-padding factors. Zero-padding factors higher than 2.5 do not substantially improve the STD value, i.e., B₀ accuracy, while a higher resolution can significantly lower this value (dash line: linear fit at zf = 2.5).

Figure 5. The standard deviation of field difference between the extended CT-derived FOV and the entire body for ten female models (left) and ten male models (right). The anatomical extension type 3 adopting the body with a similar BMI exhibited the lowest B₀ simulation error compared to other types, especially in male models. It is the optimized strategy when sufficient computation power is available while simplified anatomical extension type 4 are reasonable compromises to achieve high-resolution B₀ simulation with less discretization error under limited computation resources.

Figure 1. Sternal wire and bioprosthetic aortic valve artifact appear more prominent at 1.5T (left) compared to 0.55T field strength (right).

Figure 5. Example pacemaker leads for two patients (4-chamber on top, short-axis ventricle on bottom) imaged at 1.5T and 0.55T. The artifact caused by the lead implants was reduced at 0.55T. In addition, the de-phased blood artifact visible in the 4-chamber view at 1.5T is diminished at 0.55T.