Publications

BACKGROUND
Quantifying reproducibility of native T1 and T2 mapping over a long period (>1 year) is necessary to assess whether changes in T1 and T2 over repeated sessions in a longitudinal study are associated with variability due to underlying tissue composition or technical confounders.

OBJECTIVES
To carry out a single-center phantom study to 1) investigate measurement reproducibility of slice-interleaved T1 (STONE) and T2 mapping over 20 months, 2) quantify sources of variability, and 3) compare reproducibility and measurements against reference spin-echo measurements.

METHODS
MR imaging was performed on a 1.5 Tesla Philips Achieva scanner every 2–3 weeks over 20 months using the T1MES phantom. In each session, slice-interleaved T1 and T2 mapping was repeated 3 times for 5 slices, and maps were reconstructed using both 2-parameter and 3-parameter fit models. Reproducibility between sessions, and repeatability between repetitions and slices were evaluated using coefficients of variation (CV). Different sources of variability were quantified using variance decomposition analysis. The slice-interleaved measurement was compared to the spin-echo reference and MOLLI.

RESULTS
Slice-interleaved T1 had excellent reproducibility and repeatability with a CV<2%. The main sources of T1 variability were temperature in 2-parameter maps, and slice in 3-parameter maps. Superior between-session reproducibility to the spin-echo T1 was shown in 2 parameter maps, and similar reproducibility in 3-parameter maps. Superior reproducibility to MOLLI T1 was also shown. Similar measurements to the spin-echo T1 were observed with linear regression slopes of 0.94–0.99, but slight underestimation. Slice-interleaved T2 showed good reproducibility and repeatability with a CV<7%. The main source of T2 variability was slice location/orientation. Between-session reproducibility was lower than the spin-echo T2 reference and showed good measurement agreement with linear regression slopes of 0.78–1.06.

CONCLUSIONS 
Slice-interleaved T1 and T2 mapping sequences yield excellent long-term reproducibility over 20 months.

OBJECTIVES:

This study assessed changes in myocardial native T₁ and T₂ values after supine exercise stress in healthy subjects and in patients with suspected ischemia as potential imaging markers of ischemia.

BACKGROUND:

With emerging data on the long-term retention of gadolinium in the body and brain, there is a need for an alternative noncontrast cardiovascular magnetic resonance (CMR)-based myocardial ischemia assessment.

METHODS:

Twenty-eight healthy adult subjects and 14 patients with coronary artery disease (CAD) referred for exercise stress and/or rest single-photon emission computed tomography/myocardial perfusion imaging (SPECT/MPI) for evaluation of chest pain were prospectively enrolled. Free-breathing myocardial native T₁ and T₂ mapping were performed before and after supine bicycle exercise stress using a CMR-compatible supine ergometer positioned on the MR table. Differences in T_{1 rest}, T_{2 rest} and T_{1 post-exercise}, T_{2 post-exercise} values were calculated as T₁ and T₂ reactivity, respectively.

RESULTS:

The mean exercise intensity was 104 W, with exercise duration of 6 to 12 min. After exercise, native T₁ was increased in healthy subjects (p < 0.001). T₁ reactivity, but not T₂ reactivity, correlated with the rate-pressure product as the index of myocardial blood flow during exercise (r = 0.62; p < 0.001). In patients with CAD, T₁ reactivity was associated with the severity of myocardial perfusion abnormality on SPECT/MPI (normal: 4.9%; quartiles: 3.7% to 6.3%, mild defect: 1.2%, quartiles: 0.08% to 2.5%; moderate defect: 0.45%, quartiles: -0.35% to 1.4%; severe defect: 0.35%, quartiles: -0.44% to 0.8%) and had similar potential as SPECT/MPI to detect significant CAD (>50% diameter stenosis on coronary angiography). The area under the receiver-operating characteristic curve was 0.80 versus 0.72 (p = 0.40). The optimum cutoff value of T₁ reactivity for predicting flow-limiting stenosis was 2.5%, with a sensitivity of 83% and a specificity of 92%, a negative predictive value of 96%, a positive predictive value of 71%, and an area under the curve of 0.86.

CONCLUSIONS:

Free-breathing stress/rest native T₁ mapping, but not T₂ mapping, can detect physiological changes in the myocardium during exercise. Our feasibility study in patients shows the potential of this technique as a method for detecting myocardial ischemia in patients with CAD without using a pharmacological stress agent.

ABSTRACT:

Geometrical structure of the myocardium plays an important role in understanding the generation of

arrhythmias. In particular, a heterogeneous tissue (HT) channel defined in cardiovascular magnetic

resonance (CMR) has been suggested to correlate with conduction channels defined in electroanatomic

mapping in ventricular tachycardia (VT). Despite the potential of CMR for characterization of the

arrhythmogenic substrate, there is currently no standard approach to identify potential conduction

channels. Therefore, we sought to develop a workflow to identify HT channel based on the structural

3D modeling of the viable myocardium within areas of dense scar. We focus on macro-level HT channel

detection in this work. The proposed technique was tested in high-resolution ex-vivo CMR images in 20

post-infarct swine models who underwent an electrophysiology study for VT inducibility. HT channel

was detected in 15 animals with inducible VT, whereas it was only detected in 1 out of 5 animal with

non-inducible VT (P < 0.01, Fisher’s exact test). The HT channel detected in the non-inducible animal

was shorter than those detected in animals with inducible VTs (inducible-VT animals: 35 ± 14 mm vs.

non-inducible VT animal: 9.94 mm). Electrophysiology study and histopathological analyses validated

the detected HT channels. The proposed technique may provide new insights for understanding the

macro-level VT mechanism.

BACKGROUND:

Conduction velocity (CV) is an important property that contributes to the arrhythmogenicity of the tissue substrate. The aim of this study was to investigate the association between local CV versus late gadolinium enhancement (LGE) and myocardial wall thickness in a swine model of healed left ventricular infarction.

METHODS:

Six swine with healed myocardial infarction underwent cardiovascular magnetic resonance imaging and electroanatomic mapping. Two healthy controls (one treated with amiodarone and one unmedicated) underwent electroanatomic mapping with identical protocols to establish the baseline CV. CV was estimated using a triangulation technique. LGE+ regions were defined as signal intensity >2 SD than the mean of remote regions, wall thinning+ as those with wall thickness <2 SD than the mean of remote regions. LGE heterogeneity
was defined as SD of LGE in the local neighborhood of 5 mm and wall thickness gradient as SD within 5 mm. Cardiovascular magnetic resonance and electroanatomic mapping data were registered, and hierarchical modeling was performed to estimate the mean difference of CV (LGE+/−, wall thinning+/−), or the change of the mean of CV per unit change (LGE heterogeneity, wall thickness gradient).

RESULTS:

Significantly slower CV was observed in LGE+ (0.33±0.25 versus 0.54±0.36 m/s; P<0.001) and wall thinning+ regions (0.38±0.28 versus 0.55±0.37 m/s; P<0.001). Areas with greater LGE heterogeneity (P<0.001) and wall thickness gradient (P<0.001) exhibited slower CV.

CONCLUSIONS:

Slower CV is observed in the presence of LGE, myocardial wall thinning, high LGE heterogeneity, and a high wall thickness gradient. Cardiovascular magnetic resonance may offer a valuable imaging surrogate for estimating CV, which may support noninvasive identification of the arrhythmogenic substrate.

OBJECTIVES:

This study sought to examine the diagnostic ability of radiomic texture analysis (TA) on quantitative cardiovascular magnetic resonance images to differentiate between hypertensive heart disease (HHD) and hypertrophic cardiomyopathy (HCM).

BACKGROUND:

HHD and HCM are associated with increased left ventricular wall thickness (LVWT). Contemporary guidelines define HCM as LVWT ≥15 mm that is unexplained by other disease, which complicates diagnosis in cases of co-occurrences. Conventional global native T₁ mapping involves calculation of mean T₁ values as a surrogate for fibrosis. However, there may be differences in its spatial localization, such as diffuse and more focal fibrosis in HHD and HCM, respectively.

METHODS:

This study identified 232 subjects (53 with HHD, 108 with HCM, and 71 control subjects) for TA who consecutively underwent free-breathing multislice native T₁ mapping. Four sets of texture descriptors were applied to capture spatially dependent and independent pixel statistics. Six texture features were sequentially selected with the best discriminatory capacity between HHD and HCM and were tested using a support vector machine (SVM) classifier. Each disease group was randomly split 4:1 (feature selection/test validation), in which the reproducibility of the pattern was analyzed in the test validation dataset.

RESULTS:

The selected texture features provided the maximum diagnostic accuracy of 86.2% (c-statistic: 0.820; 95% confidence interval [CI]: 0.769 to 0.903) using the SVM. For the test validation dataset, the accuracy of the pattern remained high at 80.0% (c-statistic: 0.89; 95% CI: 0.77 to 1.00). Global native T₁, with an accuracy of 64%, separated HHD and HCM patients modestly (c-statistic: 0.549; 95% CI: 0.452 to 0.640).

CONCLUSIONS:

Radiomics analysis of native T₁ images discriminates between HHD and HCM patients and provides incremental value over global native T₁ mapping.

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.

KEYWORDS:

Automatic analysis; Convolutional neural networks; Myocardium segmentation; T1 mapping

PURPOSE:

To develop a gadolinium-free cardiac MR technique that simultaneously exploits native T₁ and magnetization transfer (MT) contrast for the imaging of myocardial infarction.

METHODS:

A novel hybrid T one and magnetization transfer (HYTOM) method was developed based on the modified look-locker inversion recovery (MOLLI) sequence, with a train of MT-prep pulses placed before the balanced SSFP (bSSFP) readout pulses. Numerical simulations, based on Bloch-McConnell equations, were performed to investigate the effects of MT induced by (1) the bSSFP readout pulses, and (2) the MT-prep pulses, on the measured, "apparent," native T₁ values. The HYTOM method was then tested on 8 healthy adult subjects, 6 patients, and a swine with prior myocardial infarction (MI). The resulting imaging contrast between normal myocardium and infarcted tissues was compared with that of MOLLI. Late gadolinium enhancement (LGE) images were also obtained for infarct assessment in patients and swine.

RESULTS:

Numerical simulation and in vivo studies in healthy volunteers demonstrated that MT effects, resulting from on-resonance bSSFP excitation pulses and off-resonance MT-prep pulses, reduce the measured T₁ in both MOLLI and HTYOM. In vivo studies in patients and swine showed that the HYTOM sequence can identify locations of MI, as seen on LGE. Furthermore, the HYTOM method yields higher myocardium-to-scar contrast than MOLLI (contrast-to-noise ratio: 7.33 ± 1.67 vs. 3.77 ± 0.66, P < 0.01).

CONCLUSION:

The proposed HYTOM method simultaneously exploits native T₁ and MT contrast and significantly boosts the imaging contrast for myocardial infarction.

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.

KEYWORDS:

compressed sensing; dictionary learning; motion correction; myocardial T1 mapping

BACKGROUND:

Left atrial ( LA ) enlargement is a marker for increased risk of atrial fibrillation ( AF ). However, LA remodeling is a complex process that is poorly understood, and LA geometric remodeling may also be associated with the development of AF . We sought to determine whether LA spherical remodeling or its temporal change predict late AF recurrence after pulmonary vein isolation ( PVI ).

METHODS AND RESULTS:

Two hundred twenty-seven consecutive patients scheduled for their first PVI for paroxysmal or persistent AF who underwent cardiovascular magnetic resonance before and within 6 months after PVI were retrospectively identified. The LA sphericity index was computed as the ratio of the measured LA maximum volume to the volume of a sphere with maximum LA length diameter. During mean follow-up of 25 months, 88 patients (39%) experienced late recurrence of AF. Multivariable Cox regression analyses identified an increased pre- PVI LA sphericity index as an independent predictor of late AF recurrence (hazard ratio, 1.32; 95% confidence interval, 1.07-1.62, P=0.009). Patients in the highest LA sphericity index tertile were at highest risk of late recurrence (highest versus lowest: 59% versus 28%; P<0.001). The integration of the LA sphericity index to the LA minimum volume index and passive emptying fraction provided important incremental prognostic information for predicting late AF recurrence post PVI (categorical net reclassification improvement, 0.43; 95% confidence interval, 0.16-0.69, P=0.001).

CONCLUSIONS:

The assessment of pre- PVI LA geometric remodeling provides incremental prognostic information regarding late AF recurrence and may be useful to identify those for whom PVI has reduced success or for whom more aggressive ablation or medications may be useful.

KEYWORDS:

atrial fibrillation; cardiovascular magnetic resonance; late recurrence; left atrial sphericity index; left atrial volume; pulmonary vein isolation