Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Aug 1.
Published in final edited form as: Magn Reson Med. 2015 Oct 19;76(2):555–565. doi: 10.1002/mrm.25907

Free-Breathing Slice-Interleaved Myocardial T2 Mapping with Slice-Selective T2 Magnetization Preparation

Tamer A Basha 1,2, Steven Bellm 1, Sébastien Roujol 1, Shingo Kato 1,3, Reza Nezafat 1
PMCID: PMC4837110  NIHMSID: NIHMS714549  PMID: 26479866

Abstract

Purpose

To develop and evaluate a free-breathing slice-interleaved T2 mapping sequence by proposing a new slice-selective T2 magnetization preparation (T2prep) sequence that allows interleaved data acquisition for different slices in subsequent heart beats.

Methods

We developed a slice-selective T2prep for myocardial T2 mapping by adding slice-selective gradients to conventional single-slice T2prep sequence. In this sequence, 5 slices are acquired during 5 consecutive heartbeats, each using a slice selective T2prep. The scheme is repeated four times using different T2prep echo times. We compared the performance of the proposed slice-interleaved T2 mapping sequence and the conventional single-slice T2 mapping sequence in term of accuracy, precision and reproducibility using phantom experiments and in-vivo imaging in 10 healthy subjects. We also evaluated the feasibility of the proposed sequence in 28 patients with cardiovascular disease and the quality of the maps was subjectively scored. Furthermore, we investigated the impact of through-plane motion by comparing T2 measurements acquired during end-systole vs. mid-diastole.

Results

T2 measurements using slice-interleaved T2 mapping sequence were correlated with spin-echo (r2 = 0.88) and single-slice T2 mapping sequence (r2 = 0.98). The mean myocardial T2 values were correlated between slice-interleaved (48 ms) and single-slice (51 ms) T2 mapping sequences. Subjective scores of T2 map quality were good to excellent in 81% of the maps in patients. There was no difference in T2 measurements between end-systole vs. mid-diastole.

Conclusions

The proposed free-breathing slice-interleaved T2 mapping sequence allows T2 measurements of 5 left ventricular slices in 20 heartbeats with similar reproducibility and precision as the single-slice T2 mapping sequence but with 4-fold reduction in acquisition time.

Keywords: Slice selective T2prep, slice-interleaved T2 mapping, Quantitative myocardial tissue characterization, myocardial T2 mapping, Precision, Reproducibility

Introduction

T2-weighted (T2W) imaging is currently the standard clinical protocol for depicting changes in myocardial T2 time for assessment of inflammation and edema (1). Commonly, a dark blood preparation pulse is combined with a turbo spin echo readout with or without fat saturation to image myocardial edema (2). However, there are several challenges in myocardial T2W imaging including regional variations of signals due to phased-array coil, stagnant blood, and low signal and contrast to noise ratios (3–6). For example, in the study by Thavendiranathan (7) only 47% of patients had diagnostic T2W image quality.

Myocardial T2 mapping is an alternative technique to T2W imaging (8–10). In this technique, several balanced steady-state free-precession images (bSSFP) are acquired immediately after a T2 magnetization preparation (T2Prep) pulses with different echo times. Subsequently, T2 value for each voxel is estimated by a fitting a recovery curve of T2 decays. To allow for full signal recovery before application of a new T2prep pulse, a rest period of several heartbeats, during which no data is acquired, is inserted (8,9). For example, a 2 heartbeats rest was used in the T2 mapping sequence proposed by Giri et. al. (9), resulting in the acquisition of only 3 T2W images over 9 heartbeat (i.e. data acquisition efficiency of 33%). However, this short rest period could results in underestimation of myocardial T2 values in patients with high heart rate (7). Therefore, higher number of rest-periods is required to reduce the sensitivity of T2 measurements to heart rate. In addition, edema and inflammation could occur regionally which necessitates whole heart coverage, which results in high scan inefficiency. For example, to acquire T2 maps from 5 short-axis slices of left ventricle, the scan time is 45 sec of which 30 sec are waiting time with no data acquisition.

3D myocardial T2 mapping is an alternative to 2D T2 mapping and have been recently proposed to improve spatial resolution and coverage of myocardial T2 mapping (11–14). 3D T2 mapping with 3D radial sampling allows myocardial T2 quantification with an isotropic resolution of 1.7 mm3 but with a scan time of 18 min (11). Instead of waiting for full magnetization recovery after each T2prep pulse, imaging every other heart-beat was proposed to reduce the scan time (13), but with reduced measurements accuracy (15). Application of a saturation pulse prior to T2prep has been recently reported to reduce the scan time as the cost of loss of signal-to-noise (12). Despite ongoing advances in 3D T2 mapping sequences, the scan duration remains long (>4 min), therefore 3D T2 mapping has not yet been adapted for routine clinical scanning.

T2prep sequence is used in myocardial T2 mapping sequences to create images with different T2 weighting (8–10,16,17). The T2prep sequence consist of a non-selective 90° radio-frequency pulse to excite the magnetization into the transverse plane followed by a series of composite or adiabatic 180° pulses and a final −90° pulse to tip the magnetization back to the longitudinal direction (18,19). All the pulses in this sequence are non-selective, resulting in magnetization excitation and T2 weighting over the entire volume of imaging. Therefore, this sequence requires a rest period after each T2prep image acquisition to allow for magnetization recovery.

In this study, we hypothesize that combining slice-selective T2prep with interleaved slice acquisition will eliminate the need for rest period after each T2prep image acquisition, thereby reducing the overall scan time of T2 mapping while increasing the spatial coverage. Based on that, we propose a novel slice-interleaved T2 mapping sequence by developing a slice-selective T2prep sequence, which allows interleaving the data acquisition of different slices in subsequent heartbeats. Phantom and in-vivo studies in healthy and patients with cardiovascular disease are performed to investigate the performance of the proposed sequence.

Methods

Pulse Sequence

Figure 1 shows the schematic of the proposed free breathing slice-interleaved T2 mapping sequence, illustrated for the acquisition of 5 slices, with the acquisition of 4 different T2prep echo times. For each T2prep echo time, each slice is acquired once using an ECG-triggered single shot acquisition. The first T2prep block usually corresponds to sampling the TET2Prep = 0 time point, with presumably no T2prep pulses applied. This block is then repeated for different TET2Prep echo times, where each image is now acquired after a slice-selective T2prep pulse. Finally, a last repetition of the block is performed by replacing each slice-selective T2prep pulses with a saturation pulse (ISAT) to simulate the effect of a very long TET2Prep (i.e. TET2Prep = ∞) (15). For TET2Prep = 0, we used T2prep pulses of zero time (i.e. 90° directly followed by −90° with no waiting time and no refocusing pulses in between) (15). In case of any flip angle imperfection, this method helps to uniformly having the same effect across all T2prep images, and thus minimizes its impact on the fitting process, and subsequently on the final estimated T2 times. For the ISAT acquisitions, we used a composite saturation pulse consisting of four rectangular 90° RF pulse train with crusher gradients in between the RF pulses (20). The RF pulses bandwidth was 1 kHz, and the composite total duration was 10 ms.

Figure 1.

Figure 1

Open in a new tab

a) Schematic of the proposed slice-interleaved T2 mapping sequence using slice-selective T2prep with b) an interleaved slice acquisition scheme. The imaging sequence uses a respiratory navigator (NAV) pulse for slice tracking by measuring respiratory motion of the right-hemi-diaphragm. The sequence consists of several T2prep blocks, each with a specific T2prep echo time. In each block, all slices are subsequently acquired in an interleaved fashion using ECG triggered T2-prepared single shot imaging. The first block usually corresponds to T2prep = 0, and the last block corresponds to T2prep = ∞, where each image is acquired immediately after a saturation pulse (ISAT). c) Schematic of the slice-selective T2 magnetization preparation composite using 2 slice-selective 90° and −90° pulses with four non-selective composite Malcolm Levitt's (MLEV) refocusing pulses. The refocusing gradient for the second 90° pulse is reversed to achieve a perfect nulling for the gradients zeroth moments in between the two 90° pulses. SS is the slice selection direction.

Figure 1.b illustrates the slice acquisition ordering, slice gaps, and slice thicknesses proposed for the T2prep and excitation pulses. The slice selective 90° pulses of the T2prep composite are applied with a slice thickness twice as large as the imaging slice to minimize the impact of slice profile imperfection. Data acquisition for different slices is interleaved to minimize slice cross-talk effects of both slice-selective T2prep pulse and excitation pulses. Furthermore, the acquisition of 5 slices provides a recovery time of 5 heart beats between two consecutive acquisitions of the same slice. Therefore, this approach removes the need for the additional 3–6 seconds rest periods commonly required in single-slice T2 mapping sequences to guarantee full spin recovery before each T2prep pulse (9,10,15). To compensate for through-plane myocardial motion, a prospective slice tracking was performed using a pencil-beam respiratory navigator (NAV) positioned on the right-hemi-diaphragm and acquired immediately before each T2prep pulse. We note that the NAV was used only for slice tracking without any gating; therefore the scan time was not increased.

Figure 1.c shows the sequence diagram for the proposed slice selective T2prep sequence. It consists of a tip-down slice selective 90° pulse, followed by four non-selective 180° refocus pulses and ends with a closing tip-up slice selective 90° pulse. A hard-pulse is used for the opening and closing 90° pulses with a bandwidth of 2.4 kHz, and duration of 0.88ms. To minimize the stimulated echo artifacts that might result from playing the slice selection gradients in between the two T2prep 90° pulses (15), we reversed the refocusing lobe of the slice-selection gradient for the second 90° pulse, similar to (21), in order to achieve a perfect nulling for the gradients zeroth moments in between the two 90° pulses. For the refocusing pulses, composite refocusing pulses (90°x, 180°y, 90°x) are used to provide second order corrections to variations in B1, and are weighted in a Malcolm Levitt's (MLEV) opposing phase pairs scheme to compensate for RF pulse shape imperfections (22). The duration of each refocus pulse is 1.75 ms.

Phantom Imaging

All imaging was performed on a 1.5T Philips Achieva (Philips Healthcare, Best, The Netherlands) MRI system using a 32-channel cardiac coil.

In a phantom imaging experiment, we compared the estimated T2 times obtained with the proposed sequence and a “conventional” single-slice sequence with no slice selective T2prep (15). Imaging was performed using 8 NiCl2 doped agarose vials, whose T2 and T1 values spanned the ranges of values found in the blood and myocardium (46, 49, 50, 51, 54, 62, 70, 80 ms). A single-shot ECG-triggered SSFP sequence was used for imaging in the proposed slice-interleaved T2 mapping sequence with the following parameters: 5-slices, FOV = 240×240 mm2, in-plane resolution = 2×2 mm2, slice thickness = 8mm, TR/TE = 2.2/1.1 ms, flip angle = 40°, 10 linear ramp-up pulses, SENSE rate = 2, acquisition window = 138 ms, number of phase encoding lines = 51, linear k-space ordering. Three different TET2Prep images were acquired at different TET2Prep = 0, 25, 50 ms, in addition to a single image after a saturation pulse to simulate TET2Prep = ∞ (total of 4 images per slice). For comparison, a 2D single-slice T2 mapping sequence was performed using the conventional non-selective T2prep sequence to image one slice of the phantom (which corresponded to the middle slice of the proposed sequence 5 slices). The same imaging and TET2Prep timing parameters were used for this non-selective T2prep sequence, but with a 4 seconds rest period after each image to allow for full spins recovery. T2 values were calculated using a 3-parameter model fit (15). To assess for measurement reproducibility, each sequence was repeated 10 times in a random order.

Additionally, a Carr-Purcell-Meiboom-Gill (CPMG) spin-echo sequence with an echo train length of 32 and TE = 10 ms (i.e. TE = 10, 20, 30‥‥320 ms) was performed as reference. The scan parameters were: FOV = 240×240 mm2, in-plane resolution = 1.25×1.25 mm2, slice thickness = 4 mm, TR = 6000 ms, flip angle = 90°, number of averages = 4. Reference T2 values were obtained from a 2-parameter model fit to the spin echo signal.

In Vivo Imaging

The study was health insurance portability and accountability act (HIPAA) compliant and the imaging protocol was approved by our institutional review board with written informed consent obtained prior to each examination. Ten healthy adult subjects (29 ± 18 years, range: 19 – 70 years, 4 men) without contraindications to MRI were recruited. Additionally, 28 patients (59±16 years, 18 males) referred for clinical CMR were recruited to undergo an additional 20 sec free-breathing slice-interleaved T2 mapping sequence in addition to clinically indicated CMR sequences to assess image quality of myocardial T2 mapping. The clinical CMR indications varied and included: atrial fibrillation (N=12), cardiac sarcoidosis (N=5), hypertrophic cardiomyopathy (N= 2), apical non-compaction (N=1), mitral valve regurgitation (N=1), amyloid (N=1), and ischemic heart disease (N= 6). Each healthy subject was imaged with the proposed slice-interleaved T2 mapping sequence with slice-selective T2prep and the “conventional” single-slice T2 mapping sequence. To assess for reproducibility, each sequence was acquired 5 times for each subject. Both sequences were performed under free breathing conditions and used a two-dimensional pencil-beam NAV positioned on the right hemi-diaphragm (RHD) to track the breathing motion and prospectively correct for the slice position during imaging. The NAV tracking used a 2D spatially selective spiral pulse (23,24), with 16 spiral excitation turns in 10 ms, and a flip angle of 90° to excite the magnetization in a circular area centered on the RHD with a diameter of 50 mm. All scans were acquired in the short axis orientation.

The proposed sequence used a free-breathing single-shot ECG-triggered slice-selective T2prep bSSFP acquisition with the following parameters: 5-slices, FOV = 320×320 mm2, in-plane resolution = 2.5×2.5 mm2, slice thickness = 8 mm, slice gap = 4mm, TR/TE = 2.2/1.1 ms, flip angle = 40°, 10 linear ramp-up pulses, SENSE rate = 2, acquisition window = 140 ms, number of phase encoding lines = 67, linear k-space ordering. Similar to the phantom experiment, 3 images were acquired with TET2Prep = 0, 25, 50 ms, in addition to the SAT image (total of 4 images per slice). The nominal acquisition time of this sequence was 20 heartbeats.

The single-slice T2 mapping sequence (15) was performed to image one 2D single-slice (corresponded to the middle slice of the proposed sequence) using the same imaging and TET2Prep timing parameters as the proposed sequence but with non-selective T2prep pulses. This sequence used a 4-seconds rest period after each T2W image acquisition to allow for a full spin recovery. The nominal acquisition time of this sequence was 13 heartbeats.

Sensitivity to cardiac motion

To assess the sensitivity of the slice-interleaved T2 mapping sequence with slice-selective T2prep to the cardiac motion during systole, we acquired additional data in healthy subjects (N= 7, 31 ± 18 years, range: 22 – 71 years, 4 men) and patients with known or suspected cardiovascular disease (N= 8, 47 ± 14 years, range: 28–65 years, 4 men). In healthy subjects, we acquired T2 maps using conventional single slice T2 mapping sequence and slice-interleaved T2 maps at end-systole and mid-diastole (4 different datasets). 5 slices in short axis view at the same location as slice-interleaved T2 maps were acquired. The imaging parameters were: FOV = 320×320 mm2, in-plane resolution = 2.1×2.1 mm2, slice thickness = 8 mm, slice gap = 4mm, TR/TE = 2.8/1.4 ms, flip angle = 55°, 10 linear ramp-up pulses, SENSE rate = 2.3, acquisition window = 185 ms, 3 TET2Prep echo images were acquired with TET2Prep = 0, 25, 50 ms, in addition to the SAT image. In patients, only two datasets were acquired using slice-interleaved T2 maps in mid-diastole and end-systole without any single-slice acquisition. Global and segmental myocardial T2 values using 16-segment AHA model were measured from each dataset.

Data Analysis

All images were transferred to a separate workstation for analysis. In the phantom data, voxel-wise T2 maps were generated using a 3-parameter model fit (15). Accuracy, precision and reproducibility were evaluated as follows. A region-of-interest (ROI) was manually defined for each vial in the spin echo data (ROISE). Similarly, a second ROI was defined for each vial in the data obtained with the two studied sequences (ROIseq). Accuracy was measured for each vial (v) as the difference between the average (over the 10 repetitions) of the mean T2 in ROIseq(v) and the mean reference T2 in ROISE(v). Precision was measured for each vial (v) as the average (over the 10 repetitions) of the T2 standard deviation in ROIseq(v). Reproducibility was measured for each vial (v) as the standard deviation (over the 10 repetitions) of the mean T2 in ROIseq(v).

For the in-vivo data, images were registered retrospectively using a non-rigid image registration algorithm (25) to compensate for residual in-plane motion. This algorithm simultaneously estimates a non-rigid motion field and intensity variations, and employs an additional regularization term to constrain the deformation field using automatic feature tracking. Upon registration, voxel-wise curve-fitting was performed, to generate T2 maps using a 3-parameter fitting model (15). Then, a myocardial segment based analysis was performed following the AHA myocardial segment model. Epi- and endocardial contours were drawn manually by an experienced reader for each T2 map in all slices. The myocardium was divided into 16 segments for the slice-interleaved results (using the three mid-ventricular slices), and 6 segments for the single-slice results. For each subject and segment, the average and standard deviation of T2 values were calculated. The standard deviation served as a surrogate of the precision. For each subject, the standard deviation (over the five repetitions) of the mean T2 values of each segment was calculated and served as a surrogate of the reproducibly. All calculations were performed using Matlab (v7.14, The MathWorks, Natick, MA).

For the patients data, the quality of the resulted T2 maps were assessed by consensus of two experienced readers using a 4-point scale that addresses the T2W image quality, registration quality, and the final T2 maps quality: 1-nondiagnostic (very low confidence in the map due to severe motion artifacts or bad image registration), 2-fair (low confidence in two or three segments of the map, but high confidence in the rest), 3-good (most of the map is excellent except for small localized areas of artifact in at most one segment), and 4-excellent (high confidence in all segments in the map).

Statistical Analysis

The quality of the curve fitting process, used to estimate the T2 values, is evaluated using the Chi-Square goodness of fit model and the corresponding p values are reported. Wilcoxon signed rank tests were used to compare the accuracy, precision and reproducibility of the proposed sequence, and the single-slice sequence in both the phantom study and the in-vivo measurements with statistical significance threshold defined at p < 0.05. In the in-vivo study, a one-way Kruskal–Wallis analysis of variance test was used to compare the T2 measurements at different myocardial levels (specifically, basal, mid and apical levels) with the same statistical significance threshold defined at p < 0.05.

Results

Phantom Imaging

Figure 2 shows the accuracy, precision and reproducibility of T2 measurements obtained in the phantom study using the single-slice and the proposed slice-interleaved T2 mapping sequences. For the vials with T2 times matching the range of normal myocardium (i.e. T2 = 40 to 60 ms), the slice-interleaved T2 values were within 5ms from the spin echo measurements and within 9 ms for the single-slice sequence. Across all vials, the slice-interleaved T2 measurements were lower than the single-slice measurements (p<0.05) but similar to the spin echo measurements (p=0.31). No significant differences were found in the precision and reproducibility between the two sequences (p = 0.09, 0.13 respectively). The goodness of fit for the fitting process was χ2 = 0.7–1.3, which corresponds to p=0.25–0.43).

Figure 2.

Figure 2

Open in a new tab

(a) Cross-sectional image for the phantom consisting of multiple cylindrical vials of different T1 and T2 values. Representative ROIs are shown to illustrate how measurements were calculated, and which vials were considered. The rest of the vials have T2 > 100 ms which is beyond the myocardium T2 range, (b) Accuracy, precision and reproducibility of the single-slice and the slice-interleaved T2 mapping sequences in a set of phantom vials with different T1/T2 values. Each sequence was repeated 10 times. Accuracy is defined as the difference between the averaged T2 times over all 10 repetitions and the spin echo T2 measurements. Precision is defined as the average over all 10 repetitions of the T2 standard deviation within each vial. Reproducibility is defined as the standard deviation over all 10 repetitions of the mean T2 times within each vial. The proposed slice-interleaved sequence yields similar absolute accuracy (p=0.915), precision (p=0.26), and reproducibility (p=0.29) compared with the single-slice sequence.

Figure 3.a shows a cross sectional image for the phantom with representative ROIs on drawn on the vials for measurements. Figure 3.b–3.d depict the regression analysis performed between the reference T2 values measured from the spin echo images (x-axis) and the mean of the estimated T2 values across repetitions using the single-slice (Figure 3.a) and the slice-interleaved (Figure 3.b) sequences. Figure 3.c shows the regression analysis between the T2 values obtained with the slice-interleaved and the singe-slice T2 mapping sequences. Figure 3.e–3.g shows the Bland-Altman plots corresponding to each of the previous regression plot. The slice-interleaved and single-slice measurements exhibit a good correlation with the spin echo measurements (r2 = 0.88 and 0.94 respectively). There is a strong correlation (r2 = 0.98) between the T2 values measured from single vs. slice-interleaved sequences with a regression slope of 0.8 indicating slightly lower T2 values when using the slice-interleaved sequence compared with the single-slice one.

Figure 3.

Figure 3

Open in a new tab

(a–c) Regression analysis between the reference T2 values measured from the spin echo images (x-axis) and the mean of the estimated T2 values across repetitions (y-axis) using a) single-slice and b) slice-interleaved sequences. c) Regression analysis obtained between the single-slice and the slice-interleaved measurements. The slice-interleaved measurements exhibit a strong correlation with both spin echo and single-slice measurements (r2 = 0.88 and 0.98 respectively), (d–f) Corresponding Bland-Altman plots.

In Vivo Imaging

Figure 4 shows example in-vivo T2 maps obtained with the five repetitions of the proposed sequence in a healthy subject. The maps visually appear homogeneous over all slices and myocardial segments suggesting a low spatial variability of T2 values. The quality of the proposed slice-interleaved T2 mapping sequence appears visually similar over the five repetition scans suggesting a good reproducibility of the method.

Figure 4.

Figure 4

Open in a new tab

Example T2 maps obtained in one subject with the slice-interleaved T2 mapping sequence. The maps are shown for all five slices and five repetitions of the sequence. All T2 maps show homogeneous T2 signal over all myocardium, slices, and repetitions.

Figure 5 shows an example of the T2 quantifications based on the proposed slice-interleaved sequence. Figure 5.a shows the bullseye of the T2 quantifications obtained in a healthy subject from one of the 5 repetitions. Figure 5.b shows the average T2 values at different slices (basal, mid and apical levels) for all healthy subjects between slice-interleaved vs. single-slice T2 mapping sequence. The Kruskal–Wallis test resulted in non-significant differences between the T2 values across the three ventricular slices (p=0.39).

Figure 5.

Figure 5

Open in a new tab

a) Example T2 maps obtained in one healthy subject with the associated bullseye of the myocardial segment-based quantification. The average T2 over the whole myocardium is shown in the center of the bullseye view. b) Basal, mid, and apical T2 values for different subjects. No significant difference was found in T2 values across the different slices (p=0.54).

Figure 6 shows an example comparison between the single and slice-interleaved sequences in 3 healthy subjects. For each subject, the T2 map generated from the single-slice sequence is compared to the T2 map generated for the corresponding slice in the slice-interleaved sequence. The T2 measurements, precision, and reproducibility are shown for each map. The goodness of fit was χ2 = 0.9–1.7 for the single-slice sequence and χ2 = 1.1–2.0 for the slice-interleaved sequence across all subjects. This corresponds to p = 0.37–0.21 and 0.31–0.17 respectively.

Figure 6.

Figure 6

Open in a new tab

Example myocardial T2 measurements, accuracy and precision of T2 maps obtained using the single-slice and the slice-interleaved sequences in 3 healthy subjects. The values in each bullseye center represent the average T2 times over the whole myocardium. Comparable image quality and T2 quantifications were obtained with the two sequences.

Figure 7 presents the subject-based analysis of in-vivo T2 quantifications over the whole myocardium in 10 healthy subjects. Over the entire myocardium, similar T2 (48±5.6 ms vs. 51±3.4 ms using single-slice sequence, p=0.07), precision (11±4.0 ms vs. 10±1.5 ms, p=0.7), and reproducibility (3±1.8 ms vs. 1.7±1 ms, p=0.19) values were obtained in the slice-interleaved sequence compared to the single-slice sequence.

Figure 7.

Figure 7

Open in a new tab

T2 measurements, precision, and reproducibility of the slice-interleaved and single-slice sequences obtained for each subject. Accuracy, precision and reproducibility were measured over the entire mid-ventricular slice in the slice-interleaved and the single-slice sequence. On average, similar T2 values measurements (48±5.6ms vs. 51±3.4ms using single-slice sequence, p=0.1), precision (11±4.0ms vs. 10±1.5ms, p=0.6) and reproducibility (3±1.8ms vs. 1.7±1ms, p=0.11) were obtained in the slice-interleaved sequence compared to the single-slice sequence.

Figure 8 shows the in-vivo myocardial segment-based analysis of T2 mapping in healthy subjects. The proposed slice-interleaved sequence provided lower segment-wise T2 values compared to the single-slice sequence in all segments (48±2.6 ms vs 51±1.8 ms, p<0.05). In average over all subjects, both sequences led to similar precision (9.8±2 ms vs. 9.5±2.5 ms, p=0.43) and reproducibility (3.3±0.5 ms vs. 3.7±1.2 ms, p=0.16) values. High precision of T2 measurements was obtained in septal segments when compared to free-wall segments for both sequences (7.5ms vs 11.5ms, p<0.05 in the single-slice, and 6.5ms vs 12ms, p<0.05 in the slice-interleaved). Reproducibility is also better in the septal segments compared to the free-wall segments in both slice-interleaved sequence (2.5ms vs 5ms, p<0.05) and single-slice sequence (3ms vs 4ms, p<0.05).

Figure 8.

Figure 8

Open in a new tab

Myocardial segment-based quantification of T2 values over all healthy subjects for a) measurements, b) precision, c) reproducibility. d) Bland-Altman plot of the T2 differences between the single-slice and the slice-interleaved sequences. The values in each bullseye center represent the average T2 times over the whole myocardium not the average of the segments values. The slice-interleaved sequence measurements led to lower segment-wise T2 values (48±3 ms vs 51±2 ms, p<0.05) and similar range of precision (9.8±2 ms vs. 9.5±3 ms, p=0.36) and reproducibility (3.3±0.5 ms vs. 3.7±1.2 ms, p=0.08) when compared with the single-slice sequence measurements.

Figure 9 shows examples of T2 maps with associated subjective scores used for the qualitative evaluation in patients. 81% of maps received a subjective score of good to excellent image quality. Only 6 slices (4%) were of poor/non-diagnosable quality, 3 of those were most apical slices. Figure 10 shows example T2 maps obtained in patients with elevated T2 times in one or more myocardial segment, which demonstrates the feasibility of this technique in detecting regional T2 elevation.

Figure 9.

Figure 9

Open in a new tab

Subjective qualitative analysis of T2 map quality in patients. a) Illustration of the four-point scale used for the assessment: 1-nondiagnostic (very low confidence in the map due to severe motion artifacts or bad image registration), 2-fair (low confidence in two or three segments of the map, but high confidence in the rest), 3-good (most of the map is excellent except for small localized areas of artifact in at most one segment), and 4-excellent (high confidence in all segments in the map). b) Overall scores on 28 patients with 140 slice maps.

Figure 10.

Figure 10

Open in a new tab

Overlaid T2 maps on magnitude images in three patients with dilated cardiomyopathy (subject 1) and suspected sarcoidosis (subject 2 and 3), which shows regional elevation of T2 values compared to normal myocardium.

Supporting Figure S1 shows example T2 maps at end-systole and mid-diastole acquired using single-slice and slice-interleaved T2 mapping sequences. Supporting Figure S2 shows Bland-Altman plots for the segment based T2 measurements at end-systole and mid-diastole in healthy subjects using the single-slice T2 mapping sequence (Fig. 12.a) and the slice-interleaved sequence (Fig. 12.b), and in patients using the slice-interleaved T2 mapping sequence (Fig 12.c). The different symbols represent different myocardial levels (i.e. basal, mid, and apical slices). There was no difference between global T2 measurements acquired using slice-interleaved T2 mapping sequence in end-systole and mid-diastole phases in healthy or patients (p=0.4, 0.35 respectively).

Discussion

In this study, we developed a novel free-breathing slice-interleaved myocardial T2 mapping sequence by implementing a slice-selective T2prep sequence in combination with an interleaved slice acquisition scheme. This significantly reduces the scan time and improves data acquisition efficiency allowing the coverage of more slices in a comparable scan time of one slice acquired using the single-slice T2 mapping sequence without slice-selective T2prep sequence. A 20 heartbeat free-breathing T2 mapping sequence with 5-slices of LV coverage can be easily integrated in any clinical CMR imaging protocol as we demonstrated in our initial clinical feasibility study.

Although, there was no statistically significant difference between T2 measurements from the single-slice vs. the slice-interleaved sequence, the proposed sequence led to reduced T2 values compared to the single-slice sequence in both phantom study (~4ms) and segment-based analysis of the in-vivo subjects (~3ms). However, “true” myocardial T2 value is unknown, therefore, it is unclear if the proposed slice-interleaved sequence under-estimate the actual T2 or it is closer to the true myocardial T2. In our phantom experiment, we found the T2 values measured by slice-interleaved sequence to be closer to spin-echo measurements and lower than single-slice sequence.

To allow for signal recovery of at least 5 heart-beat (~ 5 s for a heart rate of 60 bpm), the number of slices has to be ≥ 5 slices to maximize the throughput of the sequence in term of acquisition time vs. slice coverage. If the number of acquired slices is less than 5, rest cycles will become necessary, and the time efficiency of the sequence will be reduced. On the other hand, more slices can be acquired, which results in more time between subsequent data acquisitions for the same slice.

The proposed slice-interleaved T2 mapping sequence is acquired during free breathing. Although single-slice T2 mapping sequences are performed in one breath-hold, it has been shown over 50% of patients cannot sustain a stable 12–15sec breath-hold and retrospective motion correction is needed (25,26). Considering that we have to perform retrospective motion correction, we did not limit the scan time to the duration of a single breath-hold. A combination of prospective slice-tracking to minimize through plan motion and retrospective motion correction to reduce the impact of in-plane motion allows us to acquire free-breathing T2 mapping.

The proposed sequence can easily be adapted for segmented data acquisition for myocardial T2 mapping to acquire higher spatial resolution 2D mapping (27). The data acquisition for different k-space segments will be interleaved in exactly the same fashion the slices are interleaved in the proposed sequence, and thus can be used in a time-efficient way to acquire the entire 2D k-space data. Further studies are needed to implement and evaluate the performance of a segmented T2 mapping sequence using slice-selective T2prep sequence.

In the slice-selective T2prep composite design, we did not use slice-selective gradients for the refocusing pulses. The underlying assumption to allow this approach is that the time between the refocus pulses (i.e. TET2Prep/4 for 4 refocus pulses) is very small (e.g. 12.5ms for a TET2Prep = 50ms) compared to the myocardium T1 time (~1200ms at 1.5T), and thus the T1 recovery that occurs in the non-imaged slices, experiencing the multiple 180° refocus pulses without the 90°,−90° tipping pulses, is negligible especially that the next slice to be imaged is in the next cardiac cycle allowing enough time for the spins that might be inverted by the end of the composite to be recovered.

Slice-selective preparation pulses are typically more sensitive to through-plane motion. To minimize this impact, we have chosen a larger slab size for the T2prep pulse in addition to the prospective slice-tracking scheme. However, there may be still residual through-plane motion that could impact the T2 measurements, which might explain the lower precision values in subjects #1 and #3. To study the overall effect on the final T2 measurements, we compared between the T2 measurements at end-systole and mid-diastole on a separate group of subjects, however we did not find any differences between T2 values measured in systolic vs. diastole.

The recovery time for magnetization of each slice depends on the heart-rate. In our implementation of 5 slices, the time difference between two subsequent images for each slice is 5 heartbeats (i.e. 5 × (R-R interval) seconds). Thus, the relaxation time is reduced for elevated heart rates (i.e. shorter R-R interval), allowing less time for recovery, which in turns could affect the T2 weighting of the subsequent images. In two subjects (#1 and #9), the heart rate was varying significantly during the scans (between 60 and 90 bpm), which resulted in low reproducibility in the slice-interleaved T2 measurements. One potential solution is to increase the number of slices to allow for more relaxation time for each slice. Another alternative solution is to use a minimum rest cycle between data acquisition for the same slice.

We observed 10–20% variations in the myocardial T2 values among different myocardial segments. Similar variability ranges were reported in 3D T2 mapping (11,12). In (11), a T2 values ranges between 35 ms to 45 ms was reported across segments. In (12), the standard deviation of myocardial T2 measurements was 3.5–5 ms across the entire myocardium. We also note that this variability is in the same range as the measurements reproducibility (an average of 4 ms standard deviation among different repetitions), which suggests that these variations could be attributed to the imaging and physiological confounders of myocardial T2 mapping including sensitivity to field inhomogeneities.

We have investigated the proposed sequence only on a 1.5T system. Further studies at higher magnetic field are needed to assess the proposed slice-interleaved T2 mapping sequence in the presence of increased B1 and B0 inhomogeneities.

The slice-interleaved T2 mapping sequence uses a slice-selective T2prep sequence. Inflow and through-plane motion will adversely impact the T2prep. For myocardial T2 mapping, we have used a slice-tracking approach and a larger slice-selective gradient to minimize this impact. However, we expect blood T2 measurements to be affected with this sequence, which will result in inaccurate T2 measurements in the blood pool. Since T2 values of the left ventricular blood pool will not change at different location, a single scan using conventional single-slice T2 mapping may be sufficient if blood pool T2 measurement is needed.

Conclusion

The free-breathing slice-interleaved T2 mapping sequence allows T2 measurements of 5 left ventricular slices in 20 heartbeats with similar reproducibility and precision as a single-slice T2 mapping sequence but with 4-folds reduction in acquisition time.

Supplementary Material

Supp FigureS1-S2

Acknowledgements

The project described was partially supported by NIH R01EB008743-01A2, AHA 15EIA22710040 and Samsung Electronics, Suwon, South Korea.

References

  • 1.Eitel I, Friedrich MG. T2-weighted cardiovascular magnetic resonance in acute cardiac disease. J Cardiovasc Magn Reson. 2011;13(1):13. doi: 10.1186/1532-429X-13-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Simonetti OP, Finn JP, White RD, Laub G, Henry DA. "Black blood" T2-weighted inversion-recovery MR imaging of the heart. Radiology. 1996;199(1):49–57. doi: 10.1148/radiology.199.1.8633172. [DOI] [PubMed] [Google Scholar]
  • 3.Wince WB, Kim RJ. Molecular imaging: T2-weighted CMR of the area at risk—a risky business? Nature Reviews Cardiology. 2010;7(10):547–549. doi: 10.1038/nrcardio.2010.124. [DOI] [PubMed] [Google Scholar]
  • 4.Abdel-Aty H, Simonetti O, Friedrich MG. T2-weighted cardiovascular magnetic resonance imaging. J Magn Reson Imaging. 2007;26(3):452–459. doi: 10.1002/jmri.21028. [DOI] [PubMed] [Google Scholar]
  • 5.Arai AE. Using magnetic resonance imaging to characterize recent myocardial injury utility in acute coronary syndrome and other clinical scenarios. Circulation. 2008;118(8):795–796. doi: 10.1161/CIRCULATIONAHA.108.797373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Pennell D. Myocardial Salvage Retrospection, Resolution, and Radio Waves. Circulation. 2006;113(15):1821–1823. doi: 10.1161/CIRCULATIONAHA.105.618942. [DOI] [PubMed] [Google Scholar]
  • 7.Thavendiranathan P, Walls M, Giri S, Verhaert D, Rajagopalan S, Moore S, Simonetti OP, Raman SV. Improved detection of myocardial involvement in acute inflammatory cardiomyopathies using T2 mapping. Circulation: Cardiovascular Imaging. 2012;5(1):102–110. doi: 10.1161/CIRCIMAGING.111.967836. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Huang T-Y, Liu Y-J, Stemmer A, Poncelet BP. T2 measurement of the human myocardium using a T2-prepared transient-state trueFISP sequence. Magn Reson Med. 2007;57(5):960–966. doi: 10.1002/mrm.21208. [DOI] [PubMed] [Google Scholar]
  • 9.Giri S, Shah S, Xue H, Chung YC, Pennell ML, Guehring J, Zuehlsdorff S, Raman SV, Simonetti OP. Myocardial T2 mapping with respiratory navigator and automatic nonrigid motion correction. Magn Reson Med. 2012;68(5):1570–1578. doi: 10.1002/mrm.24139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.van Heeswijk RB, Feliciano H, Bongard C, Bonanno G, Coppo S, Lauriers N, Locca D, Schwitter J, Stuber M. Free-breathing 3 T magnetic resonance T2-mapping of the heart. JACC Cardiovasc Imaging. 2012;5(12):1231–1239. doi: 10.1016/j.jcmg.2012.06.010. [DOI] [PubMed] [Google Scholar]
  • 11.van Heeswijk RB, Piccini D, Feliciano H, Hullin R, Schwitter J, Stuber M. Self-navigated isotropic three-dimensional cardiac T mapping. Magn Reson Med. 2014 doi: 10.1002/mrm.25258. 10.1002/mrm.25258. [DOI] [PubMed] [Google Scholar]
  • 12.Ding H, Fernandez-de-Manuel L, Schär M, Schuleri KH, Halperin H, He L, Muz Zviman M, Beinart R, Herzka DA. Three-dimensional whole-heart T2 mapping at 3T. Magn Reson Med. 2014 doi: 10.1002/mrm.25458. [DOI] [PubMed] [Google Scholar]
  • 13.Yang H-J, Sharif B, Pang J, Kali A, Bi X, Cokic I, Li D, Dharmakumar R. Fast, Free-Breathing, Whole-heart cardiac magnetic resonance T2 mapping for myocardial edema imaging. Circulation. 2014;130(Suppl 2):A12805. [Google Scholar]
  • 14.Yang HJ, Sharif B, Pang J, Kali A, Bi X, Cokic I, Li D, Dharmakumar R. Free-breathing, motion-corrected, highly efficient whole heart T2 mapping at 3T with hybrid radial-cartesian trajectory. Magn Reson Med. 2015 doi: 10.1002/mrm.25576. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Akçakaya M, Basha TA, Weingärtner S, Roujol S, Berg S, Nezafat R. Improved quantitative myocardial T2 mapping: Impact of the fitting model. Magn Reson Med. 2014 doi: 10.1002/mrm.25377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Giri S, Chung YC, Merchant A, Mihai G, Rajagopalan S, Raman SV, Simonetti OP. T2 quantification for improved detection of myocardial edema. J Cardiovasc Magn Reson. 2009;11:56. doi: 10.1186/1532-429X-11-56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Blume U, Lockie T, Stehning C, Sinclair S, Uribe S, Razavi R, Schaeffter T. Interleaved T(1) and T(2) relaxation time mapping for cardiac applications. J Magn Reson Imaging. 2009;29(2):480–487. doi: 10.1002/jmri.21652. [DOI] [PubMed] [Google Scholar]
  • 18.Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishimura DG. Coronary Angiography with Magnetization-Prepared T2 Contrast. Magn Reson Med. 1995;33(5):689–696. doi: 10.1002/mrm.1910330515. [DOI] [PubMed] [Google Scholar]
  • 19.Nezafat R, Stuber M, Ouwerkerk R, Gharib AM, Desai MY, Pettigrew RI. B1-insensitive T2 preparation for improved coronary magnetic resonance angiography at 3 T. Magn Reson Med. 2006;55(4):858–864. doi: 10.1002/mrm.20835. [DOI] [PubMed] [Google Scholar]
  • 20.Oesingmann N, Zhang Q, Simonetti O. Improved saturation RF pulse design for myocardial first-pass perfusion at 3T. J Cardiovasc Magn Reson. 2004;6(1):373–374. [Google Scholar]
  • 21.Wu G, Lee S, Zhao X, Li Z. Crusher Gradient Reversal to Eliminate Stimulated Echo Artifacts in Dual Spin Echo Diffusion MRI. Proceedings of the 19th Scientific Meeting of ISMRM; Montreal. 2011. May, p. 1964. [Google Scholar]
  • 22.Levitt M, Freeman R, Frenkiel T. Broadband heteronuclear decoupling. J Magn Reson. 1982;47:328–330. [Google Scholar]
  • 23.Pauly J, Nishimura D, Macovski A. A k-space analysis of small-tip-angle excitation. J Magn Reson. 1989;81(1):43–56. doi: 10.1016/j.jmr.2011.09.023. [DOI] [PubMed] [Google Scholar]
  • 24.Hardy CJ, Cline HE. Broadband nuclear magnetic resonance pulses with two-dimensional spatial selectivity. J Appl Phys. 1989;66(4):1513–1516. [Google Scholar]
  • 25.Roujol S, Foppa M, Weingärtner S, Manning WJ, Nezafat R. Adaptive registration of varying contrast-weighted images for improved tissue characterization (ARCTIC): Application to T1 mapping. Magn Reson Med. 2014 doi: 10.1002/mrm.25270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Xue H, Shah S, Greiser A, Guetter C, Littmann A, Jolly M-P, Arai AE, Zuehlsdorff S, Guehring J, Kellman P. Motion correction for myocardial T1 mapping using image registration with synthetic image estimation. Magn Reson Med. 2012;67(6):1644–1655. doi: 10.1002/mrm.23153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Jang J, Tschabrunn C, Anter E, Basha TA, Nezafat R. High Resolution Multi-slice Myocardial T2 Mapping with Improved Scan Time Efficiency. Proceedings of the 23rd Scientific Meeting of ISMRM; Toronto. 2015. p. 2648. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp FigureS1-S2
NIHMS714549-supplement-Supp_FigureS1-S2.pdf (605.8KB, pdf)

RESOURCES