F U LL PA P ER

Black blood myocardial T2 mapping

Chengyan Wang1,2* | Jihye Jang1,3* | Ulf Neisius1 | Maryam Nezafat1 |

Ahmed Fahmy1,4 | Jinkyu Kang1 | Jennifer Rodriguez1 | Beth Goddu1 |

Patrick Pierce1 | Sophie Berg1 | Jue Zhang2 | Xiaoying Wang2,5 | Reza Nezafat1

1Department of Medicine (Cardiovascular Division), Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts

2Academy for Advanced Interdisciplinary Studies, Peking University, Beijing, People’s Republic of China3Department of Computer Science, Technical University of Munich, Munich, Germany

4Biomedical Engineering Department, Cairo University, Giza, Egypt

5Department of Radiology, Peking University First Hospital, Beijing, People’s Republic of China

CorrespondenceReza Nezafat, Beth Israel DeaconessMedical Center, 330 Brookline Ave,Boston, MA, 02215.Email: rnezafat@bidmc.harvard.edu

Funding informationThis study is supported in part by theNational Institutes of Health (NIH)1R01HL129185-01, 1R21HL127650, and1R01HL129157; and the American HeartAssociation (AHA) 15EIA22710040.Chengyan Wang is funded by theNational Science Foundation of China81627901, and 81571666; and the ChinaPostdoctoral Science Foundation17Z102060138

*Chengyan Wang and Jihye Jangcontributed equally to this study.

Purpose: To develop a black blood heart-rate adaptive T2-prepared balanced steady-state free-precession (BEATS) sequence for myocardial T2 mapping.

Methods: In BEATS, blood suppression is achieved by using a combination of pre-excitation and double inversion recovery pulses. The timing and flip angles of thepreexcitation pulse are auto-calculated in each patient based on heart rate. Numericalsimulations, phantom studies, and in vivo studies were conducted to evaluate the per-formance of BEATS. BEATS T2 maps were acquired in 36 patients referred forclinical cardiac MRI and in 1 swine with recent myocardial infarction. Two readersassessed all images acquired in patients to identify the presence of artifacts associatedwith slow blood flow.

Results: Phantom experiments showed that the BEATS sequence provided accurateT2 values over a wide range of simulated heart rates. Black blood myocardial T2

maps were successfully obtained in all subjects. No significant difference was foundbetween the average T2 measurements obtained from the BEATS and conventionalbright-blood T2; however, there was a decrease in precision using the BEATSsequence. A suppression of the blood pool resulted in sharper definition of theblood–myocardium border and reduced partial voluming effect. The subjectiveassessment showed that 16% (18 out of 108) of short-axis slices have residual bloodartifacts (12 in the apical slice, 4 in the midventricular slice, and 2 in the basal slice).

Conclusion: The BEATS sequence yields dark blood myocardial T2 maps withbetter definition of the blood–myocardium border. Further studies are warranted toevaluate diagnostic accuracy of black blood T2 mapping.

KEYWORD S

black blood, double inversion recovery, myocardial edema, myocardial T2 mapping, partial volume effect

1 | INTRODUCTION

Inflammatory processes underlie a broad spectrum of condi-tions that injure the myocardium and cause both structural*Chengyan Wang and Jihye Jang contributed equally to this study.

Magn. Reson. Med. 2018;1–14. wileyonlinelibrary.com/journal/mrm VC 2018 International Society for Magnetic Resonance in Medicine | 1

Received: 26 December 2017 | Revised: 21 April 2018 | Accepted: 23 April 2018

DOI: 10.1002/mrm.27360

Magnetic Resonance in Medicine

and functional deficits.1 In the presence of inflammation oredema, myocardial T2 increases and can be detected usingT2-weighted MRI.2,3 Myocardial T2 mapping is a quantita-tive technique that has emerged as a reproducible method toaccurately measure T2 relaxation times.4,5 Compared to con-ventional T2-weighted imaging, myocardial T2 mappingreduces the intersubject and interstudy variability. T2 map-ping has proven advantageous for visualizing myocardialedema in acute myocardial infarction (MI) cases with bothhigh sensitivity and specificity, yielding improved diagnosticaccuracy compared to T2-weighted dark-blood imaging.

T2 mapping can be performed using a single-shot turbospin-echo (SE) sequence with multiple echoes.6 However,the turbo SE sequence is limited by cardiac motion artifactsdue to a long acquisition window, which greatly affects theaccuracy of T2 measurements. An alternate method used formyocardial T2 mapping is T2-preparation (T2prep)

7-9 withbalanced steady-state free precession (bSSFP) imaging orgradient echo.4,10,11 A combination of spin echo excitationswith gradient echo readouts was also recently proposed aspart of this method.12 T2prep-bSSFP–based T2 mappingsequence provides very reproducible T2 measurements13 andhas replaced the turbo SE–based approach on clinical recom-mendation.14 In this sequence, T2prep pulses with differentTEs are used to create T2-decays of the magnetization for thegiven duration of the TE. A set of images are acquired withdifferent TEs and used to generate a T2 map through voxel-wise curve-fitting using either a 2-parameter or 3-parameterexponential model.15 However, T2 quantification in areasbordering the myocardium and the blood pool is challengingdue to partial volume errors that cause T2 measurementerrors. Although motion correction can improve measure-ment reproducibility,16 it cannot completely resolve the par-tial voluming error. One approach to resolve this limitation isto suppress the blood pool in myocardial T2 mapping. Doingso would effectively reduce the partial volume effects toimprove the image contrast at the blood–myocardium boun-daries, allowing for improved image registration and motioncorrection. However, blood suppression in myocardial T2

mapping is challenging, and there currently is no availablesequence for black blood T2 mapping.

There are numerous blood suppression preparationsequences in cardiac and vascular imaging.17-25 Doubleinversion recovery (DIR)18-25 sequencing is the most widelyused technique; a nonselective 1808 inversion pulse is fol-lowed by a slice-selective 1808 inversion pulse to invert theblood signal outside the imaging plane.25 Spatial saturationpulses26,27 in which a volume upstream from the slice is pre-excited and dephased have also been used in cardiacimaging. A motion-sensitizing magnetization preparationtechnique20,21,28,29 can also suppress the blood pool bydephasing all moving blood spins. However, these prepara-tion sequences are not directly applicable to myocardial T2

mapping. DIR sequence suffers from the problem of insuffi-cient recovery time for T2 mapping within a single R-R inter-val (the interval time between two R waves). Spatialsaturation pulses cannot robustly suppress the blood chamberdue to complex blood motion. DIR preparation also suffersfrom slow blood flow. Furthermore, motion-sensitized tech-niques cause additional signal suppression in the myocar-dium that impact T2 measurements. Therefore, there is anunmet clinical need for developing a robust black blood T2

mapping sequence.In this study, we sought to develop a black blood myocar-

dial T2 mapping sequence. Numerical simulation and phantomexperiments were used to evaluate the proposed sequence. Invivo studies in healthy subjects, patients, and an animal modelwere used to investigate the efficacy of the sequence to sup-press the blood pool in myocardial T2 mapping.

2 | METHODS

All the scans were performed on a 1.5T MR system (PhilipsAchieva, Philips Healthcare, Best, The Netherlands), withmaximum amplitude of 40 mT/m and a slew rate of 150 mT/m/ms. A 32-channel cardiac coil was used for signalreception.

2.1 | Pulse sequence design

We propose a “Black blood hEart-rate Adaptive T2-preparedbSSFP (BEATS)” sequence (Figure 1A) for myocardial T2

mapping. The pulse sequence consists of 3 preparationpulses: 1) a preexcitation nonselective RF pulse with a flipangle of b, 2) DIR pulses, and 3) T2prep pulses. Imagingwas performed using bSSFP readouts, similar to a conven-tional bright blood T2 mapping sequence.15

A DIR pulse is combined with T2 preparation pulses toobtain T2 mapping with blood suppression. However, bloodsuppression by DIR preparation pulses requires a long delaytime for blood magnetization recovery (due to the long bloodT1 relaxation time), which is not feasible within a single R-Rinterval. Therefore, we propose to insert a preexcitation pulsewith an adaptive flip angle prior to the DIR pulse (Figure1A). The preexcitation pulse is implemented as a nonselec-tive RF pulse followed by dephasing gradients. The flipangle of this preexcitation pulse is adaptively selected basedon the patient’s heart rate to ensure that the blood signal isfully suppressed within an R-R interval. After the preexcita-tion pulse, DIR pulses are performed subsequently with apair of inversion pulses: a nonselective 1808 inversion pulsefollowed immediately by a slice-selective 1808 inversionpulse. The width of the selective inversion pulse is 3 timesthat of the imaging slice thickness. Subsequently, T2preppulses, including an initial 908x pulse, followed by 2

2 | Magnetic Resonance in MedicineWANG ET AL.

composite 1808y pulses and a final 908-x flipback pulse,7,15

are applied to generate T2 weighting with variable TE.Finally, black blood T2-weighted images are acquired withsingle-shot bSSFP readouts. A rest period of 3 heart beats isused following each image acquisition to ensure sufficientmagnetization recovery.

The longitudinal signal changes during the BEATSsequence are shown in Figure 1B. After the preexcitationpulse, the longitudinal magnetizations (Mz) are reduced tocos(b) of the initial magnetization M0, and then recoveredaccording to the T1 relaxation times:

MzðTDbÞ5M01M0

�cosðbÞ21

�� e2TDb=1; (1)

where M0 is the initial longitudinal magnetization; b is theflip angle of the preexcitation pulse; and TDb is the durationbetween the preexcitation pulse and the nonselective 1808pulse in DIR. After the application of DIR pulses, the bloodsignal is immediately inverted while the myocardial signalremains. During the TI time (i.e., the time between the nonse-lective 1808 pulse and the start of T2prep pulses), blood andmyocardial signals recover as follows:

MzbloodðTIÞ5M0;blood2�M0;blood2MzbloodðTDbÞ

�� e2TI=T1 :

(2)

MzmyoðTIÞ5M0;myo2�M0;myo2MzmyoðTDbÞ

�� e2TI=T1 : (3)

According to Equation 2, if TI is selected asT1 � lnððM01MzbloodðTDbÞÞ=M0Þ, the blood signal can befully nulled before the T2prep pulses. However, because theheart rate varies among different patients, it is difficult to fixthe timing of TI. Therefore, we propose an adaptive timingdesign in which TI and b are automatically calculated basedon the patient heart rate. The TI is set to Tdia-TDb-TET2prep toacquire the data during the diastolic rest phase, for whichTdia is the trigger delay time and TET2prep is the TE of the T2

preparation pulse. The b is then calculated according toEquation 1 and Equation 2, and given by:

b5arccos�11e2TDb=T1blood �

�e2ðTdia2TDb2TEprepÞ=T1blood22

��: (4)

TI and b are adaptive to different heart rates in the BEATSsequence (Figure 1C), with T1blood equal to 1600 ms.30 A pre-excitation pulse duration (�0.2 ms), dephasing gradients (3.1ms), a nonselective 1808 inversion pulse (6.1 ms), and a TDb

of 20 ms (fixed empirically) were applied in this study.

2.2 | Numerical simulations

Bloch simulation was performed to investigate the influenceof heart rate (beats per minute [bpm] ) and tissue T1 times

FIGURE 1 (A) Pulse sequence diagram of the BEATS sequence. (B) Blood andmyocardial longitudinal signal relaxation curves during thesequence. The pulse sequence consists of the preexcitation pulse with an adaptive flip angle of b, DIR pulses, T2prep pulse, and a bSSFP acquisition. ThebSSFP readout is synchronized to the end-diastole. The TI after DIR and the flip angle of the preexcitation pulse are auto-calculated to adapt to the heartrate of individuals. (C) Calculation of TI and b as a function of different heart rates in the BEATS sequenceBEATS, Black blood hEart-rate Adaptive T2-prepared bSSFP; bpm, beats per minute; bSSFP, balanced steady-state free-precession; DIR, double inversion recovery; T2prep,T2 preparation; TDb, the duration between the preexcitation pulse and the nonselective 1808 pulse in DIR

WANG ET AL.Magnetic Resonance in Medicine | 3

both on the efficiency of blood suppression and on the recov-ery of the myocardial signal in the BEATS sequence. Theheart rates were simulated from 40 bpm to 130 bpm, whereasthe blood and myocardial T1 values were simulated rangingfrom 1200 ms to 2000 ms and from 600 ms to 1600 ms,respectively. The value of b was calculated according toEquation (4), assuming a fixed arbitrary value for TDb (20ms) and T1blood (51600 ms). TI was estimated as a functionof heart rate (HR):

TI � 60ðsÞ=HR2Trest; (5)

where Trest was the remaining time (ms) in each R-R interval(assumed to be 250 ms), including the acquisition window,trigger delay time, and pulse durations; and HR is heart rate.The blood and myocardium suppression were evaluated bycalculating SIblood/ SIblood,0 and SImyo/SImyo,0, in whichSIblood (or SImyo) was the residual blood (or myocardium)signal intensity (SI) after the inversion recovery and SIblood,0(or SImyo,0 ) represented the initial blood (or myocardium)signal intensity before suppression.

2.3 | Phantom study

To evaluate the performance of the BEATS sequence, phan-tom scans were performed on 6 NiCl2 doped agarose vials,with similar T2 values as the myocardial T2 (�40 ms) butdifferent T1 values.31 Both BEATS and conventional brightblood T2 mapping were performed on the phantom. Thescanning parameters of the phantom were as follows:FOV5 3203 320mm2, in-plane resolution5 23 2mm2,slice number5 1, slice thickness5 8mm, SENSE factor52.5, bSSFP readout with TR/TE:5 2.8/1.4 ms, flip angle5608, 10 linear ramp-up pulses, linear k-space ordering, andacquisition window5 150 ms. A total of 9 different T2prepTEs were used, including TET2prep5 0, TET2prep51, andTET2prep values ranging from 25 ms to 95 ms with linearsteps of 10 ms.15 The TET2prep of 1 was representative ofthe long T2prep TE achieved by applying a saturation pulseprior to image acquisition. Heart rates of 60 bpm were simu-lated for the conventional bright blood T2 mapping. For theBEATS scans, heart rates of 60, 70, 80, 90, 110, and 130bpm were simulated to test whether they affected T2 mea-surement accuracy. To evaluate the repeatability of the T2

measurements, the BEATS sequence scan was repeated 3times with 10min between scans.

Reference T2 relaxation times were obtained using a SEsequence of 32 echoes. The scanning parameters were as fol-lows: FOV5 3203 320mm2, in-plane resolution5 232mm2, slice thickness5 8mm, flip angle5 908, TR5 10 s,TEs from 10 ms to 320 ms with gaps of 10 ms, and numberof signal averages5 4. Reference T1 relaxation times wereacquired using the SE sequence with 16 inversion times of50, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1250,

1500, 1750, 2000, and 3000 ms. The other imagingparameters were: FOV5 3203 320mm2, in-plane resolu-tion5 23 2mm2, slice thickness5 8mm, flip angle5 908,TR/TE5 10 s/10 ms, and number of signal averages5 4.Both T1 and T2 values from the SE images were calculatedusing a 3-parameter fit model.32,33

2.4 | In vivo human study

This study was approved by our local institutional reviewboard, and written informed consents were obtained from allparticipants. In vivo human studies were performed on 8healthy adult subjects (26.06 4.7 years, range: 21–35 years,2 males) and 36 patients (55.16 16.6 years, range: 20–82years, 20 males) who were referred for clinical cardiac MRexaminations. The patient population consisted of patientswith suspected or known hypertrophic cardiomyopathy(N5 5), nonischemic cardiomyopathy (N5 9), suspected orconfirmed ischemic cardiomyopathy (N5 4), arrhythmias(N5 8), valvular heart disease (N5 2), left ventricular hyper-trophy (N5 3), lipomatous atrial septal hypertrophy (N5 1),athletic cardiomyopathy (N5 1), suspected or confirmedmyo- or pericarditis (N5 2), and cardiotoxic cardiomyopathy(N5 1). For each subject, localizer scout images were firstacquired to define the midventricular short-axis slice. Forboth BEATS and conventional bright blood T2prep-bSSFPimaging, free-breathing ECG-triggered single-shot bSSFPreadout was used for data acquisition. Sequence parameterswere as follows: FOV5 3203 320mm2; in-plane resolu-tion5 23 2mm2; slice number5 3; slice thickness/gap5 8/4mm; SENSE factor5 2.5; TR/TE5 2.8/1.4 ms; flipangle5 608; and 9 T2prep TEs, including TET2prep5 0, 25,35, 45, 55, 65, 75, 85, 95, and 1 ms. A 2D spiral-beam nav-igator was placed at the lung–liver interface of the righthemidiaphragm and was used for gating with a window of5mm. The typical scan time for each slice was �40 sec. Toassess the scan repeatability of black blood T2 imaging usingthe BEATS sequence, each scan was repeated 3 times,10min apart, in only healthy adult subjects. Becauserecruited patients included those referred for clinical cardiacMR examinations, there was not sufficient time to assessrepeatability in patients.

2.5 | Animal study

The protocol was approved by The Beth Israel DeaconessMedical Center Institutional Animal Care and Use Commit-tee and conformed to the position of the American HeartAssociation on Research Animal Use and the Declaration ofHelsinki. Ischemia–reperfusion mediated MI was induced in1 Yorkshire swine (35 kg), as previously described.34,35

Under fluoroscopic guidance, an angioplasty balloon wasinflated in the mid-left anterior descending for 180 minutes.

4 | Magnetic Resonance in MedicineWANG ET AL.

The balloon was then deflated and withdrawn to create anischemia–reperfusion mediated MI. The animal was imaged1 week after infarct creation using an imaging protocol,which included both conventional bright blood T2 mappingand the BEATS sequence with similar imaging parameters,as described above in our human study.

2.6 | Data analysis

T2 maps were reconstructed offline using a custom-writtenprogram on MatLab (MathWorks, Natick, Massachusetts). A3-parameter fitting model was applied for pixel-wise T2

quantification15:

SmyoðTET2prepÞ5A � e2TET2 prep=T21B; (6)

where TET2prep is the TE; A is a function of the signal at fullrecovery; T2 is the myocardial T2 relaxation time; and Bcharacterizes the effect of the bSSFP imaging pulses thatwere played until the acquisition of the central k-space line.Given a set of images acquired at different (known) TEs, themodel in Equation 6 is used to fit the intensities at each voxelto estimate the parameter T2.

In the phantom experiments, region-of-interest (ROI)analysis was performed for each vial, for which the meanvalue and SD were recorded. Identical ROIs were used in allthe T2 maps with different simulated heart rates. SNR wascalculated in the source images at TET2prep5 0 ms as theaverage signal in the phantom divided by the SD of the noiseregion.33 ROI for the noise measurement was selected from abackground area free of image artifacts.

For the in vivo study, the acquired images were coregis-tered using an affine coregistration algorithm16,36 before T2

fitting. Voxel-wise curve fitting was then performed on bothblack blood and conventional bright blood T2prep-bSSFPimages according to Equation (5). To avoid fitting for theresidual blood signal in black blood T2prep-bSSFP images, athreshold processing was performed. The threshold wasselected empirically as 20% of the myocardial signalintensity in the black blood T2prep-bSSFP image, withTET2prep5 0. Regions with signal intensities below thisthreshold were avoided by the T2 fitting. The myocardiumwas manually traced by drawing the epicardial and endocar-dial contours. T2 profiles of a cross section of the midventric-ular slice were obtained on black blood and conventionalbright blood T2 maps. To evaluate the effectiveness of bloodsuppression, the blood-to-myocardial signal ratio was meas-ured for each subject.

Furthermore, a segment-based analysis was performed toinvestigate the regional T2 variations. Six segments wereused in the midventricular slice, in accordance with theAmerican Heart Association segmentation guidelines.37 Ineach segment, the mean T2 value and its SD were calculated.

In clinical patients, interobserver agreement for T2 quantifi-cation was assessed by using intraclass correlation coefficientanalysis from 10 randomly selected patients.

A subjective analysis was conducted by 2 experiencedreaders (U.N. and R.N.) in consensus to assess artifacts due toblood flow from DIR pulses in BEATS. Readers evaluatedthe T2 maps (reconstructed without performing any thresh-olding of signaling, as described above) and original T2-weighted images to identify slices with artifacts associatedwith incomplete suppression of the blood pool. The numberof slices impacted by the artifacts are reported for apical,midventricular, and basal slices. Furthermore, for patientswith flow artifacts, readers indicated whether the presence offlow artifact impacted their confidence in reliably measuringmyocardial T2 values.

2.7 | Statistical analysis

Data were reported as median (25th percentile; 75th percen-tile). The accuracy of the BEATS T2 measurement wasassessed as the difference between the black blood T2 mapsand reference T2 maps acquired with the standard SEsequence. Precision was assessed as the SD within the ROIsfor the given T2 map. A higher SD indicates lower precision.Repeatability was defined as the SD over the 3 repeatedscans of the spatial average T2 values in each subject. Theaccuracy, precision, and repeatability of the values weremeasured in each segment of the myocardium. TheWilcoxon signed-rank test was conducted to compare the T2

measurements with different methods. A P value of less than0.05 was considered statistically significant.

3 | RESULTS

3.1 | Numerical simulation

In the numerical simulation, both heart rate and blood T1

relaxation time influence blood suppression (Figure 2A). Theremaining blood signal dropped as the simulated heart rateincreased due to limited T1 decay time. However, the resid-ual blood signal after inversion recovery was maintainedwithin 22.7% of the original blood signal for a wide range ofsimulated blood T1 values ranging from 1200 to 2000 ms.

Myocardial signal recovery was influenced by the heartrate and the myocardial T1 relaxation time (Figure 2B). Asexpected, the heart rate showed a strong influence on themyocardial signal intensity. For example, at T15 1200 ms,the relative myocardial signal dropped as the simulated heartrate increased (1.00, 0.91, 0.79, 0.68, 0.58, 0.51, 0.44, 0.38,and 0.33 for simulated heart rates of 40, 50, 60, 70, 80, 90,100, 110, and 120 bpm, respectively) due to reduced TI. Ingeneral, for T1< 1200 ms and heart rate< 85 bpm, morethan 69.3% of the myocardial signal was preserved.

WANG ET AL.Magnetic Resonance in Medicine | 5

3.2 | Phantom

T2 maps generated from the SE, conventional bright blood,and BEATS sequences are shown in Figure 3. The T2 meas-urements of the phantoms using conventional bright bloodand the BEATS sequence with different simulated heart ratesare shown in Supporting Information Figure S1. Comparedto T2 measured from conventional bright blood T2 mapping,the average T2 measurements in BEATS T2 maps were notaffected with increased heart rates within a range of 60 to110 bpm; however, higher heart rates decreased measure-ment precision (observed as increased SD). The SNR within

the different vials decreased as the heart rate increased (Fig-ure 4A) due to insufficient time for signal recovery. Goodrepeatability was achieved for the BEATS T2 mapping, asmeasured by the SDs over repeated scans (0.29 ms, 0.35 ms,0.35 ms, 0.29 ms, and 0.42 ms for heart rate of 60, 70, 80,90, and 100 bpm, respectively). The phantom experimentconfirmed the accuracy of the BEATS T2 mapping (Figure4B). As expected, T2prep and heart rate had no impact onthe accuracy of T2 measurements in a range of 60 to 110bpm but did have an impact at heart rates higher than 130bpm. There was no difference in T2 values in vials measuredusing the BEATS or conventional bright blood T2 mappingsequences over heart rates of 60 to 130 bpm.

3.3 | In vivo human study

Black blood myocardial T2 images were successfullyobtained in all subjects using the BEATS sequence. Theaverage heart rate was 676 5 bpm (range: 55–75 bpm) forhealthy subjects and 666 11 bpm (range: 49–96 bpm) forpatients. The blood-to-myocardial signal ratio was reducedfrom 3.03 (2.56; 3.72) to 0.15 (0.13; 0.17) (P5 0.01) inhealthy subjects.

T2-weighted images with TET2prep5 0, 25, and 55 ms, aswell as fitted T2 maps acquired by BEATS and conventionalbright-blood T2 mapping sequences, are shown in 2 healthysubjects with different heart rates: 70 bpm (subject 1) and 55bpm (subject 2) (Figure 5). In both cases, the BEATSsequence produced T2-weighted images without any visualartifact. Both images demonstrate a relatively high contrastbetween the myocardium and the blood pool, with definedborders. The partial volume effect was significantly reducedwith the suppression of intraventricular blood signals.

FIGURE 2 Numerical simulation for (A) the effect of heart rate and blood T1 on the suppression of the blood signal (SIblood/SIblood,0), and (B) theeffect of heart rate andmyocardial T1 on the recovery of the myocardial signal (SImyo/SImyo,0). SIblood,0 and SImyo,0 are the initial blood andmyocardialsignals, respectively. Note that the remaining blood signal dropped as the simulated heart rate increased due to a limited T1 decay time. Similarly, therelativemyocardial signal dropped as the simulated heart rate increased due to reduction of TISI, signal intensity; myo, myocardium

FIGURE 3 T2 maps acquired in a phantomwith 6 different vials,measured using the spin-echo, conventional bright blood, and BEATSsequencesHR, heart rate.

6 | Magnetic Resonance in MedicineWANG ET AL.

Compared to conventional bright blood T2 mapping, BEATST2 mapping successfully reduced the blood signal, leading tobetter depiction of the myocardial structures. Signal intensityprofiles were drawn in the T2 maps across the myocardium(Figure 6, red lines). The T2 profiles have a much sharpertransition between blood and myocardium compared to thosein conventional bright blood T2 maps, suggesting a decreasein the partial volume effect of myocardial T2. Segment-basedmeasurements show comparable T2 values calculated fromBEATS and bright blood T2 mapping sequences (Figure 6).Example T2 maps from our patients are presented in Figure 7and show good image quality and maps. The original T2-

weighted images are included in Supporting InformationVideos S1 and S2. Myocardial edema was visually present inonly 1 patient with a recent history of MI (Figure 8). Arrowsshow areas of higher T2 values (59–63 ms) located at areaswith recent infarct. A signal difference in this region can alsobe seen on original black blood T2 weighted images.

No significant difference was found between the averageT2 measurements obtained from the BEATS and conventionalbright-blood T2 maps (48.8 ms [48.2; 50.0] vs. 48.7 ms [48.0;49.9], P5 0.67 in healthy subjects and 48.9 ms [47.1; 51.7]vs. 50.4 ms [48.7; 51.8], P5 0.05 in patients) (Figure 9). Theinterobserver agreement for both T2 maps was strong, with an

FIGURE 4 Phantom results of (A) SNR and (B) T2 accuracy from 6 vials measured with different simulated heart rates ranging from 60 to130 bpm

FIGURE 5 Example short axis midventricular images and T2 maps obtained from conventional bright blood T2prep-bSSFP and BEATS sequences.Only 3 (out of 9) representative T2-weighted images (TE5 0, 25, 55ms) are displayed. All the grayscale images are displayed at the same scale. Note thatthe BEATS sequence provides effective suppression of the intraventricular blood and clear depiction of the ventricular walls

WANG ET AL.Magnetic Resonance in Medicine | 7

FIGURE 6 Examples of T2 mappings obtained from conventional bright-blood T2prep-bSSFP and BEATS sequences and the T2 profiles (drawnwith interpolation). Segment-based T2 measurements of both cases are also shown in the bullseye. The red lines in the T2 mappings indicate where the T2profiles are generated. T2 profiles through the T2 maps indicate substantially steeper myocardial–blood transitions in the BEATS T2maps

FIGURE 7 Examples of T2-weighted images (TE5 0, 55, 95ms) and T2 maps acquired with BEATS from 2 patients with heart rates of 60 bpm(upper) and 70 bpm (lower). Both T2-weighted images and T2 maps show good image quality

8 | Magnetic Resonance in MedicineWANG ET AL.

intraclass correlation coefficient of 0.74 (95% confidence inter-val [CI]: 0.29–0.93) for BEATS, and intraclass correlationcoefficient of 0.93 (95% CI: 0.75–0.98) for the bright-blood T2map. Homogeneous T2 maps, as quantified by precision, wereobserved throughout the myocardium using both techniques,with a decrease in precision for the BEATS sequence due toloss of signal associated with the DIR pulse (healthy subjects:6.87 ms [6.2; 7.3] vs. 4.9 ms [4.7; 6.6] P5 0.01, patients: 8.4ms [7.3; 11.1] vs. 6.7 ms [5.7; 7.8], P< 0.001). Repeatabilityanalysis in healthy subjects shows that the variability of T2

measurements was low and thus comparable between the 2techniques (BEATS, 2.0 ms [1.6; 2.5]; bright-blood T2, 1.95ms [1.6; 2.2], P5 0.58).

The subjective assessment showed that 16% (18 of 108)of short-axis slices have artifacts due to a residual blood sig-nal from the DIR pulse in BEATS (12 in the apical slice, 4in the midventricular slice, and 2 in the basal slice). How-ever, the slow-flowing blood artifact only impacted measure-ments in 10 slices (6 at the apex, 3 at the midventricle, and 1at the base). The flow artifacts often resulted in large T2

FIGURE 8 Posterior left ventricular wall edema after a myocardial infarct. The patient presented with new onset chest pain and electrocardiographicfindings consistent with a posterior myocardial infarction (A, ST-segment depression and inverted T waves in V2–3; arrows). On coronary angiogram, alarge obtuse marginal branchwas occluded (B; arrow), and flowwas reestablished with complete filling of the artery (C; arrow). On day 4 postinfarct,comparison of the myocardial T2 mappings obtained from conventional bright blood T2prep-bSSFP and BEATS sequences (D) showed left ventricularsubendocardial edema (red arrows). Segment-based T2 measurements were also shown in the bullseye. ST-segment, the segment of the ECG between theend of the S wave and the beginning of the T wave

WANG ET AL.Magnetic Resonance in Medicine | 9

values or noisy areas with random values that could be iden-tified by reviewing original T2 weighted images. The loca-tion of artifacts was random and varied among differentpatients, including the middle of the left ventricular cavity,the subendocardial region, and around the papillary musclewhere blood is trapped between the papillary muscle and themyocardium (Supporting Information Figure S2). There wasno correlation between the presence of flow-related artifactsand heart rate (Spearman’s rho5 0.19, P5 0.27), and arti-facts were present in patients with both low and high heartrates (Supporting Information Figure S3) (Supporting Infor-mation Video S3 shows images of a patient with a heart rateof 96 bpm). In 1 patient, there was a large subendocardialenhancement in BEATS T2 maps, consistent with thepatient’s clinical history; however, readers (U.N. and R.N.)could not confirm with confidence if the increased T2 wasdue to presence of edema or slow blood flow (SupportingInformation Figure S4) (Supporting Information Video S4).

3.4 | Animal study

T2-weighted images and corresponding T2 maps in swineconfirms the presence of edema (red arrows) in the anterior

septum, consistent with the location of the infarct in our ani-mal model (Figure 10). The measured T2 in an ROI withinthe edematous region was 82 ms for conventional brightblood T2 mapping and 81 ms for BEATS T2 mapping. Theextent of edema was better depicted in the BEATS T2 mapdue to the suppression of the blood signal. Furthermore, theT2 weighted images with the suppressed blood pool in theBEATS sequence yielded improved depiction of myocardialedema in the anterior septum (red arrows), further confirmingthe presence of edema.

4 | DISCUSSION

In this study, we present a black blood myocardial T2 map-ping sequence to reduce the partial voluming effect presentin conventional bright blood T2 mapping by suppressing theblood signal using a combination of preexcitations and DIRpulses. The performance of the BEATS T2 mappingsequence was evaluated by numerical simulations, phantomstudies, and in vivo studies. The results of our study demon-strated that the BEATS sequence yields images with sup-pressed blood pool signal, resulting in better definition of theblood–myocardium border and improved visualization ofedema/inflammation.

In the BEATS sequence, selection of appropriate timingsof the preexcitations and DIR pulses are essential for bloodsuppression and the minimization of SNR loss. These param-eters should be adaptively selected based on an individualheart rate to achieve blood suppression. A lower flip angle ofthe preexcitation pulse will minimize the myocardial signalloss and improve SNR, but longer TI will be required. In theBEATS sequence, we proposed an adaptive parameter designaccording to the patient’s heart rate. TI and flip angle of thepreexcitation pulse are autocalculated in the sequence toachieve optimal blood suppression with minimal myocardialsignal loss. In our study, we applied a 3-parameter fittingmodel to 9 T2-weighted images, with an additional saturatedimage acquired during free-breathing to estimate the T2

maps.15 However, T2 maps can be estimated with as low as4 images within a single breath-hold, albeit with lowerprecision.38

One of the challenges of the DIR preparation pulse inblack blood T2 mapping is the limitation of the number ofslices that can be acquired per scan. There have been recentadvances in myocardial T2 mapping to increase coverage anddata acquisition efficiency using either whole heart or slice-interleaved acquisition.39-43 Our current implementation ofBEATS limits the acquisition to a single slice per acquisition.Improved DIR for multi-slice imaging has been previouslydemonstrated22-24 and could potentially be adopted forBEATS T2 mapping and warrant further investigation.

FIGURE 9 The bullseye images of T2 (mean), precision, and repeat-ability among the 8 healthy subjects and 36 patients in a segment-basedanalysis using both conventional bright-blood T2prep-bSSFP and theBEATS sequence. The center T2 represents segmental mean T2 in all 6 seg-ments. No statistical significant differences of T2 measurements were found

10 | Magnetic Resonance in MedicineWANG ET AL.

Inflammation or edema could also occur in the thin wallof the left atrium (LA) or right ventricle. For example, radio-frequency ablation for treatment of atrial fibrillation causesLA edema, which can be detected with T2 weighted imag-ing44-46; however, T2 weighted imaging in LA or rightventricle is challenging and interpretation is subjective. Con-ventional bright blood T2 maps do not allow accurate mea-surement of LA T2 due to partial voluming. The proposedblack blood T2 mapping could potentially be used to measureLA/right ventricle T2; however, it will require higher spatialresolution. Further validation and technical improvementsare necessary to develop high-resolution black blood T2

mapping for LA/right ventricle.We used a combination of a preexcitation pulses with

DIR to suppress the blood pool in BEATS. We onlyacquired T2 maps in the short-axis view because DIR doesnot effectively null the blood signal if blood flows withinthe imaging plane (Supporting Information Figure S5).One main problem with DIR could be slow-moving bloodalong the myocardial border. This is a well-known problemin myocardial regions affected by disease (e.g., dyskineticmyocardium following MI), overall contractile motionanomalies as observed in myocarditis, or significant trabe-culations. In such cases, the stagnant blood can appearbright in each individual T2prep-SSFP image. However,blood will experience different preparation pulses in

subsequent data acquisitions, and the signal may not followa simple T2 decay. Therefore, the region may either appearnoisy in the estimated T2 maps or with very high T2 values.Our preliminary data shows these types of artifacts couldoccur in different slices and in patients with different heartrates. A combination of postprocessing to account forblood flow may be used to mitigate the issue. Further stud-ies are needed to minimize artifact in the proposed blackblood T2 mapping sequence.

Our study has limitations. We used a predetermined fixedT1 and did not measure T1 values of the blood to optimizethe imaging parameters. A patient-specific blood T1 can beacquired using the T1 mapping sequence, but this was notstudied. For high heart rates >100 bpm, the myocardial sig-nal will decrease, which could impact measurement accuracyand precision. For patients with high heart rates, DIR pulsesand imaging readouts can be applied in 2 subsequent heart-beats; however, this will make the sequence sensitive to heartrate variability. The sample size of our patient data wassmall, and the majority of patients did not have edema orinflammation. Therefore, we cannot assess the diagnosticaccuracy of BEATS; doing so requires a rigorous animalstudy with histological validation, which is not within thescope of our study. There is no gold standard noninvasiveimaging modality with sufficient spatial resolution that canbe used to assess sensitivity or specificity of black blood T2

FIGURE 10 Example of short-axis midventricular T2 maps obtained from conventional bright-blood T2prep-bSSFP and BEATS sequences in ananimal model with prior myocardial infarction. Edema was observed in anterior septum (red arrows) in bothmaps (bright-blood: 82ms, BEATS: 81ms).Signal intensity profiles drawn across the myocardium in edema (yellow lines) in BEATS show sharper transition between blood andmyocardium

WANG ET AL.Magnetic Resonance in Medicine | 11

mapping sequencing for identifying subendocardial inflam-mation in patients.

5 | CONCLUSION

The proposed black blood myocardial T2 mapping sequenceefficiently suppresses the blood signal, resulting in betterdefinition of the blood–myocardium border by reducing theimpact of the partial volume effect that is present in conven-tional bright-blood T2 mapping. The residual blood flowassociated with DIR is a limitation of the proposedtechnique.

ACKNOWLEDGMENT

Authors would like to acknowledge Dr. Warren J. Manningfor editorial assistance. This study is supported in part bythe National Institutes of Health (NIH) 1R01HL129185-01,1R21HL127650, and 1R01HL129157; and the AmericanHeart Association (AHA) 15EIA22710040. ChengyanWang is funded by the National Science Foundation ofChina 81627901, and 81571666; and the China Postdoc-toral Science Foundation 17Z102060138.

ORCID

Jue Zhang http://orcid.org/0000-0003-0440-1357

REFERENCES[1] Marchant DJ, Boyd JH, Lin DC, Granville DJ, Garmaroudi FS,

McManus BM. Inflammation in myocardial diseases. Circ Res.2012;110:126-144.

[2] Mcnamara MT, Higgins CB, Schechtmann N, et al. Detectionand characterization of acute myocardial infarction in man withuse of gated magnetic resonance. Circulation. 1985;71:717-724.

[3] Payne AR, Casey M, McClure J, et al. Bright-blood T2-weighted MRI has higher diagnostic accuracy than dark-bloodshort tau inversion recovery MRI for detection of acute myocar-dial infarction and for assessment of the ischemic area at riskand myocardial salvage. Circ Cardiovasc Imaging. 2011;4:210-219.

[4] Huang T-Y, Liu Y-J, Stemmer A, Poncelet BP. T2 measurementof the human myocardium using a T2-prepared transient-statetrueFISP sequence. Magn Reson Med. 2007;57:960-966.

[5] Baeßler B, Schaarschmidt F, Stehning C, et al. Reproducibilityof three different cardiac T2-mapping sequences at 1.5 T.J Magn Reson Imaging. 2016;44:1168-1178.

[6] Kim D, Jensen JH, Wu EX, Sheth SS, Brittenham GM. Breath-hold multiecho fast spin-echo pulse sequence for accurate R2measurement in the heart and liver. Magn Reson Med. 2009;62:300-306.

[7] Brittain JH, Hu BS, Wright GA, Meyer CH, Macovski A, Nishi-mura DG. Coronary angiography with magnetization-preparedT2 contrast. Magn Reson Med. 1995;33:689-696.

[8] Nezafat R, Ouwerkerk R, Derbyshire AJ, Stuber M, McVeighER. Spectrally selective B1-insensitive T2 magnetization prepa-ration sequence. Magn Reson Med. 2009;61:1326-1335.

[9] Nezafat R, Stuber M, Ouwerkerk R, Gharib AM, Desai MY,Pettigrew RI. B1-insensitive T2 preparation for improved coro-nary magnetic resonance angiography at 3 T. Magn Reson Med.2006;55:858-864.

[10] Giri S, Chung Y-C, Merchant A, et al. T2 quantification forimproved detection of myocardial edema. J Cardiovasc MagnReson. 2009;11:56.

[11] Foltz WD, Al-Kwifi O, Sussman MS, Stainsby JA, Wright GA.Optimized spiral imaging for measurement of myocardial T2relaxation. Magn Reson Med. 2003;49:1089-1097.

[12] Sprinkart AM, Luetkens JA, Träber F, et al. Gradient Spin Echo(GraSE) imaging for fast myocardial T2 mapping. J CardiovascMagn Reson. 2015;17:12.

[13] Bellm S, Basha TA, Shah RV, et al. Reproducibility of myocar-dial T1 and T2 relaxation time measurement using slice-interleaved T1 and T2 mapping sequences. J Magn Reson Imag-ing. 2016;44:1159-1167.

[14] Messroghli DR, Moon JC, Ferreira VM, et al. Clinical recom-mendations for cardiovascular magnetic resonance mapping ofT1, T2, T2* and extracellular volume: a consensus statement bythe Society for Cardiovascular Magnetic Resonance (SCMR)endorsed by the European Association for Cardiovascular Imag-ing (EACVI). J Cardiovasc Magn Reson. 2017;19:75.

[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. 2015;74:93-105.

[16] Roujol S, Basha TA, Weingärtner S, et al. Impact of motioncorrection on reproducibility and spatial variability of quantitativemyocardial T2 mapping. J Cardiovasc Magn Reson. 2015;17:46.

[17] Koktzoglou I, Li D. Diffusion-prepared segmented steady-statefree precession: application to 3D black-blood cardiovascularmagnetic resonance of the thoracic aorta and carotid artery walls.J Cardiovasc Magn Reson. 2007;9:33-42.

[18] Edelman RR, Chien D, Kim D. Fast selective black blood MRimaging. Radiology. 1991;181:655-660.

[19] Debbins JP, Riederer SJ, Rossman PJ, et al. Cardiac magneticresonance fluoroscopy. Magn Reson Med. 1996;36:588-595.

[20] Nguyen TD, de Rochefort L, Spincemaille P, et al. Effectivemotion-sensitizing magnetization preparation for black bloodmagnetic resonance imaging of the heart. J Magn Reson Imag-ing. 2008;28:1092-1100.

[21] Newbould RD, Owen DR, Shalhoub J, Brown AP, GambarotaG. Motion-sensitized driven equilibrium for blood-suppressedT2* mapping. J Magn Reson Imaging. 2011;34:702-709.

[22] Saranathan M, Savin GS. Multi-slice MR data acquisition in suc-cessive heartbeats with black blood contrast. US Patent7,412,277. Aug 12, 2008.

[23] Fayad ZA, Itskovich VV, Mani V, Szimtenings MM. Rapid mul-tislice black blood double-inversion recovery technique for bloodvessel imaging. US Patent 7,369,887. May 6, 2008.

[24] Kuribayashi H, Tessier JJ, Checkley DR, Wang YX, Hultin L,Waterton JC. Effective blood signal suppression using doubleinversion-recovery and slice reordering for multislice fast spin-echo

12 | Magnetic Resonance in MedicineWANG ET AL.

MRI and its application in simultaneous proton density and T2weighted imaging. J Magn Reson Imaging. 2004;20:881-888.

[25] Simonetti OP, Finn JP, White RD, Laub G, Henry DA. “Blackblood” T2-weighted inversion-recovery MR imaging of theheart. Radiology. 1996;199:49-57.

[26] Felmlee JP, Ehman RL. Spatial presaturation: a method for sup-pressing flow artifacts and improving depiction of vascular anat-omy in MR imaging. Radiology. 1987;164:559-564.

[27] Nayak KS, Rivas PA, Pauly JM, Scott GC, Kerr AB, Hu BS,Nishimura DG. Real-time black-blood MRI using spatial presa-turation. J Magn Reson Imaging. 2001;13:807-812.

[28] Srinivasan S, Hu P, Kissinger KV, et al. Free-breathing 3Dwhole-heart black-blood imaging with motion sensitized drivenequilibrium. J Magn Reson Imaging. 2012;36:379-386.

[29] Wang J, Yarnykh VL, Hatsukami T, Chu B, Balu N, Yuan C.Improved suppression of plaque-mimicking artifacts in black-blood carotid atherosclerosis imaging using a multislice motion-sensitized driven-equilibrium (MSDE) turbo spin-echo (TSE)sequence. Magn Reson Med. 2007;58:973-981.

[30] Nickander J, Lundin M, Abdula G, et al. Blood correctionreduces variability and gender differences in native myocardialT1 values at 1.5 T cardiovascular magnetic resonance–a deriva-tion/validation approach. J Cardiovasc Magn Reson. 2017;19:41.

[31] Captur G, Gatehouse P, Keenan KE, et al. A medical device-grade T1 and ECV phantom for global T1 mapping qualityassurance—the T1 Mapping and ECV Standardization in cardio-vascular magnetic resonance (T1MES) program. J CardiovascMagn Reson. 2016;18:58.

[32] Weingärtner S, Roujol S, Akçakaya M, Basha TA, Nezafat R.Free-breathing multislice native myocardial T1 mapping usingthe slice-interleaved T1 (STONE) sequence. Magn Reson Med.2015;74:115-124.

[33] Dietrich O, Raya JG, Reeder SB, Reiser MF, Schoenberg SO.Measurement of signal-to-noise ratios in MR images: influenceof multichannel coils, parallel imaging, and reconstruction filters.J Magn Reson Imaging. 2007;26:375-385.

[34] Tschabrunn CM, Roujol S, Nezafat R, et al. A swine model ofinfarct-related reentrant ventricular tachycardia: electroanatomic,magnetic resonance, and histopathological characterization.Heart Rhythm. 2016;13:262-273.

[35] Whitaker J, Tschabrunn CM, Jang J, et al. Cardiac MR Charac-terization of Left Ventricular Remodelling in a Swine Model ofInfarct Followed by Reperfusion. J Magn Reson Imaging. 2018.https://doi.org/10.1002/jmri.26005.

[36] Roujol S, Foppa M, Weingartner S, Manning WJ, Nezafat R.Adaptive registration of varying contrast-weighted images forimproved tissue characterization (ARCTIC): application to T1mapping. Magn Reson Med. 2015;73:1469-1482.

[37] Cerqueira MD, Weissman NJ, Dilsizian V, et al. Standardizedmyocardial segmentation and nomenclature for tomographicimaging of the heart. A statement for healthcare professionalsfrom the Cardiac Imaging Committee of the Council on ClinicalCardiology of the American Heart Association. Circulation.2002;105:539-542.

[38] Basha T, Akcakaya M, Roujol S, Nezafat R. Precision andreproducibility of T2 quantifications in myocardial T2 mapping:

impact of the number of echoes and reconstruction model. InProceedings of the 18th Annual SCMR Scientific Sessions,Nice, France, 2015 17(Suppl 1):W9.

[39] Basha TA, Bellm S, Roujol S, Kato S, Nezafat R. Free-breathingslice-interleaved myocardial T2 mapping with slice-selective T2magnetization preparation. Magn Reson Med. 2016;76:555-565.

[40] Kvernby S, Warntjes MJB, Haraldsson H, Carlhäll C-J, EngvallJ, Ebbers T. Simultaneous three-dimensional myocardial T1 andT2 mapping in one breath hold with 3D-QALAS. J CardiovascMagn Reson. 2014;6:102.

[41] Yang HJ, Sharif B, Pang J, et al. Free-breathing, motion-corrected, highly efficient whole heart T2 mapping at 3T withhybrid radial-cartesian trajectory. Magn Reson Med. 2016;75:126-136.

[42] Heeswijk RB, Piccini D, Feliciano H, Hullin R, Schwitter J,Stuber M. Self-navigated isotropic three-dimensional cardiac T2mapping. Magn Reson Med. 2015;73:1549-1554.

[43] Akçakaya M, Weingärtner S, Basha TA, Roujol S, Bellm S,Nezafat R. Joint myocardial T1 and T2 mapping using a combi-nation of saturation recovery and T2-preparation. Magn ResonMed. 2016;76:888-896.

[44] Hsing J, Peters DC, Knowles BR, Manning WJ, Josephson ME.Cardiovascular magnetic resonance imaging of scar developmentfollowing pulmonary vein isolation: a prospective study. PLoSOne. 2014;9:e104844.

[45] Dickfeld T, Kato R, Zviman M, et al. Characterization of acuteand subacute radiofrequency ablation lesions with nonenhancedmagnetic resonance imaging. Heart Rhythm. 2007;4:208-214.

[46] Lardo AC, McVeigh ER, Jumrussirikul P, et al. Visualizationand temporal/spatial characterization of cardiac radiofrequencyablation lesions using magnetic resonance imaging. Circulation.2000;102:698-705.

SUPPORTING INFORMATION

Additional supporting information may be found in theonline version of this article.FIGURE S1 T2 measurements of phantoms using conven-tional bright-blood T2prep-bSSFP (simulated heart rates of60 bpm) and BEATS sequences with different simulatedheart rates ranging from 60 bpm to 110 bpm. T1 values ofeach vial are shown on the left-top, denoted as ReferenceT1. All measurements are displayed with average T2 (topnumber) and standard deviation (bottom number), exceptfor the reference T1 and T2 values.FIGURE S2 Example slices from 4 different patients withdifferent degrees of blood flow related artifacts (redarrows). The residual artifacts could occur on differentlocations; however, it often has a different signal level andrecovery curve compared to the myocardium. Images at theapex show more artifacts at the mid-LV blood cavity (bot-tom right case).FIGURE S3 Example of T2-weighted images and BEATST2 maps in apical, mid-ventricular, and basal slices from a

WANG ET AL.Magnetic Resonance in Medicine | 13

patient with a heart rate of 96 bpm. Arrows show areas ofblood-related artifact.FIGURE S4 Stunned myocardium after a non-ST segmentelevation myocardial infarction (NSTEMI) affecting the leftanterior descending artery territory. The patient was admit-ted with chest pain, a small troponin rise (Troponin I0.12 ng/ml) and unspecific electrocardiographic changes(A). A coronary angiogram showed sequential severe ste-nosis in the left anterior descending artery (B, arrows). Acardiac MRI (C) on day 5 post NSTEMI showed improv-ing left ventricular systolic function (LV ejection fractionof 59% vs. 35% on day 1 echocardiogram) with apical andmid-septal hypo- to dyskinesis. As the left ventricular func-tion fully recovered 2 months later, the subendocardial totransmural late gadolinium enhancement (C, arrows) likelyreflects cardiac edema secondary to stunned myocardium.In this context, septal subendocardial signal intensityappeared increased on short axis T2-weighted images byBEATS (TE5 0, 25, 55 ms; D-F) with the slice locationand orientation depicted by green and orange lines (D).Comparison of T2 maps obtained from conventional brightblood T2prep-bSSFP (G) and BEATS (H) showed differentresults. This could potentially reflect on improved delinea-tion of the subendocardial region on BEATS, blood flow,related artifacts, or both.

FIGURE S5 Example of T2-weighted images and BEATST2 maps acquired in short axis (SHAX), 4-chamber (4CH),horizontal long axis (HLA), and 2-chamber (2CH) views.In the imaging plane where the blood flows within theplane such as the 4CH, HLA, and 2CH views, DIR doesnot effectively null the blood signal.VIDEO S1 Examples of original T2-weighted imagesacquired with BEATS from one patient with heart rates of60 bpm.VIDEO S2 Examples of original T2-weighted imagesacquired with BEATS from one patient with heart rates of70 bpm.VIDEO S3 Example of original T2-weighted images inapical, mid-ventricular, and basal slices from a patient witha heart rate of 96 bpm.VIDEO S4 Example of original T2-weighted images in apatient with suspected large subendocardial enhancementin BEATS T2 maps.

How to cite this article: Wang C, Jang J, Neisius U,et al. Black blood myocardial T2 mapping. Magn.Reson. Med. 2018;00:1–14. https://doi.org/10.1002/mrm.27360

14 | Magnetic Resonance in MedicineWANG ET AL.