Myocardial Approximate Spin-lock Dispersion Mapping using a Simultaneous T2 and TRAFF2 Mapping at 3T MRI

Ischemic heart disease (IHD) is one of the leading causes of death worldwide. Myocardial infarction (MI) represents a third of all IHD cases, and cardiac magnetic resonance imaging (MRI) is often used to assess its damage to myocardial viability. Late gadolinium enhancement (LGE) is the current gold standard, but the use of gadolinium-based agents limits the clinical applicability in some patients. Spin-lock (SL) dispersion has recently been proposed as a promising non-contrast biomarker for the assessment of MI. However, at 3T, the required range of SL preparations acquired at different amplitudes suffers from specific absorption rate (SAR) limitations and off-resonance artifacts. Relaxation Along a Fictitious Field (RAFF) is an alternative to SL preparations with lower SAR requirements, while still sampling relaxation in the rotating frame. In this study, a single breath-hold simultaneous TRAFF2 and T2 mapping sequence is proposed for SL dispersion mapping at 3T. Excellent reproducibility (coefficient of variations lower than 10%) was achieved in phantom experiments, indicating good intrascan repeatability. The average myocardial TRAFF2, T2, and SL dispersion obtained with the proposed sequence (68.0±10.7 ms, 44.0±4.0 ms, and 0.4±0.2 × 10−4 s2, respectively) were comparable to the reference methods (62.7±11.7 ms, 41.2±2.4 ms, and 0.3±0.2 × 10−4 s2, respectively). High visual map quality, free of B0 and B1+ related artifacts, for T2, TRAFF2, and SL dispersion maps were obtained in phantoms and in vivo, suggesting promise in clinical use at 3T.


I. INTRODUCTION
Magnetic resonance imaging (MRI) established itself as an important modality in the clinical and preclinical evaluation of ischemic heart disease (IHD) [1].Currently, the gold standard for the detection of chronic or irreversible injury of the myocardium is late gadolinium enhancement (LGE) MRI.However, LGE requires the use of gadoliniumbased contrast agents (GBCA).GBCAs pose the risk of deposition in the brain and nephrogenic systemic fibrosis (NSF), and are, therefore, contraindicated in patients with renal dysfunction [2].With the emergence of quantitative cardiac MR methods, quantitative myocardial tissue characterization has been proposed as a non-contrast alternative for scar assessment.For example, T 2 relaxation time mapping has been implemented to detect edema in acute myocardial infarction (MI) [3].However, detection of fibrosis and permanent tissue damage has so far remained out of reach for conventional relaxometry.Recently, rotating frame relaxometry, e.g.T 1ρ mapping [4], has revealed new insights about the macromolecular content of biological tissues and has been used to quantify myocardial fibrosis without the injection of contrast agents.However, similar to conventional relaxometry, T 1ρ is not specific and is affected by numerous underlying physical effects.
Spin-lock (SL) dispersion has recently been proposed as a new imaging biomarker for the detection of MI [5].SL dispersion has been shown to isolate chemical exchange contributions to SL relaxation, potentially presenting a more specific imaging biomarker [6].SL dispersion is usually obtained by acquiring a range of T 1ρ maps at different amplitudes (frequencies).However, at 3T, the specific absorption rate (SAR) limits the use of high-frequency SL pulses, and long SL preparations suffer from strong off-resonanceinduced artifacts.Adiabatic T 1ρ may be used to avoid off-resonance effects, but its use in dispersion mapping is hampered because variable effective field strength and sweep duration complicate the relation to a well-defined SL frequency.Relaxation Along a Fictitious Field (RAFF) is an alternative to SL pulses with lower SAR requirements [7].RAFF operates in a sub-adiabatic regime with constant effective field strength and an identical, constant fictitious field strength leading to uniform sweeps.In previous studies, RAFF has been shown to preserve excellent sensitivity to acute MI and chronic scar [8], but its use for dispersion mapping remains unstudied.
In this work, we sought to provide a method for the assessment of myocardial SL dispersion at 3T using RAFF in a single breath-hold (BH).The proposed technique is evaluated in phantom measurements and initial in vivo images of a healthy subject.

A. Sequence Design
In the proposed simultaneous T RAFF2 and T 2 mapping sequence, five balanced steady-state free precession (bSSFP) images with different magnetization preparations were acquired during a single 16s BH (Fig. 1).
A first image acquired without preparation is followed by images prepared with two RAFF2 preparation blocks [7] (T RAFF2p ) with different durations and one T 2 preparation (T 2p ) duration, each preceded by a 4 s rest period.A saturation-prepared image was acquired in the subsequent heartbeat to capture the effect of imaging pulses on the magnetization curve.Simultaneous T RAFF2 and T 2 maps were obtained using a four-parameter fit: SL dispersion maps were calculated by the difference between T 2 (zero-frequency SL) and T RAFF2 (approximation for SL relaxation) times normalized by the RAFF frequency as the assumed SL frequency.
All imaging was performed on a 3T MRI scanner (Prisma, Siemens Healthineers, Erlangen, Germany) with a 28-channel receiver coil array.This study was approved by the local institutional review board and written informed consent was obtained before each examination.

B. Phantom experiments
Phantom measurements were performed on the T1MES phantom [9] to evaluate the accuracy and precision of the proposed sequence against regular T RAFF2 and T 2 mapping.Simultaneous T RAFF2 and T 2 maps were obtained using the proposed sequence with T RAFF2p times of 12.9 and 25.7 ms (RF pulse power of 625 Hz) and T 2p of 50 ms.The SL dispersion map obtained with the proposed sequence was achieved following Eq. 1.
A reference T RAFF2 map was obtained using T RAFF2p = 0, 12.9, 25.7, and 38.6 ms and an image preceded by a saturation pulse.Reference T 2 map was acquired with T 2p of 0, 50, 50, 50 ms, followed by an image preceded by a saturation pulse [10].Reference SL dispersion map was calculated using the individual reference T 2 and T RAFF2 maps.The imaging parameters were kept identical to those of the proposed method.
Correlation and Bland-Altman plots were obtained to compare the simultaneous and the individual T RAFF2 , T 2 , and SL dispersion values.
To assess intrascan reproducibility, three scan repetitions were performed in separate acquisitions in the phantom experiments, and the coefficient of variation (CV) across all measurements was computed.

C. In vivo experiments
In vivo T RAFF2 and T 2 maps were obtained using the proposed sequence for a healthy subject (male, 35 years) with a BH of 16 s.The T RAFF2p and T 2p preparation times were identical to the sequence used in the phantom experiments.All baseline in vivo images were registered using groupwise registration [11] to minimize motion within and across the separate BH and the SL dispersion map was computed (Eq.1).Standard deviation (SD) maps were obtained from the fit residuals to obtain a spatially resolved estimation of the precision of the proposed sequence [12].In two separate 16 BH acquisitions, individual T RAFF2 and T 2 reference maps were obtained using the same sequence parameters as in the phantom experiments.Reference SL dispersion map was obtained from the individual single-parameter reference maps.Visual image quality and mean (± std) myocardial quantitative values were compared between the proposed sequence and the individual reference maps for T RAFF2 , T 2 , and SL dispersion.

A. Phantom
Joint T RAFF2 and T 2 values obtained with the proposed sequence exhibited excellent correlation (R 2 = 1.00) with individually acquired reference T RAFF2 and T 2 values for the phantom vials (Fig. 2).A bias of −7.7ms was measured for T RAFF2 , 0.8ms for T 2 , and −14ms 2 for SL dispersion.The limits of agreement were ±17.4ms (T RAFF2 ), ±5.4ms (T 2 ), and ±34ms 2 (SL dispersion).
SL dispersion maps yield good sensitivity to changes in the vials of the T1MES phantom (range 0.7 -3.3 × 10 −4 s 2 ).The three imaged biomarkers showed a coefficient of variation of less than 10% between repeated scans, indicating good short-term reproducibility (Fig. 3).

B. In vivo
In vivo T RAFF2 and T 2 maps obtained with the proposed sequence achieved high visual map quality, and no B 1 + or B 0 related artifacts were visible (Fig. 4).The average (± std) myocardial T RAFF2 and T 2 obtained with the proposed sequence (68.0 ± 10.7 and 44.0 ± 4.0 ms, respectively) were comparable to the reference methods (62.7 ± 11.7 and 41.2 ± 2.4 ms, respectively).The myocardial SL dispersion map resulted in an excellent depiction of the myocardium with a small increase in variability when compared to the SL dispersion obtained with the single-parameter reference maps (0.4 ± 0.2 and 0.3 ± 0.2 × 10 −4 s 2 , respectively).

IV. DISCUSSION
In this work, we propose a method for approximate SL dispersion mapping in the heart by measuring T RAFF2 and T 2 in a single 16s BH at 3T. Homogeneous T RAFF2 , T 2 , and SL dispersion maps of high visual quality were obtained for phantom and in vivo.
Our preliminary phantom results showed good sensitivity for a wide range of expected myocardial values.In a healthy volunteer, a clear depiction of the myocardium with no visible B 1 + or B 0 artifacts was achieved.
The proposed sequence design includes a saturation-prepared image, in addition to the RAFF2-and T 2 -prepared images.This allows the use of a four-parameter fit model, which has previously been shown to capture the effects of the image readout and, therefore, enables a more reproducible mapping across different protocol parameters [10].However, the model assumes a negligible asymptotic magnetization for the exponential decay functions.This assumption has previously been used for RAFF decay.However, experimental validation may be warranted to ascertain bias-free mapping with the proposed sequence.
Similar to conventional relaxation times, T 1ρ relaxation times are driven by several factors, including diffusion, chemical exchange, J-coupling, and spin-spin interaction.Alternatively, SL dispersion has been shown to provide a more specific measure of exchange effects by separating the relaxation rate, exchange rates, and frequency offsets of exchanging pools.Thus, SL dispersion may offer a valuable tool to increase specificity in the detection of pathological remodelling.Therefore, clinical use in ischemic and non-ischemic heart disease will be the subject of further investigation.
Previous works have demonstrated a good sensitivity of T 1ρ to fibrosis in the liver [13], however, limited studies have been performed for in vivo myocardial fibrosis, especially at high fields.This can be attributed to the high specific absorption rate (SAR) of the multiple T 1ρ RF pulse amplitude required in a conventional SL dispersion mapping.RAFF offers an attractive alternative to continuous wave SL due to lower SAR burden (30% reduction [7]) and better resilience to off-resonances.At the same time, RAFF may be more reflective of a single SL frequency component compared to adiabatic T 1ρ .Since T 1ρ relaxation depends on the orientation of the magnetization and the effective field, as well as the effective field strength, T 1ρ relaxation is non-constant throughout adiabatic T 1ρ preparation pulses.In RAFF preparations, the effective field strength, as well as the fictitious field components, giving rise to the sweep, are kept constant and identical throughout the preparation.Thus, the proposed simultaneous T RAFF2 and T 2 mapping sequence is an excellent candidate for the clinical application of SL dispersion rate due to the lower requirements in SAR while still approximating exchange rate information.The proposed approach is readily available for use in the clinical environment, and future work will include the evaluation of the proposed approximate SL dispersion method for the detection of myocardial infarction in a large cohort of patients.

V. CONCLUSION
Our results show promising potential for mapping approximate myocardial spin-lock dispersion in a single BH using T RAFF2 mapping with clinically tolerable SAR.The proposed sequence obtains simultaneous T RAFF2 and T 2 maps and yields high visual image quality, homogeneous signal, no B 1 + or B 0 artifacts, and excellent reproducibility.Thus, approximate spin-lock dispersion mapping with T RAFF2 bears great promise for clinical sensitivity to pathological remodeling and is an excellent candidate for the non-contrast assessment of scar with cardiac MRI.

Clinical relevance-
T RAFF2 and T 2 imaging promises non-contrast assessment of scar and focal fibrosis in a single breath-hold using approximate spin-lock dispersion mapping.Coefficient of variation (CV) across three measurements using the proposed simultaneous T RAFF2 and T 2 mapping sequence.All vials show a CV lower than 10% demonstrating excellent reproducibility of the proposed sequence.Standard deviation (σ) maps computed from the fit residuals to obtain a spatially resolved estimation of the T RAFF2 , T 2 , and spin-lock (SL) dispersion precision.

Fig. 1 .
Fig. 1.A) Pulse sequence diagram for the proposed simultaneous myocardial T RAFF2 and T 2 mapping sequence.B) Schematic representation of the amplitude of the RF pulses.C) The difference between the T RAFF2 and T 2 map divided by the RAFF2 pulse amplitude results in an approximation for a spin-lock (SL) dispersion map.

Fig. 2 .
Fig. 2.A) T 2 , T RAFF2 , and spin-lock (SL) dispersion maps for the T1MES phantom using the proposed simultaneous sequence (top row) and the single-parameter reference maps (bottom row).B) Correlation and Bland-Altman plots for each parameter, showing an excellent correlation between the relaxation properties in phantom, in the expected in vivo range.

Fig. 4 .
Fig. 4.A) Baseline images for the proposed simultaneous T RAFF2 and T 2 mapping sequence with the corresponding preparation times.B) In vivo myocardial T RAFF2 , T 2 , and spin-lock (SL) dispersion maps were obtained with the proposed sequence and with the single-parameter sequences (Reference).Following image registration, the dispersion map is obtained by subtracting T 2 from T RAFF2 and dividing by the RAFF2 pulse power.All maps present high visual image quality, homogeneous myocardial signal and no B 1 + or B 0 artifacts are visible.