Summary of Main Findings

RavFa-T₂ combines a transient SSFP sequence and a variable flip angle scheme for myocardial T₂ quantification at a rapid scan time (≤4 seconds). It showed accurate T₂ quantification and lower T₂ precision than T₂-Prep in a phantom study. The precision of RavFa-T₂ in-vivo was higher than T₂-Prep.

Introduction

Transverse relaxation times (T₂) have long been used for tissue characterization of the myocardium in cardiovascular magnetic resonance imaging. Increased myocardial T₂ values have been shown to correlate with myocarditis ¹ and edema. ² A conventional approach for assessing myocardial T₂ is based on a method called T₂-Prep with different delays of a single-shot steady-state free precession (SSFP) imaging sequence. ³ This technique, however, requires a long breath-hold acquisition time of 12 seconds, which can be demanding for very young or ill cardiac patients. To reduce breath-hold time, we developed a new method to quantify T₂ of the myocardium in ≤4 seconds.

Methods

The proposed RavFa-T₂ pulse sequence combines transient SSFP imaging,^4,5,6 and a variable flip angle scheme. ^7,8 Compared to T₂-Prep (Figure 1A), four images are acquired with variable flip angles ($$$\alpha_i= 30°, 67°, 103°, 140°$$$ ) at a specific cardiac phase in 4 heartbeats (Figure 1B). Assuming heartrates between 60-120 bpm, the acquisition time is 2-4 seconds. To account for partially recovered magnetization from previous images, we introduce an additional, iterative term $$$\delta_i$$$ to the transient phase SSFP relaxation model⁶:
$$$M_{RavFa} (\alpha_i) = \delta_i \times M_{trans}(\alpha_i)$$$,

where $$$M_{trans}(\alpha_i)=\left( M_0sin(\frac{\alpha_i}{2})-M_{SS}(\alpha_i)\right) \lambda(\alpha_i)^n+M_{SS}(\alpha_i)$$$

with $$$ M_{SS}(\alpha_i) = M_0 \frac{\sqrt{E_2}(1-E_1)sin(\alpha_i)}{1-(E_1-E_2)cos(\alpha_i)-E_1E_2} $$$, $$$\lambda(\alpha_i)=E_1cos^2(\frac{\alpha_i}{2})+E_2sin^2(\frac{\alpha_i}{2})$$$, $$$E_{1/2}=exp(-\frac{T_R}{T_{1/2}})$$$,

and $$$\delta_1 = M_0$$$, $$$\delta_i=M_0-(M_0-M_{RavFa}(\alpha_{i-1})cot(\frac{\alpha_{i-1}}{2}))exp(-\frac{t}{T_1})$$$
where, t describes the time between heartbeats, T_R is the repetition time, and n was fitted to match the underlying magnetization. To reduce the number of unknown parameters, we normalize the pixel intensities by dividing them by the corresponding pixel of the last image. This eliminates the term for proton density M₀.

We systematically investigated the feasibility of the novel RavFa-T₂ approach with a phantom and a two-center in-vivo study. The phantom study was performed to compare the proposed sequence to T₂ of the gold-standard spin echo (SE) (T_R/T_E: 10000/32*10ms, α: 90°, total acquisition time (TAT): 4920 seconds), and T₂-Prep (T_R/T_E: 2.65/1.33ms, α: 35°, rest period: 2 heartbeats, T₂-Prep delay: 0, 25, 50, and 75ms, TAT: ≤12s). The experiments were performed on the T1MES phantom ⁹ at 4 different heartrates (60, 80, 100, 120 bpm) with a field-of-view (FOV): 200 (RL) × 200 (AP) × 10 (FH) mm, CS-SENSE 3, and 10 startup pulses for all methods. We performed a two-center in-vivo feasibility study in 44 patients using the same imaging parameters as in the phantom study. All scans were performed on 1.5T Philips scanners. The quantification algorithms were implemented in MATLAB (MathWorks Inc., Natick, MA, USA). Method evaluation was based on accuracy (mean error compared to SE) and precision (pixel-wise standard deviation within region-of-interest). A two-tailed paired student t-test was used for statistical analysis and a p-value≤0.05 was assumed statistically significant.

Results

Figure 2 shows the T₂ maps of the phantom in all sequences. For the phantom study, the gold-standard SE yielded the following ranges for T₂ = [48-235ms]. T₂ quantification with T₂-Prep demonstrated an accuracy/precision of 12.3ms/3.9ms compared to SE. The comparison of RavFa-T₂ to SE resulted in an accuracy/precision of -9.4ms/12.5ms (Table 1). Forty-four patients at institution 1 (n= 22, 14 female, median age 28.5 [11-60] years old) and institution 2 (n = 22, 10 female, median age 19 [6-41] years old) successfully completed T₂-prep and RavFa-T₂ acquisitions in a mid-ventricular short-axis slice. Figure 3 demonstrates the T₂ maps of 2 patients calculated with T₂-Prep and RavFa-T₂. For the two-center in-vivo study, the acquisition time of RavFa-T₂ was consistently shorter than T₂-Prep (Table 2, p-value <0.001) and the precision of RavFa-T₂ was better than T₂-Prep (Table 2, all p-values <0.001).

Discussion

RavFa-T₂ allows rapid myocardial T₂ quantification in 4 heartbeats (≤4 seconds), and its acquisition time is at least 3 times faster than that of T₂-Prep. In the phantom study, RavFa-T₂ shows higher accuracy and lower precision compared to T₂-Prep. However, the precision of RavFa-T₂ is significantly better than T₂-Prep in in-vivo study partially due to the shorter acquisition time of RavFa-T_2. Compared to 12 second T₂-Prep acquisitions, patients were able to completely hold their breath during the 4 second RavFa-T₂ acquisitions, leading to a minimum drift in respiratory motion and better precision in T₂ estimation.

Conclusion

We demonstrate technical feasibility of myocardial T₂ mapping with RavFa-T₂. The estimation of myocardial T₂ in the phantom results in accurate values with acceptable precision compared to T₂-Prep, whereas in in-vivo study, results in better precession compared to T₂-Prep. Future studies aim at extending RavFa-T₂ method for simultaneously estimation of T₁ and T₂ of myocardium.

Acknowledgements

This project was supported by the Harvard Scholarship Foundation Germany e.V.

References

1.	Spieker M, Haberkorn S, Gastl M, et al. Abnormal T2 mapping cardiovascular magnetic resonance correlates with adverse clinical outcome in patients with suspected acute myocarditis. Journal of Cardiovascular Magnetic Resonance. December 2017;19.
2.	Nishii T, Kono AK, Shigeru M, et al. Cardiovascular magnetic resonance T2 mapping can detect myocardial edema in idiopathic dilated cardiomyopathy. The International Journal of Cardiovascular Imaging. April 2014;30.
3.	Giri S, Chung YC, Merchant A, et al. T2 quantification for improved detection of myocardial edema. Journal of Cardiovascular Magnetic Resonance. December 2009;11:56.
4.	Oppelt A, Graumann R, Barfuss H, Fischer H, Hartl W, Schajor W. FISP: eine neue schnelle Pulssequenz fuer die Kernspintomographie. Electromedica. 1986;54(1):15-18.
5.	Bieri O, Scheffler K. Fundamentals of balanced steady state free precession. Journal of Magnetic Resonance Imaging. 2013;38:2-11.
6.	Scheffler K. On the Transient Phase of Balanced SSFP Sequences. Magnetic Resonance in Medicine. April 2003;49(4):781-783.
7.	Worters P, Hargreaves B. Balanced SSFP transient imaging using variable flip angles for a predefined signal profile. Magnetic Resonance in Medicine. November 2010;64(5):1404-1412.
8.	Srinivasan S, Ennis DB. Variable flip angle balanced steady-state free precession for lower SAR or higher contrast cardiac cine imaging. Magnetic Resonance in Medicine. March 2014;71(3):1035-1043.
9.	Captur G, Captur G, Gatehouse PD, et al. A T1 and ECV phantom for global T1 mapping quality assurance: The T1 mapping and ECV standardisation in CMR (T1MES) program. Journal of Cardiovascular Magnetic Resonance. 2016;18(1):1-3.