Synopsis

We propose a new 3D T1 mapping technique (CS 3D T1 mapping) to acquire the entire left ventricle in a single breath-hold. The technique combines flexible saturation time sampling, sharing the saturation pulses between different RR intervals and compressed SENSE.

The proposed sequence successfully acquired a 3D-T1 map of 12 slices in a single breath-hold (14 heart beats) at 3T, achieving T1 values in good agreement with the T1 values estimated by the IR-SE sequence. Additionally, myocardium T1 values obtained with the proposed technique are similar to those already published in literature for 3T.

Introduction

Cardiac tissue characterization by T1 mapping has been widely described as important for clinical decision making in cardiovascular diseases¹. Recently, different 3D T1 mapping sequences have been proposed to cover the entire left ventricle. Most of these 3D techniques are based on saturation recovery in order to avoid T1 underestimation from partial recovery after subsequence inversion pulses. The SASHA technique acquires a proton density (PD) image and several saturation recovery images at different saturation times without recovering beats². More flexible sampling schemes share the saturation pulse between several RR beats and acquire images with longer saturation times³.

3D sequences based on saturation recovery have been developed in order to fully cover the left ventricle with high accuracy T1 estimation^4,5. However, these 3D saturation recovery sequences are free-breathing and demand a longer acquisition time. Compressed SENSE exploits image sparsity and compressibility for k-space and accelerates acquisition by undersampling without significant image degradation⁶. This method significantly reduces scan time, particularly in 3D acquisitions.

We propose a new 3D T1 mapping technique (CS 3D T1 mapping) to acquire the entire left ventricle in a single breath-hold. The technique combines flexible saturation time sampling, sharing the saturation pulses between different RR intervals and compressed SENSE.

Methods

The proposed sequence is formed by three distinct blocks (Fig1). The sequence starts with a PD image to avoid T1 effects from prior saturation pulses (first block). Full signal recovery is ensured by waiting 5 RR intervals between each PD readout. The second block consists of images acquired at TS1 and TS1+RR interval and the third block of images acquired at TS2 and TS2+RR interval.

All studies were performed on a Philips Achieva 3T-Tx using a 32 channel cardiac coil.

The proposed sequence was based on spoiled saturation recovery TFE sequence (TR/TE/FA = 2.8ms/1.32ms/5^o) with echo train length of 76, covering 60 mm in slice direction with slice thickness of 10 mm reconstructed to 5 mm (12 slices). In plane resolution was 2.0x2.0 mm, covering a FOV of 322x322 mm. Full image acquisition was accelerated using a spatial domain compressed SENSE factor of 3.4 in order to acquire a whole 3D volume in 2 independent TFE shots, resulting in a total scan time of 14 RR intervals. The compressed SENSE reconstruction was performed in real time on the scanner. The same sequence was used for phantom and in vivo data collection.

A phantom of 9 tubes with different T1s (180 ms–1900 ms) was built to validate T1 values obtained by the proposed technique. The phantom was scanned with the proposed sequence with a simulated RR interval of 1s (HR= 60 bpm). Reference T1 values were obtained from the same phantom using an inversion-recovery spin-echo (IR-SE) sequence with repetition time of 15s and 15 inversion times ranging 100-3500 ms.

In-vivo data were acquired from a healthy pig with an RR interval of 0.780s (HR= 77 bpm), resulting in a total scan time of 11 seconds.

T1 values were estimated by non-linear least square fitting following Bloch equations modelling.

Passing-Bablok regression analysis was applied to study the T1 values obtained from the phantom experiment. For the in-vivo experiment, a region of interest was placed in the septum at different levels as well as in the blood pool.

Results

Passing-Bablok regression analysis (Fig2) provides a slope of 0.99 (CI=0.97-1.01) and an intercept of 11.18 (CI=-2.14-24.64) showing no significant bias compared with the IR-SE results.

In-vivo middle slice images from PD and the 4 saturation times are presented in Fig3 while T1 maps for all 12 slices are presented in Fig4. The septal T1 values for basal, middle and apical slices were respectively: 1477±41 ms, 1450±59 ms and 1492±54 ms. A blood T1 of 1936 ± 99 ms was measured at slice 6.

Discussion

Conclusion

Acknowledgements

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement N722427.

This study forms part of a Master Research Agreement between CNIC and Philips Healthcare.

References

1. Haaf P, Garg P, Messroghli DR, Broadbent DA, Greenwood JP, Plein S. Cardiac T1 Mapping and Extracellular Volume (ECV) in clinical practice: a comprehensive review. J Cardiovasc Magn Reson. 2016;18:89.

2. Chow K, Flewitt JA, Green JD, Pagano JJ, Friedrich MG, Thompson RB. Saturation recovery single-shot acquisition (SASHA) for myocardial T(1) mapping. Magn Reson Med. 2014;71:2082–2095.

3. T. Song, J.A. Stainsby, V.B. Ho, M.N. Hood, G.S. Slavin. Flexible cardiac T1 mapping using a modified Look-Locker acquisition with saturation recovery. Magn Reson Med. 2012;67:622-627

4. Nordio G, Henningsson M, Chiribiri A, Villa ADM, Schneider T, Botnar RM. 3D myocardial T1 mapping using saturation recovery. J Magn Reson Imaging. 2017;46(1): 218–227

5. Rui Guo, Zhensen Chen, Yishi Wang, Daniel A. Herzka, Jianwen Luo and Haiyan Ding. Three-dimensional free breathing whole heart cardiovascular magnetic resonance T1 mapping at 3 T. J Cardiovasc Magn Reson. 2018;20:64

6. M. Lustig, D. Donoho, and John M. Pauly. Sparse MRI: The Application of Compressed Sensing for Rapid MR Imaging. Magn Reson Med. 2007; 58:1182–1195.