Black-blood T1 mapping at 3T: Reduced partial-voluming using adiabatic MSDE preparation

Background Myocardial T1 mapping in pathologies with decreased myocardial wall thickness such as dilated cardiomyopathy (DCM) is strongly impaired by partial-voluming from the neighboring blood pools [Kellman et al., JCMR2014]. Significant differences between the T1 times in myocardium and blood lead to decreased accuracy in the presence of partial-voluming. This causes sensitivity to the region-of-interest (ROI), compromising the inter-observer reproducibility. The aim of this work is to study the use of blood-signal suppression using a motion-sensitized driven equilibrium (MSDE) [Wang et al., MRM2007] magnetization preparation in order to reduce partial-voluming in myocardial T1 mapping.


Background
Myocardial T 1 mapping in pathologies with decreased myocardial wall thickness such as dilated cardiomyopathy (DCM) is strongly impaired by partial-voluming from the neighboring blood pools [Kellman et al., JCMR2014].
Significant differences between the T 1 times in myocardium and blood lead to decreased accuracy in the presence of partial-voluming. This causes sensitivity to the region-of-interest (ROI), compromising the inter-observer reproducibility.
The aim of this work is to study the use of blood-signal suppression using a motion-sensitized driven equilibrium (MSDE) [Wang et al., MRM2007] magnetization preparation in order to reduce partial-voluming in myocardial T 1 mapping.

Methods
An adiabatic MSDE preparation module was added directly before the imaging pulses of a SAPPHIRE sequence [Weingärtner et al., MRM2014] (Fig. 1). The preparation consists of a rectangular tip-down pulse, an adiabatic BIREF1 refocusing pulse, a composite tip-up pulse and motion-sensitizing gradients before and after refocusing. The MSDE parameters were TE MSDE = 11 ms, gradients: amplitude = 16 mT/m, duration = 2 ms. 6 healthy volunteers (25 ± 6 y; 4 M) were scanned using conventional and black-blood T 1 mapping on a 3T MR Scanner (Siemens Skyra). T 1 mapping was performed using a bSSFP imaging readout with the following parameters: TE/TR/α = 1.0 ms/2.9 ms/35°, FOV/res = 440 × 375 mm²@1.7 × 1.7 mm², sl.th. = 8 mm, GRAPPA = 2, Partial-Fourier = 6/8, bw = 1085 Hz/px. A three parameter model was used for T 1 fitting, avoiding potential quantification inaccuracies caused by the recovery curve modulation through the MSDE preparation. T 1 times, the average thickness and the apparent in-plane area of the myocardium were quantified in the T 1 maps using manually drawn ROIs. Furthermore, cross myocardial T 1 times were analyzed from the endo-to the epicardial border.

Results
Visually strong blood suppression was achieved using the adiabatic MSDE preparation (Fig. 2a). Quantitative analysis reveals increased T 1 times towards the myocardial borders in conventional T 1 mapping (Figure 2c), while consistent T 1 times through the entire myocardial thickness were measured using black-blood SAPPHIRE. No significant difference was found in the average T 1 time of the two methods (Conv.: 1574 ± 52 ms vs BB: 1593 ± 47 ms). A 25%-28% gain in apparent in-slice area of the myocardium and average wall-thickness in the T 1 maps was achieved using blood-suppression (BB: 1596 ± 266 mm 2 , 7.37 ± 1.16 mm vs. Conv.: 1278 ± 213 mm 2 , 5.72 ± 0.87 mm, p < 0.05).

Conclusions
An adiabatic MSDE preparation enables robust myocardial T 1 Mapping at 3T. The apparent myocardial in-slice area and average wall-thickness is significantly increased using a black-blood preparation. Furthermore, elevated T 1 times at the myocardial borders were eliminated. This reduces sensitivity to ROI placement and potentially benefits the reproducibility of myocardial T 1 mapping, especially in the presence of pathologies with reduced myocardial wallthickness.  An adiabatic MSDE preparation is inserted directly before the imaging pulses. In MSDE, blood-signal suppression is caused by symmetric dephasing gradients before and after a refocusing pulse, causing incomplete refocusing of moving tissue and greatly increasing the blood/myocardium contrast. b) Simulated magnetization signal at the various inversion times of a SAPPHIRE sequence. High signal is observed in conventional T 1 mapping from the blood pools despite the long T 1 time. Black-blood SAPPHIRE shows almost complete suppression of the blood signal for the trade-off against a slightly decreased dynamic range in the myocardial signal.
Figure 2 a) T 1 -weighted images of a healthy volunteer with and without MSDE magnetization preparation. Visually strong and homogenous suppression of the blood-signal can be observed. b) Corresponding T 1 quantification in the myocardium using conventional and black-blood T 1 mapping. Visually homogenous T 1 times are observed around the myocardium using both methods. c) T 1 times through the myocardial thickness analyzed separately for the anterolateral, inferolateral and septal part of the myocardium. Conventional T 1 mapping shows strongly elevated T 1 times at the endo-and/or the epicardial border. No such elevation is observed with black-blood T 1 mapping.