Spatial Modulation of Magnetization for Cardiac Tagging Using Sinc Modulated RF Pulse Train

MRI myocardial tagging offers a tremendous opportunity to quantitatively, serially and noninvasively examine the intramyocardial motion and deformation [l-51. The ideal presaturation or tagging stripe should be of rectangular intensity profile and controllable width with respect to the distance between adjacent stripes. Here we propose a sinc modulated RF pulse train that offers sharper-edged tags and flexibility to change the ratio of tag width to tag separation. The technique was demonstrated with both phantom and in vivo heart study. INTRODUCTION DANTE and SPAMM tagging techniques are highly efficient. However, DANTE tagging that consists of a finite number of RF pulses of uniform amplitude yields presaturation strips of sinc profile [5]. SPAMM produces sinusoidal modulation of magnetization [2] and high-order SPAMM improves the presaturation profile by binomial modulation of the RF pulse train [3], yet they do not provide the ability to alter the ratio of tag width to tag separation. METHODS The 1st-order approximation of presaturated magnetization profile produced by a RF pulse train in presence of constant gradient G is the Fourier transform of the RF pulse train waveform. A sinc modulated FW pulse train of finite duration (Fig.1) can be written as rf(t) = comb(t, At2) x sinc(2nt / Atl) where comb(t, At2) is the comb function of spacing At2 and sinc(2ntl Atl) the sin, function with bandwidth l/Atl. Thus the spatial modulation of magnetization can be approximated as M,,,(x) = comb{x, l/(yGAt2)} C3 rect{x/(l/yGAtl)}, which provides a tag width ll(yGAt2) and separation l/yGAtl. Fig.1 illustrates some sinc modulated RF pulse trains and corresponding magnetization saturation profiles. RI' Saturation Profile Fig.1 Sinc modulated RF pulse trains and approximate magnetization saturation profiles in presence of constant gradient G. This approach has the advantage of (1) producing better tag profile and (2) controlling the ratio of tag width to tag separation by At2:Atl ratio. It can also be extended to 2D and 3D tagging. Note sinc modulated RF pulse train can he numerically optimized by simulating Block equations. To tag large FOV with very small tag separation, as in DANTE tagging, the finite length of each individual pulse should be small in order to minimize its side-effect, i.e., sinc modulation across FOV. This may impose high RF peak power requirements for human study. To improve the utility for practical implementation, the constant gradient G can be replaced by gradient segments between RF pulses without affecting the saturation profile, thus allowing use of (a) longer pulse duration to reduce peak power requirement and (b) equalized RF pulse amplitudes to minimize the total pulse train duration. RESULTS This tagging method was implemented on a 9.4T Bruker WB400 microimaging system and demonstrated on both phantom and mouse heart. Fig.2 shows the water phantom study. Images were obtained with 2D GE sequence with TWTE=300ms/3ms, flip angle 30°, 256x256 matrix. 1 NEX. 22 mm FOV. and 2 mm slice thickness. h C -.... ..... . .. . . ,. ,. , .. ?A...A_. . c . \ v ~ . . ... .,.,