Proc.฀Intl.฀Soc.฀Mag.฀Reson.฀Med฀9฀(2001) 1905 A Fast, Multiple-Slice, Multiple-Point Method for the Determination of T1 Relaxation Time for use in Quantitative Myocardial Perfusion Measurement John P Ridgway 1 , Aleksandra Radjenovic 1 , David M Higgins 1 , Andrea Kassner 2 , Mohan Sivananthan 3 1 Department of Medical Physics, Leeds General Infirmary, Gt George Street, Leeds, UK; 2 Philips Medical Systems, 7 Killingbeck Drive, Leeds, UK; 3 Cardiac MRI Unit, Leeds General Infirmary, Leeds, UK; Introduction Quantitative measurement of myocardial perfusion based on the first pass of an extracellular contrast agent through the heart relies on the determination of contrast agent concentration during the dynamic phase. Methods th at are based on the alteration of the T1 relaxation time require knowledge of the initial T1 value of the myocardium and its subsequent variation due to the passage of the contrast agent through the myocardium. Previous methods for fast myocardial T1 measurement have been proposed [1,2] however they either only measure a single slice or the data acquisition times and slice locations differ from those of the dynamically acquired data set. This limits their use in post-processing approaches that use pixel map ping. We propose a multiple slice, multiple-point T1 measurement that is both fast and whose slice locations and data acquisition timing exactly match those of the dynamic perfusion measurement pulse sequence. Methods A multiple-slice, myocardial pe rfusion imaging pulse sequence [3,4] using a gradient-echo technique with a segmented echoplanar readout (TFE-EPI) has been implemented on a 1.5 Tesla GYROSCAN ACS NT clinical MR imaging system (Philips Medical Systems, Best, The Netherlands) with a gradient performance of maximum amplitude 30mT/m and slew rate 150mT/m/ms. T1 contrast is achieved by the use of a saturation pulse applied after the R wave. This also makes the signal intensity insensitive to different heart rates and arrhythmias [5]. Six slices are acquired in total with three slices acquired on alternate heart beats (slice order 1,5,3; 2,6,4) providing a temporal resolution of 2 heartbeats. The total acquisition time for each slice is 90ms (9 RF pulses TR 8.7ms, TE 3.7ms, α=30°, 5 EPI readouts, 'halfscan' factor of 0.71, 128x96 matrix, 6mm slice thickness, 288mm x 193mm Field of View). A linear interleaved k -space trajectory is used with ky = 0 occuring at the echo time following the fifth α pulse. Four presaturation regions are applied prior to each slice acquisition to prevent aliasing from outside the field of view. For the dynamic perfusion sequence the saturation recovery time, T SAT , is 160ms for slices 1 and 2. For the multiple-point T1 measurement the above pulse sequence is repeated using four different saturation recovery times, (T SAT = 60, 90, 120, 160 ms for slices 1 and 2; 150, 180, 210, 250ms for slices 5 and 6; 240, 270, 300, 340ms for slices 3 and 4) plus a further acquisition with no saturation pulse applied (T SAT =∞). The Cardiac trigger delay for each slice is kept constant as TSAT is varied to allow pixel based T1 maps to be generated. For each value of T SAT two acquisitions were performed and averaged resulting in an acquisition time of 4 heartb eats per measurement. Both the dynamic perfusion sequence and the T1 measurement can be performed at heart rates up to 150 bpm without modification. The signal behaviour of the Saturation recovery TFE EPI sequence was modelled to take into account the approach towards steady state during the 9-rf-pulse TFE-EPI acquisition. Simulations using signal intensities calculated from this model, performed over a T1 range from 200ms to 1000ms demonstrated a mean error of 0.44% for T1when fitted to the following empirically derived expression: Y(T SAT ) = 1 - exp[-(T SAT -delta)/T1] Y(T SAT ) is the ratio between the measured signal at recovery time T SAT and that at T SAT =∞, and delta represents the equivalent time shift of T SAT in ms after correction for the transient modification of M z during data acquisition. Both 1/T1 and delta are obtained as results of the fitting process. The accuracy of the T1 measurement sequ ence was tested on a series of test samples calibrated using a multiple-point inversion recovery method. (T1 range: 239ms - 816ms). Scaling factors taking into account the receiver gain and image reconstruction scaling were applied to the data prior to the calculation of T1. Results Good agreement was obtained between the calibrated T1 values and those obtained using the Saturation Recovery TFE -EPI data. The mean percentage difference between the two data sets was 2.34% (range 0.71%-5.18%). Discussion Saturation recovery TFE EPI provides a fast, accurate method for T1 measurement that is appropriate for myocardial perfusion studies. The measurement can be performed within a series of short breathhold periods although it is likely that image regis tration will be necessary prior to obtaining pixel maps from in vivo data. Further work is required to optimise the TFE-EPI sequence and to improve the accuracy of the model. References 1. Bellamy D, Pereira RS, McKenzie C, et al. ISMRM abstracts, 1866, 1999. 2. Fischer SE, Hofman MBM, Scott M et al, ISMRM abstracts, 848, 1997. 3. Fischer, SE, Wickline SA, Lorenz, CH, ISMRM abstracts, 682, 1996. 4. Ding S, Wolff, SD, E pstein, FH. . Magn. Reson. Med., 39, 514 - 519, 1998. 5. Tsekos, N.V., Zhang, Y., Merkle, H., et al. Magn. Reson. Med., 34, 530-536, 1995.