To minimize motion artifacts due to the beating heart, data for each image must be acquired during the same temporal phase of the cardiac cycle. The acquisition is synchronized to cardiac motion by using ECG gating. The TR is usually set to be equal to the R-R interval. For example, for a subject with a heart rate of 72 bpm, the R-R interval = (60 sec/min)/(72 beats/min) = 0.833 sec, and consequently, the TR would be 833 msec.

In most subjects, defining the TR as equal to one R-R interval will result in Tl-weighted imaging, provided a short TE is used. For subjects with slower heart rates, the contrast may be closer to proton density-weighted.

For most routine black blood imaging of the heart, the TR is equal to one R-R interval. It usually does not matter whether the image contrast is strongly Tl-weighted or
closer to being proton density-weighted. The main function of these images is to provide anatomic assessment of cardiac and vascular structures in the absence of signal from flowing blood.

T2-weighted sequences triggered off every other R wave may be important in certain settings such as evaluation of masses or vasculitis. To enhance conspicuity of pathology, fat suppression is often used.

As described in Chapter I-4, because the TE is short relative to the TR, there is ample time to collect data from multiple slice positions when performing an ECG-gated spin echo sequence (Figure III3-1).

The number of slices that can be acquired during one acquisition depends on the acquisition window and on the total number of 90°-180°-echoes that can fit into the acquisition time. The acquisition window usually corresponds to about 85% of the R-R interval (Chapter III-1).

For the subject with a heart rate of 72 bpm, the R-R interval is 833 msec, and the acquisition window is about 700 msec. If the total time from start of the 90° RF pulse to the end of the echo is 35 msec, then the number of slices that could be imaged in the same time as needed for a single slice would be 700/35 or 20 (Figure III3-1).

With an ECG-triggered spin echo sequence, the first slice position is imaged immediately after the R wave, while the last is acquired toward the end of the R-R interval. Hence, each slice will reflect the heart at a different phase of the cardiac cycle (Figure III3-1). On a series of transaxial images, the more cephalad images may be acquired during systole, and the more caudal slices may be acquired during diastole. Alternatively, the slices could be acquired in reverse order. It is important to recognize that motion artifacts are most likely to affect images collected during peak systole, while signal in the blood pool is most likely during diastole.

As with all spin echo imaging, acquisition times are dependent on TR. For cardiac applications, because TR
depends on heart rate or R-R interval, the total acquisition time can be expressed as:

depending on whether triggering occurs with every heartbeat or every other heartbeat. With conventional spin echo imaging, a single line of k-space is acquired during each heartbeat, so that the echo train length, ETL, is equal to 1. If there are 128 phase-encoding lines, then the acquisition time for one set of data (N_Acq = 1) will be 128 heartbeats. For a subject with heart rate 72 bpm, and R-R interval = TR = 833 msec, the acquisition time will be 128 × 0.833 sec = 106 sec, or nearly 2 min. If a more T2-weighted sequence is desired, then a TR equal to two R-R intervals will double the acquisition time (212 sec) to nearly 4 min.

To generate a spin echo, protons must experience both the 90° excitation pulse and the 180° refocusing pulse. If the RF pulses are slice selective and if the protons in the blood are not present within the slice long enough to experience both RF pulses, then no echo is generated, resulting in a signal void. For this reason, flowing blood is typically black with conventional spin echo imaging. As discussed in Chapter II-1, to minimize the signal from flowing blood, several strategies can be employed: using a longer TE, thinner slices, and slice positioning orthogonal to the direction of flow. Moreover, acquisitions during systole, when flow is fastest, have more complete nulling of flowing blood.