Dynamic contrast enhanced (DCE) MRI for quantifying perfusion in the myocardium is becoming a more robust and useful tool. Typically the first pass of gadolinium contrast is tracked by imaging a number of slices sequentially each heartbeat. For some applications it would be ideal to obtain multiple slices at systole and diastole that are optimized to use data from time periods when cardiac motion is small. Here we propose a unique perfusion acquisition that uses one saturation pulse per heartbeat. This combined with simultaneous multi-slice (SMS) methods allows for acquiring the same set of slices continuously through the cardiac cycle. This has a number of advantages including the ability to retrospectively reconstruct an accurate arterial input function (AIF) and optimized systolic/diastolic frames, or other portions of the cardiac cycle. This is similar to recently developed ungated methods that quantify perfusion [1-3] or that operate without a saturation pulse and do not quantify perfusion [4-5], but the ability to simultaneously acquire multiple slices opens up many new possibilities. The approach proposed here acquires both k-space rays that reflect the influence of the saturation pulse and other rays that reflect the steady-state GRE contrast, and thus is termed the “hybrid” method.A saturation recovery radial turboFLASH sequence was used in 5 subjects to investigate the potential of hybrid cardiac perfusion imaging. TR/TE=2.7/1.6msec, FOV=260mm, ~1.8x1.8x8mm pixel size on a 3T Prisma (Siemens) scanner. Scans were done at rest with ~0.05mmol/kg gadoteridol. A non-selective saturation pulse was played once per heartbeat with the readout starting 10msec after the saturation pulse. Depending on heart rate, approximately 300 rays (golden ratio ordering) were then acquired each heartbeat, simultaneously in 3 slices (multiband factor=3). Subsets of the k-space data could then be retrospectively reconstructed into different cardiac phases. Here there were two goals – to obtain an accurate AIF from rays acquired shortly after the saturation pulse, and to use all of the rays from the longest quiescent phase for systole, to maximize image quality (diastolic and other cardiac phases can be reconstructed as well). For the AIF, images were reconstructed from the first 24 rays. For the systolic frames, a cine slice was inspected and the times that appeared to be most stationary during systole were chosen for the perfusion reconstruction. The 3 slices were reconstructed jointly with an iterative compressed sensing SMS method. The resulting images were registered to compensate for breathing motion, and converted to gadolinium concentration by considering their excitation history and using measured proton density frames. The curves were fit with a compartment model to give Ktrans values. In two of the subjects, an ungated radial method that does not acquire any slices simultaneously and has been validated for quantitation of perfusion by dual bolus and gated acquisitions [2] was acquired for comparison. Fig. 1 shows an example in one subject of different reconstructions across cardiac phase and time frame (four time frames shown). The slices were acquired summed as indicated by the plus signs (although the RF excitations were modulated differently). Rays 89-124 (265 to 360ms after the R-wave trigger) were used for the systolic images. For this subject, diastolic frames were also reconstructed (using rays 241-300, 676-835ms after the R-wave). Image quality is good; the reconstruction approach is able to separate the simultaneously acquired slices well. Fig. 2 shows quantitative flow results from the five subjects. All of the subjects had mean resting flows between 0.7-1.1 ml/min/g. In the two scans that were not imaged with the hybrid method and were not SMS acquisitions, the mean flow values differed little from the new hybrid method results. This new approach is an ideal way to maximize the image quality at systole or any part of the cardiac cycle. The maximum number of rays can be determined retrospectively, which is especially important at stress. This could potentially be done with an automated method. Temporal resolution is better than 3D acquisitions but as in 3D, the slices are acquired simultaneously so have the same respiratory motion and contrast. It may also be possible to push the SMS techniques further and acquire 4 slices at the same time. The studies here show that the hybrid approach has promise for quantifying cardiac perfusion and has unique advantages. More studies are needed to further evaluate the technique.