Early in pregnancy, remodeling of placenta develops such that blood flow in the maternal-placental circulation increases significantly. Placental insufficiency is a serious complication that occurs when placental remodeling fails and oxygen and nutrients are not adequately supplied to the fetus through blood flow. While ultrasound is the current standard for examining the placenta, it provides indirect measurements for placental insufficiency, leading to low sensitivity and low predictive values. Conventional MR perfusion imaging involves the use of contrast agents, which are contraindicated during pregnancy due to the potential adverse effects to the developing fetus. In this study we examined the feasibility of non-invasive placental perfusion imaging using arterial spin labeling (ASL) and 3D image acquisition with whole placenta coverage. While all previous studies on placental ASL [1-4] have used flow-sensitive alternating inversion recovery (FAIR), we deployed velocity-selective ASL (VSASL) [5] because it is well suited for arteries with tortuous geometry like the uterine arteries and therefore enables larger imaging coverage. Velocity-selective labeling was achieved using dual sech inversion with a cutoff velocity = 2 cm/s. Image acquisition was performed using fast-spin-echo 3D stack of spirals with 8 interleaves, TR/TE = 3000/10 ms, matrix size = 64x64x36, and spatial resolution = 5.2x5.2x4.0 mm³. Pseudocontinuous ASL (PCASL) was also performed in a subset of the subjects for comparison, using labeling duration/post-labeling delay = 1.5 s/2 s, and the identical image acquisition as in VSASL. Total scan time was about 4:30 min for both VSASL and PCASL. All imaging was performed on a GE MR450 1.5 T scanner using an 8-channel cardiac coil. Ten pregnant women (GA: 33 ± 3 weeks) from an ongoing prospective study whose fetuses were diagnosed with congenital heart disease (CHD) were recruited and compared to ten healthy pregnant women (GA: 33 ± 5 weeks). Each subject provided written informed consent under local IRB guidelines. Placental perfusion was calculated using the general kinetic model. For global perfusion calculation, the placenta was manually segmented on each slice of the proton density images with T2-weighted anatomical images as a referenceFigure 1 demonstrates comparison of VSASL and PCASL in one representative subject. Compared to PCASL, VSASL showed higher sensitivity to placental perfusion and depicted sufficient perfusion signal within the whole placenta. Figure 2 illustrates the perfusion measurements averaged in the whole placenta from all twenty subjects. Mean placental perfusion averaged across subjects was 213 ± 96 ml/100 g/min in patients with CHD and 204 ± 49 ml/100 g/min in healthy pregnant volunteers, with no significant difference between the two groups (p=0.779). While these measurements are rather higher than those measured in uterine arteries using ultrasound [6-8] presumably due to the contribution of both maternal and fetal circulations with ASL, they match ranges of previous placental ASL reports [1-3]. Among the twenty subjects, five showed placental lobulation on their VSASL images similar to those that are apparent on T2-weighted placenta images. The average placental perfusion measured within these lobules in five subjects was 427 ± 216 ml/100 g/min, which was two-fold higher compared to mean global perfusion. Examples of lobulation on VSASL images are shown in Figure 3.We demonstrated the feasibility of non-invasive placental perfusion imaging using VSASL, and quantified global and regional placental perfusion in high and low-risk pregnancies. To the best of our knowledge, this is the first report of placental ASL with whole placenta coverage. Further validation of placental VSASL will be performed in pregnancies complicated by placental insufficiency such as fetal growth restriction or preeclampsia.