MR-based bone imaging is hampered by 1) low proton density (~20% relative to water) and 2) short signal live time (~0.4ms at 3T) [Du et al]. Accordingly, skeletal imaging is typically based on X-ray radiography or computed tomography (CT). Nevertheless, there are an increasing number of applications requiring MR-based bone imaging; often not for diagnostic purposes but rather for accompanying calibration or planning purposes. Examples include MR-based PET attenuation correction (MR-AC) in hybrid PET/MR imaging, MR-based radiation therapy planning and MR-guided focused ultrasound. For these applications, bone presents the tissue with the highest attenuation effect that typically cannot be neglected. In this abstract, we describe a zero TE based method for whole body skeletal imaging and demonstrate its performance in volunteer scans.Recently, we described a novel method for MR bone imaging in the head [Wiesinger et al]. Different to conventional ultra-short TE (UTE) echo subtraction (based on T2 contrast), the method explores 1) proton density (PD) contrast between soft tissues & body fluids (~80-100%) and bone (~20% relative to water) and 2) contrast inversion for bone enhancement to mimic CT image appearance. Zero TE imaging is used for efficient PD imaging; especially for short T2 tissues. Whole body, in vivo volunteer experiments were conducted on a GE M750w 3T scanner (GE Healthcare, Waukesha, WI) with zero TE imaging parameters set to: FOV=50cm, res=2.6mm, FA=0.5deg, BW=±125kHz, 110592 radial spokes, 7 bed positions acquired in 1min24sec per bed. For segmentation purposes, the zero TE images were bias corrected and normalized via division by a low-pass filtered version of the original image. The lungs were identified as the largest, low-intensity connected comment within the chest region and removed. Figure 1 shows coronal views of zero TE image in linear and inverted grayscale acquired in the pelvis comparing different imaging BW of ±125kHz (left), ±62.5kHz (middle), ±31.25kHz (right) with all other imaging parameters unchanged. For the highest imaging BW of ±125kHz soft tissue appears most uniform. Decreasing imaging bandwidth, and hence longer readout duration, off-resonance artifacts increase; especially apparent at organ/tissue interfaces and the bitmarks at the edge of the imaging FOV. Figure 2 shows a whole-body, Maximum Intensity Projection (MIP) of the skeleton using above described image processing method. Apparently, the method is able to correctly identify all major bone structures in the human; including the skull, scapula, humerus, spine, ribs and the sternum, pelvic bones, femur, tibia, fibula, among others. Remaining air cavities can be removed via thresholding or using prior knowledge (such as bone probability maps). Adding a fast and robust bone imaging method to MRI’s arsenal of contrast mechanisms will certainly help MRI to expand its role towards treatment planning in radiation oncology and high-intensity focused ultrasound. In addition it will allow more accurate PET attenuation correction in hybrid PET/MR imaging.