Ultrashort echo time (UTE) magnetic resonance imaging sequences provide direct imaging of short T₂ or T₂* tissues such as cortical bone, tendons, menisci and ligaments. Regular UTE imaging often benefits from fat suppression which improves short T₂ contrast. However, conventional spectral fat suppression may significantly reduce the signal from broad line short T₂ tissues through direct saturation or magnetization transfer (MT). UTE images can also be used with Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares estimation (IDEAL) processing to generate images with robust fat/water separation as well as T₂* and fat fraction maps. We have previously showed that combining 2D UTE sequences with IDEAL processing (2D UTE-IDEAL) produces high contrast images of Achilles tendon while preserving its signal intensity¹. In this work, we report the use of 3D UTE Cones imaging and IDEAL processing (3D Cones-IDEAL) for volumetric imaging of short T₂ tissues in the knee and ankle joints at 3T.The 3D UTE Cones sequence employs a short rectangular pulse excitation (duration = 14-26 µs) followed by Cones trajectories with a minimal nominal TE of 32 µs. With this technique, 3D k-space is divided into multiple cones with twisted radial trajectories along each cone. The 3D Cones sequence allows anisotropic fields of view and spatial resolution, resulting in vastly reduced scan times. Volumetric imaging can be achieved in a few minutes, covering the whole knee or ankle joint with high spatial resolution. This time efficiency greatly benefits clinical applications. The knee and ankle joints of five healthy volunteers were investigated in this study using three sets of dual-echo 3D Cones scans. For the knee joint the following parameters were used: TE = 0.032/2.8, 0.8/4.4 and 1.6/6.0 ms, BW = 250 kHz, FOV = 15 cm, slice thickness = 2.5 mm, matrix = 192×192, flip angle = 10°, TR = 20 ms, voxel size = 0.78×0.78×2.0mm³, scan time = 3.8 min for each dataset. Similar parameters were used for the ankle joints except for slightly different TEs (0.032/2.8; 0.8/3.6; 1.6/4.4 ms), BW = 166 kHz, FOV = 6 cm, voxel size = 0.23×0.23×2.0 mm³, scan time = 2.3 min for each dataset. Regular fat-saturated images were included for comparison. The reconstruction was done using the graphcut algorithms² available in the ISMRM Fat-Water Toolbox³. A six peaks spectrum and single T₂* decay model was applied.Excellent fat/water separation as well as R₂* and fat fraction mapping were achieved in the knee and ankle joints of all volunteers. Selected slice of water/fat images with the corresponding fat-saturated image of the knee joint in a 29 year old male volunteer are shown in Figure 1. On the water image (Fig 1A), the lateral collateral ligament (LCL, white arrow) and anterior cruciate ligament (ACL, hollow arrow) are seen with high signal intensity and good contrast to noise ratio (CNR) . The meniscus (red arrow) is also visible with high signal intensity. Residual fat signal was observed on the fat-saturated image (Fig 1C) but not on the water image (dotted arrow). Figure 2 shows in-vivo 3D Cones-IDEAL imaging of the ankle of a 28 year old healthy volunteer. The whole enthesis and surrounding Achilles tendon can be covered in a single scan with high spatial resolution and image contrast. High quality water and fat images as well as fat fraction and R₂* maps were generated. On the R₂* map (Fig 2D) the R₂* value of the enthesis is approximately 450 s^-1 while that of the Achilles tendon is about 800 s^-1, corresponding to T₂*s of 2.2 ms and 1.3 ms respectively, which are consistent with T₂* values reported in the literature.We have demonstrated that the 3D UTE Cones-IDEAL sequence can provide high resolution volumetric imaging of the knee and ankle joints, together with robust fat/water separation, field map estimation, R₂*/T₂* mapping and fat fraction mapping. Furthermore, the 3D UTE Cones-IDEAL sequence allows morphological and quantitative imaging not only of long T₂ tissues (such as the superficial layers of articular cartilage), but also of short T₂ tissues or tissue components (such as the deep layers and calcified cartilage, menisci, ligaments, tendons and bone). The latter groups of tissues are “invisible” with conventional clinical sequences. As joint degeneration is a ‘whole-organ’ disease involving all the principal components, volumetric imaging with a 3D Cones-IDEAL is likely to provide a more comprehensive examination than conventional 2D clinical sequences both in terms of coverage and visualization of short T₂/T₂* tissues and tissue components. Clinical evaluation of the 3D Cones-IDEAL sequence will be conducted in future studies.