Three-dimensional gradient echo T2-weighted sequences have a number of advantages over spin echo T2-weighted sequences (or even 2D gradient echo T2-weighted sequences) for assessment of the knee. They allow a multidimensional analysis based on a single acquisition sequence usually obtained in the sagittal plane. Image reconstructions can be performed secondarily in the coronal, axial and oblique planes, particularly along the specific path of the anterior cruciate ligament. By providing ultrathin serial sections, decreasing the partial volume effect, small lesions, such as cartilaginous fissures or flaps and radial meniscal lesions can be detected in the axial plane, for example. This advantage, combined with the marked sensitivity of gradient echo sequences to alterations in the tissue water content, allows the detection of partial tendon ruptures. The reduction of the partial volume effect and chemical shift artefact probably participate in the capacity of these sequences to visualise the two surfaces of the cartilage of the femorotibial joint. Flow artefacts are less of a problem than with 2D imaging, which eliminates the need for techniques such as saturation of the vascular signal or cardiac gating. A disadvantage of these gradient echo sequences (3D or 2D) is their sensitivity to the presence of metallic material, limiting their application in operated knees.