Ultrasound (US) is largely limited to an evaluation of the extra-articular soft tissues of the knee (Fig. 29.5). Due to the excellence of spatial resolution, US keeps its leading edge when dealing with muscle and tendon pathology around the knee (Peetrons 2002). The major indication for US in clinical practice is evaluating overuse conditions of the extensor apparatus or affirming a clinically suspected Baker’s cyst. High-frequency (13.5 MHz) linear-array probes are used to perform US examinations of the knee. US palpation is a very valuable tool, trying to find the point of maximal tenderness or swelling (Peetrons 2002). Dynamic US study may be very helpful to assess a “snapping” sensation that is felt by the patient during knee movement. Major advantages of US are its low cost, wide availability, short examination times, and lack of radiation exposure. An addition of color power Doppler imaging to US has allowed for the noninvasive study of blood flow and vascularization within anatomic structures and lesions. Although initial studies (Weinberg et al. 1998) reported that increased tendon vascularization is typically correlated with clinical symptoms, more recent studies have reported the opposite (Tol et al. 2012). Furthermore, US provides image guidance for interventional procedures such as drainage of fluid collections and cysts. The trade-off for high-frequency, linear transducers is their small, static scan field. Extended field-of-view ultrasound (EFOVS) overcomes this disadvantage by generating a panoramic image (Fig. 29.6). With this technique, sequential registration of images along a broad examination region and their subsequent combination into an image of larger dimension and format is obtained (Weng et al. 1997). Disadvantages of US include operator dependency, selective and often incomprehensible documentation, and the inability to evaluate the articular structures of the knee joint.

The noninvasiveness of magnetic resonance imaging (MRI) combined with its relatively low cost has led to its acceptance by the orthopedic community, and MRI of the knee is the most frequently ordered MR examination of the musculoskeletal system (Farjoodi et al. 2010). Numerous diagnostic studies have compared MRI and arthroscopy of the knee, and most have shown good diagnostic performance of MRI in detecting internal derangements of the knee (Oei et al. 2003). MRI not only contributes to diagnosis but also serves as an important guide to treatment planning and prognostication (Milewski et al. 2011).

Diagnostic knee MRI can be performed using a variety of magnet designs (closed bore, whole body, open whole body, dedicated extremity). Also, a variety of magnetic field strengths can be used but the current standard for clinical knee imaging is 1.5 tesla (T). The patient is positioned supine with the affected knee completely or near completely extended. A local coil is positioned to provide adequate anatomic coverage. Typically, an extremity knee coil that completely surrounds the knee is used to ensure maximal and uniform signal-to-noise ratio (SNR) across the entire image.

A complete MR exam of the knee requires that images be obtained in the axial, sagittal, and coronal planes. Although techniques for knee MRI vary widely, a typical imaging protocol uses a field of view (FOV) of 160 mm or smaller, a slice thickness in the sagittal and coronal plane of 3 mm or less, and an imaging matrix with at least 140 steps in the phase direction and 256 steps in the frequency direction for 2D imaging (Tuite et al. 2012). The use of a contrast agent (e.g., DOTAREM®) is only indicated when dealing with a tumoral or tumorlike mass, in cases of knee inflammation, or as part of arthrography (performed only for postoperative knee imaging when conventional MRI reveals equivocal findings).

Knee MRI can be performed with a wide variety of pulse sequences. Most commonly, the proton-density (PD) and T2-weighted 2D turbo spin-echo (TSE) sequence and the T1-weighted 3D gradient-recalled (GRE) sequence are used for routine imaging of the knee (Recht et al. 2005; Kijowski 2010; Trattnig et al. 2009).

TSE sequences are advantageous in daily clinical practice because they allow assessment of the menisci, ligaments, and bony structures in addition to the articular cartilage, in a relatively short time (Recht et al. 2005). Fat suppression (FS) TSE images are most sensitive to bone marrow and soft tissue edema or joint effusion. Different FS techniques are available (spectral FS, Dixon method, inversion recovery technique), but in our institution, we prefer the spectral FS technique because of its better SNR and spatial resolution compared to the other FS techniques (Fleckenstein et al. 1991). Also, for good detection of fluid with preservation of anatomic detail and good differentiation between joint fluid and articular cartilage, we include an FS TSE intermediate-weighted sequence (TR/TE = 3,500/30–35 ms) in at least one imaging plane in our routine knee MR protocol (Fig. 29.7). Although 2D TSE acquisitions provide excellent tissue contrast and high in-plane spatial resolution, they may suffer from partial volume averaging artifacts because they use relatively thick slices with small gaps between the slices, possibly obscuring pathology (Hargreaves et al. 2003; Link et al. 2007). To overcome this disadvantage, 3D TSE sequences with isotropic resolution have recently been developed and are now commercially available on many MR vendor platforms. These sequences include 3D fast spin-echo (FSE) cube (GE Healthcare), 3D Fourier transform (FT, Philips Medical Systems), and sampling perfection with application-optimized contrasts using different flip-angle evolutions (SPACE, Siemens Medical Systems) (Kijowski and Gold 2011; Gold et al. 2007; Notohamiprodjo et al. 2009; Ristow et al. 2009). The advantage of these new 3D TSE acquisitions is their capability of mimicking the contrast properties of conventional 2D TSE PD-weighted acquisitions (Gold et al. 2007). In addition, high quality MPR images may be created in any orientation from the volumetric source data (Fig. 29.8). Although the first clinical results on the diagnostic performance of 3D isotropic resolution TSE sequences were encouraging, it remains uncertain whether a single 3D TSE acquisition has potential for replacing the multiple conventional 2D acquisitions currently used (Notohamiprodjo et al. 2009; Jung et al. 2009; Kijowski et al. 2009, 2012b; Subhas et al. 2011; Van Dyck et al. 2012b).

Thin-partitioned, 3D gradient-recalled echo (GRE) sequences, either using selective FS (FS 3D SPGR) or selective water excitation (WE) (WE 3D FLASH; WE 3D DESS), also have been used to assess the knee. These sequences provide higher spatial and contrast resolution but require longer acquisition times (7–10 min) and are more vulnerable to magnetic susceptibility and metallic artifacts. Moreover, GRE images can be used for evaluating the articular cartilage of the knee joint but the menisci, ligaments, and bone marrow are only poorly visualized with this sequence (Kijowski and Gold 2011).

MRI has the disadvantage of not always being accepted well by patients, of being incompatible with dynamic maneuvers, and of not always being possible in emergency conditions. Also, specific MRI safety measures and contraindications for the use of MRI (e.g., claustrophobia, filters, stents, clips, cardiac pacemakers, etc.) must be considered.