Familiarity with the wide range of appearances of the normal thymus gland on chest radiography is important for all radiologists. In infants and young children, the thymus can be impressively large on frontal radiographs. The classic “sail” sign, a sharply demarcated pointed base seen most frequently along the right inferior margin of the thymus gland, is actually found in only 5 % of children (Fig. 3a). The wave sign, the smooth undulation of the thymus as it interdigitates along the anterior costal margin, is more frequently seen but is less striking (Fig. 3b). To potentiate the wave sign, decubitus views can be used as a problem-solving tool for determining the etiology of anterior mediastinal mass. In most instances the thymus is seen simply as a smoothly marginated, low-density soft tissue causing widening of the superior mediastinum. This is in contrast to pathologic masses, which commonly have irregular or lobular margins and cause mass effect on adjacent structures, rather than smoothly conforming to the available space in the mediastinum. The thymus becomes less evident on chest radiographs beginning around 2 years of age and typically is not seen in children over 8 years of age (Arthur 2000). In cases of a particularly unusual appearance of the thymus, further evaluation can be obtained with other imaging modalities as described below.

Ultrasound is a useful tool in the evaluation of the thymus in infants and young children due to its lack of ionizing radiation and its real-time imaging capabilities. Even after the thymus is no longer apparent radiographically, it may still be visible with ultrasound (Adam and Ignotus 1993). A high frequency, linear transducer works best in the neonatal population with lower frequency transducers used in the older age groups. A variety of approaches can be used in the infant including subxyphoid, parasternal, suprasternal, and transcostal windows (Mong et al. 2012). However, as ossification of the manubrium and sternum progresses, a suprasternal approach provides the best acoustic window.

In the neonate and young child, the thymus has a distinctive sonographic signature making it identifiable when present in abnormal locations. The thymus has a heterogeneous, but uniform echotexture, and is predominantly hypoechoic relative to the liver, thyroid, and spleen, with punctate and linear echogenic foci secondary to the fat present (Fig. 4) (Mong et al. 2012; Newman 2011; Kim et al. 2000). As the patient ages, the thymus becomes more homogeneous and echogenic due to increased fat content. It should have a smooth contour due to the capsule surrounding each lobe with no compression of the surrounding structures. During real-time scanning, the thymus should change shape in response to respiratory and cardiac motion reflecting its soft, pliable nature, which further helps to delineate it from a pathologic mediastinal mass.

Particularly in younger children, ultrasound can also be used in the evaluation of thymic masses. Ultrasound can also be used for real-time guidance of biopsies of suspected thymic masses.

Fluoroscopy is of limited use in the dedicated evaluation of the thymus. However, knowledge of the expected fluoroscopic appearance of the thymus can be useful during routine examinations such as airway fluoroscopy and upper GI examinations. As on chest radiography, the thymus is seen as a soft tissue density in the anterior mediastinum. Its size and shape should vary with respiration, and it should never exert substantial mass effect on any of the mediastinal structures, i.e., displacement of a contrast filled esophagus should not be attributed to a normal thymus.

CT provides excellent visualization of the thymus and thymic pathology. However, given concerns about radiation risk from CT scanning, this modality should be used judiciously in characterizing abnormalities of the thymus. The thymus can typically be seen in the anterior mediastinum on CT imaging throughout childhood, and is still visible in most patients up to the age of 30 years. On CT imaging performed without intravenous contrast administration, the thymus has a homogeneous appearance that, compared to cardiac and chest wall musculature, is hyperdense in infancy and nearly isodense later in childhood (mean Hounsfield Unit values ranging from 80.8 in infancy to 56.4 at 14 years of age) (Sklair-Levy et al. 2000). It demonstrates homogeneous enhancement on post-contrast images. The attenuation of the thymus declines with increasing age as it becomes increasingly replaced with fat (Sklair-Levy et al. 2000).

On cross-sectional imaging, the appearance of the thymus is consistent with its radiographic appearance. The thymus has a quadrilateral configuration with bulging contours in infants and young children, becoming progressively more triangular in shape with concave margins in older children and adolescents (Fig. 5). The relationship of the thymus to its surrounding structures, particularly the presence of compression or invasion in the setting of thymic pathology, is well seen with CT. Internal heterogeneity, calcifications, fat, or abnormal enhancement are also well depicted with CT. For operative planning purposes, the location of the feeding and draining vessels can be delineated on contrast-enhanced CT imaging.

MRI can be a valuable tool in the evaluation of patients with suspected thymic pathology. However, due to its relatively greater expense, study length, and variable need for sedation, MRI is not typically the initial choice of imaging modality for evaluation of the thymus. The MRI appearance of the thymus varies as it involutes. Prior to fatty infiltration, the signal intensity of the thymus is slightly greater than that of muscle but less than that of fat on T1-weighted images, while on nonfat suppressed FSE-T2 weighted images (Henkelman et al. 1992), the signal intensity is increased relative to muscle, but lower than subcutaneous fat (Fig. 6) (Molina et al. 1990). As the thymus becomes progressively more fatty its signal intensity approaches with that of pure fat (Ackman and Wu 2011; Siegel et al. 1989). As on CT, there is homogeneous enhancement of the normal thymus on post-contrast images. Due to the presence of fat within non-neoplastic thymus, opposed phase imaging has been studied to differentiate between normal thymus/thymic hyperplasia and a solid mass lesion (Inaoka et al. 2007). This imaging technique has not been studied on the thymus of younger (<16 years of age) patient populations. However, it may be of value since small quantities of fat coexisting with nonfatty soft tissue in the same imaging voxel can be detected by opposed phase imaging, and could provide a sensitive means of characterizing normal thymic tissue in younger age groups, where the quantity of thymic fat is less abundant.

Coil selection depends on the child’s size. Both T1- and T2-weighted images in at least two planes should be used for routine evaluation of the thymus. The sagittal plane in particular can nicely demonstrate the anteriorly positioned normal thymus. In order to achieve high-quality MRI images of the thymus, one must keep in mind the proximity of the thymus to the moving heart and lungs. If close scrutiny of the thymus is desired, cardiac gating and breath-hold sequences may be needed. The phase-encoding direction should be in the transverse medial–lateral rather than anterior–posterior direction to prevent motion artifact from interfering in the evaluation of the thymus as shown in Fig. 7.

Thymic cysts can either be congenital or acquired. As described previously, congenital thymic cysts arise from a patent thymopharyngeal duct and reside along the carotid sheath. If large enough, they can be detected in utero on both ultrasound and MRI. These cysts are lined by ciliated epithelium, typically unilocular, and contain simple fluid if drained (Restrepo et al. 2005; Petroze and McGahren 2012; Mortelmans and Hermans 2005). They rarely measure greater than 6 cm (Hegde et al. 2012), and in most cases they are asymptomatic and are discovered incidentally on imaging. They may present as a base of neck mass if there is sufficient suprasternal extension of the cyst, as shown in Fig. 13. If they are large enough, they can present with chest pain, hoarseness, or dyspnea from mass effect. These masses are often occult on chest radiographs; however, if they are large enough they can be seen as a smoothly marginated mass arising from the upper mediastinum. On ultrasound imaging, they will present as an anechoic mass and may contain a variable amount of debris depending on the presence of inflammation or prior hemorrhage. CT will show a round, fluid-attenuation mass with no solid components, as shown in Fig. 14. On MRI, simple thymic cysts will typically have high-signal intensity on fluid sensitive sequences (Fig. 13). No solid components should be present, and there should be no internal contrast enhancement.

Acquired cysts vary in etiology. Typically, they are postinfectious but can be seen in the setting of prior neoplasm, radiation, or autoimmune diseases such as myasthenia gravis. Children with HIV infection can also present with massive cystic enlargement of the thymus related to aberrant immunoregulation (Kontny et al. 1997). These acquired cysts are more complex than the simple congenital cysts described above. They often have thicker cyst walls, internal septations, and contain a more gelatinous liquid (Restrepo et al. 2005). They tend to be multilocular and can range in diameter from a few millimeters to over 15 cm (Hegde et al. 2012). The complex nature of the acquired cysts is reflected in their appearance on imaging studies. On ultrasound imaging they range from anechoic to hypoechoic, with a variable number of septations. CT or MRI will also demonstrate a variable number of septations that may enhance. On MRI, these lesions may be higher in signal intensity on T1-weighted images and lower signal intensity on T2-weighted images, as compared to simple cysts, reflecting the complex nature of the fluid, often due to higher protein content. Regardless of the etiology of these cysts, they are uniformly benign in the pediatric population (Petroze and McGahren 2012).