Introduction

The sternal lymph nodes are located immediately dorsal to the sternum, along the internal thoracic artery and vein in the cranial mediastinum. There is usually a single node on each side, dorsal to the second or third sternebra in dogs. Occasionally, there may be a single or more than two nodes on either one side or the other.^1,2 The sternal lymph nodes receive drainage from a broad area of the thoraco-abdominal region. Afferent vessels receive tributaries from the ribs, sternum, thoracic and abdominal walls, thymus, abdominal and pelvic cavities, and cranial and caudal thoracic mammary glands.^1,3,4 The sternal lymph nodes can be enlarged in response to conditions such as infection, inflammation, and malignant tumors in this region. It has also been suggested that hemoperitoneum can cause sternal lymphadenopathy.²

It is important to distinguish whether or not a lymph node is enlarged. On radiographic images, a sternal lymph node appears as an isolated, small, soft tissue mass when enlarged sufficiently to distort the contour of the mediastinum on lateral radiographs.⁵ The radiographic characteristics of enlarged sternal lymph nodes in dogs have been previously reported, and the size, location, and diseases associated with sternal lymphadenopathy have been described.⁶ Normal sternal lymph nodes can be seen as fusiform opacities with a mean length of 30 mm, but they are only visible in a few large-breed dogs.⁷ Ultrasonography provides useful information regarding the size, echogenicity, shape, border definition, and blood flow of lymph nodes.^2,8,9 Although various normal superficial and deep lymph nodes have been evaluated in dogs, there are few descriptions of normal sternal lymph nodes.^8–11

Several reports have indicated that computed tomography (CT) is a valuable modality to assess the tracheobronchial, cervical, and abdominal lymph nodes in dogs.^12–16 To the authors' knowledge, sternal lymph nodes in dogs have not been investigated. For clinical use, the purpose of this present retrospective study was to evaluate the characteristics of sternal lymph nodes using CT and to establish criteria for the upper limit of normal size, independent of body weight, in dogs considered unlikely to have lymphadenopathy.

Materials and Methods

Sternal Lymph Nodes Evaluation on CT

CT scans were acquired using a 4-detector-row CT unit^a. The scanning parameters were as follows: rotation time, 0.75 s; slice thickness, 1–2 mm; field of view, 160–340 mm; matrix dimensions, 512×512; reconstruction interval, 0.5–1 mm; detector pitch, 3.5; collimator pitch, 0.875; X-ray tube potential, 120 kV; and X-ray tube current, 30–100 mA. Post-contrast images were acquired 1–5 min after 450 mg I/kg of iodinated contrast medium^b was administered intravenously using a power injector^c. All dogs were scanned in the ventral recumbency.

CT scans of the thorax were reviewed retrospectively and reconstructed using a mediastinum algorithm (window level 10, window width 300) for the sternal lymph nodes and a bone algorithm (window level 300, window width 1,500) for the sternum. All CT images were reviewed by a single observer (R.I.) using DICOM viewer software^d. 3D multi-planner reconstruction was created for all patients, and the number, size, shape, X-ray attenuation (pre- and post-contrast administration), and uniformity (pre-contrast administration) of the sternal lymph nodes were recorded. Three measurements for analysis were made by two authors (R.I., T.M.). These were: (1) long axis dimension of the sternal lymph node, (2) short axis dimension of the sternal lymph node, and (3) dimension of the thickness of the middle portion of the second sternebra (Figure 1). The long axis of the sternal lymph node was defined as the longest dimension in the craniocaudal plane, and the short axis dimension was defined as the length perpendicular to the long axis in the same plane. The second sternebra thickness was defined as the length perpendicular to the craniocaudal line of the second sternebra. The measurement of the long axis dimension of the sternal lymph node was made in the craniocaudal plane, and the short axis dimension of the sternal lymph node and second sternebra thickness were made in the transverse plane using an electronic caliper. The ratio of the short axis dimension of the sternal lymph node to the thickness of the second sternebra was calculated.

View Figure

• Download as Powerpoint
• Download Figure

The sternal lymph node shape was classified as (1) oblong, (2) oval, or (3) rounded, based on the craniocaudal plane. As previously reported for abdominal lymph nodes, oblong was defined as having a short axis/long axis dimension ratio ≤0.5.⁹ Rounded was defined as short axis/long axis dimension ratio >0.8. Oval was numerically classified as between oblong and rounded. As previously described, when a lymph node is visible on three or more consecutive transverse slices and is large enough to fit a region of interest of at least 4 mm², the X-ray attenuation before and following contrast medium administration is measured in Hounsfield units (HU) in the central slice.¹⁶ The uniformity of the sternal lymph node is classified as either homogeneous or heterogeneous.

Results

Case Population

Following a review of 1,575 medical records of dogs that had undergone a thoracic CT scan, 152 dogs met the inclusion criteria. The included dogs had the following conditions: subcutaneous or intraperitoneal lipomas or infiltrative lipomas (n = 27), various epulides (n = 16), prostate hyperplasia (n = 7), perianal gland adenomas (n = 7), salivary gland cysts (n = 6), benign mammary gland tumors (n = 6), uterine or vaginal leiomyomas (n = 4), intervertebral disc herniations (n = 4), gingivitis (n = 3), prostate cysts (n = 3), bladder calculi (n = 2), perineal hernias (n = 2), rectal leiomyomas (n = 2), gastrointestinal foreign bodies (n = 2), and sebaceous epitheliomas (n = 2). Thirty-five other dogs were affected by 35 single conditions, and 24 were classified as “no abnormalities” (Table 1). None of the included diseases were associated with sternal lymphadenopathy. The body weights of the dogs ranged from 2.0 to 36.3 kg (mean 13.9 kg, median 10.1 kg). The ages ranged from 1.3 to 15.3 years (mean 8.8 years, median 9.4 years). There were 69 females (38 intact, 31 neutered) and 83 males (60 intact, 23 neutered). Both purebred dogs and mixed breeds were included.

Table 1 Diagnosis of 152 Dogs Included in the Sternal Lymph Node Evaluation

View Fullscreen

CT Characteristics of the Sternal Lymph Nodes

One or more sternal lymph nodes was detected on CT images in all 152 dogs (Table 2). When the largest sternal lymph node was measured in each dog, the long axis dimension ranged from 0.216 to 1.657 cm (mean 0.700 cm, median 0.621 cm), and the short axis dimension ranged from 0.149 to 0.782 cm (mean 0.368 cm, median 0.333 cm). Most of the lymph nodes were oval or oblong in shape. While oval and oblong sternal lymph nodes were seen in various breeds, 11 of 15 dogs (73.3%) that had rounded sternal lymph nodes weighed <10 kg. The X-ray attenuation was measured in 100 lymph nodes and ranged from 14.3 to 59.3 HU before contrast administration (mean 30.2 HU, median 29.0 HU). In only five lymph nodes after contrast administration, the attenuation ranged from 56.2 to 69.5 HU (mean 61.7 HU, median 62.0 HU). In 147 lymph nodes before contrast administration, most of the lymph nodes were homogeneous.

Table 2 Characteristics of Presumptively Normal Sternal Lymph Nodes in 152 Dogs

View Fullscreen

Normalization of the Sternal Lymph Nodes

Figure 2A shows that there is a positive correlation between the short axis dimension of the sternal lymph node and body weight (r_s = 0.740, P < .0001). To evaluate the size of the sternal lymph node independent of physical constitution, the ratio of the sternal lymph node to the thickness of the second sternebra, which also had a positive correlation with body weight, was calculated (Figure 2B; r_s = 0.866, P < .0001). The ratio ranged from 0.278 to 0.665 (Figure 2C; mean 0.457, median 0.451, 95 % prediction interval 0.317 to 0.596). The regression line of this result was represented as: y = −0.0007 x + 0.466 (x; body weight, y; sternal lymph node: second sternebra ratio).

View Figure

• Download as Powerpoint
• Download Figure

Discussion

In this retrospective study, at least one sternal lymph node could be identified and measured in all cases on CT scans by using 3D multi-planner reconstruction. On radiographic images, a sternal lymph node is detected when it is enlarged sufficiently to distort the contour of the mediastinum and it is evaluated as lymphadenopathy.⁵ However, lymphadenopathy may be present even if the sternal lymph nodes are undetectable, limiting the value of radiography to differentiate between normal and abnormal. The criteria we established for this study was aimed at excluding patients with lymphadenopathy, so it is obvious that CT has a higher detection sensitivity than radiographs for normal sternal lymph nodes. However, the identification of the soft tissues in the mediastinum using either CT or radiography depends on the amount of fat present.^5,7,17 The number of identified abdominal lymph nodes is also dependent on the amount of intra-abdominal fat.¹⁶ In this study, it was more difficult to detect sternal lymph nodes in dogs with a thin versus a thick mediastinum because of the close proximity of the surrounding soft tissues, such as the internal thoracic artery and vein and the intercostal muscles. Therefore, in dogs with lower amounts of mediastinal fat, it is possible that some sternal lymph nodes were overlooked.

We evaluated the ratio of the short axis dimension of the sternal lymph node to the thickness of the second sternebra in addition to the long axis and short axis dimension to determine the size of the sternal lymph node. In dogs, several reports have shown that the lymph node size varies widely between breeds.^{10–12,15,16,18} In this study, sternal lymph node size showed a positive correlation with body weight (minimum: 0.149 cm with a body weight of 1.98 kg; maximum: 0.782 cm with a body weight of 36.0 kg). The ratio of the short axis of the lymph node to the thickness of the second sternebra was 0.457 ± 0.07 (mean ± SD), which was almost a constant value unrelated to body weight. The reasons we used the short axis dimension in the transverse plane were (1) to reduce the measurement error by reconstructing images, (2) to distinguish the sternal lymph node from the internal thoracic artery and vein, and (3) because the short axis dimension compared to the long axis dimension tends to increase when the lymph node is affected by a malignant tumor. Using ultrasonography, it has been observed that the short axis to long axis ratio of the superficial and deep lymph nodes exceeds 0.5 in abnormal lymph nodes.^8,9 It is suggested that changes in the short axis dimension reflect differences between normal and malignant lymph nodes using CT as well. There are few CT studies that have determined lymph node size by evaluating an adjacent skeletal structure. ¹² However, it would be effective to use this method, especially in dogs, in which conformation varies according to breed. The thickness of the second sternebra also showed a positive correlation with body weight in our population. Our results suggest that the sternal lymph node size can be evaluated correctly by using the second sternebra. The upper 95% prediction interval value of the sternal lymph node to the second sternebra ratio was 0.596 in normal dogs. However, pathological examination is needed when the ratio is over 0.596 to determine whether the sternal lymph is abnormal.

In addition to the size, the shapes of the sternal lymph nodes were also evaluated. The results were similar to other reports, which indicated that most lymph nodes are oblong or oval in shape. ^8,10,19 In our study, small lymph nodes tended to be round, resulting in a relative increase in the short to long axis ratio. However, this may not be true for abnormal lymph nodes. Evaluating the ratio of the short axis dimension of the sternal lymph node to the thickness of the second sterneba can avoid this problem by not using the long axis dimension.

The mean X-ray attenuation of the sternal lymph nodes was 30.2 ± 10.4 HU before contrast administration and 61.7 ± 5.2 HU after contrast administration. This attenuation before contrast administration was similar to previous reports; however, it was hypodense after contrast administration.^15,16 Another study of normal tracheobronchial lymph nodes reported that seven of nine lymph nodes had contrast enhancement and two of nine did not; however, the X-ray attenuation was not described.¹² Hypodense attenuation after contrast administration in our population was probably due to using a lower contrast medium dose than in the previous two reports. There were few dogs in our retrospective study in which contrast administration was used in the thoracic region. Further investigation is needed to compare other normal and abnormal lymph nodes.

Evaluating the uniformity of HU values is thought to be important, especially when assessing enlarged lymph nodes. In our study, most sternal lymph nodes appeared homogeneous before contrast administration; however, the heterogeneous appearance of the lymph node cortex and medulla suggests greater hyper attenuation in the periphery than centrally.¹⁶ In our study population, 22 lymph nodes (14.5%) were heterogeneous, and 20 of 22 lymph nodes were hyper attenuating in the periphery. These dogs were relatively large breeds and had larger lymph nodes, so the lymph node structure was more clearly visible. There can be both homogeneous and heterogeneous X-ray attenuation in the normal lymph node, and this would make interpretation difficult using only the uniformity of the HU values. The evaluation of the post-contrast administration uniformity was excluded because the number (five lymph nodes) was too low. In this retrospective study, contrast medium was administered only in the region of the focal lesion, just after thoracic or whole-body CT scans were finished, which was a limitation of this study. Abnormal patterns such as rim and heterogeneous contrast enhancement are often seen in malignancies and may be related to the pathological change in angioarchitecture associated with intranodal tumor infiltration.^12,19,20 Further investigation, especially of enlarged sternal lymph nodes, is needed to evaluate the correlation between the CT contrast enhancement pattern and the pathological diagnosis.

Conclusion

Sternal lymph nodes are easily measured in presumptively normal dogs using CT, and the second sternebra ratio can provide a constant value unrelated to body weight. The upper 95% prediction interval value was 0.596. Further investigation using a pathological approach is needed to ensure whether this value is significant. Not only the size of the sternal lymph nodes but shape, X-ray attenuation, and contrast enhancement pattern would also be helpful for diagnosis.