Evaluation of SLNs during physical examination and survey radiography of dogs poses a diagnostic challenge because of their location and size. On the basis of results for anatomic gross dissection, SLNs in clinically normal dogs range from 2 to 20 mm in length and are located dorsal to the first 3 cranial sternebrae, typically with a single node on each side of the thorax closely associated with the internal thoracic artery and veins.1 The SLNs are part of the parietal group of thoracic lymph nodes and receive afferent lymphatics from the mammary glands, thymus, ribs, sternum, serous pleural membranes, and adjacent muscles as well as the peritoneal and pelvic cavities.1,2 Although it is not frequently mentioned in the veterinary literature, lymphatic drainage from the pelvic cavity is similar to that of the peritoneal cavity, including drainage to the SLNs, but with greater drainage via the thoracic lymphatic trunk and thoracic duct.2 Diseases affecting these anatomic structures may result in pathological changes of the SLNs, and evaluation of SLNs can be important for staging purposes. Because of the small size and anatomic location of SLNs, sample collection via fine-needle aspiration or biopsy may be technically difficult and is not frequently performed, as evidenced by the lack of peer-reviewed veterinary literature on the topic.3 Therefore, noninvasive, real-time imaging techniques are important for the evaluation of SLNs.
Imaging modalities commonly used in contemporary veterinary practice for the evaluation of SLNs include radiography, ultrasonography, and CT.3–6 Computed tomography is favored for imaging of intrathoracic structures because of excellent contrast resolution and a lack of anatomic superimposition of cranial mediastinal, pleural, and pulmonary structures.6,7 Despite the popularity of CT in veterinary medicine, to the authors’ knowledge, the CT appearance of SLNs in clinically normal dogs has not been reported. Provision of such reference data is required for identification of SLNs with an abnormal CT appearance, which may be used to identify patients in which fine-needle aspirates or biopsy samples of the SLNs should be obtained and examined for staging purposes.
The objective of the study reported here was to determine dimensions and attenuation of SLNs in healthy adult dogs by use of CT. Our hypothesis was that CT would allow identification and measurement of the SLNs in all dogs. We further intended to describe the enhancement patterns of the SLNs. These results could provide objective reference data for evaluation of SLNs in healthy adult dogs. This information could be useful for staging purposes in clinical patients and for comparative purposes for future research studies of dogs with abnormal SLNs.
Materials and Methods
Animals
Twelve healthy adult purpose-bred research dogs (8 mixed-breed dogs and 4 Beagles) were used in the study. Median age was 1.5 years (IQR, 1.0 to 3.8 years; range, 1.0 to 4.0 years). Median body weight was 23.7 kg (IQR, 10.4 to 25.6 kg; range, 9.5 to 26.0 kg), and all dogs had a body condition score of 4 on a scale of 1 to 9. These 12 dogs were previously used in an unpublished unrelated study of cardiac imaging involving thoracic CT scans. To ensure the study population was healthy, a physical examination, CBC, serum biochemical analysis, arterial blood gas analysis, echocardiography, and ECG were performed on each dog; all results were within reference limits. Collection of urine samples and analysis were not performed for any dog. There were no additional inclusion or exclusion criteria. The research protocol was approved by the Oregon State University Institutional Animal Care and Use Committee.
CT
Images were obtained by use of a 64-detector helical CT scanner,a which was calibrated daily. Preventative maintenance was performed on the scanner quarterly, and scanner function was assessed annually by a medical physicist. Nonionic iodinated contrast mediumb was used for contrast enhancement. Contrast medium (300 mg of I/kg) was injected IV by use of a power injectorc through a catheter inserted in a cephalic vein. Images were acquired before and 1 minute after injection of contrast medium. All helical images were acquired as volume data with 0.5-mm voxels, rotation speed of 0.5 seconds, helical pitch of 0.828, and a 512 × 512 matrix. Variable imaging settings were as follows: 120 to 135 kVp, 40 to 450 mA, slice thickness of 0.5 to 3.0 mm, field of view of 18 × 18 cm to 32 × 32 cm (depending on size of the dog), and pixel width of 0.35 to 0.625 mm (depending on field of view). Volume data were reconstructed by use of a soft tissue algorithm with a 0.5-mm reconstruction diameter in isovolumetric transverse, sagittal, and dorsal planes. The CT images were available in a soft tissue window (width, 400 HU; level, 40 HU). However, the observer was able to adjust the window and level of the images.
Collection of data
The DICOM CT images were viewed on diagnostic imaging workstations by use of commercially available software.d All measurements were obtained for contrast-enhanced images by a board-certified veterinary radiologist experienced in evaluating thoracic CT images (SN). Archived thoracic CT images obtained for 32 client-owned dogs with clinical diseases of areas with afferent lymphatics to the SLNs were included among the images evaluated to help prevent bias of the veterinary radiologist, who was unaware of the medical record data for each dog. A total of 44 sets of images (12 healthy study dogs and 32 client-owned diseased dogs) were reviewed in series. All measurements for all images were obtained again by the same veterinary radiologist 6 months after the initial measurements to determine intraobserver variability.
Only data for the healthy research dogs were used in the present study. The number of SLNs detected in each dog was recorded. Anatomic landmarks useful for finding the SLNs for each dog were recorded. When > 1 SLN was identified, the largest lymph node was measured. Location of the SLNs with respect to the sternebrae (position) was recorded as a number by use of a method described elsewhere.8 Briefly, the location of the maximum height of the SLNs was recorded as the corresponding sternebra number, with a decimal corresponding to the location within that sternebra. For example, an SLN with a maximum height at the midpoint of the second sternebral segment would have the location recorded as a value of 1.5.
Dimensions of SLNs were measured. This included length (defined as the maximum distance from the cranial border to the caudal border of the SLN as measured on sagittal images), height (defined as the distance from the dorsal to the ventral border of the SLN), and width (defined as the distance from the medial to the lateral border of the SLN); both height and width were measured on transverse images (Figure 1). To account for differences in body size, the T4 vertebral body length was measured in each dog, and ratios of SLN length, height, and width to T4 length were calculated. All measurements were recorded in millimeters. Additional calculated values included SLN volume and short-to-long axis ratios, including the height-to-length ratio, height-to-width ratio, and width-to-length ratio.9,10 Volume was approximated by use of the equation for the volume of an ellipse as follows: volume = 4/3π × (length/2) × (width/2) × (height/2).
Postcontrast CT images in a soft tissue algorithm (window width, 320 HU; window level, 30 HU) of the SLN of a healthy adult purpose-bred research dog. The white bar in each panel indicates an SLN dimension. A—The width is indicated; width was defined as the maximum distance from the medial to the lateral border of an SLN on a transverse image. B—The height is indicated; height was defined as the maximum distance from the dorsal to the ventral border of an SLN on a transverse image. C—The length is indicated; length was defined as the maximum distance from the cranial border to the caudal border of an SLN on a sagittal image. Notice that the SLN in this figure has an oval shape as defined by the short-to-long axis ratio (value ≤ 0.5).
Citation: American Journal of Veterinary Research 78, 3; 10.2460/ajvr.78.3.289
Attenuation and contrast enhancement of SLNs on CT images were assessed as described elsewhere for other lymph nodes.9,11 Attenuation was determined by placing a circular or oval region of interest on the SLN of interest and recording the mean HU value on precontrast and postcontrast images. The region of interest was as large as possible while remaining within the SLN, was in the same location on precontrast and postcontrast images, and excluded the hilus. Heterogeneity of SLNs was quantified as the SD of the HU value within the region of interest on the precontrast and postcontrast images.11,12 The degree of contrast medium enhancement was calculated as the difference in HU between precontrast and postcontrast images. Contrast medium enhancement heterogeneity was defined as the difference between precontrast and postcontrast heterogeneity.
Presence of a lymph node hilus, hilar fat, and perinodal fat was determined as described elsewhere.9 A hilus was defined as a focal fat-attenuating to hypoattenuating region in the middle to cranial aspect of a lymph node. Specifically, attenuation of this region was fat attenuating (−1 to −100 HU) or soft tissue attenuating (HU values that were positive but less than the HU values of the surrounding lymph node parenchyma). Whether a hilus was fat attenuating or hypoattenuating was determined by use of a point-of-interest tool. When a negative HU value was detected within the hilus, it was recorded as fat attenuating. The presence or absence of perinodal fat separating the lymph node from the surrounding cranial mediastinal structures was recorded.
Statistical analysis
Continuous data were reported as median, IQR, and range, whereas categorical data were reported as the absolute number of dogs. Statistical analyses were performed by use of a commercially available computer software package.e Precontrast and postcontrast mean attenuation was compared by use of a 2-tailed paired t test. Precontrast and postcontrast mean heterogeneity was compared by use of a nonparametric Mann-Whitney U test. Parametric Pearson and nonparametric Spearman correlation coefficients were calculated between body weight, T4 vertebral body length, and SLN dimensions. Intraobserver variability was assessed by calculation of intraclass correlation coefficient values for SLN length, height, width, precontrast and postcontrast attenuation, and position. The Cohen unweighted κ was calculated for presence of a lymph node hilus and whether it was fat attenuating. Significance for all tests was set at values of P < 0.05.
Results
By use of contrast-enhanced CT, SLNs were identified in all 12 dogs. The SLNs were identified in the cranial mediastinum of the ventral aspect of the thorax, cranial to the heart, and dorsal to the sternebral segments. Two SLNs were identified in 10 dogs, and only 1 SLN was identified in 2 dogs. Perinodal fat was seen in all 12 dogs, and it provided a partial to complete separation of the SLNs from the surrounding cranial mediastinal structures. The most useful anatomic landmarks for locating the SLNs were the internal thoracic artery and vein and the second sternebra. Median position of SLNs was 1.3 (IQR, 1.1 to 1.6; range, 0.9 to 1.8), which indicated a location dorsal to the midbody of the second sternebrae. All SLNs had a visible hilus, which was fat attenuating in 8 dogs and hypoattenuating in 4 dogs.
Objective measurements of the largest SLN for each of the 12 study dogs were summarized (Table 1). Median dimensions of SLNs were 8.5 mm (IQR, 7.0 to 12.3 mm; range, 4.0 to 22.0 mm) for length, 6.0 mm (IQR, 4.3 to 7.0 mm; range, 3.0 to 10.0 mm) for height, and 5.0 mm (IQR, 4.0 to 6.5 mm; range, 3.0 to 10.0 mm) for width, which resulted in a calculated volume of 148 mm3 (IQR, 74 to 250 mm3; range, 38 to 484 mm3). Volume was highly correlated with SLN length (Pearson r = 0.95; P < 0.001), less highly correlated with SLN width (Pearson r = 0.65; P = 0.02), and not correlated with SLN height (Pearson r = 0.28; P = 0.38). Of the 3 measured SLN dimensions, length was the largest dimension for 10 of 12 dogs, and width or height was the smallest dimension for 7 and 5 dogs, respectively. For 9 dogs, the shortest-to-longest ratio (height-to-length ratio or width-to-length ratio for each SLN, depending on which was the lower value) was ≤ 0.5, which indicated an oval shape, whereas the remaining 3 dogs had a ratio of 0.56, 0.57, and 0.71, respectively, which indicated a circular shape (Figure 2). Median shortest-to-longest ratio for the 12 dogs was 0.41 (IQR, 0.33 to 0.51; range, 0.27 to 0.71). The T4 vertebral body length was highly correlated with body weight (Pearson r = 0.94; P < 0.001) as well as SLN height (Pearson r = 0.76; P = 0.004) and SLN volume (Pearson r = 0.64; P = 0.03). Body weight was highly correlated with SLN height (Pearson r = 0.72; P < 0.001). Other correlations were not significant. Similar correlation coefficients were obtained nonparametrically by use of Spearman correlation (data not shown).
Postcontrast transverse (A) and sagittal (B) CT images in a soft tissue algorithm (window width, 320 HU; window level, 30 HU) of the SLN of a healthy adult purpose-bred research dog. Notice the circular shape of the lymph node (arrow), which has a short-to-long axis ratio > 0.5.
Citation: American Journal of Veterinary Research 78, 3; 10.2460/ajvr.78.3.289
Objective measurements for SLNs obtained by use of CT for 12 healthy adult purpose-bred research dogs.
| Variable | Median | IQR | Range |
|---|---|---|---|
| Body weight (kg) | 23.7 | 10.4 to 25.6 | 9.5 to 26.0 |
| T4 vertebral body length (mm) | 17.0 | 12.0 to 17.8 | 12.0 to 19.0 |
| SLN length (mm) | 8.5 | 7.0 to 12.3 | 4.0 to 22.0 |
| SLN length-to-T4 length ratio | 0.64 | 0.50 to 0.76 | 0.24 to 1.22 |
| SLN height (mm) | 6.0 | 4.3 to 7.0 | 3.0 to 10.0 |
| SLN height-to-T4 length ratio | 0.37 | 0.33 to 0.42 | 0.25 to 0.53 |
| SLN width (mm) | 5.0 | 4.0 to 6.5 | 3.0 to 10.0 |
| SLN width-to-T4 length ratio | 0.29 | 0.24 to 0.41 | 0.19 to 0.67 |
| SLN volume (mm3) | 123 | 74 to 250 | 38 to 484 |
| SLN height-to-SLN length ratio | 0.57 | 0.50 to 0.81 | 0.27 to 1.75 |
| SLN width-to-SLN length ratio | 0.51 | 0.34 to 0.70 | 0.31 to 1.25 |
| SLN height-to-SLN width ratio | 1.27 | 0.78 to 1.85 | 0.50 to 2.50 |
| SLN precontrast attenuation (HU) | 18.3 | 9.7 to 29.9 | 4.4 to 36.9 |
| SLN precontrast heterogeneity (HU) | 6.8 | 5.7 to 8.2 | 4.0 to 14.2 |
| SLN postcontrast attenuation (HU) | 41.3 | 27.3 to 63.6 | 24.0 to 77.4 |
| SLN postcontrast heterogeneity (HU) | 9.0 | 4.9 to 11.4 | 3.6 to 21.7 |
| CME (HU) | 22.4 | 15.5 to 33.7 | 8.3 to 64.2 |
| CME heterogeneity (HU) | 1.4 | −2.0 to 3.8 | −5.4 to 14.8 |
CME = Contrast medium enhancement, which was defined as the change in HU from precontrast to postcontrast images.
All SLNs were soft tissue attenuating before administration of contrast medium and mildly contrast enhancing after IV administration of contrast medium (Figure 3; Table 1). Attenuation values increased significantly (P < 0.001) between precontrast and postcontrast images (median, 18.3 and 41.3 HU, respectively). The SLN parenchyma was homogeneous before and after contrast medium administration, with no significant change in the amount of SLN heterogeneity (median, 6.8 and 9.0 HU, respectively). No hypoattenuating, non-contrast-enhancing regions were observed in any SLN.
Precontrast (A) and postcontrast (B) transverse CT images in a soft tissue algorithm (window width, 320 HU; window level, 30 HU) of the SLN of a healthy adult purpose-bred research dog. Notice the uniform and mild contrast enhancement (arrow) and an adjacent internal thoracic vein (asterisk).
Citation: American Journal of Veterinary Research 78, 3; 10.2460/ajvr.78.3.289
Intraobserver variability was low. Intraclass correlation coefficient values for SLN length, height, width, precontrast attenuation, postcontrast attenuation, and position were 0.98, 0.92, 0.93, 0.87, 0.93, and 0.95, respectively. Cohen unweighted κ values for determining whether an SLN hilus was present and whether it was fat attenuating were 1.0 and 0.82, respectively.
Discussion
The primary objective of the present study was to determine quantifiable measurements of SLN dimensions and attenuation in healthy adult dogs by use of CT. The resultant data may provide reference values for the CT appearance of SLNs in healthy dogs. The reference values and data ranges reported in the present study may be compared with those for abnormal SLNs and provide objective variables for veterinarians interpreting thoracic CT images.
The SLN imaging characteristics described here shared some similarities with those of other lymph nodes, with one notable difference being that SLNs of the healthy dogs were sometimes circular. Frequently, these lymph nodes were rotated on their axis such that the smallest dimension was width (7/12 dogs) and height (5/12 dogs). For mandibular, superficial cervical, axillary, inguinal, and popliteal lymph nodes of clinically normal dogs and medial retropharyngeal lymph nodes of clinically normal cats, the short-to-long axis ratio is < 0.5.9,13 Some of the SLNs of the dogs in the present study were more circular, as reflected by the fact that the short-to-long axis ratio was > 0.5 in 2 dogs and > 0.7 in a third dog, with a median of 0.41 and range of 0.27 to 0.71. In humans, a short-to-long axis ratio > 0.5 is associated with malignancy in cervical lymph nodes, whereas in dogs, a value > 0.7 is useful for differentiating between benign and malignant lymphadenopathy of the head and neck.13,14 For tracheobronchial lymph nodes, a transverse maximum diameter of 12 mm or node-to-thoracic body ratio of 1.05 for CT images has been proposed as a cutoff to identify lymph node metastasis.4 On the basis of results for the study reported here, caution should be exercised in the use of these cutoff values for short-to-long axis ratios of canine SLNs because these values may be greater than the values for other lymph nodes, given that 3 of 12 healthy dogs in the present study had a short-to-long axis ratio > 0.5.
Consistent with CT findings for the present study, previous studies15–17 conducted with ultrasonography have revealed that lymph node size is proportional to body weight in dogs. Analysis of data for the present study also revealed strong correlations between T4 vertebral body length and SLN dimensions. These results suggested that T4 vertebral body length may provide a reference measure for normalizing objective SLN measurements among dogs of various sizes and allow for more accurate detection of abnormal SLN size. For example, investigators of another study4 found that an increased ratio of tracheobronchial lymph node transverse diameter, compared with the ventrodorsal diameter of a thoracic vertebral body within 1-mm slices of the lymph node, was associated with disease states in dogs. A larger number of dogs with a wider range of body weights would be required to validate these relationships.
In the present study, all 12 dogs had homogenous contrast medium enhancement of the SLNs. Investigators of a study4 conducted to evaluate CT of the tracheobronchial lymph nodes in dogs reported that dogs without thoracic disease had homogenous contrast enhancement or a lack thereof (7/9 and 2/9 dogs, respectively). In that study,4 postcontrast images were acquired 5 to 20 minutes after manual IV administration of an ionic iodinated contrast medium, whereas in the study reported here, images were obtained 1 minute after IV injection of a nonionic iodinated contrast medium with a power injector. We suggest that differences between that previous study4 and the present study may be attributable to differences in these factors. Characterizing the magnitude and homogeneity of CT contrast enhancement may be important for detecting abnormal lymph nodes, which has been determined for medial retropharyngeal lymph nodes of cats with neoplasia.11
In the present study of healthy dogs, most CT-based SLN dimensions were less than the reported detection limits for radiography.1,8,18 Furthermore, CT can be used to assess the presence of a hilus, hilar fat, perinodal fat, contrast-enhancement patterns, and parenchymal heterogeneity, whereas none of these variables can be evaluated by use of radiography. This is important because changes to these nodal imaging characteristics are associated with reactive or metastatic lymphadenopathy.4,6,11,19,20 Collectively, results of the present study provided clinicians with CT imaging characteristics of SLNs in heathy dogs, many of which have been reported for other intrathoracic lymph nodes.4–6,8,19,20
The most important limitations for the present study were the modest number of dogs and lack of histologic confirmation of no abnormalities in the SLNs. Dogs used in this study were healthy adult purpose-bred research dogs, so the likelihood that occult disease affected the SLNs was relatively low. All dogs were thoroughly evaluated through a combination of a complete physical examination, CBC, serum biochemical analysis, arterial blood gas analysis, echocardiography, and ECG, with all results being within reference limits. Urinalyses and microbial cultures were not performed and may have potentially revealed a source of occult disease in the study population, although none of the dogs had clinical signs of lower urinary tract disease throughout the study period. Regardless, the possibility that diseased SLNs were included in the study remains a consideration. Future larger studies are warranted to determine whether the findings for the present study remain valid for a larger population.
Only 1 board-certified veterinary radiologist evaluated all CT images. However, in other studies9,11 conducted by our laboratory group, we established that there was a good correlation between the reviewer (SN) and other radiologists for assessing lymph node images, which suggests that other veterinary radiologists would obtain similar measurements. Furthermore, the use of a single reviewer allowed for maximizing consistency in measurements, as reflected by the fact that the intraclass correlation indicated low variability between measurements obtained at 6-month intervals.
Finally, measurements of SLN dimensions were derived from postcontrast images. This potentially could have yielded values that would have differed slightly from values obtained on precontrast images.
Data obtained for the present study may serve as baseline reference values for the CT appearance of SLNs in healthy dogs. All SLNs were readily identified and measured in 3 dimensions and were homogenously contrast-medium enhancing. Two SLNs were seen during CT of most dogs (10/12). The values and data ranges reported here may provide veterinarians interpreting thoracic CT images with objective information for the evaluation of SLNs, which should facilitate comparisons with results for abnormal SLNs in future studies.
ABBREVIATIONS
| HU | Hounsfield unit |
| IQR | Interquartile range (25th to 75th percentile) |
| SLN | Sternal lymph node |
Footnotes
Toshiba Aquilion, Tochigi, Japan.
Isovue (300 mg of I/mL), Bracco Diagnostics Inc, Princeton, NJ.
E-Z EM, Bracco Diagnostics Inc, Monroe Township, NJ.
Merge E-film software, Merge Healthcare, Milwaukee, Wis.
GraphPad Prism, version 6.07 for Windows, GraphPad Software Inc, San Diego, Calif.
References
1. Bezuidenhout A. The lymphatic system. In: Evans HE, ed. Miller's anatomy of the dog. 3rd ed. Philadelphia: WB Saunders Co, 1993;730–732.
2. Shibata SJ, Hiramatsu Y, Kaseda M, et al. The time course of lymph drainage from the peritoneal cavity in Beagle dogs. J Vet Med Sci 2006; 68:1143–1147.
3. De Swarte M, Alexander K, Rannou B, et al. Comparison of sonographic features of benign and neoplastic deep lymph nodes in dogs. Vet Radiol Ultrasound 2011; 52:451–456.
4. Ballegeer EA, Adams WM, Dubielzig RR, et al. Computed tomography characteristics of canine tracheobronchial lymph node metastasis. Vet Radiol Ultrasound 2010; 51:397–403.
5. Fina C, Vignoli M, Terragni R, et al. Computed tomographic characteristics of eosinophilic pulmonary granulomatosis in five dogs. Vet Radiol Ultrasound 2014; 55:16–22.
6. Reetz JA, Buza EL, Krick EL. CT features of pleural masses and nodules. Vet Radiol Ultrasound 2012; 53:121–127.
7. Nemanic S, London CA, Wisner ER. Comparison of thoracic radiographs and single breath-hold helical CT for detection of pulmonary nodules in dogs with metastatic neoplasia. J Vet Intern Med 2006; 20:508–515.
8. Smith K, O'Brien R. Radiographic characterization of enlarged sternal lymph nodes in 71 dogs and 13 cats. J Am Anim Hosp Assoc 2012; 48:176–181.
9. Nemanic S, Nelson NC. Ultrasonography and noncontrast computed tomography of medial retropharyngeal lymph nodes in healthy cats. Am J Vet Res 2012; 73:1377–1385.
10. Ying M, Pang BSF. Three-dimensional ultrasound measurement of cervical lymph node volume. Br J Radiol 2009; 82:617–625.
11. Nemanic S, Hollars K, Nelson NC, et al. Combination of computed tomographic imaging characteristics of medial retropharyngeal lymph nodes and nasal passages aids discrimination between rhinitis and neoplasia in cats. Vet Radiol Ultrasound 2015; 56:617–627.
12. Uosyte R, Shaw DJ, Gunn-Moore DA, et al. Effects of fluid and computed tomographic technical factors on conspicuity of canine and feline nasal turbinates. Vet Radiol Ultrasound 2015; 56:494–502.
13. Nyman HT, Kristensen AT, Skovgaard IM, et al. Characterization of normal and abnormal canine superficial lymph nodes using gray-scale B-mode, color flow mapping, power, and spectral Doppler ultrasonography: a multivariate study. Vet Radiol Ultrasound 2005; 46:404–410.
14. Steinkamp HJ, Cornehl M, Hosten N, et al. Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol 1995; 68:266–270.
15. Burns GO, Scrivani PV, Thompson MS, et al. Relation between age, body weight, and medial retropharyngeal lymph node size in apparently healthy dogs. Vet Radiol Ultrasound 2008; 49:277–281.
16. Agthe P, Caine AR, Posch B, et al. Ultrasonographic appearance of jejunal lymph nodes in dogs without clinical signs of gastrointestinal disease. Vet Radiol Ultrasound 2009; 50:195–200.
17. Mayer MN, Lawson JA, Silver TI. Sonographic characteristics of presumptively normal canine medial iliac and superficial inguinal lymph nodes. Vet Radiol Ultrasound 2010; 51:638–641.
18. Kirberger RM, Avner AVI. The effect of positioning on the appearance of selected cranial thoracic structures in the dog. Vet Radiol Ultrasound 2006; 47:61–68.
19. Fukuda S, Kobayashi T, Robertson ID, et al. Computed tomographic features of canine nonparenchymal hemangiosarcoma. Vet Radiol Ultrasound 2014; 55:374–379.
20. Tsai S, Sutherland-Smith J, Burgess K, et al. Imaging characteristics of intrathoracic histiocytic sarcoma in dogs. Vet Radiol Ultrasound 2012; 53:21–27.
