The medial retropharyngeal lymph nodes are the largest lymph nodes in the head and cranial cervical region of animals. These lymph nodes receive lymph via afferent lymphatic ducts from the oropharynx, nasopharynx, oral cavity, tongue, salivary glands, deep aspects of the external ear, larynx, and esophagus.1 They also indirectly receive lymph from other structures in the head via efferent lymphatics of the mandibular and parotid lymph nodes. Afferent lymphatic ducts to medial retropharyngeal lymph nodes can cross the midline of the head in domestic animals and humans, as indicated by results of studies2–4 in which metastatic disease was found in medial retropharyngeal lymph nodes contralateral to oral neoplasms. Medial retropharyngeal lymph nodes are located ventral to the tympanic bullae and first 2 cervical vertebrae, medial to mandibular salivary glands, and lateral to the common carotid arteries and trachea. These lymph nodes are only palpable in humans, dogs, and cats when they are severely enlarged, necessitating diagnostic imaging for noninvasive evaluation.5,6
In humans, diseases causing reactive or metastatic lymphadenopathy in the head and cervical region alter the size and shape of lymph nodes, causing them to become enlarged and round rather than fusiform. Such changes in size and shape can be objectively evaluated via determination of the ratio of the short-axis length to the long-axis length of a lymph node. Results of a study7 in which lymphadenopathy in the cervical region of humans was investigated indicate a short-axis–to–long-axis length ratio > 0.5 is predictive of metastatic lymphadenopathy with a diagnostic accuracy of 95%. For dogs, a lymph node short-axis–to–long-axis length ratio > 0.7 is useful for distinguishing between malignant and benign mandibular and superficial cervical lymph nodes; however, up to 23% of lymph nodes are incorrectly classified when multiple imaging characteristics, including lymph node size, distribution of blood flow, and pulsatility index, are used simultaneously.8 To the authors' knowledge, similar studies have not been conducted to determine criteria for evaluation of lymph nodes in the head and cervical region of cats.
Metastases in lymph nodes may extend into perinodal tissues, resulting in altered ultrasonographic appearance of lymph node margins and loss of adjacent fat.8,9 Neoplastic infiltration of lymph nodes may be associated with heterogeneity of parenchyma and loss of a central hyperattenuating hilus in ultrasonographic images.10–12 Results of recent studies indicate medial retropharyngeal lymph nodes in cats with fungal rhinitis13 and oral squamous cell neoplasia14 are subjectively enlarged and heterogeneously contrast enhancing in CT images; however, the degree of lymph node enlargement was not quantified in these studies.
Medial retropharyngeal lymph node measurements for healthy dogs have been determined via ultrasonography,15 CT,16 and MRI.17 However, to the authors' knowledge, no such studies have been conducted to determine medial retropharyngeal lymph node measurements for cats. An objective of the study reported here was to determine various measurements of medial retropharyngeal lymph nodes in healthy adult cats via CT and ultrasonography. Another objective was to determine whether lymph node size differs on the basis of imaging modality or age, sex, body weight, or body condition score of cats. The hypothesis was that medial retropharyngeal lymph nodes of cats would have similar sizes as determined with ultrasonography or CT. If this hypothesis was supported, then direct comparisons could be made between lymph node measurements acquired with either imaging modality during monitoring of cats for disease progression or response to treatment.
Materials and Methods
Animals—By use of commercially available software,a a sample size of 35 cats was determined to be sufficient to achieve adequate statistical power. Values for this power calculation were obtained via measurement of medial retropharyngeal lymph nodes in CT images of cats that were patients at the Veterinary Teaching Hospital at the Michigan State University College of Veterinary Medicine and from mlrp published data for medial retropharyngeal lymph nodes in clinically normal dogs determined via ultrasonography.15 Adult cats (2 to 8 years old) owned by students and staff of the hospital were recruited for inclusion in the study from September 21 through December 4, 2009.
Forty-five cats were screened for inclusion in the study because it was expected that some cats would not meet criteria for inclusion. Physical and oral examinations, serologic testing to detect FeLV and FIV infection, a CBC, and serum biochemical analyses were performed for each cat. Age, sex, breed, body weight, and body condition score of each cat were recorded. Cats were required to have unremarkable results of physical examination for inclusion in the study. Cats with positive results for FeLV or FIV were excluded from the study. Cats with CBC variable values outside the reference ranges (other than results indicative of a stress leukogram) or with serum biochemical analysis results indicative of systemic disease were also excluded from the study. During oral examinations, gingival index scores18 (0 [no gingivitis], 1 [mild gingival inflammation, redness, and edema with no bleeding], 2 [moderate gingival inflammation, redness, and edema with some bleeding], or 3 [severe gingival inflammation, edema, and ulceration with profuse bleeding and periodontal pocketing]) were determined; the worst gingival index score determined for any tooth in a cat was the score assigned for that cat. Cats with a grade 2 or 3 gingival index score were excluded from the study. Cats with evidence of rhinitis, neoplasia, periodontal disease, tooth root abscess, tooth fracture, or odontoclastic resorptive lesions in CT images were excluded. Owner consent was obtained for participation of cats in the study. The Institutional Animal Care and Use Committee of Michigan State University approved the study.
CT and ultrasonography—For CT, cats were sedated with dexmedetomidine (0.01 mg/kg, IM) and butorphanol (0.3 mg/kg, IM). Each cat was positioned in sternal recumbency, and its head was placed on a pad so that the neck was in a neutral position (ie, not flexed or extended) and the hard palate was perpendicular to the CT gantry and scan plane. For each cat, CT was performed to obtain images from the rostral aspect of the nose to the midcervical region. A 16-slice helical CT scannerb with a low-pass filter, slice thickness of 0.625 mm, collimator pitch of 0.9375, matrix of 512 × 512, peak kilovoltage of 120 kVp, and amperage of 120 to 430 mA was used. Voxel size was approximately 0.2 × 0.2 × 0.625 mm. Landmarks for identification of medial retropharyngeal lymph nodes in CT images were mandibular salivary glands, common carotid arteries, and tympanic bullae (Figure 1).
Representative transverse CT image (window width, 300; window level, 30) of the head of a healthy cat (A) and parasagittal long-axis power Doppler ultrasonographic image of the cervical region of a healthy cat (B) indicating anatomic landmarks for identification of medial retropharyngeal lymph nodes. In panel A, the left medial retropharyngeal lymph node is surrounded by a dashed white line, and left and right medial retropharyngeal lymph nodes are labeled with white arrows. The right mandibular salivary gland (asterisk), common carotid artery (white triangle), and tympanic bulla (black triangle) are visible. In panel B, a medial retropharyngeal lymph node (white arrow), mandibular salivary gland (asterisk), and common carotid artery (white triangle) are visible. Notice the power Doppler ultrasonographic signal in the common carotid artery. The gray scale on the left side of the ultrasonographic image represents the range of strength of the power Doppler signal. The numbers in the scale on the right side of the ultrasonographic image indicate centimeters.
Citation: American Journal of Veterinary Research 73, 9; 10.2460/ajvr.73.9.1377
Immediately following CT, ultrasonographyc was performed separately by each of the authors to obtain images of the right and left medial retropharyngeal lymph nodes of each cat. The order in which authors performed ultrasonography for each cat and the lymph node that was imaged first during each ultrasonographic examination were determined via a randomization procedure (coin toss). Ultrasonography was performed with a linear transducer operating at 14 MHz with the harmonic mode activated. Each cat was positioned in dorsal recumbency in a padded v-shaped trough with its head and neck extended. Hair was clipped over the region of interest, and transducer contact with skin was maintained with alcohol or acoustic coupling gel. Left and right medial retropharyngeal lymph nodes were identified in parasagittal plane long-axis views; the transducer was rotated slightly until the length of the lymph node on the image was maximized. Short-axis images were obtained by rotating the transducer approximately 90° to the plane used to acquire long-axis images. Short-axis images were acquired at the rostral, middle, and caudal aspects of each lymph node. Therefore, 4 images were acquired by each author for each lymph node (long-axis and rostral, middle, and caudal short-axis images). After ultrasonography, an oral examination was performed for each cat, and effects of dexmedetomidine were reversed with atipamazole (0.3 mg/kg, IM).
Image analysis—Distances on CT and ultrasonographic images were automatically calibrated by the imaging machines and via image archiving and communication systemd software. Each author determined measurements for left and right lymph nodes (7 measurements/lymph node) in CT and ultrasonographic images of each of the 38 cats (total number of measurements, 2,128). Each author only evaluated images they had acquired (ie, authors did not evaluate images acquired by the other author). Measurements were performed with an electronic caliper tool by use of imaging software.d Each author measured the height and width of the rostral, middle, and caudal aspects and the length of each lymph node in ultrasonographic images (Figure 2). Each author measured the width and height of the rostral, middle, and caudal aspects of each lymph node in transverse CT images (Figure 3). Width was defined as the distance from the medial to the lateral border of a lymph node, and height was defined as the distance from the dorsal to the ventral border of a lymph node. Each author independently decided which CT image best represented rostral, middle, and caudal aspects of lymph nodes. Length of lymph nodes in CT images was determined via 2 methods; length was calculated by multiplying the number of transverse images in which the lymph node could be observed by the CT slice thickness (calculated CT lymph node length) and was measured in sagittally reformatted CT images with an electronic caliper tool by use of imaging softwared (measured CT lymph node length). Sagittally reformatted images were generated with commercially available softwaree; lymph node margins were easily identified in these images despite the fact that the voxels were not isotropic. Rostral and caudal limits of lymph nodes were determined via comparison of transverse and reformatted CT images. Lymph node length measurements were determined by use of ultrasonographic images acquired in a plane that resulted in the maximal long-axis lymph node length. Attenuation (Hounsfield units) was determined by placing a circular region of interest over the largest portion of each medial retropharyngeal lymph node in transverse CT images; the mean attenuation value for all voxels within the region of interest was recorded. Ratios of the rostral, middle, and caudal lymph node widths to the lymph node length (ie, short-axis–to–long-axis length ratios) were determined for images acquired by use of each modality. The measured CT lymph node length was used for calculation of CT lymph node short-axis–to–long-axis length ratios.
Long-axis (A), longitudinal (B), and transverse short-axis (C) ultrasonographic images of medial retropharyngeal lymph nodes in representative healthy cats indicating electronic caliper tool placement for measurement of various dimensions. Notice electronic caliper tool placement for measurement of medial retropharyngeal lymph node length (arrow; A); rostral, middle, and caudal heights (left, middle, and right arrows, respectively; B); and rostral height (solid arrow) and width (dashed arrow; C). Notice moderate heterogeneity of lymph node parenchyma. The gray scale on the left side of the ultrasonographic image represents echogenicity (white represents hyperechogenicity and black represents hypoechogenicity). See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 73, 9; 10.2460/ajvr.73.9.1377
Representative transverse CT image (window width, 150; window level, 30) of the head of a healthy cat indicating electronic caliper tool placement for measurement of medial retropharyngeal lymph node height (solid arrow) and width (dashed arrow; A) and placement of region of interest (circle; B [same CT image as in panel A]) for measurement of medial retropharyngeal lymph node attenuation. In this CT image, attenuation is 44.3 Hounsfield units. Notice fat surrounding the ventral, medial, dorsal, and dorsolateral aspects of medial retropharyngeal lymph nodes. Also notice mild to moderate heterogeneity of lymph node parenchyma, which is similar to that of the adjacent longus coli muscle.
Citation: American Journal of Veterinary Research 73, 9; 10.2460/ajvr.73.9.1377
Each lymph node was subjectively evaluated in CT and ultrasonographic images by both authors to determine parenchymal heterogeneity; ratings of none, mild, moderate, or severe were assigned. For ultrasonographic images, heterogeneity was defined as multiple hypoechoic and hyperechoic regions in close proximity that had greater differences in echogenicity than were observed in surrounding tissues. For CT images, heterogeneity was defined as differences in attenuation in adjacent regions of parenchyma that were greater than those observed in the longus coli muscle.
The presence of a hilus in each lymph node was determined by consensus between the authors for each imaging modality. For ultrasonographic images, a hilus was defined as a hyperechoic band within the parenchyma of a lymph node. For CT images, a lymph node hilus was initially defined as a fat-attenuating structure (as determined by a slightly negative Hounsfield unit region-of-interest value) entering a lymph node from the periphery (Figure 4). Because initial results indicated a large difference between imaging modalities regarding numbers of lymph nodes with a hilus, CT images were reevaluated by use of an alternate definition for a lymph node hilus (lymph node hilus defined as a focal hypoattenuating region in the middle to rostral aspect of the lymph node [whether fat attenuating or not]).
Representative transverse CT images (window width, 400; window level, 90) of the heads of 2 different healthy cats (panels A and B) and a representative transverse short-axis ultrasonographic image of the cervical region of the same cat as in panel A (C). In panel A, notice the hilus in each medial retropharyngeal lymph node (white arrows), identified as fat-attenuating structures entering the medial aspects of the lymph nodes and extending dorsally and ventrally in the centers of the lymph node parenchyma. In panel B, notice the rounded hypoattenuating regions in the centers of the rostral aspects of both medial retropharyngeal lymph nodes (white arrows). In panel C, notice a medial retropharyngeal lymph node hilus (white arrow), identified as a broad hyperechoic region in the center of the lymph node parenchyma. See Figures 1 and 2 for remainder of key.
Citation: American Journal of Veterinary Research 73, 9; 10.2460/ajvr.73.9.1377
Appearance of lymph node margins was determined by consensus between the authors. Ultrasonographic and CT images were evaluated to determine whether lymph node margins were smooth or irregular. By use of CT images, amount of fat separating a lymph node from the digastricus muscle (dorsally), sternocephalicus muscle (laterally), and common carotid artery (medially) was determined. Amount of fat was rated as small if there was insufficient fat to separate the margins of adjacent structures, moderate if fat was observed along the dorsolateral to ventromedial aspects of a lymph node, and large if fat completely encircled a lymph node.
Statistical analysis—Medial retropharyngeal lymph node length, width, and height were analyzed via a split-plot ANOVA with sex as a grouping factor and imaging modality and the 7 lymph node measurements as repeated factors.f A split-plot ANOVA was also performed with side (right vs left) as a grouping factor and the 7 lymph node measurements as repeated factors. Values of P < 0.05 were considered significant. If a significant interaction was detected between factors, a post hoc Bonferroni test was performed for each of the 7 lymph node measurement locations with a corrected P value (P = 0.05/number of post hoc comparisons). Post hoc analyses were not performed if the interaction between factors was not significant. A paired t test was performed to compare short-axis–to–long-axis length ratios calculated by use of lymph node width measurements at rostral, middle, and caudal locations in CT and ultrasonographic images; mean values of ratios determined by both authors for left and right lymph nodes were used for the analysis.
Pearson correlation coefficients were calculated to determine linear associations between medial retropharyngeal lymph node volume and age, body weight (in kilograms), and body condition score of cats. Lymph node volume was calculated by the formula for the volume of an ellipsoid ([4/3π] × length × rostral width × rostral height),19 by use of the mean values determined by both authors for left and right lymph nodes. Pearson correlation coefficients were also calculated to determine associations between Hounsfield unit values and body weight (in kilograms) and body condition score of cats. Values of P < 0.05 were considered significant.
Interobserver variability was determined via calculation of an intraclass correlation coefficient, which classified correlations as slight (intraclass correlation coefficient, 0 to 0.2), fair (0.2 to 0.4), moderate (0.4 to 0.6), substantial (0.6 to 0.8), or almost perfect (0.8 to 1.0).20 The purpose of this test was to determine the reliability of quantitative measurements determined by different observers. The analysis was performed by use of the means of values determined by both authors for left and right lymph nodes. An overall correlation coefficient was determined by calculation of the mean of the correlation coefficients of each measured lymph node location with each imaging modality (CT and ultrasonography).
Results
Six of 45 cats screened for inclusion in the study were excluded because of positive results for FIV (n = 1) or FeLV (1), heart murmur detected via physical examination (1), or patient noncompliance at the time of collection of blood samples for hematologic analysis (3). Thirty-nine cats underwent CT and ultrasonography. Data for 1 cat were excluded from analysis because that cat had evidence of nasal disease in CT images. Therefore, data for 38 cats (20 neutered males and 18 spayed females) were included in the study. Mean age of these cats was 4.6 years (range, 2 to 8 years). Mean body weight was 5.3 kg (range, 2.9 to 8.1 kg). Mean body condition score was 5.3 out of 9 (range, 4 to 8 out of 9). Breeds included domestic shorthair (n = 29), domestic medium-hair (4), domestic longhair (2), Maine Coon (1), Ragdoll (1), and Siamese (1).
Means, SDs, and ranges for medial retropharyngeal lymph node measurements and short-axis–to–long-axis length ratios determined by use of CT and ultrasonographic images were summarized (Table 1). Mean dimensions (length × rostral height × rostral width) of medial retropharyngeal lymph nodes were 20.7 × 12.4 × 3.7 mm and 20.7 × 13.1 × 4.7 mm in ultrasonographic and CT images, respectively. Maximum medial retropharyngeal lymph node dimensions (length × rostral height × rostral width) determined by use of either imaging modality were approximately 32 × 20 × 7 mm. There was a significant (P < 0.001) difference between imaging modalities for most lymph node measurements (rostral width, middle height and width, and caudal height) and short-axis–to–long-axis length ratios. The differences between CT and ultrasonographic width and height measurements were small (0.2 to 2.1 mm). Lymph node measurements were significantly larger as determined with CT images at some lymph node locations and significantly larger as determined with ultrasonographic images at other lymph node locations (interaction of lymph node measurement location and imaging modality, P < 0.001). Results of Bonferroni post hoc analysis with a corrected value of P < 0.002 indicated measurements obtained by use of CT images were significantly larger than those obtained by use of ultrasonographic images for rostral width (P < 0.001), middle height (P < 0.001), and middle width (P < 0.001) of lymph nodes. Measurements obtained by use of ultrasonographic images were significantly larger than those obtained by use of CT images for caudal height (P < 0.001) of lymph nodes. There was no significant difference between imaging modalities for rostral height (P = 0.004), caudal width (P = 0.3), or length (P = 0.9) of lymph nodes. The mean medial retropharyngeal lymph node short-axis–to–long-axis length ratios were 0.23 (rostral), 0.21 (middle), and 0.15 (caudal) as determined by evaluation of CT images and were 0.18 (rostral), 0.16 (middle), and 0.13 (caudal) as determined by evaluation of ultrasonographic images. Significant differences were detected between imaging modalities for rostral (P < 0.001), middle (P < 0.001), and caudal (P = 0.005) short-axis–to–long-axis length ratios. Lymph node lengths determined by use of sagittally reformatted images (ie, measured CT lymph node lengths) were used for calculation of short-axis–to–long-axis length ratios.
Mean, SD, and range of various measurements and rostral-short-axis–to–long-axis length ratios of medial retropharyngeal lymph nodes of 38 healthy cats determined by use of CT and ultrasonography.
| Variable | CT | Ultrasonography | ||||
|---|---|---|---|---|---|---|
| Mean | SD | Range | Mean | SD | Range | |
| Rostral height (mm) | 13.1 | 2.4 | 7.2–19.0 | 12.4 | 3.0 | 6.0–20.0 |
| Rostral width (mm)* | 4.7 | 0.9 | 3.0–7.0 | 3.7 | 0.8 | 1.0–6.0 |
| Middle height (mm)* | 13.1 | 2.4 | 6.4–18.0 | 11.0 | 2.4 | 6.0–18.0 |
| Middle width (mm)* | 4.2 | 0.8 | 2.3–6.4 | 3.3 | 0.8 | 2.0–6.0 |
| Caudal height (mm)* | 7.8 | 2.1 | 2.8–14.0 | 8.9 | 2.0 | 4.0–13.0 |
| Caudal width (mm) | 2.9 | 0.6 | 1.7–4.8 | 2.7 | 0.8 | 1.0–7.0 |
| Length (mm) | 20.7 | 3.6 | 11.0–29.0 | 20.7 | 4.0 | 13.0–32.0 |
| Rostral-short-axis–to–long-axis length ratio*† | 0.23 | 0.04 | 0.16–0.34 | 0.18 | 0.02 | 0.14–0.24 |
Values were calculated by use of data determined by the 2 authors of the present study for left and right medial retropharyngeal lymph nodes. Width and height of the rostral, middle, and caudal aspects of each lymph node were determined by use of transverse ultrasonographic and CT images. Length of lymph nodes in ultrasonographic and CT images was determined by use of longitudinal and sagittally reformatted images, respectively.
Mean values are significantly (P < 0.05) different between imaging modalities.
Values are the ratio of the rostral width to the length of lymph nodes.
Mean lymph node length in ultrasonographic and sagittally reformatted CT images (ie, measured CT lymph node length) was 20.7 mm, whereas mean lymph node length determined via multiplication of the number of CT slices on which a lymph node was observed by CT slice thickness (ie, calculated CT lymph node length) was 14.8 mm. Lymph node length determined by use of ultrasonographic images was not significantly (P = 0.9) different from measured CT lymph node length; however, lymph node length determined by use of ultrasonographic images was significantly (P < 0.001) different from calculated CT lymph node length. This finding indicated length of the medial retropharyngeal lymph nodes determined by use of sagittally reformatted CT images was similar to that obtained by use of ultrasonographic images.
No significant (P = 0.9) differences were detected between measurements for left and right lymph nodes of cats determined by use of either imaging modality. Mean differences between left and right lymph node values were 2.6 mm (length), 1.8 mm (rostral height), and 0.65 mm (rostral width). Maximum differences between left and right lymph node values for any individual cat were 8.0 mm (measured length), 8.0 mm (rostral height), and 3.0 mm (rostral width).
No significant effects of sex of cat on lymph node volume (P = 0.16) were detected. Interactions between sex and lymph node location; sex and imaging modality; and sex, lymph node location, and imaging modality were not significant (P > 0.3). These results indicated there were no differences between male and female cats regarding medial retropharyngeal lymph node dimensions.
All lymph nodes were mildly to moderately heterogeneous in ultrasonographic (Figure 2) and CT (Figure 3) images. Lymph node heterogeneity in ultrasonographic images was typically characterized by marked differences in echogenicity between adjacent regions of lymph node parenchyma. In CT images, lymph node heterogeneity was characterized by mild variations in attenuation of adjacent regions of parenchyma; this attenuation was similar to that observed in longus coli muscles. In 1 cat, one of the medial retropharyngeal lymph nodes was bilobed, with a focally thin region in the middle aspect of the node in sagittal plane CT images. The mean ± SD Hounsfield unit value for medial retropharyngeal lymph nodes was 40.2 ± 5.3.
A lymph node hilus was identified in ultrasonographic images for 36 of 38 (95%) cats (Figure 4). A lymph node hilus was identified in CT images in only 9 of 38 (24%) cats by use of the initial definition (lymph node hilus defined as a fat-attenuating structure [as determined by a slightly negative Hounsfield unit region-of-interest value] entering a lymph node from the periphery). Some lymph nodes had a central region within the parenchyma that was hypoattenuating, compared with the rest of the lymph node parenchyma, but had a positive Hounsfield unit value. By use of the alternate definition (lymph node hilus defined as a focal hypoattenuating region in the center of the lymph node [whether fat attenuating or not]), a lymph node hilus was identified in CT images of an additional 26 of 38 (68%) cats. By use of these 2 definitions, a lymph node hilus was identified in CT images in 35 of 38 (92%) cats. No lymph nodes had both a fat-attenuating hilus and a hypoattenuating region.
Margins of lymph nodes were smooth in CT images for all cats. Lymph nodes were surrounded by a small amount of fat ventrally, medially, dorsally, and dorsolaterally (Figure 3). Fat was not observed along the lateral or rostrodorsal surfaces of lymph nodes in the cats; lymph nodes were in contact with the sternocephalicus and digastricus muscles, respectively, in these locations. Medially, medial retropharyngeal lymph node margins were indistinct because they were adjacent to a common carotid artery. Lymph node margins could be identified in regions lacking adjacent fat because lymph nodes were mildly hypoattenuating, compared with muscle and vascular structures. The amount of perinodal fat seemed to be similar among cats, regardless of body condition score. In ultrasonographic images for all cats, there was a thin hyperechoic band surrounding lymph nodes that separated lymph node parenchyma from perinodal tissues. The margins of this hyperechoic band were mildly irregular in shape in all cats (Figure 2).
A significant negative linear correlation was detected between volume of medial retropharyngeal lymph nodes and age of cats for both imaging modalities (CT, P = 0.002; ultrasonography, P = 0.008). This finding indicated medial retropharyngeal lymph nodes were larger in young adult cats than they were in old cats (Figure 5). No significant linear correlations were detected between volume of medial retropharyngeal lymph nodes and weight or body condition scores of cats.
Scatterplot indicating relationship between medial retropharyngeal lymph node volume and age for 38 healthy cats that underwent CT and ultrasonography for determination of various lymph node measurements. The black line indicates the best-fit slope line of the linear regression equation.
Citation: American Journal of Veterinary Research 73, 9; 10.2460/ajvr.73.9.1377
The overall intraclass correlation coefficient was moderate (0.40). The intraclass correlation coefficients for each measured lymph node location in CT images were as follows: rostral width, 0.63; rostral height, 0.24; middle width, 0.58; middle height, 0.62; caudal width, 0.27; caudal height, 0.30; measured length, 0.83; and calculated length, 0.67; the mean of these values was 0.52. The intraclass correlation coefficients for each measured lymph node location in ultrasonographic images were as follows: rostral width, 0.41; rostral height, 0.18; middle width, 0.37; middle height, 0.11; caudal width, 0.10; caudal height, 0.19; and length, 0.43; the mean of these values was 0.26. Thus, intraclass correlation coefficients were higher for each measured lymph node location as determined by use of CT images compared with those determined by use of ultrasonographic images.
Discussion
Significant differences were detected between data determined with ultrasonographic and CT images for some of the lymph node measurements (rostral width, middle height, middle width, and caudal height) but not others (rostral height, caudal width, and length [comparison between ultrasonographic lymph node length and measured CT lymph node length]) in the present study. Cats were in different positions during CT and ultrasonography, which may have caused lymph nodes to have different shapes in images acquired with each modality. We positioned cats for CT and ultrasonographic examinations in accordance with common clinical practices. Imaging planes for the 2 modalities may have differed slightly. Because medial retropharyngeal lymph nodes in cats have a curved shape in a rostrodorsal to caudoventral direction, transverse CT images did not represent true short-axis sections but instead represented slightly oblique sections of lymph nodes. This could have caused a slight difference in measurements between CT and ultrasonographic methods because the ultrasound transducer could be manipulated to acquire true long-axis and short-axis images. However, differences between measurements determined by use of ultrasonographic and CT images were small (≤ 2.1 mm) and would likely be clinically irrelevant for comparison of measurements acquired with each of these modalities, as may be necessary for comparison of data for initial and follow-up examinations in a clinical setting.
For humans and dogs, short-to-long–axis length ratios are helpful in discriminating between reactive cervical lymph nodes and those with metastases.5,9 In the present study, short-axis–to–long-axis length ratios for medial retropharyngeal lymph nodes of cats were 0.15 to 0.23 as determined by use of CT images and 0.13 to 0.18 as determined by use of ultrasonographic images. Future studies would be needed to determine whether short-axis–to–long-axis length ratios are useful for discriminating between reactive medial retropharyngeal lymph nodes and lymph nodes with metastases in cats.
No significant differences were detected between measurements of left and right medial retropharyngeal lymph nodes determined by use of either imaging modality. This indicated medial retropharyngeal lymph nodes in cats in the present study were symmetric. Other authors have reported that medial retropharyngeal lymph nodes are asymmetric in cats with bilateral fungal rhinitis13 or oral squamous cell neoplasia14; however, authors of those studies did not provide discussion regarding those findings and histologic analysis of those lymph nodes was not performed. Further studies in which histologic analysis of asymmetric medial retropharyngeal lymph nodes is performed would be needed to determine the importance of such findings for evaluation of inflammatory and neoplastic diseases in the heads of cats.
Hounsfield unit values for medial retropharyngeal lymph nodes of cats in the present study were similar to those determined for retropharyngeal lymph nodes of dogs via CT without administration of contrast media in another study.16 Hounsfield unit values in the present study were determined by use of images in which a fat-attenuating or hypoattenuating lymph node hilus was not observed. No significant linear relationships were detected between Hounsfield unit values and weight or body condition score of cats in the present study. As such, lymph node attenuation did not seem to be affected by weight or body condition score of cats.
All lymph nodes in cats in the present study were mildly to moderately heterogeneous as determined with each imaging modality. Lymph node heterogeneity in CT images was mild and similar to heterogeneity of longus coli muscles. Heterogeneity is an atypical finding for lymph nodes evaluated via ultrasonography. Most lymph nodes have a homogeneous appearance in clinically normal animals and are heterogeneous or hypoechoic in animals with disease.9,21 Cats in the present study were free of systemic disease as determined via evaluation of results of CBCs and serum biochemical analyses and were free of oronasal diseases as determined via evaluation of results of CT and dental examinations. Cats were determined to be healthy before inclusion in the study. However, medial retropharyngeal lymph nodes were heterogeneous in all cats in the present study. Therefore, healthy cats may have ultrasonographically heterogeneous medial retropharyngeal lymph nodes, and such heterogeneity may not indicate pathological changes in lymph nodes of cats with sinonasal disease.
Identification of a lymph node hilus is important because the loss of a hilus has been associated with neoplastic infiltration of lymph nodes in regions other than the heads of animals.10–12 In the present study, a lymph node hilus was identified more frequently in ultrasonographic images (95% of lymph nodes) than in CT images (24% of lymph nodes; by use of the initial definition [lymph node hilus defined as a fat-attenuating structure {as determined by a slightly negative Hounsfield unit region-of-interest value} entering a lymph node from the periphery]); however, this difference may be attributable to differences in the definitions of the appearance of a hilus for each imaging modality. For ultrasonographic images, a hilus was defined as a hyperechoic band in the parenchyma of a medial retropharyngeal lymph node. For CT images, a hilus was defined as a fat-attenuating region (negative Hounsfield unit value) within a medial retropharyngeal lymph node. Many of the cats did not have a fatty hilus in a lymph node but nevertheless had a rounded, slightly hypoattenuating region in the middle and rostral aspects of that lymph node in CT images; these areas typically had attenuation that was 20 to 30 Hounsfield units less than that of the rest of the lymph node parenchyma. The finding that Hounsfield unit values for these lymph node regions were > 0 may have been attributable to volume averaging of fat-attenuating and soft tissue–attenuating regions. Thus, such slightly hypoattenuating regions in lymph nodes in CT images may have corresponded to regions with hyperechoic bands in ultrasonographic images. When a lymph node hilus was identified by use of the initial definition or the alternate definition (lymph node hilus defined as any hypoattenuating region in the lymph node [whether it was fat attenuating or not]) for CT images, a hilus was identified in 92% of medial retropharyngeal lymph nodes. Further studies would be needed to determine whether the hilus of a medial retropharyngeal lymph node is altered during disease and whether a hypoattenuating, non–fat-attenuating region in a lymph node in CT images corresponds to a lymph node hilus identified via histologic evaluation.
To the authors' knowledge, the CT and ultrasonographic appearance of the margins of medial retropharyngeal lymph nodes of cats has not been previously described. In CT images in the present study, margins of the medial retropharyngeal lymph nodes of cats were smooth, whereas they were slightly irregular in ultrasonographic images. The interfaces between a lymph node and adjacent tissues were easily identified with each imaging modality, because of differences in attenuation of tissue types in CT images and thin hyperechoic bands surrounding lymph nodes in ultrasonographic images. Metastatic lymphadenopathy in humans is detected in CT images by identification of irregular, thick lymph node margins with infiltration of lymph node tissues into perinodal fat.22 In ultrasonographic images of dogs, metastatic disease in lymph nodes appears as either an irregular to lobular lymph node margin or as a large difference in echogenicities between a lymph node and perinodal tissues.10 Further studies would be needed to determine whether similar changes in appearance of the margins of medial retropharyngeal lymph nodes develop in cats with diseases of the head.
A significant negative linear association was detected between age of cats and lymph node volume in the present study; medial retropharyngeal lymph nodes were typically larger in young adult cats than they were in old cats. This finding was similar to that of another study15 in which a negative correlation was detected between age and size of medial retropharyngeal lymph nodes of dogs. That finding of the present study was important because many cats with neoplastic diseases of the nasal cavity are old; thus, detection of enlarged medial retropharyngeal lymph nodes in an old cat may be clinically important. Cats < 2 years old were not included in the present study. Results of other studies23–25 indicate abdominal lymph nodes are larger in juvenile dogs than they are in adult dogs as determined with ultrasonography. Medial retropharyngeal lymph nodes may be larger in cats < 2 years old, compared with those in cats in the present study; however, determination of medial retropharyngeal lymph node measurements in cats < 2 years old would require further investigation.
The overall interobserver correlation coefficient for lymph node measurements determined by use of CT images was moderate, and that determined by use of ultrasonographic images was fair. Each author independently identified the rostral, middle, and caudal aspects of lymph nodes, and there may have been variation in location of measurements between these observers. Such variation in measurements would be expected between observers in a clinical setting. As such, measurements determined in the present study would likely be clinically useful as references for measurements of lymph nodes in cats. These results also indicated multiple observers should be used in future studies in which medial retropharyngeal lymph node measurements are determined for cats with neoplastic or inflammatory diseases of the head.
A limitation of the present study was that histologic analysis of lymph nodes was not performed. Each cat was screened for diseases via evaluation of results of a CBC, serum biochemical analyses, physical and dental examinations, and CT of the head. Cats included in the study were determined to be free of hemolymphatic, dental, and nasal diseases. Therefore, we considered lymph nodes in these cats to be clinically normal. Future studies are warranted in which medial retropharyngeal lymph nodes of healthy and diseased cats are histologically evaluated. An additional limitation of this study was that iodinated contrast medium was not administered IV during CT examinations. Contrast medium was not administered because of the risk of adverse effects in the healthy cats. Future studies are warranted in which patterns of contrast medium enhancement of medial retropharyngeal lymph nodes of cats are determined.
In the present study, medial retropharyngeal lymph nodes in cats were easily imaged via ultrasonography and CT. Mean retropharyngeal lymph node dimensions for cats in the present study was approximately 21 × 13 × 4.5 mm (maximum dimensions, approx 32 × 20 × 7 mm; length × height × width). Left and right lymph nodes were symmetric. Lymph node parenchyma was mildly to moderately heterogeneous in the healthy cats in this study. A lymph node hilus was identified in ultrasonographic images more often than in CT images. Young adult cats had larger lymph nodes than did old cats. Data determined in the present study were intended to serve as a reference for measurements and CT and ultrasonographic appearance of medial retropharyngeal lymph nodes of healthy cats.
SigmaStat, Systat Software Inc, Chicago, Ill.
GE Brightspeed 16-slice helical CT scanner, General Electric Medical Systems, Milwaukee, Wis.
GE Logiq 9, General Electric Medical Systems, Milwaukee, Wis.
Horizon Rad Station Distributed, McKesson, San Francisco, Calif.
Voxar3D, Barco NV, Kortrijk, Belgium.
SAS, version 9.1, PROC GLM, SAS Institute Inc, Cary, NC.
References
- 2.
Herring ESSmith MMRobertson JL. Lymph node staging of oral and maxillofacial neoplasms in 31 dogs and cats. J Vet Dent 2002; 19:122–126.
- 3.
Saito HSato TYamashita Y, et al. Topographical analysis of lymphatic pathways from the meso- and hypopharynx based on minute cadaveric dissections: possible application to neck dissection in pharyngeal cancer surgery. Surg Radiol Anat 2002; 24:38–49.
- 4.
Lurie DMSeguin BSchneider PD, et al. Contrast-assisted ultrasound for sentinel lymph node detection in spontaneously arising canine head and neck tumors. Invest Radiol 2006; 41:415–421.
- 5.
Castelijns JAvan den Brekel MW. Imaging of lymphadenopathy in the neck. Eur Radiol 2002; 12:727–738.
- 6.
Murakami SYamanishi MWAzuma R. Lymph node abscess due to Actinomyces viscosus in a cat. J Vet Med Sci 1997; 59:1079–1080.
- 7.↑
Steinkamp HJCornehl MHosten N, et al. Cervical lymphadenopathy: ratio of long- to short-axis diameter as a predictor of malignancy. Br J Radiol 1995; 68:266–270.
- 8.↑
Nyman HTKristensen ATSkovgaard IM, et al. Characterization of normal and abnormal canine superficial lymph nodes using greyscale B-mode, color flow mapping, power and spectral Doppler ultrasonography—a multivariate study. Vet Radiol Ultrasound 2005; 46:404–410.
- 9.
Nyman HTO'Brien RT. The sonographic evaluation of lymph nodes. Clin Tech Small Anim Pract 2007; 22:128–137.
- 10.↑
Nyman HTKristensen ATFlagstad A, et al. A review of the sonographic assessment of tumor metastases in liver and superficial lymph nodes. Vet Radiol Ultrasound 2004; 45:438–448.
- 11.
Salwei RMO'Brien RTMatheson JS. Characterization of lymphomatous lymph nodes in dogs using contrast harmonic and power Doppler ultrasound. Vet Radiol Ultrasound 2005; 46:411–416.
- 12.
Sato AFSolano M. Ultrasonographic findings in abdominal mast cell disease: a retrospective study of 19 patients. Vet Radiol Ultrasound 2004; 45:51–57.
- 13.↑
Karnik KReichle JKFischetti AJ, et al. Computed tomographic findings of fungal rhinitis and sinusitis in cats. Vet Radiol Ultrasound 2009; 50:65–68.
- 14.↑
Gendler ALewis JRReetz JA, et al. Computed tomographic features of oral squamous cell carcinoma in cats: 18 cases (2002–2008). J Am Vet Med Assoc 2010; 236:319–325.
- 15.↑
Burns GOScrivani PVThompson 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.↑
Kneissl SProbst A. Comparison of computed tomographic images of normal cranial and upper cervical lymph nodes with corresponding E12 plastinated-embedded sections in the dog. Vet J 2007; 174:435–438.
- 17.↑
Kneissl SProbst A. Magnetic resonance imaging features of presumed normal head and neck lymph nodes in dogs. Vet Radiol Ultrasound 2006; 47:538–541.
- 18.↑
Wiggs RBLobprise HB. Veterinary dentistry: principles and practice. Philadelphia: Lippincott-Raven Publishers, 1997;196.
- 19.↑
Ying MPang BFS. Three-dimensional ultrasound measurement of cervical lymph node volume. Br J Radiol 2009; 82:617–625.
- 21.
Kinns JMai W. Association between malignancy and sonographic heterogeneity in canine and feline abdominal lymph nodes. Vet Radiol Ultrasound 2007; 48:565–569.
- 22.↑
Sakai OCurtin HDRomo LV, et al. Radiologic evaluation of the neck lymph node pathology benign proliferative, lymphoma, and metastatic disease. Radiol Clin North Am 2000; 38:979–998.
- 23.
Pugh CR. Ultrasonographic examination of abdominal lymph nodes in the dog. Vet Radiol Ultrasound 1994; 35:110–115.
- 24.
Stander NWagner WMGoddard A, et al. Normal canine pediatric gastrointestinal ultrasonography. Vet Radiol Ultrasound 2010; 51:75–78.
- 25.
Agthe PCaine ARPosch B, et al. Ultrasonographic appearance of jejunal lymph nodes in dogs without clinical signs of gastrointestinal disease. Vet Radiol Ultrasound 2009; 50:195–200.
