Materials and Methods

Criteria for selection of cases—Computerized and hard copy medical records of the University of California-Davis Veterinary Medical Teaching Hospital were searched for dogs with a clinical diagnosis of zygomatic sialadenitis from January 1, 1990, to December 31, 2009. Inclusion criteria were findings of fine-needle aspiration or tissue core biopsy evaluation of the zygomatic salivary gland consistent with zygomatic sialadenitis and contemporaneous diagnostic imaging that included some combination of MRI, CT, and ultrasonography. Clinical signs of retrobulbar disease included exophthalmos, decreased or absent retropulsion of the globe or signs of pain during retropulsion, strabismus, protrusion of the third eyelid, signs of pain on opening the mouth, and enophthalmos.

The medical records were also searched for dogs without evidence of retrobulbar disease and for which results of an MRI or CT examination including the zygomatic salivary glands and orbits was available (ie, control dogs). Brachycephalic and toy breed dogs were excluded to more closely reflect the population of affected dogs. Minimum requirements for MRI sequences were transverse unenhanced T1W and T2W and contrast-enhanced T1W images. For CT, transverse unenhanced and contrast-enhanced images were required. Ten cases that included CT images and 10 that included MRI images were selected; these included the most recently evaluated dogs that fit the criteria for each diagnostic imaging method.

Medical records review—Information regarding signalment, ophthalmic and systemic clinical signs, results of clinicopathologic tests, cytologic and histologic diagnosis, treatment, qualitative features of zygomatic salivary gland sialadenitis, and disease course was collected from the medical records of affected dogs, and images obtained via MRI, CT, and ultrasonography were evaluated. For control dogs, signalment information was recorded and MRI and CT images were evaluated.

Diagnostic image acquisition and analysis—Magnetic resonance imaging and CT examinations were performed by MRI-CT technologists at the teaching hospital. Zygomatic salivary gland signal and area measurements were determined via the same methods for affected and unaffected dogs. Magnetic resonance images were obtained with a 1.5-T MRI system.³ Noncontiguous transverse and dorsal plane images with a 3-to 5-mm slice thickness and an interslice gap of 0.5 to 1.0 mm were generated by use of T1W (transverse, TR = 450 to 485 milliseconds and TE = 11 to 15 milliseconds; dorsal, TR = 550 to 600 milliseconds and TE = 11 to 16 milliseconds) and T2W (transverse, TR = 2,900 to 2,960 milliseconds and TE = 80 to 87 milliseconds; dorsal, TR = 3,400 milliseconds and TE = 80 to 83 milliseconds) spin-echo pulse sequences. The field of view was 12 to 24 cm. Additional transverse and dorsal T1W images were acquired after bolus administration of gadopentetate dimeglumine^b (0.1 mmol/kg [0.045 mmol/lb], IV). The CT images were acquired with a helical CT scanner.^c Examinations consisted of contiguous, 5-mm, collimated transverse images obtained before and after administration of ionic^d or nonionic^e iodinated contrast agent (approx 880 mg of I/kg [400 mg/lb], IV). Quantitative and qualitative analysis of digital imaging and communication in medicine (ie, DICOM)–formatted MRI and CT images was performed by a board-certified radiologist (MSC) by use of a dedicated image-viewing station and commercially available viewing and analysis software.^f

Zygomatic salivary gland density, intensity, and cross-sectional area were quantitatively estimated from CT and MRI data. Transverse unenhanced T1W and T2W and contrast-enhanced T1W MRI sequences were used for intensity estimates,^9,10 and unenhanced and contrast-enhanced transverse CT images were used for glandular density estimates. Circular 0.2-cm² regions of interest were drawn within the parenchyma of the zygomatic salivary gland and adjacent medial pterygoid muscle on an image containing the largest glandular cross-sectional area. The medial pterygoid muscle was selected because of its location medial to the zygomatic salivary gland, which allowed all measurements to be made in the same image. If an abnormality of the medial pterygoid muscle was identified, the temporal muscle was used as the basis for comparison. Major salivary ducts, fluid-filled tissue components, and blood vessels were avoided when defining the regions of interest. Normalized signal intensity ratios (for MRI) or signal attenuation ratios (for CT) were calculated for each gland as zygomatic salivary gland signal to medial pterygoid or temporal muscle signal.

Glandular cross-sectional area was estimated from the transverse contrast-enhanced T1W image sequence (for MRI) and the transverse contrast-enhanced series (for CT). For each series, the image that subjectively had the largest zygomatic salivary gland area was used for measurement. Maximum height and width measurements of the gland were determined by use of electronic calipers. An approximation of gland area was calculated as maximum height × maximum width. Dorsal plane measurements were not performed because of variation in dorsal plane orientation in relation to the optic nerve and retrobulbar space among dogs.

Qualitative MRI and CT assessment was also performed. This included evaluation of the optic nerve, extraocular and masticatory muscles, and periocular tissues as well as globe position (eg, displacement) and size of regional lymph nodes. Fluid-filled tissue components suggestive of sialoceles were subjectively classified as small (< 1/3 of the gland cross-sectional area), medium (1/3 to 2/3), and large (> 2/3).

Ultrasonographic examinations^g were performed by use of 8-5-MHz curvilinear, 12-5-MHz linear, or 15-7-MHz linear transducers, with the exception of 1 scan performed with a 10-MHz component of a multiple-frequency mechanical sector transducer.^h All scans were performed by a board-certified radiologist or by a radiology resident under the direct supervision of a board-certified radiologist. Scan planes were at the discretion of each examiner. Qualitative assessment was performed by retrospectively reviewing static images and written reports to determine lesion size, echogenicity, presence of fluid-filled tissue components, and zygomatic salivary gland involvement.

Statistical analysis—A Shapiro-Wilk W test was used to verify that data were normally distributed. Mean, SD, and 95% CIs were calculated for intensity and attenuation ratios for MRI and CT data, respectively. To ensure independence of the data, means of bilateral measurements from control dogs and from affected dogs with bilateral disease were calculated; the mean of these 2 measurements from each control dog and from each bilaterally affected dog was used as the sample value for that dog. A Student t test was used to compare intensity and attenuation ratios between affected and control dogs.

Zygomatic salivary gland cross-sectional areas were compared between MRI and CT images in control dogs. Linear regression analysis was used to determine the relationship between body weight and cross-sectional gland area. Logistic regression was used to assess the predictive value of cross-sectional area on disease status. Size and body weight of affected versus control dogs were compared by use of the same descriptive statistics used for the intensity and attenuation ratios.

For all statistical evaluations, values of P < 0.05 were accepted as significant. All analysis was performed by use of statistical software.ⁱ

Results

Eleven dogs with zygomatic sialadenitis met the study inclusion criteria. Two additional dogs with clinical and diagnostic imaging findings suggestive of zygomatic sialadenitis were excluded because cytologic or histopathologic confirmation was lacking. Mean age was 8 years old (range, 3 to 12 years). Body weight ranged from 8.7 to 45.0 kg (19.1 to 90.0 lb) with a mean weight of 23.4 kg (51.5 lb). Nine dogs were neutered males, 1 was a sexually intact male, and 1 was a neutered female. Breeds included mixed (3 dogs), and Belgian Malinois, Newfoundland, English Springer Spaniel, Australian Shepherd, German Shepherd Dog, Welsh Terrier, Labrador Retriever, and Pekingese (1 of each). Seven dogs underwent MRI and ultrasonography; 3 dogs underwent CT and ultrasonography, with CT images available for use in the present study for 2 of these dogs; and 1 dog underwent ultrasonography alone. Zygomatic sialadenitis was unilateral in 9 of 11 dogs and bilateral in the remaining 2.

Mean age of the 20 dogs without retrobulbar disease was 7.7 years (range, 2 to 12 years). Body weight ranged from 8.6 to 50.2 kg (18.9 to 110.4 lb) with a mean weight of 30.7 kg (67.5 lb). Nine dogs were neutered males, 8 were neutered females, and 3 were sexually intact males. Breeds included mixed (6 dogs), Labrador Retriever (5), Golden Retriever (3), German Shepherd Dog (2), and Viszla, Australian Shepherd, Cairn Terrier, and Brittany (1 of each).

Clinical evaluation and treatment—Ophthalmic clinical signs in affected dogs were evaluated. All affected dogs had clinical evidence of retrobulbar disease; commonly detected signs included protrusion of the third eyelid (in 11/13 orbits), exophthalmos (9/13), decreased or absent retropulsion (7/13), and signs of pain on opening the mouth (7/11 dogs). Other signs of retrobulbar disease were detected less commonly. These included zygomatic papilla swelling (detected during evaluation of 4/13 orbits), periocular swelling (3/13), enophthalmos (1/13), signs of pain during retropulsion of the globe (1/13), and apparent difficulty prehending food (1/11 dogs).

Nonspecific ophthalmic signs detected in the 13 affected orbits included conjunctival hyperemia (7 orbits); chemosis (4); blepharospasm, ocular discharge, and episcleral injection (3 each); strabismus, epiphora, eyelid swelling, exposure keratitis, and lagophthalmos (2 each); and ptosis and eyelid erythema (1 each). All dogs remained sighted except for 1 dog that was initially determined to be sighted but became nonsighted 3 days after medical treatment was initiated. This dog regained sight after surgical removal of the affected gland.

Clinical signs of systemic illness were identified in 8 of 11 dogs. Common findings included lethargy (5 dogs), anorexia (4), and fever (3). One dog with signs of retrobulbar disease concurrently developed skin lesions on the muzzle; a diagnosis of adult-onset sterile granulomatous dermatitis with primarily histiocytic and lymphocytic inflammation was made by means of analysis of a biopsy sample. Biopsy of a swollen third eyelid in this dog revealed inflammation of a similar character. Results of a CBC were available for 9 of 11 dogs, and mild to moderate neutrophilia was identified in 2 of these (22,860 and 25,576 cells/μL; reference interval, 3,000 to 13,000 cells/μL); 1 other dog had a small number (222 cells/μL) of band cells with slightly toxic changes.

Ultrasound-guided fine-needle aspirates of the zygomatic salivary gland were obtained in 9 of 11 dogs (11 orbits), and surgical excisional biopsy of the affected gland was performed in the remaining 2 dogs. Cytologic evaluation of fine-needle aspirate samples revealed inflammatory cells as well as ductal epithelial cells, mucin, or both, consistent with inflamed salivary tissue in all 9 dogs. Inflammation subjectively ranged from mild to marked and included neutrophilic (n = 5 dogs), mixed (2), granulomatous (1), and lymphocytic-histiocytic (1) cell types. Aerobic and anaerobic microbial culture was performed on aspirate samples from 5 dogs; 4 of 5 samples yielded no growth. Small numbers of Peptostreptococcus anaerobius were cultured from the remaining sample. Small numbers of gram-positive coccobacilli were identified in a direct smear from 1 sample, but results of aerobic and anaerobic microbial cultures were negative. No antimicrobials were administered to this dog prior to sample collection from the gland, and contamination of the direct smear sample could not be ruled out. Histologic evaluation of the excised zygomatic salivary gland in 1 dog revealed chronic sialadenitis and periadenitis consisting of mixed inflammatory cells, with sialocele formation and intraductal mucus. In the other dog that underwent surgical excision of the affected gland, a large sialocele was evident with mild sialadenitis consisting of mixed inflammatory cells, moderate lobular degeneration, and necrosis with intraductal and interstitial mucus. Sialocele content in the 2 examined glands was gelatinous and tenacious in character.

Medical treatment was initiated in all affected dogs. Antimicrobials were administered to 10 of 11 dogs; treatments included amoxicillin-clavulanic acid (approx 13.75 mg/kg [6.25 mg/lb], PO, q 12 h, for 2 to 4 weeks) in 8 dogs, enrofloxacin (5 to 7 mg/kg [2.3 to 3.2 mg/lb], PO, q 12 h for 2 weeks) in 2, amoxicillin (50 mg/kg [22.7 mg/lb], PO q 12 h, for 5 days) in 1, and tetracycline (20 mg/kg [9.1 mg/lb], PO, q 12 h for 2 weeks) in 1. Two dogs received both amoxicillin-clavulanic acid and enrofloxacin. Five dogs received antimicrobials before gland sample collection; drug administration was initiated 3 to 14 days prior to the procedure. Systemic anti-inflammatory medications were administered in 6 of 11 dogs, including prednisone (0.5 to 1.0 mg/kg [0.23 to 0.45 mg/lb], PO, q 12 h for 7 days, followed by 7 days of tapering doses) in 2 dogs, carprofen (1.5 to 2.5 mg/kg [0.68 to 1.14 mg/lb], PO, q 12 h for 2 weeks) in 2, prednisolone (0.6 mg/kg [0.27 mg/lb], PO, q 12 h for 5 days, followed by 10 days of gradual tapering) in 1, and deracoxib (1.0 mg/kg, PO, q 24 h for an unreported duration) in 1. Tramadol (2.2 mg/kg [1.1 mg/lb], PO, q 12 h for 1 week) and butorphanol tartrate (0.2 mg/kg [0.09 mg/lb] PO, q 12 h for an unknown duration) were each administered in 1 dog. Topically administered ophthalmic medications included neomycin-polymyxin-dexamethasone in 4 dogs (3 received ointment [approx 0.25 in/application] and 1 received solution [1 drop/application], q 6 to 12 h for 1 to 2 weeks), petrolatum ophthalmic ointment (approx 0.25 in/application, q 6 to 12 h for an unknown duration) in 3, and neomycin-polymyxin-bacitracin ointment (approx 0.25 in/application, q 12 h for 2 weeks) in 1. Lateral orbitotomy with zygomatic salivary gland removal and temporary tarsorrhaphy was performed in 2 dogs that had large sialoceles and in which clinical signs were refractory to medical treatment.

Follow-up visit information that indicated whether clinical signs resolved was available for 8 of 11 dogs. Resolution of clinical signs occurred within 3 weeks after initiation of medical or surgical treatment in 6 of 8 dogs. In 1 of the other 2 dogs for which follow-up was available, exophthalmos resolved within 2 weeks after treatment was initiated, but third eyelid protrusion persisted for the next 3 months. In the remaining dog, no follow-up was available until 1.5 years later, at which time the signs had resolved.

MRI findings—Seven of 11 affected dogs and 10 of 20 control dogs underwent MRI (Figures 1 and 2). Two of the affected dogs had bilateral zygomatic sialadenitis, and in total, 9 affected orbits and associated tissues were examined. Zygomatic salivary gland signal-to-medial pterygoid or temporal muscle signal intensity ratios were determined for affected and control dogs (Table 1). Affected zygomatic salivary glands were T1W hypointense, T2W hyperintense, and more contrast enhancing, compared with those of control dogs (P < 0.05 for all comparisons). Six orbits had fluid-filled tissue components suggestive of sialoceles, including 2 each of small, medium, and large sizes.

View Full Size

Magnetic resonance images at the level of the zygomatic salivary gland in a dog without retrobulbar disease (ie, control dog). A—Transverse T1W image showing the zygomatic salivary gland (arrow) and medial pterygoid muscle (asterisk). B—TransverseT1W postcontrast image. C—TransverseT2W postcontrast image. L = left. In panels B and C, note the relative intensities of the zygomatic salivary gland and medial pterygoid muscle.

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Magnetic resonance images of the head of a dog with unilateral zygomatic sialadenitis. A—TransverseT1W image showing borders of the enlarged, affected left zygomatic salivary gland (arrowheads), which is hypointense relative to the unaffected right gland. B—Transverse T1W postcontrast image showing moderate dorsal displacement of the left globe consistent with exophthalmos; the periocular tissues (arrow) are abnormally prominent and, together with the affected gland and adjacent masticatory muscles (arrowheads), contrast enhancing. A fluid-filled tissue component is evident within the affected gland (asterisk). C—TransverseT2W image showing hyperintensity of the affected gland, compared with the unaffected contralateral gland. D—Dorsal T1W postcontrast image showing the same fluid-filled tissue component as in B (asterisk) and abnormal enhancement of soft tissues rostral to the affected orbit (arrow).

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

Globe displacement was evident in all 7 affected dogs that underwent MRI and in 7 of 9 affected orbits, including dorsal (6/7), rostral (6/7), and lateral (3/7) displacement. Globe displacement was not detected in the less severely affected orbits of the 2 bilaterally affected dogs. The optic nerve and extraocular muscles had mildly to moderately increased T2W hyperintensity and contrast enhancement in 3 orbits, including that of 1 dog in which a sialocele completely encircled the optic nerve and extraocular muscles (Figure 3). Other affected muscles or muscle groups with similar intensity and enhancement patterns included the ipsilateral medial pterygoid (3/9 orbits evaluated), temporal (3/9), and masseter (2/9) muscles. The ipsilateral temporal muscle was atrophied in 1 dog. Periocular tissues (soft and fatty tissues immediately surrounding the globe) in 4 affected orbits were abnormally T2W hyperintense and contrast enhancing. Ipsilateral mandibular lymph nodes were subjectively assessed as mildly to moderately enlarged during evaluation of 4 of 9 affected orbits (lymph nodes were affected bilaterally in 1 dog and unilaterally in 2), and contrast enhancement was heterogeneously increased in 1.

View Full Size

Postcontrast magnetic resonance images of a dog with unilateral zygomatic sialadenitis that was refractory to medical management; the dog subsequently became nonsighted. A and BDorsal and transverse T1W images, respectively, showing a large sialocele of the affected right zygomatic salivary gland (arrowheads) encircling the optic nerve and extraocular muscles. The optic nerve and extraocular muscles are abnormally thickened and contrast enhancing (arrows), compared with those of the unaffected left side. The unaffected gland is indicated (asterisk) in panel B. Within 3 weeks after surgical excision of the affected gland, the patients clinical signs resolved and sight returned.

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

The MRI appearance of affected zygomatic salivary glands was consistent between dogs, with 1 exception. One dog had clinical signs of retrobulbar disease, with a grossly enlarged zygomatic salivary gland evident on MRI and T2W hyperintensity and contrast enhancement of the extraocular muscles and optic nerve, but did not have altered signal intensity of the affected gland, compared with that of control dogs.

CT findings—Computed tomography images for 2 of 11 affected dogs (both with unilateral zygomatic sialadenitis) and 10 of 20 control dogs were evaluated (Figures 4 and 5). Attenuation ratios for control dogs were summarized (Table 1); in these dogs, mean unenhanced signal attenuation of the zygomatic salivary gland and that of the medial pterygoid muscle were approximately equal. Mean contrast-enhanced attenuation of the zygomatic salivary gland in control dogs was slightly > 1.5 times that of the medial pterygoid muscle. Statistical evaluation was not performed for CT values in affected dogs because of the small sample size. However, in these 2 dogs, the unenhanced attenuation ratios of 0.57 and 0.64 were > 2 SDs below the mean value for control dogs (1.04 ± 0.09). Contrast-enhanced attenuation ratios for the 2 dogs were 1.71 and 1.42 (within approx 1 SD of the mean value for control dogs [1.60 ± 0.17]). A large, fluid-filled tissue component confirmed as a sialocele via histologic analysis was present in 1 affected gland. Dorsal globe displacement and mild ipsilateral mandibular lymph node enlargement were detected in both dogs. No abnormalities of the optic nerve or surrounding soft tissue and musculature could be discerned.

View Full Size

Transverse CT images of the head of a dog without retrobulbar disease (window level = 40 Hounsfield units; window width = 350 Hounsfield units). A—Image obtained at the level of the zygomatic salivary gland. Note the ventrally distributed soft tissue-attenuating material in the right nasal cavity and rostral nasopharynx, indicative of nasal disease. The zygomatic salivary gland (arrow) and medial pterygoid muscle (asterisk) are shown. Attenuation of the gland and that of the muscle subjectively appear to be equal. B—Postcontrast image of the same region shown in panel A. Note the increased attenuation of the zygomatic salivary gland in relation to the medial pterygoid muscle.

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Computed tomography and ultrasonographic images of a dog with unilateral zygomatic sialadenitis. A—Transverse postcontrast CT image (window level = 40 Hounsfield units; window width = 350 Hounsfield units). The affected right zygomatic gland is markedly enlarged, compared with the unaffected contralateral gland, and a prominent sialocele (arrowheads) and moderate dorsal displacement of the globe are evident. B—Static ultrasonographic image obtained in a sagittal plane with the probe oriented dorsal to the globe of the same dog in panel A. A large echogenic fluid-filled pocket (arrowheads), which could not be identified via ultrasonography alone as originating in the zygomatic gland, is shown in the far field.

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

Size measurements—Mean CT- and MRI-estimated zygomatic salivary gland areas for control dogs were 2.97 ± 0.99 cm² and 3.14 ± 1.08 cm², respectively, and no significant (P = 0.70) difference was detected between these CT and MRI data. Thus, gland area data measured via MRI and CT for control dogs were combined, and correlation with body weight was assessed. Strong positive correlation was revealed between body weight and estimated zygomatic salivary gland area (R² = 0.65; Figure 6).

View Full Size

Correlation between body weight and estimated zygomatic gland cross-sectional area for 20 control dogs determined by use of combined MRI and CT measurements. A strong positive correlation (R² = 0.65) was detected.

Citation: Journal of the American Veterinary Medical Association 239, 9; 10.2460/javma.239.9.1211

• Download Figure
• Download figure as PowerPoint slide

Body weight and combined data for zygomatic salivary gland cross-sectional area measured via MRI and CT were compared between groups. Mean body weight of affected dogs was 28.06 ± 13.50 kg (61.73 ± 29.70 lb; 95% CI, 21.75 to 34.37 kg [47.85 to 75.61 lb]), and that of control dogs was 30.68 ± 7.57 kg (67.50 ± 16.65 lb; 95% CI, 24.86 to 36.50 kg [54.69 to 80.30 lb]); these values were not significantly (P = 0.59) different. Affected zygomatic salivary glands (mean area, 7.67 ± 6.20 cm²; 95% CI, 2.91 to 12.44 cm²) were significantly (P = 0.003) larger than those of control dogs (3.05 ± 1.01 cm²; 95% CI, 2.58 to 3.53 cm²). Cross-sectional area was predictive of disease status after accounting for body weight by use of logistic regression (P = 0.01; odds ratio, 4.01; 95% CI, 1.40 to 11.51).

Ultrasonography findings—Ultrasonographic examination results were reviewed for 10 of 11 affected dogs. For the 2 dogs with bilateral disease, only the more severely affected side underwent this diagnostic scan; thus, a total of 10 orbits were evaluated. A retrobulbar mass was identified in 9 of 10 orbits, with variable echogenicity that ranged from uniformly hyperechoic to mixed echogenicity to hypoechoic. The zygomatic salivary gland was identified as the origin of the mass in 4 orbits, with fluid-filled tissue components identified in 5 orbits (Figure 5). Results of ultrasonographic examination in 1 dog were unremarkable.

Discussion

In the study reported here, clinical and diagnostic imaging findings in dogs with zygomatic sialadenitis were described, and the size and appearance of zygomatic salivary glands in clinically normal dogs assessed by means of MRI and CT were established for purposes of comparison. The clinical signs and imaging characteristics of diseased glands were consistent and, when taken together, suggested zygomatic sialadenitis as a primary diagnosis.

Affected dogs in the present study were typically middle-aged to older, medium- or large-breed dogs. All but one of these dogs were males, suggesting a possible sex predisposition as has been previously reported for sialadenitis,² although the sample size was too small to allow definitive conclusions to be made. Affected dogs typically had clinical signs of retrobulbar disease, in combination with nonspecific systemic signs such as lethargy, anorexia, or fever. Interestingly, 4 dogs had gross swelling of the zygomatic papilla. It is possible that this abnormality may have been present in additional dogs, but because of the retrospective nature of this study, it could not be determined in all instances whether this area was closely evaluated. This abnormality has been briefly mentioned in other reports^11,12 and may represent a finding specific for zygomatic salivary gland disease, possibly attributable to associated ductal inflammation or infection.

Results of analysis of fine-needle aspirates and histologic evaluation of tissue samples from affected dogs were indicative of inflammation, but the degree and type of inflammation varied, similar to findings in a previous case series.² One dog also had a positive microbial culture result. Several contributing factors may explain this variability, including medical treatment prior to sample collection, variation in stages of disease, and nonrepresentative sample collection from glands. The etiology of zygomatic sialadenitis has not been well characterized,¹ and the possibility of different underlying causes of inflammation should be considered. Additional credence is lent to this possibility by the presence of 1 dog in this series with signs of multi-organ inflammation consistent with adult-onset sterile granulomatous dermatitis, in which similar findings of lymphocytic-histiocytic inflammation were revealed in samples from skin lesions of the muzzle, a zygomatic salivary gland, and regional lymph nodes. Overall, findings of the present study are in agreement with those in previous literature^1,3 regarding a possible immune-mediated or infectious etiology in many of these cases.

Although clinical signs resolved rapidly with medical treatment in most dogs, it is of interest that large sialoceles were detected in the 2 dogs that had disease refractory to medical management, possibly contributing to the lack of resolution of clinical signs and subsequent need for surgical intervention.^6,13,14 The fact that 1 dog that was determined to have lost sight regained sight after surgery emphasizes the importance of timely diagnosis and aggressive treatment as well as the contribution of diagnostic imaging in diagnosis and treatment determination.

Results of the study reported here revealed consistent imaging features of zygomatic sialadenitis. Altered signal intensities in affected zygomatic salivary glands were consistent with glandular interstitial edema caused by inflammation. In comparison with other causes of orbital disease⁸ and consideration of the authors' previous experiences, MRI findings of zygomatic sialadenitis in the present study were considered to be highly characteristic for this disorder. There may be some overlap of the described features with zygomatic salivary gland neoplasia (an uncommon disease for which imaging appearance is not well described)^15,16; however, the presence of bony destruction, irregular margination, or heterogenous signal intensity may help to differentiate gland neoplasia from inflammation.

One dog in this study had an enlarged zygomatic salivary gland detected via MRI without appreciably altered gland signal intensity. The dog was treated systemically with antimicrobials and steroids for 2 weeks prior to this evaluation, which caused partial resolution of retrobulbar signs, and cytology results after the MRI revealed only mild inflammation. Thus, it is likely that the stage of disease contributed substantially to the lack of altered signal intensity. Nevertheless, this result suggests that a lack of altered MRI signal intensity does not rule out a diagnosis of zygomatic sialadenitis, particularly if the gland is enlarged and there is clinical evidence of retrobulbar disease.

Because only 2 CT examinations of affected dogs were reviewed, definitive conclusions could not be made regarding the appearance of diseased zygomatic salivary glands in CT images. Both examinations revealed enlargement of the affected gland and precontrast hypoattenuation in relation to the mean value for control dogs, as expected. However, the affected glands in these 2 dogs did not contrast enhance noticeably more than did glands of control dogs. The reason for this finding is unknown, but considerations include region-of-interest sampling error, influence of prior medical treatment, stage of disease, or gland avascularity (infarction or necrosis). Additionally, differences in glandular contrast enhancement may not have been appreciable on CT images because the sensitivity of detection of iodinated CT contrast agents is 2 to 3 orders of magnitude lower than that for gadolinium magnetic resonance agents on a per-molecule basis.^17,18

Ultrasonography provided only limited information regarding zygomatic sialadenitis in comparison with MRI and CT. In affected dogs, a mass was usually (9/10 orbits) detected in the retrobulbar space, but could only occasionally (4/9 orbits) be identified as originating from the zygomatic salivary gland. This may have been attributable to poor image quality (caused by increased attenuation secondary to inflammation) or to lack of experience of the sonographer.

Although all affected dogs in the present study had clinical and cytologic or histologic evidence of zygomatic salivary gland inflammation, it was not possible to differentiate zygomatic sialadenitis with secondary sialocele formation from primary zygomatic sialocele with secondary gland inflammation. These diseases are considered separate entities, but are often detected in association with one another.² There is considerable overlap in their clinical and histologic appearances, and the distinction was not thought to influence case management.

Visualization of anatomic structures for diagnosis of zygomatic sialadenitis and for involvement of adjacent structures such as the optic nerve and extraocular and masticatory muscles was excellent with MRI and CT. These imaging findings can enable confirmation of a clinical diagnosis and aid in optimal treatment and potential surgical case management. Ultrasonography was less definitive for recognition of zygomatic gland–specific disease, but was a useful tool for tissue sample collection.