History

An 8-year-old castrated male Bouvier des Flandres was referred for evaluation of swelling in the mandibular region on the right side. No abnormalities were detected during a dental procedure performed 2 months earlier. A mass measuring 5 × 3 × 2cm³ was palpated at the angle of the right ramus of the mandible. No lesions were visible in the oral cavity.Microscopic examination of fine-needle aspirates obtained from the mandibular mass revealed abundant melanin granules in a sparse, undefined cellular environment. These findings were interpreted as malignant melanoma of the mandibular salivary gland or regional lymph nodes. No abnormalities were detected on radiographs of the thorax. Because of the peculiar location of this mass, magnetic resonance imaging (MRI) of the dog's head was performed with a 1.0-T superconducting magnet and knee coil to characterize the extent of the initial lesion (Figure 1).

View Full Size

Precontrast T1- (A) and T2-weighted (B) axial magnetic resonance images of the head of an 8-year-old castrated male Bo uvier des Flandres evaluated for a mass located in the mandibular region on the right side.

Citation: Journal of the American Veterinary Medical Association 229, 2; 10.2460/javma.229.2.209

• Download Figure
• Download figure as PowerPoint slide

Diagnostic Imaging Findings and Interpretation

The right mandibular salivary gland is large and mildly distorted (Figure 2). The capsule appears intact. On the T1-weighted precontrast axial image, there is a central region of poorly defined hyperintense tissue and a similar thin hyperintense subcapsular axial margin. A hypointense central region comparable in signal intensity to cortical bone or air can be seen on the T2-weighted image. Most of the remaining tissue is hyperintense relative to the contralateral mandibular salivary gland. The central hyperintense region on the T1-weighted image is hypointense on the T2-weighted image, which is highly compatible with the presence of melanin granules. Differential diagnoses included incidental melanosis of an inflamed or neoplastic salivary gland, a pigmented cyst, or a solitary melanoma.

View Full Size

Same magnetic resonance images as in Figure 1. Notice that the right mandibular salivary gland (arrows) is enlarged, compared with the left. Centrally located hyperintensity on the T1-weighted image (A) with hypointensity on the T2-weighted image (B) has paramagnetic signal characteristics compatible with melanin.

Citation: Journal of the American Veterinary Medical Association 229, 2; 10.2460/javma.229.2.209

• Download Figure
• Download figure as PowerPoint slide

Comments

Malignant melanomas in dogs typically originate in the oral mucosa or skin and, to the authors' knowledge, have yet to be reported as originating in a salivary gland. Malignant melanomas focally infiltrate tissues; may reoccur; and may metastasize to the lungs, liver, kidneys, brain, or lymph nodes.¹ The possibility of a salivary gland melanoma warranted further evaluation of the head and cervical region with MRI for detection of a primary tumor or other metastatic foci. Most soft tissue tumors, including those of the salivary gland, have low signal intensity on T1-weighted images and an increased signal on T2-weighted images. Conversely, melanotic melanomas have a hyperintense signal on T1-weighted images and hypointense signal on T2-weighted images. This striking pattern of signal characteristics defines deposits of melanin granules, but does not differentiate benign from malignant lesions. Additionally, intracellular methemoglobin from hemorrhage degradation will cause a similar pattern. However, as a clot evolves, this pattern changes as hemoglobin by-products are removed. Distinguishing between the 2 substances can be accomplished by temporal imaging or by use of high field-strength scans.²

Because the lesion was localized to the salivary gland, the right mandibular salivary gland, associated mandibular lymph nodes, retropharyngeal lymph nodes, and a portion of the maxillary vein were excised. Ductal epithelial cells had characteristics compatible with squamous metaplasia and occasionally dyskeratosis, which is indicative of squamous cell carcinoma. Melanophages and large amounts of extracellular melanin were detected; however, melanocytes were not definitively identified histologically or by immunostaining with melan-A. Regional lymph nodes did not contain neoplastic cells.

In the absence of hemorrhage, the signal pattern observed on magnetic resonance images was attributed to melanin granules. Histologic examination did not support a diagnosis of a melanoma or a cyst; therefore, the melanin was considered an incidental finding. Nine months after surgery, the dog was doing well, which added further support for our diagnosis. Nonpathologic deposition of melanin in various soft tissues can occur and is thought to be caused by production of melanin by unusual cell types as well as atypical migration of melanocytes derived from the neural crest.

Although melanin was detected prior to imaging, MRI helped define the location and rule out other sources of disease. The case reported here indicates the importance of recognizing the signal pattern of melanin on MRI for identification of melanin-pigmented lesions. Because this pattern is distinctive but not unique, further diagnostic tests are always warranted to exclude the signal pattern derived from hemorrhage.