Malignant Peripheral Nerve Sheath Tumor of the Diaphragm in a Dog

Introduction

Malignant peripheral nerve sheath tumors are mesenchymal neoplasms originating from cells surrounding the axons of neurons in peripheral nerves.^1–5 Tumors of the peripheral nervous system (PNS) are uncommon in domestic animals.^1,5,6 In dogs, the spinal nerves of the brachial plexus (between the sixth cervical [C₆] and second thoracic [T₂] vertebrae) and nerve roots are the most common sites for peripheral nerve sheath tumors, although origins from other peripheral nerves, spinal roots, and cranial nerves have been documented.^5–8 Although most peripheral nerve sheath tumors involve the extremities,^5,9–11 other sites have been reported. ^{4,6,8,12–17}

The terminology and classification of peripheral nerve sheath tumors are confusing in both the human and veterinary literature.^4,18,19 Numerous terms have been given to peripheral nerve tumors in animals, including schwannoma, neurilemmoma, neurinoma, neurofibroma, and neurofibrosarcoma.^1–5,9,11 Tumors are identified by the cell of origin of the neoplasm, such as schwannoma for a Schwann cell tumor, neurilemmoma for a tumor of the cells that ensheath or cover a nerve fiber, and perineurioma for a tumor of perineurial cells.^1,19

In animals, peripheral nerve sheath tumors of controversial ontogeny are simplified as benign or malignant.^1,4,10,18 Many canine peripheral nerve sheath tumors are malignant as evaluated by both cytological criteria (e.g., anaplasia, abundant mitosis, necrosis) and biological behavior (i.e., invasion of adjacent tissues, including the spinal cord).¹ In such cases, it is often impossible to definitively identify the principal cell type in routine histological sections.¹ Poorly differentiated peripheral nerve sheath tumors are best designated as malignant peripheral nerve sheath tumors, and, if necessary, they are modified when further studies (e.g., immunohistochemistry, electron microscopy, tissue culture) have been performed.^1,13,14 Definitive identification of the principal cell type in peripheral nerve sheath tumors becomes very difficult without the use of tissue culture, electron microscopy, or immunohistochemical staining. ^1,6,13,14 Peripheral nerve sheath tumors have also been confirmed using a specialized genetic marker.¹⁸ A single-point mutation of the neu oncogene has been identified in both malignant peripheral nerve sheath tumors and malignant schwannomas.¹⁰ Peripheral nerve sheath tumors are difficult to classify definitively, because they share similar histological characteristics with multiple other types of mesenchymal neoplasms.⁹ Malignant peripheral nerve sheath tumors are pleomorphic with cells arranged in fascicles, sheets, and whorls. The individual cells often include fusiform, oval to round, and multinucleated cells in the same neoplasm.⁹

The approach for the diagnosis of a malignant peripheral nerve sheath tumor should begin with routine histopathology. Once the neoplastic cell population is identified, immunohistochemical staining is employed to further characterize the cell type.^1,14,17 Immunohistochemical staining is often helpful in the diagnosis of peripheral nerve sheath tumor, as it allows verification of the neoplastic cell type through intermediate filament identification.²⁰ Vimentin is an intermediate filament protein. The use of the antibody to vimentin in immunohistochemical procedures labels a wide variety of tumors of mesenchymal origin.²¹ Cell types that are clearly labeled with the antibody include lymphoid cells, endothelial cells, fibroblasts, and smooth muscle cells.²¹ A positive vimentin in this case indicated that further characterization was needed. An S-100 protein is present in glial cells and ependymal cells and Schwann cells in the PNS.^14,22 In the skin, melanocytes and Langerhans cells are labeled;²² however, 50% of malignant peripheral nerve sheath tumors may lack S-100 protein.^1,14,20

The purposes of this report are to present an atypical location of a malignant peripheral nerve sheath tumor in a dog and to emphasize that this tumor type may be misdiagnosed if immunohistochemistry is not routinely performed on tumors of mesenchymal origin.

Case Report

An 11-year-old, 32.7-kg, castrated male greyhound was presented to the referring veterinarian at the Michigan State University Veterinary Teaching Hospital for panting of several days’ duration. A biochemical profile and complete blood count were normal. On thoracic radiographs, a mass was visualized in the left caudal thorax. Pulmonary neoplasia was suspected, and the dog was referred for further evaluation and surgical removal of the thoracic mass.

On initial presentation, the dog was panting, had a body condition score of 4/9, and was thin from persistent weight loss over the past year. Additional thoracic radiographs were obtained. On the right lateral radiographic view, a smoothly marginated, soft-tissue mass approximately 5.5 cm in diameter was visualized along the ventral aspect of the left hemidiaphragm [Figure 1A], and it was located in the left caudal thorax on the ventrodorsal projection [Figure 1B]. No evidence of pleural effusion was seen. On abdominal ultrasound, the mass was visualized cranial to the liver and compressing the diaphragm caudally. The mass was hypoechoic with a heterogeneous, internal echo pattern [Figure 2]. An ultrasound-guided fine-needle aspiration sample was nondiagnostic.

The dog was premedicated with hydromorphone^a (0.1 mg/kg) and midazolam^b (0.2 mg/kg) injected intramuscularly. Induction was achieved by injecting propofol^c (0.6 mg/kg) intravenously; it was maintained with 2.0% isoflurane^d in 0.7 liter per minute of oxygen. With the dog under general anesthesia, computed tomography was performed using a thirdgeneration scanner.^e Contiguous transverse images of 5 mm-slice thickness were obtained after intravenous injection of diatrizoate meglumine contrast medium.^f The mass had patchy contrast enhancement, a hypoattenuating center most likely from necrosis, and mild rim enhancement [Figures 3A, 3B]. No evidence of pulmonary metastatic disease was seen.

A left lateral thoracotomy was performed, and the mass was visualized attached to the left hemidiaphragm without extension into the abdomen. The mass incorporated the left phrenic nerve, and both were removed in situ and submitted for histopathology. The dog was discharged from the hospital 6 days after surgery and had a normal respiratory rate. A recheck examination was performed 4 days after discharge, and the owners noted that panting was no longer a problem.

Tissue samples reviewed histologically consisted of skeletal muscle, dense collagenous stroma, and adipose tissue (consistent with diaphragm). The sections were characterized by a poorly circumscribed, multinodular proliferation of neoplastic mesenchymal cells amidst a fine fibrovascular to focally myxomatous stroma. The mass was composed of a population of neoplastic spindle cells present in interwoven bundles and dissected by regions of dense collagen [Figure 4]. The neoplastic cells exhibited mild cellular atypia and mild nuclear pleomorphism. Mitotic figures were infrequent (<1 per high-power field). Only the collagenous stroma of the tumor was focally necrotic. The neoplastic cell population exhibited positive immunoreactivity with vimentin [Figure 5] and S-100 [Figure 6]. The combined histological features (routine and immunohistochemical) were considered to be most consistent with malignant peripheral nerve sheath tumor.

Discussion

Peripheral nerve sheath tumors in dogs are most commonly located on the extremities and associated with lameness.^9,12 Surgery is the treatment of choice for peripheral nerve sheath tumors and may be curative if the tumor is benign. These tumors are locally infiltrative and frequently recur following removal unless wide surgical excision is achieved.^5,9 One study reported that peripheral nerve sheath tumors are rarely completely resectable and have a high rate of recurrence, because complete surgical resection may not be possible due to the location of the tumor.⁹ Therefore, the prognosis for dogs with malignant peripheral nerve sheath tumors is poor, and the local recurrence rate after surgery is high.⁹ In this case, the tumor was noninvasive and completely resectable from the diaphragm.

The dog was presented for panting of several days’ duration. The differential diagnoses for panting include elevated ambient temperature, hypocalcemia, fever, endocrine diseases (e.g., hyperthyroidism, hyperadrenocorticism, pheochromocytoma), pain, anxiety, cardiac disease, drug administration (e.g., narcotics, glucocorticoids), and brain disease.²⁰ Elevated ambient temperature was an unlikely cause for panting, as the dog’s panting persisted despite living in an air-conditioned environment. The biochemical evaluation ruled out hypocalcemia. Rectal temperature was normal; therefore, fever was not a cause of panting. Other than panting, no clinical signs were consistent with hyperthyroidism. On abdominal ultrasound, both adrenal glands were enlarged; however, no clinical signs of polyphagia, polyuria, or polydipsia were seen, and the liver enzymes were within normal range. The sonographic findings were not consistent with pheochromocytoma. No pain was exhibited on physical examination. Anxiety is a diagnosis of exclusion that is unlikely, because a cause for panting was determined. The dog had normal sinus rhythm with no pulse deficits or murmurs, and no radiographic evidence of cardiac disease was found. The dog had no history of administration of drugs that might be associated with panting, and no discernable evidence of brain disease was seen. Cranial nerve examination and proprioception were normal. The dog was mentally appropriate and had no reported history of seizures.

The phrenic nerve provides the major motor nerve supply to the diaphragm. The panting exhibited was perhaps secondary to phrenic nerve involvement from the tumor.²¹ Phrenic nerve irritation in people has been shown to cause referred pain in the cervical area.²² Possibly this dog had referred cervical pain secondary to the tumor, and the panting might have been a result of the pain. However, no pain was elicited on physical examination, so pain seemed an unlikely cause of panting in this case. The phrenic nerve involvement from the tumor was the most likely cause of the dog’s panting.

Conclusion

This case is unusual, because malignant nerve sheath tumors more commonly involve the extremities. This should raise awareness that where a nerve is, there could be a peripheral nerve sheath tumor. If immunohistochemical studies are not performed, tumors that arise from mesenchymal tissues may be inappropriately classified. An incorrect identification of tumor type would likely result in less effective management decisions and prognostic outcome.