Osteosarcoma of the Humeral Head Associated With Osteochondritis Dissecans in a Dog

Introduction

Osteosarcoma is the most prevalent primary bone tumor in dogs, accounting for 2% to 7% of all canine malignancies.¹ Approximately 75% to 95.3% of spontaneously occurring canine osteosarcomas originate in the metaphysis of the long bones of the appendicular skeleton, with the distal radius, proximal humerus, distal femur, and proximal tibia being the most common sites.²³ A review of 222 cases at the University of Florida Veterinary Medical Teaching Hospital (UFVMTH), between February 1987 and February 2002, revealed that 79.3% of all osteosarcomas were localized to the appendicular skeleton. Of these cases, 25% were in the humerus, 22.7% were in the radius, 19.3% were in the femur, and 15.9% were in the tibia. Diaphyseal osteosarcomas are less common and have been associated with fractures and metallic implants.^4–6 Epiphyseal involvement and transphyseal spread of osteosarcoma have been reported in humans.^7–9 Tumors originating in the epiphysis account for <1.4% of all cases.⁹ The purpose of this report is to describe an osteosarcoma involving the caudal aspect of the proximal humeral epiphysis of a dog that was previously treated surgically for osteochondritis dissecans (OCD) at the same location.

Case Report

A 6-year-old, castrated male, Labrador retriever was referred to the UFVMTH for evaluation of right forelimb lameness. The lameness had progressed from mild to nonweight bearing over the 1 month prior to presentation. At 8 months of age, the dog was examined by the referring veterinarian because of a 4-month history of intermittent lameness of the right forelimb. Findings at that time included mild weight-bearing lameness and pain with flexion and extension of the scapulohumeral joint. Radiographs of the right scapulohumeral joint revealed an area of sclerosis on the caudal humeral head, flattening of the subchondral bone, and a calcified cartilaginous flap, substantiating the diagnosis of OCD. Medical therapy was not initiated. At 11 months of age, a scapulohumeral arthrotomy was performed to remove the cartilaginous flap. The free-floating flap was removed, and the cartilage defect was debrided via minimal curettage.

Upon presentation to the UFVMTH, the dog was bright, alert, responsive, and in good body condition. Gait analysis revealed a nonweight-bearing lameness of the right forelimb. Pain was elicited with palpation of the right proximal humerus and scapulohumeral joint and upon flexion and extension of the joint. A 7-cm, curvilinear scar was located across the skin of the right shoulder, consistent with the history of a scapulohumeral arthrotomy. No neurological abnormalities were noted, and all other physical examination findings were within normal limits.

Radiographs of the right humerus revealed an expansile, ill-defined osteolytic lesion occupying the entire caudal one-half of the humeral head [Figure 1]. The osteolytic lesion had poorly defined margins and extended across the growth plate into the metaphysis of the humerus. Radiographs of the thorax were normal. Differential diagnoses for the lesion included primary bone neoplasia (e.g., osteosarcoma, chondrosarcoma, fibrosarcoma, and hemangiosarcoma), metastatic neoplasia, and bacterial and fungal osteomyelitis.

An ultrasound-guided, fine-needle aspirate of the lesion was obtained, and cytopathological analysis revealed low cellularity and marked hemodilution. A low number of atypical mesenchymal cells with mild anisokaryosis and variable nucleus-to-cytoplasm ratios were seen. Occasional binucleation and distinct nucleoli were detected. No infectious agents were observed. Because of the low cellularity, a malignant neoplastic process could not be confirmed.

A complete blood count, serum biochemical profile, and urinalysis were performed. A mild leukocytosis (19.4 × 10³ cells/μL; reference range, 6.0 to 17.0 × 10³/μL) with a mature neutrophilia (15.3 × 10³ cells/μL; reference range, 3.0 to 11.5 × 10³/μL) was present. Abnormalities on the serum biochemical panel included elevated aspartate transaminase (357 U/L; reference range, 14 to 38 U/L) and mildly decreased potassium (3.4 mEq/L; reference range, 3.9 to 4.2 mEq/L). Urinalysis was unremarkable.

The dog was subsequently anesthetized, and fluoroscopy was utilized to obtain a bone biopsy using a Jamshidi needle.^a Six biopsies were obtained; three were submitted for histopathology, and the others were submitted for aerobic and anaerobic bacterial and fungal cultures. Histopathology of the obtained bone cores revealed fragments containing moderately cellular, unencapsulated masses of neoplastic cells arranged in sheets. Neoplastic cells were markedly pleomorphic, polygonal to fusiform, and contained dense, eosinophilic, abundant cytoplasm. Nuclei had prominent nucleoli and marked anisokaryosis and anisocytosis. One to two mitotic figures per 40× field were seen. These results were compatible with osteosarcoma. None of the cultures yielded growth.

Several treatment options were discussed with the owners (e.g., amputation, amputation plus chemotherapy, radiopharmaceutical therapy^b), and they elected amputation with chemotherapy.

Following amputation, gross examination of the humerus revealed a poorly defined, soft erosion of the head, with the greatest defect located at the caudal aspect [Figure 2]. The humeral head measured 5 × 6 × 5 cm, and it was almost entirely occupied by tumor. The articular cartilage on the humeral head was red-brown, soft, and pliable. On cut section, the epiphyseal region of the humeral head was soft and mottled tan to dark red. Normal trabecular bone could not be recognized.

Histopathology of the proximal humerus revealed an expansile and infiltrative mass composed of neoplastic mesenchymal cells within the medullary cavity. The cells formed streaming sheets and whorls and dissected through necrotic bone fragments. The cells were oval to spindle shaped, with a moderate eosinophilic to amphophilic cytoplasm. Nuclei were plump, round to oval, and euchromatic with one to three prominent nucleoli. Cells exhibited moderate anisocytosis and anisokaryosis. Neoplastic cells were present in close apposition to moderate amounts of angular, eosinophilic, extracellular material (i.e., osteoid). Bi- and multinucleate cells were frequently observed, and the mitotic index was two to three per 100×. There was regionally extensive hemorrhage admixed with fibrin, necrotic cellular debris, and necrotic bone fragments. Fibrocartilage and fibrous tissue extended into the bone beneath the defect. Neoplastic osteoblasts extensively infiltrated the articular cartilage and the marrow space adjacent to fibrocartilaginous scar tissue [Figure 3]. Histopathology of the right axillary lymph node revealed mild lymphoid follicular hyperplasia, without evidence of metastasis.

Discussion

Numerous reports exist in the veterinary literature of sarcomas developing in association with bone/medullary infarction, total hip arthroplasty, with metallic implants or following fractures.^4–610–13 To the authors’ knowledge, this is the first report of an epiphyseal osteosarcoma involving the site of a previous OCD lesion in a dog. In people, one case of sarcoma arising from an area of OCD of the medial femoral condyle has been reported.¹⁴ The tumor was an incidental finding within an area of osteochondrosis that was diagnosed at the time of surgical total knee replacement.

In dogs, osteochondrosis is defined as an abnormality of endochondral ossification, where the cartilage of the epiphysis fails to form subchondral bone, resulting in a thick, abnormal articular cartilage.¹⁵ In people, Jaffe defined OCD as “a condition in which a small osteochondral body (composed of articular cartilage along with a fragment of aseptically necrotic subchondral bone) is delimited and loosened from an articular end of bone (usually a femoral condyle) and often finally extends into the joint space.”¹⁶ The case of osteosarcoma associated with OCD reported in a human was compatible with the definition proposed by Jaffe.¹⁴ Jaffe’s definition encompasses subchondral bone infarcts, which have previously been associated with sarcomas in both humans and animals.¹¹¹⁷ Clinical differences between osteochondrosis in humans and dogs also exist and include age of onset, prevalence of bilateral lesions, and sites most commonly affected. While the pathogenesis of OCD in dogs has not been fully elucidated, suggested mechanisms for the formation of osteochondrosis lesions include trauma, hereditary factors, rapid growth, nutritional factors, and ischemia.¹⁵

Osteosarcoma confined to the epiphysis is extremely rare in humans and accounts for <1.4% of all osteosarcomas.⁹ Canine appendicular osteosarcoma usually affects the metaphyseal portion of the bone.² In cases of fracture-related or implant-induced osteosarcoma, the diaphyseal portion may be affected.⁵ To the authors’ knowledge, there are no reports in the veterinary literature of osteosarcoma confined to the epiphyseal region. Because both the epiphysis and metaphysis were involved in the case presented here, transphyseal spread probably occurred in one direction. Given the location of the lesion, the extent of epiphyseal involvement, and the prior history of OCD surgery, it is likely that the tumor originated in the subchondral bone of the epiphysis and spread across the physeal area distally into the metaphysis. Admittedly, this cannot be proven. Historically, transphyseal spread of osteosarcoma in humans was not believed to occur because of antiangiogenic properties of the physeal cartilage.¹⁸ More recently, transphyseal spread has been reported frequently in children (approximately 80%) with open physes.⁷⁸ In the case reported here, it is likely that the replacement of physeal cartilage with bone during skeletal growth had occurred prior to tumor development, especially given the typically aggressive nature of canine appendicular osteosarcoma and the length of time between diagnosis of the OCD lesion and the osteosarcoma.

Several hypotheses have been proposed to explain the association of sarcomas with bone infarcts and metallic implants. Predisposing factors for the formation of bone infarcts in humans include a dysbaric condition (i.e., physiological alterations resulting from changes in barometric pressure), sickle-cell hemoglobinopathy, alcohol consumption, steroid administration, hereditary bone dysplasia, and Gaucher disease.¹⁴ Initiating factors for implant-associated sarcomas include the type of metal used, corrosion of the implant, electrolysis between dissimilar implanted metals, tissue damage resulting from the initial trauma or subsequent surgical approach, altered cellular activity associated with fracture healing, and osteomyelitis.¹⁹ The dog in this report had no known predisposing factors for a bone infarct and did not receive a metallic implant. During the scapulohumeral arthrotomy, a curette was used to debride the affected articular cartilage from the subchondral bone of the caudal humeral head. The curette was constructed of stainless steel, and there was no detected fragmentation of the metal and no metal particles left in the joint after curettage. Therefore, possible etiologies of the osteosarcoma in this dog included altered cellular activity associated with tissue injury, surgical debridement, and healing of the OCD lesion. It is also possible that the osteosarcoma was not related to the prior OCD lesion or surgery.

Conclusion

This is the first report of an epiphyseal osteosarcoma in a dog that occurred in the area of a previous OCD lesion. The incidence rates of OCD of the humeral head in dogs have been reported to be 0.22% for males and 0.09% for females.²⁰ Osteosarcoma accounts for 2% to 7% of all canine malignancies.¹ The presence of these two lesions in the same area is extremely rare and raises questions about a causal relationship; however, no cause-effect relationship could be determined by this single case report.