Introduction

Osteosarcomas (OSAs) are the most common malignant bone tumors in dogs, representing 85–98% of primary bone tumours.^1,2 There are six histological subtypes of central OSA in dogs: poorly differentiated, osteoblastic, chondroblastic, fibroblastic, telangiectatic, and giant cell. They can arise from within, on the surface, or outside bones (termed extraosseous).^3,4 OSA commonly presents with lameness, local pain, swelling or a soft-tissue mass, and pathological fractures.^2,5,6 Telangiectatic osteosarcomas (TOSAs) are an uncommon, aggressive subtype that present similarly but have differing histological features. Although the etiology and pathophysiology of the condition is not fully understood, it is believed to originate from transformed osteoblasts, or mesenchymal stem cells, which replace normal bone with weaker tumorous bone.⁷

OSAs are most prevalent in the long bones of the forelimbs in dogs, which support ∼60% of the total body weight.^4,8 The common predilection sites include the proximal humerus, distal radius, distal femur, and proximal tibia, although the incidence of axial OSAs increases as the size of the dog decreases.^3,4 Only a small proportion of OSAs affect the axial skeleton, accounting for 22–28% of cases.^3,9 In humans, vertebral TOSAs are incredibly rare, representing only 2% of cases of vertebral OSAs.⁷ As there are only two previously reported cases of vertebral TOSA in dogs, the incidence is likely similar to that in people.^9,10 In the first case, the TOSA was localized to the cervical vertebrae, but the dog was euthanized before MRI could be performed.¹⁰ In the second case, the primary lesion affected the sacrum, and the dog was euthanized because of extreme pain.⁹ In both cases, diagnosis was achieved postmortem, with no treatment (conservative or surgical intervention) having been attempted.

In this case report, we describe a case of a vertebral TOSA in a dog, describing both MRI and computed tomography (CT) imaging features, as well as describing the outcome following surgical intervention. We will discuss the imaging features and the viability and prognosis of treating cases of TOSAs localized to the spinal column.

Case Report

A 5 yr old, female, neutered Staffordshire bull terrier was referred for a 2 wk history of stiffness and pain following exercise and occasionally lifting the right hind limb. Before referral, she was treated with meloxicam (0.1 mg/kg per os q 24 hr) and restricted exercise. The analgesics caused some initial improvement; however, on the day before referral, the dog became nonambulatory.

On presentation, general clinical examination was unremarkable. Neurological examination revealed paraplegia with absent nociception on the left hind limb, with a clonic patellar reflex. Nociception was reduced but present in the right hind limb. Segmental spinal reflexes were otherwise within normal limits. There was a cutaneous trunci reflex cut-off at the level of the thoracolumbar junction on the left. Pain was elicited on palpation of the thoracolumbar junction. Neurolocalization was to the T3–L3 spinal cord segments.

Investigations

The results of a complete blood count and serum biochemistry profile were unremarkable. The dog was anesthetized and a full-body CT without contrast was performed with a helical, multidetector, 16-slice system^a, and MRI of the thoracolumbar region was performed with a low-field 0.25 T permanent magnet^b. Acquired MRI sequences were sagittal and transverse T2-weighted (T2W), dorsal and transverse short TI inversion recovery, transverse fluid attenuation inversion recovery, transverse T1-weighted (T1W) images before and after contrast administration, and a balanced steady-state free precession sequence to obtain a detailed volumetric data set of the thoracolumbar region.

CT images revealed a well-defined, focal, osteolytic lesion centered at the left T9 vertebral lamina with destruction of both cortical and cancellous bone. No surrounding sclerosis or irregular periosteal proliferation was noted, and no soft-tissue abnormalities were seen (Figure 1).

View Figure

• Download as Powerpoint
• Download Figure

MRI identified a large (37 mm × 6 mm), well-defined and well-demarcated dorsal extradural mass lesion overlying the spinal cord along the T8–T10 vertebral bodies. The lesion was hyperintense to the spinal cord on T2W and short TI inversion recovery images, and hypointense to the spinal cord on T1W images with severe homogeneous contrast enhancement. The mass caused severe (>50%) dorsal spinal cord compression (Figure 1).

Treatment

The dog’s owners elected for decompressive surgery to be performed. A left-sided hemilaminectomy at the level of T7–T11 was performed. The T7–T11 vertebral canal was exposed and a mass of reddish coloration with a mild focal region of hemorrhage identified dorsal to the spinal cord. The friable mass was overlying and masking the spinal cord but was able to be removed in several portions to relieve the compression (Figure 2). As the mass was invading/compressing the overlying base of the spinous process, this was removed with the drill to try to obtain clear margins. Macroscopically, the totality of the soft-tissue component of the tumor appeared to be resected, and the extradural material was sent for histopathology. The bone component was not sampled, as drilling at that level took place during the surgical approach.

View Figure

• Download as Powerpoint
• Download Figure

There was quick improvement following surgery, with the dog having recovered nociception 24 hr after surgery and being discharged with nonambulatory paraparesis 3 days after surgery. The dog was prescribed 4 wk of cage rest at discharge, followed by an increasing exercise regime. There was continuous initial neurological improvement, with the dog becoming ambulatory within the first 3 wk postoperatively.

Histopathological examination (Figure 2) revealed a pleomorphic malignant mesenchymal cell population organized into two distinct growth patterns. The predominant pattern was a solid growth with irregular pleomorphic cells producing an eosinophilic-hyaline matrix consistent with osteoid spicules, that were often mineralized. Anisokaryosis was present, and nuclei were large and irregular with reduced chromatin. Mitotic figures were abundant, with 2–3 mitotic figures per low-power field (×40). Often, there were scattered multinucleate giant cells mixed in with the previously described neoplastic cell population. The second pattern featured neoplastic cells lining small and large blood-filled spaces throughout the tumor and was associated with hemorrhagic foci. Multiple foci of fibrinous eosinophilic material were surrounded by neoplastic cells and activated macrophages accumulating a brownish pigment in their cytoplasm. Some pycnotic cells were present. The neoplastic cell population was found to be penetrating adipose tissue, the extent of which indicated the aggressiveness of the tumor. Margin status was presumed not possible to evaluate, as the lesion invaded the vertebra. A definitive diagnosis of a telangiectatic osteosarcoma was made. Because the boundary of the osteosarcoma entered the vertebra, this was not considered an extraskeletal osteosarcoma.

The owner elected for the dog to have chemotherapy treatment, which was offered as carboplatin 192 mg dose, to be repeated every 3 wk, for at least four times. The first dosage was administered 2 wk after surgical removal and induced a strong emetic response that improved with symptomatic treatment. The owners refused to repeat chemotherapy administration following these side effects. Five wk postoperatively, the dog acutely deteriorated and was again nonambulatory and unable to stand. Considering the poor diagnosis, further investigations were refused by the owners and euthanasia was performed.

Discussion

This is the first ever reported case of a vertebral TOSA in a dog where both MRI and CT images were obtained. Similar to the two previously reported cases of vertebral TOSAs, patients presented with neurological deficits including paresis or plegia and ataxia and were euthanized within weeks of diagnosis because of debilitating clinical signs and poor quality of life.¹⁰ As TOSAs account for only 8–10% of OSAs in dogs,^11,12 and are incredibly rare in the vertebrae, there is little information available about the treatment of this condition in the literature, with no established therapeutic approach or prognosis.

The gold standard method of diagnosis for TOSA is achieved through bone biopsy and histopathology,^6,11 which may potentially be combined with immunohistochemistry to rule out other conditions that present similarly.¹⁰ However, biopsy of a vertebral site can be difficult; therefore, imaging interpretation is pursued in the first instance to create a list of differential diagnoses. Radiography, MRI, and CT imaging features can be used to produce a tentative diagnosis of OSA but are not sensitive for ruling out the other imaging differentials.^13–17 Based on the imaging features found in our case, this might also be the case in vertebral TOSAs in dogs. The main imaging differential for TOSA is an aneurysmal bone cyst, which can be ruled out based on imaging features.^5,13,15 However, other imaging differentials present very similarly to TOSA; therefore, it is important to rule out hemangiosarcomas, giant cell tumors, and metastatic carcinomas based on their different histopathological characteristics.^5,10,13

Radiographic features of all categories of skeletal OSA include osteolytic or osteoblastic lesions, loss of cortical bone, possible mineralization, periosteal proliferation, periosteal lifting as a result of subperiosteal hemorrhage (Codman’s triangle), and palisading cortical bone (sunburst effect).¹⁸ Interestingly, two reviews on the radiographic features of TOSAs have identified some minor distinctions, with TOSAs being purely lytic and having little or no matrix mineralization or new periosteal bone formation.^15,17 With imaging alone, these features can be confused with aneurysmal bone cysts, chondrosarcomas, and fibrosarcomas.¹⁸ CT imaging shows the same above features with hypoattenuation in osteolytic regions, and hyperattenuation where there is matrix mineralization.¹⁵

Less information is available about the MRI features of OSAs. However, one review identified some common features of vertebral OSA in dogs.¹⁹ In all canine cases, the lesion is heterogeneous in T2W sequences and postcontrast contrast enhancing in T1W sequences.¹⁹ These features are consistent with our case. Other common features include osteolysis, osteosclerosis, aggressive periosteal proliferation, better margin delineation with fat saturation, T1W isointensity, and spinal cord compression.¹⁹ Common but less frequent features include T2W hyperintensity¹⁹ and fluid-fluid levels.^7,17 It should be noted that having multiple lesions in the same bone does not rule OSA out, as skip metastases were observed in some cases.¹⁹ In TOSAs, sometimes there are T1W high-intensity regions within hemorrhagic spaces,¹⁷ which was not observed in our case. Gradient echo sequences would have given further information regarding the presence of hemorrhage; however, this was not available on the MRI equipment used to acquire the images in this case.

The current gold standard treatment for appendicular OSA is amputation with adjunctive chemotherapy.^3,4,6,11 In one study, median survival times for appendicular TOSAs were 3 mo without treatment, 4.5 mo with amputation, and 7 mo with amputation and chemotherapy, which are reported to be shorter than for other forms of OSA.¹¹ Although it is unclear whether there is a difference between survival times in canine patients with appendicular and axial OSA, research on human cases has established a significant decrease in prognosis and survival time associated with axial OSA.²⁰

Literature suggests that the sooner chemotherapy is instated after amputation, the longer the expected survival times, with the best results being achieved when chemotherapy is started within 5 days of surgery.¹⁴ Additionally, metastasis may be a negative prognostic indicator that warrants a guarded prognosis.^4,6 The primary cause of death in cases of OSA is metastatic disease, which commonly affects the lungs and other organs.^4,14 Therefore, it is recommended to perform thoracic imaging before commencing treatment to rule out lung metastases.⁶

The vertebral location of the tumor in our dog made removal of the whole bone involved challenging, even if, in hindsight, a more aggressive removal of the dorsal spinous process could have been possible with chemotherapy having also been attempted. Chemotherapy alone would likely have a poorer prognosis, as obtaining tumor-free margins during excision is an important prognostic indicator.³ Our case did not have tumor-free margins and had a survival time of only 5 wk after surgery. Only one dose of chemotherapy was provided, with the patient deteriorating soon after.

It is interesting to note that the prognosis of TOSAs in humans was believed to be poor in the past, but new evidence suggests that new chemotherapy protocols could achieve the same prognosis as in other subtypes of OSA.¹⁷ Poor prognosis is expected in TOSA in dogs; however, it is possible this suspicion could be biased by the limited number of cases reported and the rarity of this condition. Less likely, selection bias for euthanasia as a result of negative expectations by the attending veterinary surgeon could possibly make owners less likely to pursue chemotherapy, even if a reliable TOSA diagnosis relies on histopathology performed following a definite surgical treatment. More research is required to determine whether the prognosis for TOSAs in dogs may be better than originally assumed.

Conclusion

This is the first reported case of a vertebral TOSA in a dog where MRI and CT images were obtained. Suspicion of an OSA could arise when CT images show hypoattenuation in regions of osteolysis and hyperattenuation in areas of mineral deposition. On MRI, lesions can appear heterogeneous in T2W sequences and postcontrast enhancing in T1W sequences, and intralesional hemorrhagic regions would be expected. Currently, there are no established treatment options for vertebral TOSA and the prognosis appears to be poor. Further research on a larger number of cases would be required to establish more effective treatment protocols and investigate the outcome of vertebral TOSAs in dogs.

The authors declare that informed consent was obtained from the pet owners. The patient was managed according to the standard of care described in the JAAHA policy statement.