Osteosarcoma is the most common primary bone tumor in dogs. Lesions typically involve the metaphyseal region of the distal radius, proximal humerus, distal or proximal femur, and distal or proximal tibia. In most instances, both osteoproductive and osteolytic processes are occurring simultaneously. The tumors typically arise from within the bone; however, once cortical bone destruction occurs, portions of the tumor may extend outside the bone, causing soft-tissue swelling. When lesions are left untreated, severe pain and eventually pathologic fracture may occur.
The most commonly recommended treatments for canine appendicular osteosarcoma include medical management (alleviation of associated bone pain), amputation (+/- chemotherapy), limb-sparing surgery and chemotherapy, radiation therapy (palliative, fractionated therapy and stereotactic radiosurgery [SRS]).
Prior to performing limb amputation or limb-sparing surgery, it is best to complete staging including thoracic radiographs (or ideally CT), a bone scan to make sure there is no evidence of occult skeletal metastasis or synchronous primary tumor.
Amputation
Amputation remains the "gold standard" form of treatment for appendicular osteosarcoma, particularly for greyhounds given their musculature, lack of obesity, and low propensity for developmental orthopedic and arthritic conditions. Advantages of amputation include complete tumor removal with almost no chance of recurrence, elimination of pain associated with the tumor, low rate of complications, and relatively low cost. Guidelines are as follows:
Given the possibility of multifocal skeletal OSA (i.e., metastatic or synchronous primary bone lesions), my personal preference is to remove as much bone as possible with amputation. For example, proximal to scapula for forelimb tumors and at the level of the coxofemoral joint (or higher for proximal femoral lesions) for all hindlimb tumors. Although amputation has become the most common form of therapy, some dogs are not considered suitable candidates for amputation due to concurrent orthopedic or neurological conditions or being owned by people that are categorically opposed to the idea of amputation. Thus, there is growing interest among owners of dogs affected with osteosarcoma for alternatives to surgery.
Limb-Sparing Surgery
Limb-sparing is most commonly performed on tumors affecting the distal radius. There have been some reports of other anatomic locations being treated (e.g., proximal humerus, distal tibia); however, implant failure and poor limb function are common. Cost of limb-sparing is very high (ranging from $8,000–$12,000 when adjunctive chemotherapy is factored in).
Canine limb-sparing surgery typically involves performing an osteotomy in the central area of the radial diaphysis and excising the distal portion of the radius and associated bone. The dissection is marginal around the tumor pseudocapsule and is wide with respect to diaphyseal bone. The resultant bone defect is filled with some weight-supporting material and a long bone plate that spans the entire length of the radius, implant and carpal joint and the majority of the 3rd metacarpal bone. Materials that have been used thus far include cortical allografts, cement spacers (PMMA), autoclaved autogenous bone, pasteurized autogenous bone, and a metal radial endoprosthesis (available through VOI - Veterinary Orthopedic Implants). The endoprosthesis was developed as an alternative to the allograft and advantages include lack of a need for a bone bank and shorter operative times due to its ready-to-implant nature. The radial endoprosthesis was originally manufactured as a solid construct of lengths of 122 and 145 mm. A VSSO Endoprosthesis Working Group has made several modifications to the initial endoprosthesis including a locking plate version and a shorter endoprosthesis (98 mm).
Biomechanical testing has shown the endoprosthesis to be superior to the allograft in a cadaveric model that was loaded to failure (Liptak). In that study, limbs were initially cycled from 30% to 100% of body weight to simulate a clinical setting, but there was no evidence of failure in any of the 4 pilot limbs tested after 100,000 cycles. As a result, the limbs were loaded to failure. Fixations with the endoprosthesis had significantly greater yield load, energy at yield and ultimate load when compared to limbs reconstructed with a cortical allograft. In the same study, the impact of ulnar resection was also examined. Preservation of the ulna did increase the load to yield and ultimate failure by 41% and 29%, respectively, in limbs reconstructed with cortical allografts. However, this was not significant and there were no significant differences in stiffness, yield load and energy, and ultimate load and energy at failure with preservation or resection of the ulna in either of the constructs. The clinical implications of this finding are important. First, for dogs being treated with limb-sparing techniques that do not utilize the ulna for replacement of the radius (i.e., ulnar rollover technique - Seguin), resection of the ulna along with the tumor-bearing radius would simplify the surgery, as separation of the distal radius from the ulnar styloid process without disrupting the tumor pseudocapsule can be difficult. This could reduce the rate of local recurrence, which is reported as high as 28% (LaRue, Morello, Withrow 2004). Along the same lines, it can be difficult to determine whether there is ulnar involvement in some distal radial cases, and costly imaging such as MRI is sometimes necessary to determine ulnar involvement. With less concern for ulnar involvement, routine resection of the ulna could be performed in questionable cases without the use of advanced imaging. While these results provide some insight into the importance of ulnar preservation, further evaluation is needed to confirm these cadaveric acute loading findings.
Infection is the most significant complication encountered with limb-sparing surgery. The etiology of infection is multifactorial with factors including extensive soft tissue resection with vascular compromise to a poorly perfused site, poor soft tissue coverage, implantation of large orthopedic implants, non-vascularized bone (allograft), possibly immunogenic cortical bone, and administration of local and/or systemic chemotherapy. Infection occurs in 30–70% of dogs and approximately two-thirds are diagnosed six months or more after surgery. It was hoped that the all-metal nature of the endoprosthesis would reduce the risk of infection or at least make infections easier to resolve than with allografts. However, in a recent report the infection rate was comparable to that of the allograft (Liptak). A number of different bacterial organisms have been cultured with monomicrobial and polymicrobial infections occurring in approximately 50% of cases each. Initially, infections are treated with culture-directed antibiotics, isotonic saline lavages, and wet-to-dry bandages. If unresponsive or recurrent, then antibiotic-impregnated beads (polymethyl methylmethacrylate or absorbable calcium sulphate) can be surgically implanted adjacent to the infection site. Limb amputation is a salvage procedure and is used in a small percentage of dogs with uncontrollable infection. On the positive side, infection has been associated with an increase in survival time, with dogs with infected limbs living nearly twice as long as those with non-infected limbs.
Implant failure occurs in approximately 40% cases, but is catastrophic in only 10% of cases. Implant failure is commonly caused by cycling associated with everyday activity. PMMA injection into the medullary canal of cortical allografts increases screw pullout strength and reduces the incidence of screw loosening, implant failure, and allograft fracture. The Veterinary Society of Surgical Oncology (VSSO) has created an Endoprosthesis Working Group which has recommended modifications to the endoprosthesis.
Local tumor recurrence is caused by incomplete resection or, more commonly, residual neoplastic cells in the soft tissue adjacent to the tumor capsule following marginal resection of the primary bone tumor. The rate of local recurrence is reported as high as 28%; however, this rate has been reduced to less than 10% with the use of locally released chemotherapy agents, such as cisplatin from open-cell polylactic acid biodegradable implants (not commercially available) and appropriate case selection. Local recurrence may either have no effect or a negative impact on survival time, depending on how the owner handles the information. Local recurrence can be managed with a second limb-sparing surgery, amputation, or palliative radiation therapy.
Another surgical limb-sparing technique described by Dr. Nicole Ehrhart utilizes the phenomenon of distraction osteogenesis to fill the bone defect. With this technique, a segment of bone is cut from the remaining bone end and is "transported" (bone transport osteogenesis - BTO) across the defect via circular external skeletal fixation (i.e., Ilizarov).
Stereotactic Radiosurgery
Through a collaborative effort between the University of Florida's College of Veterinary Medicine, Department of Small Animal Clinical Sciences and the UF McKnight Brain Institute, Department of Neurosurgery, the use of stereotactic radiosurgery (SRS) was explored in the late 90s for single fraction treatment of various canine and feline malignancies. Out of this initial investigation, treatment of canine osteosarcoma has emerged as a viable treatment alternative. Unlike standard fractionated radiation therapy, with stereotactic radiosurgery the entire radiation treatment is delivered as a single, large dose in a single setting or up to three doses within a single week (e.g., M, W, F). Treatment delivery involves multiple, non-coplanar beams of radiation that are stereotactically focused on the target, using an image-based system. Unlike fractionated radiation therapy which is delivered via a limited number of conventionally simulated, static fields, stereotactic radiotherapy relies on extreme accuracy of radiation delivery to reduce normal tissue effects, rather than the radiobiological differences in tissue sensitivity and repair capacity utilized with fractionated protocols. This precision allows tumors to be treated with very high doses of radiation while reducing the radiation delivered to the surrounding non-neoplastic tissues.
Techniques vary between institutions. Historically, the University of Florida (UF) was the first to treat dogs with osteosarcoma with SRS (Farese et al. JAVMA, 2004) but the treatment approach has become more widespread across the U.S. over the past 10 years and is now offered in a number of facilities, including private veterinary hospitals. In general, treatments are based on a steep dose gradient, with plans generated at UF treating the center of the lesion with 45–60 Gy and the periphery 30–35 Gy in a single fraction. Complete coverage of the tumor periphery and 2- to 3-cm margins of normally appearing bone proximally and distally with the 30 Gy line is more feasible with smaller lesions and with areas with more muscle mass between the tumor and the skin, such as the proximal humerus. Typically, carboplatin (300 mg/m2) is infused intravenously over 20 minutes just after stereotactic radiosurgery therapy for potential radiosensitizing effects. Following reattachment of the targeting array to the biteplate and tumor localization, radiation therapy is performed with a 6 million electron volt (MeV) linear accelerator. Immediately following treatment, the localizing array, biteplate, and associated pins are removed, the limb wrapped with a soft-padded bandage, and the dog recovered from anesthesia. Adjunctive chemotherapy (single-agent carboplatin or alternating carboplatin/doxorubicin) is continued for treatment of micrometastatic disease.
Nearly any bone can be targeted with this technique, but some adjacent tissues cannot tolerate such high doses (e.g., spinal cord with vertebral OSAs). Skin desquamation is sometimes observed 3–4 weeks after treatment. Pathologic fracture has also been observed when lesions have significant osteolysis, as the bone loses regenerative capacity following such a high radiation dose. Because the treatment plan is different for every case, the treatment is not uniform from dog to dog. Smaller lesions with little extracortical involvement allow greater coverage with radiation without injuring the skin while large lesions with extensive extracortical and soft tissue involvement are more difficult to treat with a dose sufficient for tumor kill without injuring the skin. As we have gained more experience with the technique we have become more selective about the candidates. Though SRS for canine OSA has become an attractive new limb-salvage technique, fracture following SRS continues to be a real challenge and CT algorithms are now being tested to determine whether fractures can be predicted and patient selection can be improved.