Use of the IlluminOss Photodynamic Bone Stabilisation System Following Stereotactic Body Radiation Therapy for Limb‐Sparing in Six Dogs With Presumptive Appendicular Osteosarcoma

ABSTRACT

Osteosarcoma is the most common primary bone tumour in dogs, with approximately 75% of cases occurring in the appendicular skeleton. Standard‐of‐care therapy consists of limb amputation followed by carboplatin chemotherapy. A minority of dogs have limb‐sparing treatments including stereotactic body radiation therapy (SBRT) or surgery. SBRT has a high incidence of pathologic fracture (41%–80% of cases) following treatment, which often results in amputation or euthanasia. Prophylactic bone stabilisation following SBRT has not been recommended due to high complication rates. The IlluminOss system (IS) is an intramedullary implant consisting of a light‐curable polymer within an inflatable balloon catheter, which is approved in humans for traumatic fractures and prevention of fractures secondary to bone metastases. We performed a study to assess the safety and feasibility of using the IS following SBRT in six dogs with suspected appendicular osteosarcoma. The IS was placed the day after completion of SBRT. Five of six dogs received adjuvant carboplatin chemotherapy. The IS was well tolerated with no implant‐associated infections or immediate post‐operative complications occurring. Pathologic fractures occurred in four dogs with a median fracture‐free interval of 107 days. Three dogs ultimately underwent amputation due to pathologic fractures. No other post‐operative complications occurred. The median survival time was 156 days. These findings suggest that the IS may be a palliative option for dogs following SBRT, but further research to optimise use in the dog and evaluation of the use of the IS with ancillary fixation is necessary.

1. Introduction

Osteosarcoma is the most common primary bone tumour in the dog, accounting for approximately 85% of canine primary bone tumours [1, 2, 3]. Approximately 75% of osteosarcoma tumours develop in the appendicular skeleton and occur twice as frequently in the thoracic limbs compared to the pelvic limbs [2, 3, 4]. The distal radius is the most commonly affected site, followed by the proximal humerus; the remainder of appendicular cases are relatively evenly distributed among the proximal and distal tibia and the proximal and distal femur [3, 4]. Although only 5% of dogs present with radiographic evidence of metastasis, more than 90% have micrometastases at diagnosis [2, 5, 6], and treatment outcomes remain disappointing with only approximately 50% of dogs alive one year after diagnosis [1, 6, 7, 8].

Standard‐of‐care curative‐intent treatment consists of local tumour control via surgical excision, most commonly limb amputation, followed by adjuvant chemotherapy typically consisting of carboplatin and/or doxorubicin [2, 7, 9, 10, 11]. Although many dogs adapt well to amputation, with up to 86% of owners reporting they would make the same decision again [12], some dogs are poor candidates for amputation due to orthopaedic or neurologic conditions [9, 13, 14, 15]. Numerous limb‐sparing techniques have been described in dogs at various anatomic sites [7, 16, 17, 18, 19, 20, 21, 22, 23, 24]; however, their routine use has been severely limited by high rates of major complications, including infection (30%–70%) [7, 16, 17, 25, 26], implant failure (40%) [25], tumour recurrence (20%–28%, but reported as high as 60% in one study) [16, 17, 25, 27] and pathologic fractures [16, 21, 27].

In recent years, stereotactic body radiation therapy (SBRT) has been utilised as a minimally‐invasive, limb‐sparing treatment [22, 28, 29, 30]. SBRT is advantageous in that it can treat osteosarcoma in any anatomic site; however, it bears a high rate of pathologic fracture secondary to bone integrity compromise from tumour lysis, with fractures reported in 41%–80% of treated dogs [29, 30, 31]. Open surgical prophylactic stabilisation using an intramedullary nail or plating following SBRT has been reported, but was associated with an 88% complication rate, likely due to periosteal vascular disruption and soft tissue trauma from the combination of SBRT and the surgical approach, and is not recommended [32]. The high rate of pathologic fracture following SBRT supports the need for bone stabilisation after treatment; however, the high complication rate with previously reported methods of bone stabilisation has limited available options for these patients.

The IlluminOss System (IS; IlluminOss Medical, USA) is a minimally invasive intramedullary device approved in humans for treatment of traumatic fractures and pathologic fractures in osteoporotic bone secondary to metastatic disease [33, 34, 35]. The IS was initially approved for use in the humerus, radius and ulna [34]; more recently, it received expanded clearance for use in treatment of humeral, radial, ulnar, clavicular, pelvic, fibular, metacarpal, metatarsal and phalangeal pathologic fractures or impending pathologic fractures in skeletally mature patients. Placement involves percutaneous placement of a balloon catheter into the medullary canal, which is filled with a biocompatible liquid monomer that conforms to the canal that is cured using blue light delivered via an optical fibre [33, 36, 37]. The device is radiolucent, permitting accurate visualisation of the cortex on follow‐up imaging and minimising scatter if subsequent radiotherapy is elected. Additionally, supplementary fixation methods, such as screws and bone plates, can be applied anywhere along the device for augmentation, particularly in weight‐bearing bones [33, 35]. Because placement is percutaneous, soft tissue trauma is markedly reduced compared to traditional open stabilisation procedures. Initial clinical trials in humans indicate that the IS provides robust fixation and stability in compromised bone and has low complication rates [33, 36, 37].

The primary objective of this study was to assess the feasibility and safety of the use of the IlluminOss photodynamic bone stabilisation system following SBRT in dogs with suspected appendicular osteosarcoma.

2. Methods

2.1. Case Selection

Dogs presenting for evaluation of a presumptive appendicular primary bone tumour were screened for enrollment between 22 January 2020 and 15 June 2020 at the Ohio State University Veterinary Medical Center. Dogs were eligible for inclusion if they had a body weight greater than or equal to 15 kg and a presenting complaint of a single appendicular lesion suspected to be osteosarcoma based on cytology and radiographs (proliferative, osteolytic or combination of proliferative and osteolytic lesion). Eligible dogs had no history of prior neoplastic processes or known current neoplastic processes other than the primary bone lesion. Dogs were enrolled if owners declined amputation and elected to proceed with SBRT and adjuvant chemotherapy. All prospective dogs received standard‐of‐care diagnostic screening consisting of a complete blood count, a serum biochemistry, thoracic radiographs or computed tomography (CT) and either abdominal ultrasound or abdominal CT. Dogs were excluded from the study if another neoplastic process was identified during screening, pulmonary metastases were detected at the initial appointment on thoracic radiographs or CT, or if there was radiographic evidence of a pathologic fracture. The study protocol was approved by the Institutional Animal Care and Use Committee (IACUC #2025A00000013). Informed and written owner consent was obtained for each dog before enrollment in the study.

2.2. Clinical Data

Clinical information compiled included breed, age, sex, neuter status, body weight at presentation, tumour location and pre‐operative cytologic or histopathologic diagnosis. Pre‐operative serum chemistries were performed on all dogs and alanine aminotransferase (ALT) and alkaline phosphatase (ALP) were recorded. Complete blood counts (CBCs) were performed prior to administration of each carboplatin dose. Additional information collected included the time to documented local progression or development of metastatic disease, whether a pathologic fracture was detected on limb radiographs and, if so, characteristics of the fracture, whether fractures or local progression were treated with amputation, and overall survival. The study period was defined as the time from initial presentation to the Integrated Oncology Service at The Ohio State University Veterinary Medical Center to the time of euthanasia. If a necropsy was performed, data compiled included the final histopathologic diagnosis, presence of metastatic disease and location of metastases, and whether the IS implant remained intact.

2.3. SBRT

Prior to SBRT, all dogs underwent an SBRT‐simulation computed tomography (CT) scan under general anaesthesia with a non‐standardised anaesthetic protocol. Patients were placed in lateral recumbency, affected limb down, in a moldable bag (Vac‐Lock Cushions; CIVCO Medical Solutions, Coralville, IA) that immobilised the body and affected limb for the CT simulations. Treatment plans were created using the 1.25 mm pre‐contrast and post‐contrast CT scans using the Varian Eclipse treatment planning system (Varian Medical Systems Inc. Palo Alto, California). The gross tumour volume (GTV) was identified and contoured and included all identifiable tumour based on contrast‐enhancement of the medullary cavity and any adjacent periosteal proliferation. The GTV was expanded proximally and distally from identifiable tumour 1.5 to 2.0 cm as a clinical target volume (CTV) was not used. A 3 mm isotropic planned target volume (PTV) expansion was created around the GTV to account for daily set‐up positioning error. The treatment plans consisted of 7–9 coplanar 6MV beams aligned to a single isocenter. Patients were treated with 3 daily fractions of 8 Gy (1 patient) or 10 Gy (5 patients). In all cases, the dose to 95% of the GTV exceeded 95% and the dose to 95% of the PTV exceeded 95%.

2.4. IS Placement

Placement of the IS device was performed on the day following the final SBRT fraction. In all cases, the surgical sites were routinely aseptically prepared and draped. For humeral lesions, an approximately 1 cm incision was made at the craniolateral aspect of the humerus over the greater tubercle and a 2 mm intramedullary pin was inserted at the level of the greater tubercle and directed distally along the medullary cavity. For one radial lesion, an approximately 3 cm incision was made on the proximolateral antebrachium and a 2 mm intramedullary pin was inserted distal to the humeral condyle and directed distally along the length of the radius. The other radial lesion was approached via an approximately 1 cm incision on the cranial aspect of the distal radius. A 2 mm intramedullary pin was inserted into the cranial aspect of the distal radial metaphysis and directed proximally along the medullary canal. For tibial lesions, an approximately 1 cm incision was made medial to the patellar tendon on the proximal tibia and a 2 mm intramedullary pin was inserted through the medial ridge of the tibial plateau and directed distally along the medullary canal. In all cases, insertion of the intramedullary pin was followed by a straight or curved cannulated awl, which was used to widen the initial entry portal. A guide wire was inserted into the medullary canal, and flexible cannulated reamers were used to clean the canal prior to insertion of the IS sheath and balloon catheter. Positioning was confirmed by fluoroscopy prior to removal of the sheath. The IS monomer was injected until the balloon was fully inflated. An external blue light source was applied for 500–800 s as recommended by the manufacturer for the designated implant size to polymerise the monomer prior to cutting the balloon catheter off the implant. The cortical portal access was filled with the balloon. After polymerisation of the monomer, the balloon was then cut flush to the cortex. The incision was closed routinely. All dogs received cefazolin 25 mg/kg intravenously at induction and every 90 min intra‐operatively and were discharged on cephalexin 22–30 mg/kg by mouth every 12 h and carprofen 2.2 mg/kg by mouth every 12 h for 5 days post‐operatively. All dogs were discharged to their owners for ongoing care and monitoring within 24 h of the IS implantation procedure. A summary of the implant placement procedure is provided in Figure 1.

FIGURE 1.

Schematic of placement of the IlluminOss device in a human humerus. A small skin incision is performed and an entry portal on the target bone is made using a cannulated awl. The medullary canal is prepared using a flexible cannulated reamer. (A) The balloon catheter is inserted into the prepared medullary canal. The desired placement is confirmed using fluoroscopy. (B) The balloon catheter is filled with monomer until a constant pressure is maintained and no additional monomer can be infused. (C) The light fibre is connected to the light console and light is applied to cure the monomer. The duration of light curing is dependent on the size of the implant. The light fibre is disconnected and discarded following completion of curing. (D) The catheter tube is separated from the implant and discarded. Following skin closure, the implant can be assessed for proper deployment radiographically.

2.5. Post‐Operative Chemotherapy and Monitoring

Intravenous carboplatin at a dose of 300 mg/m [2] was initiated 2 weeks following implantation of the IS. Carboplatin was given every 3 weeks for a maximum of six treatments. Chemotherapy was discontinued if local progression or pulmonary metastases were noted. Starting 3 weeks after implantation of the IS, limb radiographs and standard three‐view thoracic radiographs were performed every 3 months to monitor for local progression, pathologic fracture and pulmonary metastases.

2.6. Statistical Analysis

Descriptive statistics were calculated to summarise dog signalment information. Time to first event was calculated as the time between the first day of treatment with SBRT and documentation of progressive local disease or metastatic disease. Events included local progression of the primary tumour, fracture of the treated limb and metastatic progression. Survival time was calculated as the time between the first day of treatment with SBRT and death. Dogs that were lost to follow‐up were censored at the last date of owner contact. One dog was censored from median time to fracture and median survival time after an external trauma caused multiple comminuted fractures of both the tumour‐bearing radius and non‐tumour‐bearing ulna.

2.7. Cell Line Validation Statement

No cell lines were used in the current study.

3. Results

3.1. Study Population

A total of six dogs with suspected appendicular osteosarcoma were enrolled. Pre‐operative complete blood counts and serum biochemistries revealed no clinically significant abnormalities in any dog. The tumour locations consisted of the distal tibia (n = 2), distal radius (n = 2) and proximal humerus (n = 2). One dog with a distal radius tumour was excluded from analysis following SBRT and IS implantation due to experiencing a major trauma falling down a mobility ramp 17 days post‐operatively that resulted in comminuted diaphyseal fractures of the osteosarcoma‐bearing radius and tumour‐free ulna. Amputation was recommended and declined by the owner, at which time euthanasia was recommended, and the dog was lost to follow‐up. Due to the known trauma, the dog was censored from the study. The signalment information for all dogs is presented in Table 1.

TABLE 1.

Signalment information of enrolled dogs.

Dog number	Breed	Sex	Age at presentation (years)	Body weight at presentation (kg)	Tumour location
1	German Shorthaired Pointer	FS	11	24.9	Right distal tibia
2	Great Pyrenees	FS	7	74.3	Left distal radius
3	Great Dane	MN	8	60.6	Right mid‐distal radius
4	German Shepherd Dog	MN	10	42.4	Left distal tibia
5	Greyhound	FS	7	23.2	Left proximal humerus
6	Greyhound	MN	6	34.2	Right proximal humerus

Note: The median age at enrollment was 7.5 years. The median body weight at enrollment was 38.3 kg, with a mean of 43.3 kg. The primary tumour location was evenly distributed between the distal radius (n = 2), distal tibia (n = 2) and proximal humerus (n = 2).

3.2. IS Placement & Immediate Post‐Operative Period

In all dogs, the implant was placed successfully (Figure 2) and no intra‐operative complications were experienced. The procedure was well‐tolerated, and all dogs were able to be discharged for ongoing care and monitoring within 24 h of surgery. No dogs developed post‐operative surgical site infections or implant‐associated infections.

FIGURE 2.

Post‐operative radiographs of IlluminOss implants in the dog. (A) Post‐operative placement in the humerus; (B) radius and (C) tibia. In all radiographs, the implant fills the medullary canal, shown by the radiopaque light fibre cable evenly extending across the medullary canal. No fractures are evident in any of the radiographs.

3.3. Outcomes

The outcomes for each dog are summarised in Table 2. Of the five dogs that received carboplatin, one received the maximum of six doses without any documented local or metastatic progression (Dog #1). The dog had no evidence of disease progression until 744 days after starting SBRT, at which time the dog became acutely mildly lame in the right pelvic limb. Radiographs revealed a closed, complete, pathologic fracture of the right distal tibia with an intact IS. An amputation was performed and the dog continued to do well until the owner reported that the dog was diagnosed with osteosarcoma in another limb 890 days after starting SBRT. The dog was subsequently lost to follow‐up.

TABLE 2.

Outcomes of enrolled dogs.

Dog number	Body weight (kg)	Tumour location	Number of carboplatin treatments	Event‐free interval (days since first SBRT)	Event type	Implant failure (yes/no)	Survival time (days)
1	24.9	Right distal tibia	6	744	Pathologic fracture	No	Lost to follow‐up after 890 days
2	74.3	Left distal radius	0	17	Multiple comminuted fractures and implant failure (trauma)	Yes	Lost to follow‐up after 17 days; euthanasia was recommended.
3	60.6	Right distal radius	3	60	Local progression + pathologic fracture	No	747
4	42.4	Left distal tibia	4	106	Pathologic fracture	No	156
5	23.2	Left proximal humerus	2 a	19	Pathologic fracture	No	102
6	34.2	Right proximal humerus	3	73	Pulmonary metastasis	No	95

Note: Dog #2 was censored from further analyses due to experiencing a trauma 2 weeks post‐operatively that resulted in mid‐diaphyseal comminuted fractures of the left radius and ulna. Amputation was recommended and declined by the owner; euthanasia was recommended if not proceeding with amputation. The dog was lost to follow‐up. Dog #3 had local progression noted 60 days post‐SBRT; a pathologic fracture was identified 108 days post‐SBRT, at which time the dog received an amputation. Dog #4 received 4 doses of carboplatin before owners elected to discontinue chemotherapy.

^a Dog #5 experienced a pathologic fracture 19 days after starting SBRT and received an amputation prior to receiving 2 doses of carboplatin.

One dog received four doses of carboplatin before chemotherapy was discontinued at the owner's request (Dog #4). Two dogs received three carboplatin doses prior to progression (local progression without fracture, Dog #3; pulmonary metastatic progression, Dog #6). Dog #5 was presented to the Emergency Service 16 days post‐operatively for acute onset of vocalising and severe lameness. Radiographs of the treated limb performed at another emergency clinic immediately prior to presentation revealed a long, spiral, pathologic fracture of the humeral diaphysis and the dog underwent a limb amputation. Following recovery from amputation, Dog #5 was treated with two doses of carboplatin before pulmonary metastases were observed on thoracic radiographs.

Pathologic fractures ultimately developed in Dog #1, Dog #3, Dog #4 and Dog #5 over a wide range of time (744, 108, 106 and 19 days post‐SBRT, respectively). The fractures occurred in both dogs with tibial lesions (Dogs #1 and #4), one dog with a radial lesion (Dog #3) and one dog with a humeral lesion (Dog #5). Dog #3 developed locally progressive disease that was first noted 60 days post‐SBRT, and a pathologic fracture was first noted on radiographs 108 days post‐SBRT. Three of the fractures (Dog #1, Dog #3 and Dog #5) occurred along the diaphysis of the bone and were over a portion of the IS. Dog #4's fracture was located in the epiphysis distal to the end of the IS. In Dogs #1 and #3, the medullary canal was accessed distally, and the fractures occurred proximal to or at a distance from the insertion site. In Dog #4, the medullary canal was accessed proximally, and the fracture occurred at the distal epiphysis, at a distance from the insertion site. In Dog #5, the canal was accessed through the proximal greater tubercle, and the fracture occurred within the proximal to mid diaphysis, meaning a stress riser contribution at the insertion site cannot be excluded. The fractures can be viewed in Figure 3. Dogs #1, #3 and #5 ultimately underwent limb amputation with a median limb‐spare duration of 108 days. Dog #4 had a pathologic fracture identified on limb radiographs on day 106 post‐SBRT. On examination, the dog showed no clinical lameness or pain on palpation, and the owner reported no observable lameness at home during normal activity. Amputation was recommended, but the owner elected to monitor for development of clinical signs and declined amputation. Dog #4 developed a fracture of the right eighth rib that was presumed due to metastatic disease, as well as an additional pathologic fracture of the distomedial aspect of the distal tibia on day 150 post‐SBRT. The dog was euthanised 156 days post‐SBRT. The overall median event‐free‐survival time was 73 days (range: 19–744 days) and the median survival time was 156 days (range: 95–890 days).

FIGURE 3.

Lateral radiographs of pathologic fractures that occurred in the study population. (A) Dog #1 presented with a closed, mildly cranially displaced oblique fracture of the distal tibial diaphysis 744 days post‐SBRT. (B) Dog #3 developed a closed, minimally displaced fracture of the proximal diaphysis of the radius and a closed, incomplete fracture of the mid‐diaphysis of the ulna 107 days post‐SBRT. (C) Dog #4 developed a minimally displaced, complete articular fracture of the distal tibial epiphysis 106 days post‐SBRT. The dog had no apparent lameness at that time. (D) Dog #5 developed a closed, spiral pathologic fracture of the proximal‐mid humeral diaphysis following standing up and shaking off.

4. Discussion

There is an unmet need for additional limb‐salvage options for dogs with appendicular osteosarcoma. The present case series is the first to report the implantation of the IS in dogs with suspected appendicular osteosarcoma receiving SBRT. All six of the dogs in this study received the IS with no intra‐operative or immediate post‐operative complications. No dogs developed surgical‐site infections or implant‐associated infections, unlike high rates of post‐operative infections that have been previously reported with other stabilisation methods following SBRT [32]. However, the high long‐term fracture incidence in our study demonstrates that further optimisation of the IS in dogs would be necessary for fracture prevention.

Of the five dogs in our study cohort, four dogs ultimately developed pathologic fractures; notably, three of the dogs were doing well at home and presented for only mild (grade one out of five) weight‐bearing lameness. The fractures in Dog #1, Dog #3 and Dog #4 were located markedly distal to the insertion site and thus it is considered unlikely that these fractures occurred secondary to iatrogenic stress riser formation during implant placement. The fracture in Dog #5 occurred within the proximal to mid diaphysis of the humerus and the medullary canal was originally accessed through the greater tubercle; in this case, stress riser contribution to the fracture cannot be ruled out. Dog #5's fracture occurred secondary to large rotational forces that the IS does not counteract; the owner reported that the dog stood and shook immediately prior to vocalising. Although intramedullary implants provide stabilisation against bending and shear forces, they do not mitigate rotational forces, and application of ancillary stabilisation that does protect against rotational forces, such as plating along the implant, may be necessary. Additionally, implantation of an intramedullary implant may further weaken the bone through the access point; the IS device attempts to mitigate this risk by filling the cortical defect in the bone with the balloon. Three of the four fractures in this case series occurred at the opposite end of the diaphysis from the access point and do not support that the fractures occurred secondary to device implantation.

In our case series, the tumour‐bearing bone was stabilised in all dogs perceived to be at risk of fracturing based on radiographs. However, more sensitive methods of predicting fracture risk and identifying suitable candidates for intramedullary implants could be instrumental in identifying patients that could benefit from this intervention and warrant further investigation. In humans, fracture risk is objectively determined using methods like CT‐based finite element analysis (CT‐FEA) for structural assessment of neoplasia in long bones [38, 39, 40, 41, 42]. CT‐FEA has high reported sensitivity (100%) and specificity (90%–94%) in predicting pathologic fractures in human patients [39]. Use of CT‐derived structural analysis to affect clinical decision‐making for veterinary patients has not been reported; however, retrospective analysis of dogs with antebrachial osteosarcoma demonstrated that CT‐FEA can be used to determine fracture risk successfully [42]. Future investigations into bone stabilisation in dogs with osteosarcoma may benefit from utilising fracture risk assessments such as CT‐FEA for assessing the benefit of an implant and are currently ongoing using data from this preliminary study with a new cohort of dogs.

All of the dogs in this case series received the IS with no ancillary fixation applied. It is likely that for optimal use in weight‐bearing bones in the dog, augmentation with Kirschner wires, screws or plate and screws is necessary. The need for ancillary fixation with the IS in weight‐bearing bones has been recognised in the human literature [33, 37, 43]; the IS received FDA clearance for use in the femur and tibia specifically as supplemental fixation to existing hardware systems [44]. Terek et al. reported in 2024 that standalone IS use warrants caution in complete pathological fractures due to device breakage, and suggested that supplemental plate or screw fixation may be necessary in these cases [37]. However, robust clinical outcome data specifically evaluating adjunctive fixation with the IS in weight‐bearing long bones remain limited, and its benefit in this context has yet to be formally established even in humans. Three dogs in this case series ultimately required limb amputation due to pathologic fracture (n = 2) or local progression with subsequent pathologic fracture (n = 1) (Table 3). The median survival time of all dogs was 156 days, which is comparable to or slightly lower than what is reported for dogs treated with SBRT alone [30, 45]. Our data indicate that the IS as a sole, prophylactic stabilisation technique for SBRT‐treated osteosarcoma does not prevent pathologic fractures; however, use of the IS with ancillary fixation methods, as recommended in humans, has not been assessed in the dog. Outcomes may be improved with the use of structural analysis and predictive modelling to identify suitable candidates and recommend supplementary fixation methods to augment the stabilisation. In comparison with other reports of limb‐sparing techniques in dogs that have high post‐operative infection rates [32], there were no surgical site or implant‐associated infections in any of the dogs treated in this study. This may be in part due to the small sample size, but suggests that the IS may have a lower risk of post‐operative infection, which may be due to the percutaneous approach and small incision.

TABLE 3.

Summary of amputations performed.

Dog number	Reason for amputation	Time to fracture (days)	Implant status (intact/fractured)	Duration of limb spare (days)
1	Pathologic fracture; mildly cranially displaced oblique fracture of distal tibial diaphysis	744	Intact	745
3	Local progression with pathologic fracture; minimally displaced oblique fractures of proximal radial diaphysis and ulnar diaphysis	108	Intact	108
5	Pathologic fracture; long oblique humeral diaphyseal fracture	19	Intact	23

Note: Three of six dogs ultimately received limb amputations due to pathologic fracture (n = 2) or local progression (n = 1). In both cases with fractures and subsequent amputations, the fractures were long, oblique fractures. 1 additional dog experienced a tibial articular fracture that was incidentally identified at a routine recheck; due to lack of clinical signs, the dog's owner declined to proceed with amputation and continued to monitor. The dog was ultimately euthanised 106 days post‐SBRT due to development of an additional pathologic fracture.

Because this study was assessing initial safety and feasibility, the sample size was small. Dogs enrolled in this study were not required to have a histopathologic diagnosis of osteosarcoma, and five dogs were enrolled based on cytology consistent with sarcoma and ALP‐positive on immunocytochemistry along with radiographic confirmation of a lytic, proliferative or combination lytic and proliferative appendicular lesion. Although this represents clinical practice, there was the chance of inclusion of other tumour types. This was the case for Dog #6, where histopathology of the primary tumour performed at necropsy was inconclusive, but suspected to be a hemangiosarcoma or myxosarcoma rather than an osteosarcoma. It is not known how different sarcoma types affect the loading strength of the surrounding bone. Further evaluation of the effects of tumour type on loading strength would be needed to optimise the use of intramedullary implants in these patients.

In conclusion, the present study found that the IS was well‐tolerated in dogs and was associated with no post‐operative infections, but the high failure rate demonstrates that the IS, as used here without ancillary fixation or fracture risk assessment‐guided patient selection, is not a viable standalone technique for prevention of pathological fractures following SBRT. There are currently no limb‐sparing techniques widely recommended for dogs due to unacceptably high complication rates; because the IS was otherwise associated with few complications and has shown tolerability in humans when used in appropriate cases [33, 35, 37], further investigation into optimising its use in dogs could be warranted. This could specifically evaluate (1) the use of CT‐FEA or other structural analysis tools to identify dogs most likely to benefit from intramedullary stabilisation; and (2) the use of ancillary fixation methods (e.g., Kirschner wires, cortical screws or plate and screw constructs) in combination with the IS to address the rotational and torsional forces that the device alone does not counteract. These modifications are being evaluated in an ongoing prospective cohort to determine whether the IS can be optimised for use as a limb‐sparing treatment option in the dog.

Funding

This research was funded by funding provided to The Ohio State University College of Veterinary Medicine Canine Research Fund by Ohio Revised Code 955.14.

Disclosure

The images used in Figure 1 are copyrighted material held by IlluminOss Medical and were used with written permission. This work was presented as an abstract at the combined Veterinary Cancer Society/Veterinary Society of Surgical Oncology conference in March 2026.

Conflicts of Interest

Robert Rabiner is the founder and Chief Technical Officer of IlluminOss Medical. The other authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.