A diagnosis of cancer affecting the bone can be devastating for dogs and their owners.1–3 In most instances, the affected bone and associated muscular envelope needs to be removed to achieve the necessary oncologic margins required for a curative-intent resection. This loss of skeletal support will usually have a structural and a functional impact on the patient.4–6 Although amputation is considered the optimal treatment when cancer disrupts the appendicular skeleton,2,3,7 the presence of orthopedic or neurologic deficits affecting other limbs may prevent the patient from adapting to acceptable mobility.
For cancer affecting the appendicular skeleton, various limb-sparing techniques have been used with good success over the past 30 years.2,3,7,8 Although dogs appear to tolerate these techniques well,9 complication rates are high. In a multi-institutional retrospective case series,10 clinically important infection developed in up to 75% of dogs in which limb sparing was performed with a metallic endoprosthesis, and 36% of dogs developed implant failure.10 Because > 50 % of dogs with osteosarcoma will succumb to metastatic disease within a year after treatment is initiated, this high complication rate erodes the enormous benefits that limb-sparing surgery can otherwise provide these patients for their remaining lifetime.
Traditional manufacturing methods for orthopedic implants require a precise mold or tooling apparatus to create a generic design that must fit a wide range of patient conditions.11,12 The surgeon must manipulate the implant at the time of surgery to achieve an optimal fit for a specific patient. By comparison, CAD allows complex and patient-specific 3-D geometries to be designed that are optimally configured to the needs of an individual patient.13–15 Additive manufacturing, otherwise known as 3-D printing, allows realization of the computer-designed implant. Powder-bed fusion technologies (involving lasers or electron beams) allow fabrication of orthopedic implants from various biocompatible materials such as 316L stainless steel, titanium-6 aluminium-4 vanadium, and cobalt-chromium.11 Powder-bed fusion involves use of an energy source to melt and fuse particles of metal powder at precise points as determined by the computer design. When the selective melting of one layer is completed, the building platform is then lowered by a predetermined distance before another layer of powder is applied. This process is repeated until the entire 3-D structure is complete.
Selective laser melting or electron-beam melting has been used for creation of orthopedic implants, each involving a different energy source and different build conditions.11,12,14–17 The CAD and additive manufacturing provide freedom from the design constraints of conventional manufacturing processes,13–15 while keeping build time and costs to a realistic amount.
An advantage of additive manufacturing over traditional manufacturing processes is the ability to incorporate a design-dependent porosity into the structural component of the implant.11,12,16 The porous scaffold enables vascular and cellular ingrowth into the implant, such that new bone tissue may ultimately form directly within the implant itself.18 The porous implant can also be engineered to have an elastic modulus that more closely mimics the actual properties of bone.12,15,16,18–20 The optimal geometry and architecture of the scaffold structure for supporting osteointegration along a length of a large defect remains unknown, but this approach offers the promise of overcoming the issues of biofilm generation and stress protection, complications that are current limitations of wholly metallic implants.10,13,18,21
In human surgery, custom-designed implants22,23 are rapidly gaining acceptance and have been used in various situations, particularly for craniofacial reconstruction and joint reconstruction.15,23,24 Owing to the precise geometric fit of a custom implant, surgical complexity and operating times are reduced, with improved cosmetic outcomes.15 However, even in human surgery, 3-D–printed patient-specific implants remain a novelty, with almost all reports18,24–27 to date consisting of single case examples, with no population studies or long-term analysis available.
Veterinary experiences with CAD-additive manufacturing of orthopedic implants have largely mirrored those of human medicine, with custom-designed implants being used to manage complex surgical problems.14,28 Existing publications13,14,28,29 are mostly limited to conceptual reviews, isolated case reports, or media releases. In an extensive review14 of their activities in this field, Harryson et al allude to several clinical examples in which patient-specific implants were used, including 1 case involving a distal radial endoprosthesis similar in style to that used in the study reported here. Customized endoprostheses have also reportedly been developed for a dog requiring a total knee prosthesis after a gunshot injury28 and for humeral head resurfacing due to severe pathological change affecting the shoulder joint.30 Another publication13 describes the use of custom-printed implants for the reconstruction of craniofacial deficits following trauma, infection, or oncological surgery. The purpose of the study reported here was to describe our experiences with the use of custom-designed patient-specific implants to provide functional replacement of skeletal structures in dogs with tumors of the mandible, radius, or tibia.
Materials and Methods
Case selection criteria
Dogs brought to the Massey University Veterinary Teaching Hospital, Palmerston North, New Zealand; the Centre for Animal Referral & Emergency, Melbourne, Australia; or Veterinary Specialists Auckland, Auckland, New Zealand with a destructive disease process affecting the mandible or distal aspect of the radius or tibia during the period June 2013 to September 2016 were considered eligible for the study. Dogs were considered suitable candidates for a function-sparing endoprosthesis because either their body size or comorbidities made them poor candidates for amputation7,9 or their owners were unwilling to consider amputation or hemimandibulectomy as a treatment option.
All dogs received a complete physical examination prior to study entry. A diagnosis was made by cytologic or histologic examination of tissue obtained directly from the mass or radiography to identify the radiographic appearance typical of osteosarcoma.2 When the primary bone disease was suspected or confirmed to be a cancerous process, tumor staging was completed for all patients, including palpation or aspiration of draining lymph nodes and CT examination of the thorax. Routine hematologic and serum biochemical analysis was performed for all dogs. All owners provided informed consent for the procedure and were advised that although the techniques performed were not novel, the true rate of complications and long-term outcomes for the implant were unknown.
For implant development, CT scans of both the affected and contralateral musculoskeletal compartment were obtained. Computed tomographic data acquisition was performed with between 1.5- and 3-mm-thick slices, with a 1-mm reconstruction interval and a standard filter for soft and bone tissues. For cancer affecting the appendicular skeleton, imaging extended from the phalanges to the elbow or stifle joints. For lesions of the skull, the entire head to the level of vertebra C2 was included.
CAD of the implant
The surgeon reviewed the CT scans with a DICOM viewera and, using the measuring and annotation tools, marked the optimal resection margins directly onto each digital file. For tumors of the radius and tibia, resection margins were marked 5 cm from any periosteal or medullary disruption by the tumor. Because of the more limited bone stock in the mandible, resection margins of at least 1 cm of radiographically normal bone from the tumor border was the goal; typically all of the premolar and molar dentition was removed in the planned resection. The surface volume data from the CT scans, together with the planned resection notations, were exported as an STL file, which was then imported into an appropriate CAD software package by the engineering team. Various software packages were used,b,c which allowed for different surface or solid and mesh modeling tools to be used.
The resection marks made previously were reviewed and confirmed with the surgeon, with mild adjustments made to accommodate the optimal size, geometry, and fixing areas of the implant. The planned resections were then made in the CAD software by use of Boolean subtraction tools to create a mesh model that reflected the bone geometry after tumor removal. Through the use of solid and surface CAD modeling techniques, the base geometry of the implant was overlaid on this mesh file. The base geometry was designed to reflect the surface bone geometry prior to tumor development. Fixation points were created to penetrate remaining bone areas, with the aim of obtaining 4 to 6 screw holes/segment. The design was reviewed with each surgeon prior to finalizing to confirm implant geometry and fixation points.
Once the implant design had been finalized, the areas of the implant to be applied with porous features for osteointegration were identified and split away from other bodies of the digital model. Lattice structures, designed and created in the CAD software, were then applied to this section using Boolean subtraction tools.
The remaining bodies of the implant model were rejoined to form a single implant body and converted into a mesh file format. The resected bone geometry model was then removed from the implant to provide accurate fitting surfaces between the implant and bone. This precise alignment of bone geometry to implant also provides self-alignment during surgery. The final digital implant model was then exported as an STL file ready for additive manufacturing.
A surgical cutting jig was also created by means of CAD to ensure the planned osteotomies could be accurately applied to the affected bone at the time of implant placement. The jig was designed from the original STL file (with affected areas still intact), which was overlaid on the final resection file. Bony landmarks or other anatomic features that ensure the cutting jig could be accurately oriented onto the bone at the time of surgery were identified, and these location points were added to the jig. Cutting templates were then added at the line of planned osteostomy; when necessary, supporting struts were added to the final jig design to tie the various components together into a stable device. The digital jig model was then exported as an STL file for additive manufacturing in thermoplastic.
Implant production
The final STL file of each digital implant model was reviewed by the manufacturing team, and an optimal build orientation was determined. Sacrificial support structures were then added to the digital model in a CAD packaged to ensure the implant would be able to support its own weight on the base plate during the build process. The build file was loaded into a selective laser melting machinee that used a medical-grade titanium Ti64 powder. The build was undertaken in an inert gas atmosphere, by use of preassigned and proprietary laser parameters. After completion of the build process, the physical implant was removed and any residual powder removed, including from within any cavities. The implant was then heat treated for stress relief before being removed from the build plate and cleaned up ready for sterilization. The physical supports added as part of the build process were removed and the edges smoothed with a polishing burr.
Surgery
On receipt of the physical implants, surgeons reviewed their surgical plan for tumor removal and implant placement. In preparation for surgery, anesthesia was provided to each patient as determined by the anesthesiologist or veterinarian supervising the case. Analgesia was also provided, including a combination of regional nerve blocks and parenteral opiate administration. All dogs received NSAIDs and additional opioid medications in the postoperative period. The surgical plan for tumor removal and implant placement for each tumor type was as follows. For this study, surgery reports for all patients were reviewed by the investigators to identify any variations in this plan or the occurrence of any intraoperative complications.
Mandible—Dogs were positioned in dorsal recumbency, and a sterile mouth gag was placed to maintain separation between the dentition of the mandible and maxilla. Soft pads were placed under the neck to keep the body of the mandible parallel with the operating table. Mandibulectomy was performed via a ventral approach. A skin incision was made along the length of the body of the mandible, extending from the mandibular lymph node to the mucocutaneous junction of the lip. The incision was continued through the platysma muscle to expose the underlying bone. The insertion points of the digastricus muscle were sharply elevated from the ventral surface of the mandible. The inferior alveolar artery, vein, and nerve were isolated as they entered the mandibular foramen medially; bupivacaine hydrochloride (1 mL) was instilled around the nerve before ligation and transection of this neurovascular bundle. A periosteal elevator and self-retaining retractors were used to gently elevate the soft tissues from the medial and lateral surfaces of the mandible.
As the tumor was approached, tissues were carefully elevated around the circumference of the pseudocapsule. The gingiva was excised from the symphysis along the edge of the dental arcade; a narrow margin (2 to 3 mm) of tissue was maintained around the expansile border of the tumor. With the oral cavity open, further dissection or elevation of the digastricus or masseter muscles was performed until the tumor was isolated and resection boundaries were clearly identified (Figure 1). In all dogs, preservation of the symphysis cranially and preservation of the temporomandibular joint and angular process of the mandible caudally were achieved. The CAD-designed cutting jig was molded to fit along the full length of the dental arcade.
Representative photographs of the surgical approach to placement of a patient-specific porous titanium mandibular implant in a dog with a mandibular tumor (A to D) and postoperative lateral radiographic image (E). A—A ventral approach to the mandible was used, providing good visibility of the tumor and neurovascular bundles and allowing secure closure of the oral cavity to avoid exposing the implant to contaminated tissues. B—A cutting jig made in thermoplastic was used to ensure the computer-planned osteotomies could be accurately transferred to the actual patient. C—With the cutting jig in place, the planned resection lines could be cut with an oscillating saw. D—After closure of the oral tissues and restoration of the surgical sterile field, the implant was introduced into the defect and secured with screws. E—Postoperative radiography reveals good alignment of the implant and screws.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
After osteotomy of the mandible, the oral mucosa was closed in 2 layers with a simple continuous pattern of 3-0 poligecaprone. When necessary, bone tunnels were used in the symphysis or remaining cortical bone to achieve robust suture security. Once the oral cavity was closed, the surgical wound was lavaged copiously with sterile saline (0.9% NaCl) solution. The sterile surgical field was then restored; the surgeons reapplied gowns and gloves, all surgical drapes were replaced, and a fresh set of sterile instruments was introduced. The sterile titanium implant was then positioned into the defect and secured with titanium screws of the appropriate size (Figure 1). The surgical wound was lavaged with saline solution prior to closure of the subcutaneous tissues and skin in 3 layers. Postoperative radiographs were obtained to confirm implant positioning and accuracy of screw placement.
Radius—Dogs were positioned in dorsal or lateral recumbency. The entire limb from the digits to the proximal aspect of the scapula was clipped of hair and aseptically prepared. Cancellous graft material was collected from the proximal aspect of the ipsilateral humerus. A sterile, waterproof plastic adhesive drapef was secured to the limb from the phalanges to the elbow joint prior to making the first incision. Tumor excision was performed as previously described.7,9,10,31 Briefly, a cranial incision was made from the proximal aspect of the antebrachium to the distal end of the metacarpal bones. Subcutaneous tissues were excised around the circumference of the pseudocapsule. The cephalic vein was preserved and reflected laterally. The abductor pollicis longus and extensor carpi radialis muscles were isolated and transected.
Once the soft tissue envelope around the entire tumor had been isolated, the cutting jig was placed onto the cranial surface of the bone, with the reference hole aligned with the center of the radiocarpal bone. The proximal osteotomy of the radius was performed with a sagittal saw along the line defined by the cutting jig. Osteotomy of the ulna was also performed at this time if this was deemed necessary from the preoperative CT scans or if there was any evidence of possible tumor invasion at the time of surgery. Remaining joint capsule and ligamentous attachment to the distal aspect of the radius (and ulna, if this was included) were excised, and the ostectomized section of radius was removed (Figure 2). Radiographs of the excised bone section were obtained to confirm complete tumor excision. An air-powered oscillating burr was used to remove cartilage from the proximal surface of the radiocarpal bone. Bone graft was packed into these spaces.
The sterilized implant was then introduced into the defect. The proximal section of radius was elevated slightly from the wound to allow the intramedullary peg of the implant to be pushed into the medullary canal; the radial section of the plate was then sitting flush with the bone. The limb was straightened; skin and soft tissues were reflected until the implant was correctly aligned with metacarpal bones III and IV (Figure 2). Two 22-gauge 1-inch hypodermic needles were placed on the medial and lateral sides of these bones and through the slit in the plate to confirm correct alignment. Starting with the single screw in the radiocarpal bone, then the most proximal screw in the radius, and then either of the most distal screws in the metacarpal bones, all planned screw holes were drilled, measured, and secured with an appropriate gauge and length of titanium screw. A drill guide was used to ensure accurate alignment. Care was taken to avoid incorporating the proximal aspect of the ulna into the screw holes.
After all screws were placed, each was individually tightened commencing from the proximal aspect of the radius. The wound was then flushed with warmed sterile saline solution. A closed suction drain was placed next to the prosthesis. The incision was closed in 3 layers (fascia, subcutaneous tissue, and skin). A modified Robert Jones bandage was placed after surgery. The excised tissues were submitted for light microscopic examination. Postoperative radiographs were obtained to confirm implant positioning and accuracy of screw placement (Figure 2).
Photographs showing the resected portion of bone (A) from a dog with a tumor of the radius and the patient-specific porous titanium implant created to replace it (B) as well as a postoperative lateral radiographic image confirming proper implant placement (C). In this dog, 60% of the total radius length has been removed. The implant fit securely in place, allowing rapid surgical placement.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
Tibia—The dog was positioned in lateral recumbency. The entire limb from the digits to the hip joint was clipped and aseptically prepared. Cancellous graft material was collected from the proximal aspect of the ipsilateral humerus. A sterile, waterproof plastic adhesive drape was secured to the limb from the phalanges to the stifle joint prior to making the first incision. A cranial incision was made from just distal to the stifle joint to the distal end of the metatarsal bones. Subcutaneous tissues were excised around the cranial, medial, and lateral surfaces of the pseudocapsule. The proximal osteotomy of the tibia and fibula was performed with a sagittal saw along the line defined by the cutting jig. With the distal aspect of the tibia now more mobile, remaining joint capsule and ligamentous attachments to the talus and calcaneus were excised, taking care to preserve deep vascular structures. The ostectomized section of tibia and fibula were then removed. Radiographs of the excised bone section were obtained to confirm complete excision of the tumor.
The ridges of the talus were excised with an oscillating saw. The sterilized implant was then introduced into the defect (Figure 3). The proximal section of tibia was reflected slightly from the wound to allow the intramedullary peg of the implant to be pushed into the medullary canal; the tibial section of the plate was then sitting flush with the bone. The limb was straightened; skin and soft tissues were reflected until the implant was correctly aligned with metatarsal bones III and IV. Two 22-gauge 1-inch hypodermic needles were placed on the medial and lateral sides of these bones and through the slit in the plate to confirm correct alignment. Starting with the single screw in the talus, then the most proximal screw in the tibia and then either of the most distal screws in the metatarsal bones, all planned screw holes were drilled, measured, and secured with appropriate gauge of titanium screws. A drill guide was used to ensure accurate alignment.
Photograph of a patient-specific porous titanium tibial implant designed to replace the resected tumorous distal portion of the tibia in a dog with a tibial tumor (A) and postoperative lateral radiographic image confirming proper implant placement (B). The distal tibial implant was 31 cm in length and designed to conform precisely to the underlying bone profile. Tibiotarsal arthrodesis was required.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
After all screws were placed, they were individually tightened commencing from the proximal aspect of the tibia. The wound was then flushed with warmed sterile saline solution. Bone graft material was packed into the interface between tarsal bone and the implant. A closed suction drain was placed next to the prosthesis. The incision was closed in 3 layers (fascia, subcutaneous tissue, and skin). A modified Robert Jones bandage was placed after surgery. The excised tissues were submitted for light microscopic examination. Postoperative radiographs were obtained to confirm implant positioning and accuracy of screw placement (Figure 3).
Postoperative care
All dogs, including those that received mandibular implants, were encouraged to eat as soon as they had recovered suitably from general anesthesia or the sedative effects of opiate infusions. For the mandibular implant recipients, only soft canned food was provided for periods of up to 4 weeks, but after that period dogs were allowed to return to their usual diet.
All dogs with osteosarcoma received adjuvant chemotherapy. This was delivered by SC administration via a wound-diffusion catheter surgically placed onto the lateral wall of the thorax at the time of implant placement as described elsewhere.32 Carboplatin (300 mg/m2) was delivered through the diffusion catheter via a volumetric infusion pump at a rate of 0.5 mL/h. Chemotherapy began within 24 hours after surgery and was typically completed within 4 days after surgery. The wound diffusion catheter was then removed.
All dogs remained in the hospital until they were considered comfortable and were receiving only orally administered anti-inflammatory drugs or had completed their adjuvant chemotherapy protocol. Clinical notes of all patients were reviewed for details regarding their clinical function and any postoperative manipulations required.
Follow-up data collection
Dogs returned for a routine examination and repeated radiography of the thorax and implant between 6 and 9 weeks after surgery. Follow-up at the end of the study was performed by means of a questionnaire administered via telephone to the referring veterinarian and owner. Respondents were asked about any complications that might have developed since the dog was discharged from the referral center; any medication provided; the degree of exercise or activity the dog was able to perform; any change in the level of clinical activity or exercise possible, compared with the preoperative level; and current status of the dog (alive or dead) and, if dead, the date and cause of death.
Complications
Complications were assessed as described elsewhere10 and classified as related to infection, implant failure, or functional disruption to the patient. Complications were considered major if they resulted in revision surgery or removal of the implant, amputation of the limb, or patient death. Minor complications comprised events that were resolved conservatively, with or without medications. Time to complication was defined as the number of days from surgery to documentation of infection, implant failure, or functional disruption.
Statistical analysis
Because of the small number of dogs in the study, only descriptive statistics of each treatment group were calculated. The Kaplan-Meier method was used to determine times from surgery to local tumor recurrence and tumor-related death. Dogs that remained alive on the date of analysis were right censored.
Results
Dogs
Twelve dogs were included in the study: 6 treated for conditions affecting the mandible, 5 with cancer of the distal aspect of the radius, and 1 with cancer of the distal aspect of the tibia (Table 1). All dogs were in good physical condition at the time of surgery. No evidence of regional or distant metastasis was detected.
Characteristics of 12 dogs with tumors of the mandible, tibia, or radius that underwent tumor resection and subsequent placement of a patient-specific porous titanium mandibular implant.
| Object length (cm) | Nature of complication | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Location | Breed | Sex | Age (y) | Body weight (kg) | Diagnosis | Implant | Tumor | Intercalary porous segment | %OBLR | Intraoperative | Postoperative | Type | Cause | Survival time (d) |
| Radius | German Shepherd Dog | NM | 2.5 | 34 | OSA | 23 | 3.8 | 9 | 40 | Unable to fill 1 screw hole | Broken plate | Major | Implant design or production problem | 101* |
| Great Dane | NM | 6.5 | 75 | OSA | 35 | 7.7 | 13 | 50 | NA | Local recurrence | Major | Local recurrence | 147* | |
| Rottweiler | NM | 7 | 52 | OSA | 30 | 6 | 10.5 | 50 | Implant too large for tibial medullary canal | Infection that resolved with treatment | Minor | Infection | 207* | |
| Implant design or production problem | ||||||||||||||
| Mix | IM | 7.5 | 31.5 | OSA | 22 | 4.5 | 8.5 | 50 | NA | NA | NA | NA | 169 | |
| Samoyed | NM | 12 | 31 | OSA | 22 | 5.5 | 11.3 | 60 | NA | NA | NA | NA | 689* | |
| Mandible | Boxer | IM | 5 | 35 | OSA | 13 | 6 | 8.5 | 60 | Tumor size progressed, requiring more extensive surgery | Local recurrence | Major | Local recurrence | 92* |
| Cairn Terrier | NM | 10.5 | 13 | Amelo-blastoma | 9 | 3 | 5.6 | 50 | NA | Mucosal ulceration | Minor | Implant design or production problem | 653 | |
| Doberman | NM | 10 | 34 | FSA | 17 | 5.5 | 10 | 70 | NA | None | NA | NA | 776 | |
| Labrador cross | NM | 10 | 26 | OSA | 13 | 7.5 | 9 | 60 | Tumor size progressed, requiring more extensive surgery | Infection; plate removed | Major | Implant less secure due to more extensive bone resection performed | 65* | |
| Bichon Frise | NM | 5 | 8.7 | Osteomyeltitis; epithelial cysts | 11 | 5 | 6.5 | 60 | NA | Residual infection and malalignment | Major | Preexisting infection | 36 | |
| Rhodesian Ridgeback | IM | 9 | 45 | OSA | 16 | 4 | 7 | 40 | Tumor size progressed, requiring more extensive surgery | Local recurrence | Major | Local recurrence | 47* | |
| Tibia | Rottweiler | NM | 8.5 | 52 | OSA | 31 | 4.5 | 10 | 50 | No cutting jig used; implant did not sit flush with talus | Signs of digit pain from implant; implant breakage | Major | Implant design or production issue | 280* |
Deceased.
%OBLR = Percentage of original bone length resected. FSA = Fibrosarcoma. IM = Sexually intact male. NA = Not applicable. NM = Neutered male. OSA = Osteosarcoma. To convert kilograms to pounds, multiply by 2.2.
Mandibular tumors
Dogs—Median age of the 6 dogs with mandibular tumors was 8 years (range, 5 to 10 years). All dogs were males (4 neutered and 2 sexually intact) and of different breeds (Table 1). Median body weight was 30 kg (66 lb; range, 8.7 to 45 kg [19.1 to 99.0 lb]).
Five dogs had an oral mass affecting the body of the mandible. One dog (a Boxer) also had a large swelling affecting the masseteric area. Tumor types included osteosarcoma (n = 3), ameloblastoma (1), and fibrosarcoma (1). Tumor lengths ranged from 3.0 to 7.5 cm. The sixth dog had a history of chronic osteomyelitis that had been nonresponsive to repeated curettage and chronic antimicrobial treatment. Histologic examination of the mandible following excision revealed multiple epithelial cysts. Active growth of Actinomyces spp was obtained via microbial culture of a sample from the medullary cavity of the angular process of the retained mandible.
For 3 dogs, the tumor had grown considerably since the initial examination and CT scanning for implant design. This necessitated a more extensive dissection than for the others and an unplanned osteotomy different from that originally planned. As a result, some screw holes were no longer aligned over bone, so fewer screws could be placed than had been planned. In 1 dog, only 2 screws could be placed into the residual angular process and vertical ramus. In the dog with a previous history of osteomyelitis, an extensive periosteal reaction several millimeters thick distorted the underlying skeletal anatomy of the angular process, such that an optimal location for the implant could not be easily identified. Although the implant was ultimately securely fitted, some malalignment of the symphysis was present.
Four dogs were able to eat without apparent difficulty on the morning after surgery. In 2 dogs, extensive paraglossal submucosal edema was present, causing some distortion of the tongue, and 1 dog appeared reluctant to eat in the initial postoperative period as a result. The mucosal edema in both dogs resolved within 5 days after surgery without specific treatment.
Complications—Two-thirds (4/6) of the dogs developed major complications. Two dogs developed local tumor recurrence 30 or 47 days after surgery. In 1 of these 2 dogs, palliative radiotherapy (4 doses at 8 Gy) was provided, but the tumor continued to progress rapidly. This dog was euthanized 92 days after surgery owing to signs of unremitting pain from the tumor. For the dog with preexisting osteomyelitis, the implant was removed 4 weeks after initial surgery because microbial culture of bone tissue revealed a persistent Actinomyces spp infection. This dog also had an extensive periosteal reaction that affected the fit of the implant and caused subsequent malalignment of the symphysis. The decision was made to remove the implant to provide the best opportunity for treating the residual infection. If required, the implant could have then been redesigned to adjust for the extensive soft tissue distortion of the underlying skeletal anatomy and retrofitted. The fourth dog developed halitosis and dehiscence of the oral wound 65 days after surgery, necessitating implant removal.
One dog developed a minor complication 18 months after surgery. This dog had periods of apparent oral pain, with teeth chattering observed when at rest. Some response to NSAIDs was reported. At a subsequent examination, the upper fourth premolar tooth on the side in which the implant was placed was impacting on the lower plate edge, causing some ulceration of the mucosal surface at this site (Figure 4). Although a small section of the implant was exposed, radiography revealed no evidence of implant loosening or infection. The upper fourth premolar tooth was removed, leading to resolution of the clinical signs.
Photograph showing mucosal ulceration and exposed implant in a dog that had received a patient-specific porous titanium implant to replace the resected tumorous portion of the mandible. It was found that the opposing maxillary carnassial tooth was impinging on the dorsal edge of the implant, leading to ulceration and signs of oral pain, 18 months after implant placement. The tooth was removed, allowing resolution of this complication.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
Survival time—Three dogs remained alive with functional implants at the time of analysis. One dog was euthanized owing to local tumor recurrence. In 2 dogs, implants were removed because of complications, but these dogs survived with conventional hemimandibulectomy and remained alive at the time of analysis. A median survival time of 92 days (95% confidence interval, 65 days to infinity) was calculated, with the follow-up period ranging from 65 to 698 days.
Radial tumors
Five dogs had tumors affecting the distal aspect of the radius. Median age was 7 years (range, 2 to 12 years). All dogs were male (4 neutered and 1 sexually intact) and of various breeds (Table 1). Median body weight was 44 kg (96.8 lb; range, 31 to 75 kg [68.2 to 165.0 lb]).
On final histologic examination, all tumors were considered to be osteosarcoma. Tumor lengths ranged from 4.5 to 7.7 cm. In 1 dog, approximately 65% of the radius was excised to maintain a minimum of 5 cm beyond the radiographic margins of the tumor.
Modifications to the standard surgical plan were required for 4 dogs. For 1 dog, the intramedullary peg was too thick and would not fit into the medullary canal of the radius. The internal cortex was reamed with a bone rasp, but owing to concerns that the thickness of the bone was becoming too fragile, a scallop-shaped section of the cranial portion of cortex was removed from the radius to allow the implant to sit against the remaining caudal portion of cortex. The most distal screw in the radial section of the implant was placed to ensure the hole was filled, but no bone stock could be incorporated. In another dog, the most distal screw hole in the porous peg of the radial segment was blocked with some stray titanium fragments that had not been removed after printing, meaning this screw hole could not be used and had to be left unfilled. For the third dog, appropriate lengths of titanium screws were not available at the time of surgery, necessitating the use of 316L stainless steel screws instead. Modifications to the surgical plan in the fourth dog included the use of an experimental membrane impregnated with bone morphogenetic proteins that was wrapped around the bone and implant interfaces.
All 5 dogs were able to bear weight on the affected limb within 24 hours after surgery. One dog developed swelling and lymphedema of the limb despite the use of a soft-padded dressing. Another dog also had mild to moderate persistent edema of the manus that worsened after exercise, with these signs persisting 5 weeks after surgery. Although that dog remained quite active, the edema was associated with some discomfort and worsening of lameness, which was moderately controlled with NSAID administration.
Complications—Major complications developed in 2 of the 5 dogs. One dog was returned to its primary care veterinarian 101 days after surgery when the owners noted a small lump on the affected limb. The dog was still ambulating on the limb but had developed a mild lameness in the previous few days. Palpation of this site provoked a painful reaction from the dog. Radiography revealed breakage of the implant through the screw hole that had been left empty (Figure 5). The dog was subsequently euthanized without prior contact with the study investigators to discuss options for management. The second dog developed local tumor recurrence 76 days after surgery. The owners declined amputation or radiotherapy, and the dog was managed palliatively with activity restriction, orally administered analgesics (meloxicam at 0.1 mg/kg [0.045 mg/lb], PO, q 24 h; codeine at 1 mg/kg [0.45 mg/lb], PO, q 12 h; and gabapentin at 20 mg/kg [9.1 mg/lb], PO, q 12 h), and an osteoclastic inhibitor (pamidronate at 2 mg/kg [0.9 mg/lb], q 28 d). With this regimen, the dog remained ambulatory on the affected limb but was ultimately euthanized because of sudden signs of lethargy and worsening lameness 157 days after surgery.
Lateral radiographic view of the radius of a dog in which implant failure occurred. The second-to-last screw in the radius could not be placed owing to stray metal fragments obstructing the screw hole in the intramedullary porous component of the implant, leading to a stress riser.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
Minor complications developed in 1 of the 5 dogs. This dog developed lameness, pyrexia, and swelling of the affected limb 57 days after surgery. Radiography revealed evidence of an osteomyelitis with periosteal new bone formation around the proximal portion of the implant. The 2 most proximal screws were loose, and the implant had begun to pull away from the cortical surface. The screws were replaced with 4.5-mm-diameter 316L stainless screw screws during a short anesthetic episode. Then, following tourniquet placement at the level of the elbow joint, regional perfusion of the limb with antimicrobial solution (cephazolin at 22 mg/kg [10 mg/lb]) was performed, delivered via catheter into the cephalic vein.33,34 The catheter was placed prior to exsanguination of the limb with an Esmarch bandage applied to the level of the elbow joint. After IV antimicrobial administration by means of an infusion pump, the Esmarch bandage was maintained for 30 minutes to allow antibiotic perfusion of the limb. This dog was discharged from the hospital and continued to receive orally administered antimicrobials for an additional 4 weeks. Although follow-up visits were encouraged, no further radiographs were obtained. However, the dog was reportedly ambulatory after treatment and was euthanized because of the development of pulmonary metastatic disease 207 days after the initial surgery.
Survival time—Of the 5 dogs that received radial implants, 4 were eventually euthanized because of metastatic disease (n = 2), local tumor recurrence (1), or implant failure (1). Median survival time as calculated through Kaplan-Meier analysis was 207 days (95% confidence interval, 101 days to infinity), with the follow-up period ranging from 101 to 689 days.
Tibial tumor
Dog—One dog, an 8-year-old neutered male Rottweiler, had a tibial tumor. Surgery was performed as reported; however, a cutting jig was not created owing to time constraints. The osteotomy sites were approximated by placing the implant alongside the exposed bone and marking the bone with a cautery tip. The bone was then cut with an oscillating saw. Once the distal end of the tibia was removed, the ridges of the talus were then removed to create a flat surface for the implant. Because of the free-hand nature of the ostectomies, slightly more bone length was removed than planned. Although the porous intercalary section sat flush with the proximal end of the tibia, a 4-mm gap existed at its junction with the talus. There was no way of correcting this error, so the defect was packed with bone graft material and fragments of the resected talus that were broken down into small pieces with a rongeur.
This dog was able to walk and bear weight with only mild lameness within 24 hours after surgery. However, apparent comfort deteriorated after 3 to 4 days because of progressive swelling of the distal portion of the limb caused by lymphoedema. This swelling was managed with a soft-padded dressing for 2 weeks, by which time it had resolved. The edema recurred after removal of the bandage, and the dog developed a moist exudative discharge from the surface of the hock joint with some excoriations of the skin. Radiography revealed good implant security and no evidence of infection. Bandaging of the limb was reinstituted for a further 6 to 8 weeks, during which the edema gradually resolved. Pain management with NSAIDs, tramadol, and gabapentin was instituted.
Complications—Although the owners were pleased with their dog's outcome, its clinical function following limb-sparing surgery was considered poor on the basis of clinical and video analysis. The dog would only bear weight on the limb intermittently when walking and would generally carry the limb when trotting or running, occasionally placing the digits for balance only. Examination revealed sensitivity around the digits, and radiography revealed extensive new bone formation around the end of the implanted plate, where it was impinging on the phalanges. The remainder of the implant was stable, and there was no evidence of infection or instability at other locations. However, when the paw was weight-bearing, the distal end of the plate impacted on dorsal surface of the first phalanx, leading to radiographic evidence of periosteal reaction and underlying bone lysis (Figure 6). This likely contributed considerably to the noted signs of pain in the paw and led to the reluctance to use the limb. Options for shortening the plate were discussed with the owners, but no action was taken. Catastrophic failure of the implant occurred 280 days after surgery. This failure occurred through the proximal section of porous matrix.
Postoperative lateral radiographic view of the pes of the dog in Figure 3 that received a tibial implant. Inaccurate location of the osteotomy in the tibia resulted in the implant overhanging the metatarsal bones, causing some impingement on the phalanx during weight bearing. The resultant reaction visible in this radiograph was a cause of signs of postoperative pain and disability in this dog.
Citation: Journal of the American Veterinary Medical Association 251, 5; 10.2460/javma.251.5.566
Survival time—The dog was euthanized 280 days after surgery as a result of implant failure. The owners declined revision surgery or amputation.
Discussion
To the authors' knowledge, the study reported here represents the largest case series of patient-specific endoprosthetic implants reported for any species. Although previous publications13,14,28 have alluded to individual examples, none have provided any prolonged follow-up or survival information on the patients involved. The follow-up periods for many dogs in the present study exceeded 12 months, providing an opportunity to assess the long-term impact of the implants on the patients.
Results of this study suggested that patient-specific implants could provide rapid restoration of clinical function in dogs with bone cancer, with potentially low rates of complication. Although major complications occurred in 7 of 12 patients, for at least 3 dogs these complications were considered a consequence of implant design or manufacture or the surgical plan, reflecting the challenges with any novel procedure. The authors are confident that these development challenges have been overcome and adjustments made to the design and implant processing cycle to prevent their recurrence in the future. If patients with these developmental issues were censored from the analysis, the major and minor complication rates for the study would be 44% (4/9) and 22% (2/9), respectively.
Importantly, rates of infection in the present study were low, with infection occurring with only 1 distal radial implant and 1 mandibular implant. The infection associated with the mandibular implant may have developed as a result of suboptimal fixation of the implant caused by the larger-than-planned ostectomy of the caudal portion of the mandible; any resultant instability would have increased the opportunity for infection to develop.
Infection has represented a considerable impedance to the progression of the limb-sparing solutions to date. Traditional limb-sparing solutions involving cortical allografts or wholly metallic implants fabricated in 316L stainless steel are associated with unacceptably high rates of infection.10,35 In a retrospective study,10 infection developed in 78% of 45 dogs following limb-sparing surgery, leading to implant revision, amputation, or euthanasia. Similar results have been reported by others.7,9
Microbial contamination of biomedical implants is a substantial challenge for human and veterinary medical communities, with implants or medical devices reportedly responsible for 60% to 70% of all hospital-acquired infections.36 Bacteria can adhere to the surface of virtually any implanted device and then remain isolated and immune from destruction within an extracellular polysaccharide biofilm. Bacterial colonization within the biofilm can remain viable and inert for months to years after implantation, and the colonization contributes to septic or aseptic loosening of the implant.
In the present study, a porous architecture was applied to the section of the implant that bridged the large skeletal defect with the intent of encouraging tissue ingrowth directly into the implant. It has been shown that if integration of the implant can be enabled before the surface bacteria have a chance to multiply and develop a protective extracellular polysaccharide barrier, biofilm development may be inhibited and the risk of implant-associated infection minimized.36 Various surface topologies and coatings may also influence the extent of achieved osteointegration or reduce the risk of infection.11 Although the number of dogs in the present study was limited, the comparatively low rates of infection in light of the prolonged follow-up period is encouraging support for the potential of porous titanium implants to minimize the risk of this devastating complication. Limbs from some of the dogs that died during the course of the study have been harvested to allow for follow-up histologic examination of the quality of integration and osteosynthesis achieved within the implants.
Incomplete surgical margins or local tumor recurrence occurred in one-third of dogs in the present study and was the most common major complication reported. This is a high rate of local failure9,10,31 and directly influenced the success of the implant procedure and survival time for these dogs. In all 4 dogs, a delay of up to 4 weeks from the time of original examination for implant design to surgery allowed considerable progression of the tumor to occur. In 2 dogs, more bone had to be resected than was originally planned, which led to less-than-perfect implant positioning. Any delay in the definitive treatment of the primary tumor can be unacceptable in the management of cancer patients. This will likely be one of the challenges for patient-specific implants in the future, as there will inevitably be some lag while an implant is designed, printed, and readied for implantation.
Experiences in the present study suggested that when a delay in definitive treatment is anticipated because of the logistics of implant design and production, steps should be taken to try and slow inevitable progression of the tumor. Neoadjuvant chemotherapy may help ensure the clinical status of the patient does not deteriorate from the time of initial evaluation. Indeed, neoadjuvant treatment may help sterilize and consolidate the reactive zone that extends beyond the pseudocapsule of the tumor, thereby improving the success of local resection.37,38,g
One of the most evident findings in the study reported here was the relative simplicity of the surgical protocol for most dogs. Given that the implants were designed to fit each individual patient exactly, the implants were a snug fit and literally snapped into place. The use of a cutting jig helped ensure the osteotomies were performed exactly as planned, and where this process was followed, implantation was typically successful.
In human implant surgery, the use of 3D-cutting jigs improves the accuracy of fit for customized implants and minimizes the deviation in osteotomy plane, compared with freehand cutting.15,39 However, this precise fit created challenges for several dogs in the present study when the surgical plan deviated from that envisaged in the CAD or when excessive periosteal thickening distorted the underlying skeletal architecture. Because the implant is precisely contoured to the bone, it will typically only fit in 1 location. In the affected dogs, this perfection of fit became imperfect and led to some screw holes being no longer available for fixation, the implant impinging on adjacent structures, or the implant causing misalignment. On the basis of similar experiences in human medicine, it has been stated that “the need for precise computer assisted planning (and) pre and postoperative 3D model simulation cannot be over emphasized.”15
Breakage of the implants occurred in 2 dogs of the present study. In 1 dog, implant failure could potentially be explained by the presence of a stress riser that resulted after 1 screw hole could not be filled owing to stray metallic fragments obstructing the lumen. However, in the second dog, catastrophic implant failure occurred through the porous section of the implant. The current implant involves a diamond-shaped lattice structure, with diamonds ranging in size from 1.5 mm to 2 mm and porosities ranging between 59% and 66%. Stiffness and strength of a porous array structure is known to decrease with increasing porosity.23 Given the proportions of the critical defect in this 52-kg (114.4-lb) dog, the implant may have lacked sufficient strength in a portion of the implant where strain forces were the highest. A focal weakness may also have existed if the laser sintering process created a flaw in the metallic structure.
Accurate prediction of the stresses and strains on a metallic implant across a critical defect of 10 to 12 cm will be essential if the intercalary porous segment of the implant is to adequately support weight bearing for the life of the patient. A more heterogeneous pore array will likely be required to mimic the stress and strain characteristics of the bone it has replaced. Research continues to help validate the correlation between computer modeling and the actual behavior of the porous architecture in 3-D printed implants.22,23
Alternative surgical strategies for limb sparing have been reported, including pasteurized or irradiated autografts,31 allograft,9 ulna transposition,40,41 or distraction osteogenesis.42,43 All of these procedures are associated with their own disadvantages, including the need for multiple procedures, periodic hospitalization, prolonged rehabilitation, high financial cost, or unacceptably high complication rates. Because patients with skeletal osteosarcoma are ultimately destined to die of their cancer, it is essential that any treatment solution provides the opportunity for prompt discharge from hospital care, a rapid return to function, and minimal risk of complications.
Nonsurgical options for enabling limb-sparing management of dogs with bone cancer includes stereotactic radiosurgery, which allows delivery of a curative-intent radiation using a high-dose-per-fraction, precision-controlled strategy that helps limit toxic effects to surrounding tissues.44 Clinical reports on the use of stereotactic radiosurgery for dogs with cancer are limited. Investigators in a study44 involving 46 dogs suggested that this modality will certainly be of value, with survival times comparable to those of surgical strategies (median survival time, 9.7 months). Unfortunately, in that study,44 fractures occurred in 39% to 77% dogs (a 63% known fracture rate), with this event being a cause of early euthanasia. Prediction of which patients are susceptible to fracture is not currently possible, although findings suggested that fracture was more likely to occur when CT revealed evidence of tumor lysis affecting the subchondral bone.44 Another limitation of stereotactic radiosurgery is the restricted availability of this modality to the veterinary community.
As an alternative to amputation, patient-specific intraosseous transcutaneous amputation prostheses have been evaluated for outcome in 4 dogs.45 Each prosthesis was fitted to the remaining skeletal structure at the time of amputation, but instead of being wholly contained within the body, a short stump protruded beyond the skin. Once skeletal and dermal integration of the prosthesis had occurred, a weight-bearing exoprosthesis could be fitted to the stump. In that study,45 all dogs regained excellent limb function within 8 weeks after surgery, with continued good function for their remaining life (follow-up period, 8 to 17 months). Fracture of the prosthesis interface occurred in 1 dog, but revision surgery was successful.
Clinical descriptions of limb sparing at sites other than the distal aspect of the radius are limited, reflecting the challenge of conforming generic implants to anatomic locations that are complex in shape or function.8,28 The CAD design and additive manufacturing approach provides some freedom from these constraints, allowing implants to be developed for virtually any location, limited largely by the imagination of the surgeon and the engineer.15 However, because osteosarcoma of the appendicular skeleton typically develops within the metaphysis of long bones,2,3 resection of the primary tumor will always result in disruption of joint function. Unless implants are developed that can integrate with joint replacement technologies, or can bridge the joint with integrated articulating segments, limb sparing will require ankyloses or arthrodesis of the relevant joint. This requirement will influence the potential success of limb sparing at other sites, with clinical function dictated by tolerance of the dog to the arthrodesis. On the basis of the results of joint arthrodesis for nonneoplastic causes, it can be predicted that dogs will tolerate fusion of the carpus and shoulder joints with only a small degree of clinical impact,46 but fusion of the elbow, stifle, and hock joints will lead to a variable degree of disability.47–51
Existing treatment paradigms for mandibular tumors in dogs involve variations in subtotal, total, or bilateral hemimandibulectomy.4–6,52–59 Dogs generally tolerate such radical excision well, with most owners stating satisfaction with the procedure. However, in 1 study,4 difficulty eating was noted for 12 of 27 (44%) dogs following hemimandibulectomy. The most common long-term complication following extensive mandibular resection is medial drifting of the remaining hemimandibular section.56 The potential consequences of this mandibular drift include malocclusion of the dentition, ulceration of the hard palate, and pain from the misaligned contralateral temporomandibular joint.
Prevention of mandibular drift after surgery can be achieved by elastic training with orthodontic buttons and a power chain, but this approach requires good client compliance.60 A series of reports61–63 describe regeneration of the mandibular bone by use of a compression-resistant matrix infused with recombinant bone morphogenic protein-2. Although good results were achieved in the associated studies, descriptions of these techniques suggest that considerable intraoperative contouring of generic implants is required.
The results of the present study suggested that patient-specific porous titanium implants may provide an effective solution for function-sparing surgery in dogs with bone cancer at multiple skeletal sites, with a relatively low risk of infection-related complications. As with any novel procedure, several technical challenges remain to be overcome to ensure these implants can achieve their promised potential for all patients.
Acknowledgments
Supported in part by the Massey University Research Fund.
Presented in abstract form at the Veterinary Cancer Society Annual Conference, Orlando, Fla, October 2016.
ABBREVIATIONS
| CAD | Computer-aided design |
| STL | Stereolithography |
Footnotes
Osirix Lite, version 8.0.2, Pixmeo SARL, Bernex, Switzerland.
Delcam for SolidWorks, SOLID Applications Limited, Oldbury, West Midlands, England.
SolidWorks, Dassault Systèmes SOLIDWORKS Corp, Waltham, Mass.
Materialise Magics, Materialise HQ, Leuven, Belgium.
SLM Solutions 280HL, SLM Solutions Group AG, Lübeck, Germany.
Ioban, 3M, Saint Paul, Minn.
Wodajo FM, Wittig J, Kumar D, et al. Successful treatment of high grade soft tissue sarcoma with induction chemotherapy: clinicopathological analysis of thick capsule formation allowing less extensive “marginal” surgical resection (abstr), in Proceedings. Annu Meet Am Soc Clin Oncol, 2001;2912.
References
1. Thompson JP, Fugent MJ. Evaluation of survival times after limb amputation, with and without subsequent administration of cisplatin, for treatment of appendicular osteosarcoma in dogs: 30 cases (1979–1990). J Am Vet Med Assoc 1992; 200: 531–533.
2. Straw RC, Withrow SJ, Powers BE. Management of canine appendicular osteosarcoma. Vet Clin North Am Small Anim Pract 1990; 20: 1141–1161.
3. Szewczyk M, Lechowski R, Zabielska K. What do we know about canine osteosarcoma treatment? Vet Res Commun 2015; 39: 61–67.
4. Fox LE, Geoghegan SL, Davis LH, et al. Owner satisfaction with partial mandibulectomy or maxillectomy for treatment of oral tumors in 27 dogs. J Am Anim Hosp Assoc 1997; 33: 25–31.
5. Kosovsky JK, Matthiesen DT, Marretta SM, et al. Results of partial mandibulectomy for the treatment of oral tumors in 142 dogs. Vet Surg 1991; 20: 397–401.
6. Straw RC, Powers BE, Klausner J, et al. Canine mandibular osteosarcoma: 51 cases (1980–1992). J Am Anim Hosp Assoc 1996; 32: 257–262.
7. Withrow SJ. Surgery for skeletal sarcomas. Clin Tech Small Anim Pract 1998; 13: 53–58.
8. Liptak JM, Pluhar GE, Dernell WS, et al. Limb-sparing surgery in a dog with osteosarcoma of the proximal femur. Vet Surg 2005; 34: 71–77.
9. Straw RC, Withrow SJ. Limb-sparing surgery versus amputation for dogs with bone tumors. Vet Clin North Am Small Anim Pract 1996; 26: 135–143.
10. Mitchell KE, Boston SE, Kung M, et al. Outcomes of limb-sparing surgery using two generations of metal endoprosthesis in 45 dogs with distal radial osteosarcoma. A Veterinary Society of Surgical Oncology retrospective study. Vet Surg 2016; 45: 36–43.
11. Sing SL, An J, Yeong WY, et al. Laser and electron-beam powder-bed additive manufacturing of metallic implants: a review on processes, materials and designs. J Orthop Res 2016; 34: 369–385.
12. Wang X, Xu S, Zhou S, et al. Topological design and additive manufacturing of porous metals for bone scaffolds and orthopaedic implants: a review. Biomaterials 2016; 83: 127–141.
13. Devasconcellos P, Balla VK, Bose S, et al. Patient specific implants for amputation prostheses: design, manufacture and analysis. Vet Comp Orthop Traumatol 2012; 25: 286–296.
14. Harrysson OLA, Marcellin-Little DJ, Horn TJ. Applications of metal additive manufacturing in veterinary orthopedic surgery. JOM 2015; 67: 647–654.
15. Parthasarathy J. 3D modeling, custom implants and its future perspectives in craniofacial surgery. Ann Maxillofac Surg 2014; 4: 9–18.
16. Razi H, Checa S, Schaser KD, et al. Shaping scaffold structures in rapid manufacturing implants: a modeling approach toward mechano-biologically optimized configurations for large bone defect. J Biomed Mater Res B Appl Biomater 2012; 100: 1736–1745.
17. Van der Stok J, Van der Jagt OP, Amin Yavari S, et al. Selective laser melting-produced porous titanium scaffolds regenerate bone in critical size cortical bone defects. J Orthop Res 2013; 31: 792–799.
18. Shah FA, Omar O, Suska F, et al. Long-term osseointegration of 3D printed CoCr constructs with an interconnected open-pore architecture prepared by electron beam melting. Acta Biomater 2016; 36: 296–309.
19. Brown I, Smith G, McGavin P, et al. Optimizing cellular structures for titanium implant design, in Proceedings. Cell Mater Conf 2014.
20. Liu Y, Lim J, Teoh SH. Review: development of clinically relevant scaffolds for vascularised bone tissue engineering. Biotechnol Adv 2013; 31: 688–705.
21. Liptak JM, Dernell WS, Ehrhart N, et al. Cortical allograft and endoprosthesis for limb-sparing surgery in dogs with distal radial osteosarcoma: a prospective clinical comparison of two different limb-sparing techniques. Vet Surg 2006; 35: 518–533.
22. Brown I, Smith G, Ryan M, et al. Design and properties of cellular structures for titanium alloy implants, in Proceedings. MetFoam Conf 2015.
23. Smith G, Brown I, Kirchner A, et al. Engineering properties of cellular Ti-6Al-4V structures fabricated by electron beam melting. Key Eng Mater 2016; 704: 318–324.
24. Imanishi J, Choong PF. Three-dimensional printed calcaneal prosthesis following total calcanectomy. Int J Surg Case Rep 2015; 10: 83–87.
25. Huang YH, Seelaus R, Zhao L, et al. Virtual surgical planning and 3D printing in prosthetic orbital reconstruction with percutaneous implants: a technical case report. Int Med Case Rep J 2016; 9: 341–345.
26. Aranda JL, Jimenez MF, Rodriguez M, et al. Tridimensional titanium-printed custom-made prosthesis for sternocostal reconstruction. Eur J Cardiothorac Surg 2015; 48: e92–e94.
27. Toso SM, Menzel K, Motzkus Y, et al. Patient-specific implant in prosthetic craniofacial reconstruction: first report of a novel technique with far-reaching perspective. J Craniofac Surg 2015; 26: 2133–2135.
28. Liska WD, Marcellin-Little DJ, Eskelinen EV, et al. Custom total knee replacement in a dog with femoral condylar bone loss. Vet Surg 2007; 36: 293–301.
29. Brannam E. Custom titanium implants a Kiwi first. Newshub. Available at: www.newshub.co.nz/nznews/custom-titanium-implants-a-kiwi-first-2016040913#axzz4FkHhEqdM. Accessed Jun 12, 2017.
30. Sparrow T, Fitzpatrick N, Meswania J, et al. Shoulder joint hemiarthroplasty for treatment of a severe osteochondritis dissecans lesion in a dog. Vet Comp Orthop Traumatol 2014; 27: 243–248.
31. Morello E, Vasconi E, Martano M, et al. Pasteurized tumoral autograft and adjuvant chemotherapy for the treatment of canine distal radial osteosarcoma: 13 cases. Vet Surg 2003; 32: 539–544.
32. Simcock JO, Withers SS, Prpich CY, et al. Evaluation of a single subcutaneous infusion of carboplatin as adjuvant chemotherapy for dogs with osteosarcoma: 17 cases (2006–2010). J Am Vet Med Assoc 2012; 241: 608–614.
33. Fiorello CV, Beagley J, Citino SB. Antibiotic intravenous regional perfusion for successful resolution of distal limb infections: two cases. J Zoo Wildl Med 2008; 39: 438–444.
34. Rubio-Martínez LM, Cruz AM. Antimicrobial regional limb perfusion in horses. J Am Vet Med Assoc 2006; 228: 706–712.
35. Morello E, Buracco P, Martano M, et al. Bone allografts and adjuvant cisplatin for the treatment of canine appendicular osteosarcoma in 18 dogs. J Small Anim Pract 2001; 42: 61–66.
36. Veerachamy S, Yarlagadda T, Manivasagam G, et al. Bacterial adherence and biofilm formation on medical implants: a review. Proc Inst Mech Eng H 2014; 228: 1083–1099.
37. Bray J, Polton G. Neoadjuvant and adjuvant chemotherapy combined with anatomical resection of feline injection-site sarcoma: results in 21 cats. Vet Comp Oncol 2016; 14: 147–160.
38. Hohenberger P, Wysocki WM. Neoadjuvant treatment of locally advanced soft tissue sarcoma of the limbs: which treatment to choose? Oncologist 2008; 13: 175–186.
39. Helguero CG, Kao I, Komatsu DE, et al. Improving the accuracy of wide resection of bone tumors and enhancing implant fit: a cadaveric study. J Orthop 2015; 12: S188–S194.
40. Pooya HA, Seguin B, Mason DR, et al. Biomechanical comparison of cortical radial graft versus ulnar transposition graft limb-sparing techniques for the distal radial site in dogs. Vet Surg 2004; 33: 301–308.
41. Hodge SC, Degner D, Walshaw R, et al. Vascularized ulnar bone grafts for limb-sparing surgery for the treatment of distal radial osteosarcoma. J Am Anim Hosp Assoc 2011; 47: 98–111.
42. Jehn CT, Lewis DD, Farese JP, et al. Transverse ulnar bone transport osteogenesis: a new technique for limb salvage for the treatment of distal radial osteosarcoma in dogs. Vet Surg 2007; 36: 324–334.
43. Ehrhart N. Longitudinal bone transport for treatment of primary bone tumors in dogs: technique description and outcome in 9 dogs. Vet Surg 2005; 34: 24–34.
44. Kubicek L, Vanderhart D, Wirth K, et al. Association between computed tomographic characteristics and fractures following stereotactic radiosurgery in dogs with appendicular osteosarcoma. Vet Radiol Ultrasound 2016; 57: 321–330.
45. Fitzpatrick N, Smith TJ, Pendegrass CJ, et al. Intraosseous transcutaneous amputation prosthesis (ITAP) for limb salvage in 4 dogs. Vet Surg 2011; 40: 909–925.
46. Fitzpatrick N, Yeadon R, Smith TJ, et al. Shoulder arthrodesis in 14 dogs (Erratum published in Vet Surg 2013; 42:229). Vet Surg 2012; 41: 745–754.
47. Petazzoni M, Nicetto T. Stifle arthrodesis using a locking plate system in six dogs. Vet Comp Orthop Traumatol 2015; 28: 288–293.
48. Cofone MA, Smith GK, Lenehan TM, et al. Unilateral and bilateral stifle arthrodesis in eight dogs. Vet Surg 1992; 21: 299–303.
49. Roch SP, Clements DN, Mitchell RA, et al. Complications following tarsal arthrodesis using bone plate fixation in dogs. J Small Anim Pract 2008; 49: 117–126.
50. Harasen G. Arthrodesis—part II: the tarsus. Can Vet J 2002; 43: 806–808.
51. Choate CJ, Arnold GA. Elbow arthrodesis following a pathological fracture in a dog with bilateral humeral bone cysts. Vet Comp Orthop Traumatol 2011; 24: 398–401.
52. Domnick ED, Smith MM. Bilateral rostral mandibulectomy in the dog. J Vet Dent 2015; 32: 134–139.
53. Carvalho CM, Rahal SC, Dos Reis Mesquita L, et al. Mandibulectomy for treatment of fractures associated with severe periodontal disease. Can Vet J 2015; 56: 292–294.
54. Arzi B, Verstraete FJ. Mandibular rim excision in seven dogs. Vet Surg 2010; 39: 226–231.
55. Walker KS, Reiter AM, Lewis JR. Marginal mandibulectomy in the dog. J Vet Dent 2009; 26: 194–198.
56. Verstraete FJ. Mandibulectomy and maxillectomy. Vet Clin North Am Small Anim Pract 2005; 35: 1009–1039.
57. White RA, Gorman NT. Wide local excision of acanthomatous epulides in the dog. Vet Surg 1989; 18: 12–14.
58. Lantz GC, Salisbury SK. Partial mandibulectomy for treatment of mandibular fractures in dogs: eight cases (1981–1984). J Am Vet Med Assoc 1987; 191: 243–245.
59. Harvey CE. Oral surgery. Radical resection of maxillary and mandibular lesions. Vet Clin North Am Small Anim Pract 1986; 16: 983–993.
60. Bar-Am Y, Verstraete FJ. Elastic training for the prevention of mandibular drift following mandibulectomy in dogs: 18 cases (2005–2008). Vet Surg 2010; 39: 574–580.
61. Spector DI, Keating JH, Boudrieau RJ. Immediate mandibular reconstruction of a 5 cm defect using rhBMP-2 after partial mandibulectomy in a dog. Vet Surg 2007; 36: 752–759.
62. Arzi B, Verstraete FJ, Huey DJ, et al. Regenerating mandibular bone using rhBMP-2: part 1—immediate reconstruction of segmental mandibulectomies. Vet Surg 2015; 44: 403–409.
63. Arzi B, Cissell DD, Pollard RE, et al. Regenerative approach to bilateral rostral mandibular reconstruction in a case series of dogs. Front Vet Sci 2015; 2: 4.
