Patient-Specific Instruments for Forearm Sarcoma Resection and Allograft Reconstruction in Children: Results in 4 Cases

For pediatric malignant bone tumors located in the limbs, limb salvage surgery is the gold standard, but it requires adequate resection margins to avoid local recurrence. Primitive bone sarcomas of the forearm (radius or ulna) are very rare and the reconstruction remains challenging. We describe a method to ensure minimal but adequate resection bone margins with precision in four consecutive patients with primitive bone sarcomas of the forearm. During the preoperative planning, magnetic resonance imaging (MRI) was used to delineate the tumor and the tumor volume was transferred to computerized tomography (CT) by image fusion. A patient-specific instrument (PSI) was manufactured by 3D printing to allow the surgeon to perform the surgical cuts precisely according to the preoperative planning. The first PSI was used for the resection of the tumor, which adopted a unique position at the bony surface. A second PSI was intended for the cutting of the bone allograft so that it fitted perfectly with the bone defect. In all four cases, the safe margin obtained into the bone was free of tumor (R0: microscopically margin-negative resection). The functional result was very good in all four patients. This limb salvage surgical technique can be applied in forearm bone sarcoma and improves surgical precision while maintaining satisfactory local tumor control. It can also reduce the surgical time and allow a stable osteosynthesis.


Introduction
Primary bone tumors are rare in the European population (<0.2% of malignant neoplasms registered in the EUROCARE (European Cancer Registry)), and their incidence is 3 per 1,000,000 inhabitants per year [1]. Osteosarcoma and Ewing's sarcoma are the main malignant primary bone tumors in children and young adults [1]. Boys are more frequently afected than girls [1]. Te forearm bones (radius and ulna) are rarely afected by osteosarcomas and Ewing's sarcomas. More exactly, the localization for Ewing's sarcomas is only 0.3% for the radius and 1.0% for the ulna [2].
Advances in diagnostic and therapeutic techniques in bone tumors have enabled the development of limb salvage surgery, particularly in the pediatric population [3]. In the case of limb salvage surgery, inappropriate resection margins are associated with an increased local recurrence rate [4]. Terefore, it is imperative to have adequate margins (free of tumor) to obtain local control of the tumor and improve oncological outcomes. However, large resection margins involve signifcant bone defects with a signifcant impairment of the limb function [5].
Tis article discusses a new surgical technique to decrease the resection margins of these tumors by using a patient-specifc instrument (PSI). Te PSI helps to intraoperatively reproduce resection plans that are planned before surgery. Tey are based on computed tomography (CT) and magnetic resonance imaging (MRI) sequences of the patient. Tis technique also helps in the reconstruction of the limb in order to preserve the function of the limb. Particularly, in young patients, limb function is of major interest. A lasting reconstruction must be carried out because the life expectancy of these patients has greatly increased [5]. Diferent reconstruction techniques exist for such bone defects: reconstruction by bone allograft, vascularized autograft, endoprosthetic reconstruction, and reimplantation of the tumor bone segment after being devitalized [5][6][7].

Sarcoma
Te massive bone allograft reconstruction technique ofers the advantages of tendon and ligament reattachment [8,9]. Other advantages are the biological incorporation of the graft (at least partial) and preservation of growth cartilage and joint [8,10]. Te 5-year survival rate is 80% for massive allografts [11].
A high rate of complications is, however, associated with allograft reconstruction, such as allograft fractures, infections of the operatory site, and nonunion between the allograft and the host bone [10,11].
Despite these drawbacks, intercalary reconstruction by allograft bone grafting should be favored for the upper limb [12,13].
PSIs have already been used in pedicle screw insertion, hip resurfacing, total knee arthroplasty, and long bones corrective osteotomy with more accuracy [14,15]. In surgical oncology, the use of the PSI concept has been developed as a new surgical technique for bone tumor resection, allowing the surgeon to perform the trajectories of the cutting tool around the tumor [16].
Te use of PSIs on bone allografts has previously been described in tibial reconstructions [17] and pelvic reconstruction [18]. Little data are available in the literature on tumor resection of the forearm, but studies on pelvic and tibial tumors have shown that PSIs guarantee good resection margins while reducing the risk of local recurrence [17,18].
Te aim of this article is to detail the surgical management of forearm primary bone sarcoma in young patients who require resection/reconstruction by using a PSI. (Table 1). Four patients were operated on for primary bone tumor of the forearm by tumor resection and allograft reconstruction with PSIs. Te primary bone tumor was Ewing's sarcoma in 3 cases (one radius and two ulnae) and one telangiectatic osteosarcoma of the radius. One patient had pulmonary metastases and one patient had two skip metastases into the proximal radius.

Patient Series
At the time of the resection, the age of patients ranged from 7 years and 5 months to 16 years and 3 months. Te radiological evaluation consisted of radiography, CT scan, MRI, and fuorodeoxyglucose PET (positron emission tomography)/CT. Neurological defcits were absent. Te diagnosis was confrmed by a surgical biopsy in all cases. Te patients received multiagent neoadjuvant and adjuvant chemotherapy, according to EurAmos, Euro-Ewing 99, Euro-Ewing 2008, and Euro-Ewing 2012 protocols.

Preoperative
Planning. Te imaging techniques (CTscan and MRI) used for tumor diagnosis were used for the planning: Brilliance 40 CT-scanner (Philips, the Netherlands; 0.5 mm spacing between slices, 1 mm slice thickness, 120 kV peak voltage, and 99 mA tube current) and 1.5 TNT Scan Intera MRI (Philips, the Netherlands, 4 mm spacing between slices, 3 mm slice thickness, 550 ms TR, and 14 ms TE).
Based on the prechemotherapy MRI sequences (Figure 1), the tumor was manually delineated by the surgeon by using a web-based segmentation tool (Customize V1; 3Dside). Tis software was developed by the 3D side and was used online by the surgeon. Te tumor volume was drawn on each slice by using a simple polygon function. Te tumor volume was automatically saved on the web server.
By using a multimodal registration algorithm, MRI images and their associated tumor volume were merged with CT images. Te tumor volume thus shared the same coordinate space as the CT. Te CT image was segmented by using customized software to obtain a 3D representation of the bone. Te tumor volume was converted into a 3D model. A combined 3D representation of the forearm bones and the tumor was therefore obtained (Figure 2). Tese 3D models were used to defne the resection planes (target planes), which revealed the trajectories of the saw blade into the bone. Te resection planes were frst placed in contact with the tumor, and then translated with a safe margin defned by the surgeon. Te usual safe margin was 10 mm, but in some cases, it could be decreased to preserve relevant structures such as physis or epiphysis.

Preoperative Planning for Allograft
Cutting. Te local bone bank scans all harvested allografts (SOMATOM Defnition AS, Siemens; slice thickness of 0.35 mm, 0.7 mm spacing between slices, 120 kV peak voltage, and 99 mA tube current).
CT scans were sent to the 3D side for processing. After segmentation, the 3D models of all allografts were matched with the patient's bone to select the most suitable radius or ulna allografts (Figure 3). Te objective was to fnd the best-ftting allograft for the postresection bone defect. For radius osteochondral allografts, particular attention was paid to the articular surface of the radial head. Te resection plans were applied to the matched allograft. Tis  Figure 1. Top left and right: the segmented tumor on MRI was merged with the CTscan. Te tumor is in red (there were 3 small separated tumoral areas in the proximal ulna). Te picture gives the detail of the preoperative planning for the proximal section. A step cut was chosen to improve stability. Te chosen margin was 5 mm in order to preserve the joint. Bottom: the green zone in the ulna is the area that is planned to be resectioned. Left: medial view and right: lateral view. 4 Sarcoma process allowed us to obtain an identical cutting plan for the patient and the allograft and therefore, an optimal flling of the bone defect.

Patient-Specifc Instruments.
Two PSIs were created for each patient: one for tumor resection and a second for allograft cutting. Te PSIs were initially modeled by software (Blender 2.63.11). Each PSI was designed to accommodate a unique bone surface position (Figure 4). Tey held stainless steel cylinders to insert 2 mm Kirschner wires (K-wires) that fx the guide on the bone surface. Te PSI was equipped with a fat surface representing the resection planes, guiding the saw blade during osteotomies.
Te PSI was manufactured in a biocompatible material (polyamide 2200) by the technique known as selective laser sintering (SLS). It is a layer-by-layer additive manufacturing process. A laser sinters free powder by drawing a 2D shape. A new layer of powder is spread before drawing a second 2D shape. Te process is repeated until the fnal 3D object is obtained.
Te minimal time for virtual creation and for 3D printing of the guides is 15 days. However, the time is not critical due to the time of the presurgical chemotherapy in Ewing's sarcoma and osteosarcoma (several months).

Assisted Surgery.
During the surgery, the patients were placed in the decubitus position with an arm table. An anterior surgical approach (Henry's approach) was used in the two patients with radius sarcoma and a direct ulnar approach in the two patients with ulnar sarcoma ( Figure 5).
Te biopsy site was excised as an ellipse with the tumor. A progressive soft tissue dissection was performed to isolate the tumor with the expected margin.
Te two PSI (one for the patient, one for the allograft) were previously sterilized by standard autoclave the day before the operation. Once the PSI has been placed in a unique position, it is fxed using K-wires. Using a surgical saw, the osteotomy is then performed with the chosen safety margin. Resection lengths ranged from 14.0 cm to 15.7 cm.
Extemporaneous pathology analyses were performed on the distal and proximal resection sites to ensure adequate  margins (free of tumors). After confrmation of a healthy margin, a reconstruction was undertaken. Te second PSI was fxed to the allograft, taking into account the cutting thickness of the saw blade (1.5 mm). Te adjusted allograft was placed on the bone defect and osteosynthesis was performed using plates and screws ( Figure 6).
A long arm cast was put in place for all patients for 6 weeks.

Functional Evaluation.
Te functional outcome was evaluated by using the Musculoskeletal Tumor Society (MSTS) scoring system [19]. Tis functional classifcation system is scored out of 30 points established by the MSTS. Tis classifcation takes into account function, emotional acceptance, pain, hand positioning, dexterity, and lifting ability. Each variable is rated on a scale ranging from 0 to 5 points.

Results
Te average surgical time was 176 minutes from the time of skin incision to the end of skin closure. Histological examination of the removed sarcoma confrmed the diagnosis of Ewing's sarcoma in three patients and telangiectatic osteosarcoma in one patient.

Obtained Margins.
All the planned margins were obtained with less than 3 mm of error. All bone resection margins were assessed R0. All soft tissue resection margins were also R0, except one patient for whom the resection margin was R1 in the proximal radius. Tis patient required additional adjuvant radiotherapy. One patient developed a tumor recurrence within the soft tissue (nodule located in the fexor carpi ulnaris muscle requiring resection of this nodule 22 months postoperatively). Tere was no local recurrence into the bone for none of the patients. Physiotherapy was authorized after 6 weeks of cast immobilization for all patients.

Reconstruction Result.
We used the classical radiograph, MRI, and PET/CT performed in the follow-up to assess the host-graft union. Te results of the postoperative radiograph, CT, and MRI revealed a satisfactory host-graft contact, except in one patient. Tis last patient developed pseudarthrosis at the level of the proximal junction of the allograft and needed additional autologous bone graft at 12 months postoperatively (cancellous bone grafts harvested in the right iliac crest). For the other three patients, the  6 Sarcoma radiological union was obtained at the graft-host junction at 6, 9, and 12 months. Patient 2, after initial graft-host junction healing, sustained 2 successive allograft fractures, needing osteosynthesis the frst time and fnally allograft replacement with fnal graft-host junction healing.

Functional Result.
Tree patients had an MSTS score of 30/30 and one patient had an MSTS score of 23/30. Tis gives an average of 28/30 for all of our patients.

Discussion
Tis article reports the use of PSI for forearm bone sarcoma resection and reconstruction in children. PSI assistance is used for tumor resection and for cutting the massive allograft, allowing optimal reconstruction. We applied this technique to our 4 patients.
Surgical excision of the tumor requires obtaining a wide margin to avoid local recurrences. However, limb salvage surgery requires the preservation of a functional limb at the expense of obtaining safe margins [4]. Te precise preoperative localization of the tumor allows precise planning, and the PSI improves the precision of the resection during the surgery. Te combination of these techniques allows resection with adequate (free of tumor) but minimal safe margins, thus avoiding unnecessary resection and preserving, if necessary, articular cartilage in young patients. We generally plan a minimal margin of 10 mm, but we sometimes decrease it to 5 mm if needed for physis or joint sparing.
In the specifc anatomic location of the forearm, one of the challenges was the relatively smooth surfaces of the radius and ulna. Given the triangular section of the radius and ulna, it was necessary to have a PSI with a surface straddling a crest of the bone to increase the stability of the PSI. Moreover, the proximal PSI and the distal PSI were connected by a bar. Te fact that they were attached to each other further increased the precision in the positioning of the PSI. Another challenge was the size of the bones in our youngest patients (7,9, and 10 years, respectively) and the need to respect the anatomical radial curves (supinator and pronator curvatures) which are important to keep the pronation and supination movements of the forearm. Te most corresponding allograft was chosen in all cases respecting this radial curvature and the size of the bone as much as possible. Te surgical approach was, in all cases, communicated to the engineer, and the PSI was created to be positioned according to the most accessible side of the bone communicated by the surgeon. Tanks to these precautions, during the surgery, no positioning problem was encountered nor was there confict between PSI and the soft tissues.
In our series, we obtained adequate bone margins in the 4 cases (R0: microscopically margin-negative resection). Te R1 margin was in the soft tissue in only one case.
Primary bone tumors localized into the forearm are extremely rare. A structural or functional defcit of an upper limb can signifcantly afect the quality of life.
Te reconstruction by allograft of the forearm is extremely poorly published in the literature. Diferent techniques have been described such as the olecranonization of the radius initially described by Rydholm and then by Duncan [20,21] or even a fbula autograft in the event of radial damage [22].
Te absence of precise correction, after osteotomy, can lead to poor clinical results, especially in the upper limbs, where the anatomical confguration of the bones is of considerable importance for the function [23].
Tis is why the PSI technique allows the allograft to be cut with a high degree of precision, producing a graft of the optimal shape to fll the bone defect. Te operating time is decreased as all the planning is carried out before the surgery.
Precise tumor and allograft sections obtained by the PSI yield a strong contact at host-graft junctions, resulting in stable osteosynthesis. Mechanical stability of the graft facilitates improved and faster bone healing and fusion due to increased blood vessel growth in the graft [24].
Open communication between surgeon and engineer is important. Te engineer must have a strong clinical experience to understand the medical context and the prerequisites of the PSI that will be generated. Te PSI should be localized to a single site on the bone surface that will be exposed without adding additional surgical approaches, unnecessary skin incisions, or dissection. Surgeons will be asked to anticipate the constraints of the surgery (for example, surgical approach, access to the bone surface, and presence of soft tissues). Te engineer must inspect these clinical data and determine a compromise between the invasive character of the PSI and its stability on the bone surface. In our series of patients, precise positioning of the instruments was easily achieved in each case and the instruments were stable.
Te preliminary results obtained from our four patients indicate that this technique is a clinically reliable method. Allograft reconstruction of the forearm can be performed with precision by using the PSI.
Other publications have already shown a decrease in local recurrence rate [18] by using the PSI in bone sarcoma resection and a decrease in operative time.

Conclusions
Limb preservation is the rule in oncological surgery. A structural or functional defcit of an upper limb can signifcantly afect the quality of life. For upper extremity sarcomas, the surgical team must fnd a balance between sufcient resection margins and reliable and durable reconstruction. Given the spatial constraints of the child's anatomy, it is impossible to obtain margins of several centimeters around the tumor. Te techniques described in this article may help improve patient safety and surgical precision for the resection of forearm bone sarcomas and the accompanying reconstruction at these sites [26].

Data Availability
Underlying data can be obtained upon request from the corresponding author.