Based on a retrospective search of our hospital records from January 2014 to December 2019, we found that a total of 13 patients diagnosed with primary malignant or local aggressive bone tumor lesions were treated with en bloc resection and reconstructed by bone cement prosthesis, nine men and four women, with an average age of 50.2 years (range, 17–70 years) at the time of surgery. All patients underwent radiography, CT, and MRI preoperatively. Plain radiographs or CT scans of the chest were required to rule out pulmonary metastasis. A needle biopsy specimen was obtained preoperatively to make a correct diagnosis for each patient. Pathological classifications of the above patients were as follows: 6 patients were diagnosed with osteosarcoma in the proximal humerus, 1 was diagnosed with invasive chondroblastoma in the proximal humerus, 1 was diagnosed with recurrent GCT in the proximal humerus, 4 were diagnosed with recurrent GCT in the distal radius, and 1 was diagnosed with recurrent aneurysmal bone cyst in the distal radius. All cases in this series were staged according to the Enneking System for Staging Benign and Malignant Musculoskeletal Tumors(15). The cases of GCT were further staged according to the Campanacci staging system(3) (Table 1).Patients with osteosarcoma received neoadjuvant chemotherapy preoperatively. This research has been approved by the IRB of the authors’ affiliated institutions. All patients were informed and consent with the risks and benefits of this reconstruction method.
Table 1
Clinical data of all patients
Cases | Gender | Age(years) | Location | Diagnosis | Staging | Length of resected bone (cm) |
1 | Male | 17 | Proximal humerus | Osteosarcoma | ⅡA | 13.5 |
2 | Male | 19 | Proximal humerus | Invasive Chondroblastoma | ⅡA | 13.0 |
3 | Male | 56 | Proximal humerus | Osteosarcoma | ⅡA | 12.5 |
4 | Male | 61 | Proximal humerus | Osteosarcoma | ⅡA | 11.0 |
5 | Male | 50 | Proximal humerus | Recurrent GCT | Aggressive (Campanacci grade Ⅲ) | 14.0 |
6 | Male | 45 | Proximal humerus | Osteosarcoma | ⅡA | 14.5 |
7 | Male | 67 | Proximal humerus | Osteosarcoma | ⅡA | 9.0 |
8 | Male | 70 | Proximal humerus | Osteosarcoma | ⅡA | 14.0 |
9 | Male | 42 | Distal radius | Recurrent GCT | Aggressive (Campanacci grade Ⅱ) | 6.5 |
10 | Female | 60 | Distal radius | Recurrent GCT | Aggressive (Campanacci grade Ⅱ) | 6.0 |
11 | Female | 55 | Distal radius | Recurrent GCT | Aggressive (Campanacci grade Ⅲ) | 8.0 |
12 | Female | 53 | Distal radius | Recurrent GCT | Aggressive (Campanacci grade Ⅲ) | 7.0 |
13 | Female | 57 | Distal radius | Recurrent Aneurysmal bone cyst | Aggressive | 7.0 |
Table 2
Cases | MSTS score | | VAS score | | Metastasis | Recurrence | Oncologic outcomes | Follow up (months) |
Score | Percentage (%) | Preoperative | Postoperative |
1 | 22 | 73.3% | | 8 | 1 | | None | None | CDF | 64 |
2 | 20 | 66.7% | | 7 | 1 | | None | None | CDF | 38 |
3 | 21 | 70.0% | | 9 | 2 | | None | None | CDF | 51 |
4 | 24 | 80.0% | | 7 | 1 | | None | None | CDF | 38 |
5 | 24 | 80.0% | | 8 | 1 | | None | None | CDF | 15 |
6 | 21 | 70.0% | | 7 | 2 | | None | None | CDF | 30 |
7 | 24 | 80.0% | | 6 | 1 | | None | None | CDF | 24 |
8 | 22 | 73.3% | | 6 | 1 | | None | None | CDF | 33 |
9 | 23 | 76.7% | | 7 | 1 | | None | None | CDF | 39 |
10 | 25 | 83.3% | | 5 | 2 | | None | None | CDF | 22 |
11 | 24 | 80.0% | | 5 | 1 | | None | None | CDF | 24 |
12 | 26 | 86.7% | | 7 | 1 | | None | None | CDF | 35 |
13 | 25 | 83.3% | | 6 | 2 | | None | None | CDF | 36 |
Oncologic outcomes: CDF (complete disease free) |
MSTS score is calculated as percentage of the maximum possible score of 30. |
We detail this procedure in a case of reconstructing the bone defect in the proximal humerus. (The procedure for reconstructing the bone defect in the distal radius was similar to that): A 50-year-old man had a local recurrence of GCT at the right proximal humerus, who underwent intralesional curettage at the local site and had a defect filled with a bone cement spacer one year ago (Fig. 2a). As the bone cortex and marrow cavity in the proximal humerus had been damaged and their data could not be used to design the mold, we based on the CT scan data mirrored from the left humerus to complete the design. The details are presented as follows (Fig. 1):
①The CT scan data of the left humerus in DICOM 3.0 format was imported into the software of Mimics 17.0 (Materialise company, Belgium) to reconstruct the model of the humerus. Mirror the image of the normal (left) humerus to the right side. ②Geomagic Studio(Raindrop, USA) and Materialise 3-MATIC 12.0 (Materialise, Belgium) were applied to complete the virtual repair of the humerus bone model by filling holes and redrawing mesh to make the image smooth. ③The distance between the apex of the humeral head and the expected osteotomy plane on the affected side was measured. The osteotomy plane was designed with a 2 cm safe surgical margin, and the length of the resected segment was 14 cm. ④The virtually repaired humeral bone model in STL format was imported into the Geomagic Studio software for surface reconstruction and then saved in IGES format. The SoildWorks2014 software (Dassault Systemes, France) was used to complete the mold design. ⑤The mold was designed detachable in order to make the bone cement easy to take out (Fig. 2. b1). The final mold design was imported into the industrial-grade 3D printer (3D-AZSL500, ZRapid Tech, China) to conduct printing.
All surgeries were performed by an experienced musculoskeletal oncologist (corresponding author).
Section 1: The resection and reconstruction of the proximal humerus.
Following general anesthesia, the patient was positioned in a “beach-chair” position. An incision was made from the acromioclavicular joint along the deltopectoral groove and the lateral border of the biceps muscle to an appropriate level in the arm. The cephalic vein was bluntly dissected and protected by a Pancoast's drain, then the deltoid muscle and the pectoralis major muscle were separated and retracted bilaterally. After separating the pectoralis major from its insertion into the proximal humerus, good exposure of the proximal humerus were acquired. A needle was inserted into the cortex of the proximal humerus to mark the osteotomy plane. The osteotomy plane should be at a minimal 2.0 cm from the lower boundary of the tumor lesion to ensure a wide margin, and this procedure was confirmed by intraoperative fluoroscopy. Then, the en block resection was performed with a Gigli saw at the intended osteotomy plane. The mean length of the resected proximal humerus was 12.7 cm (range, 9.0 cm to14.5 cm) in this case series. Tissues obtained from the distal medullary canal were sent for intraoperative frozen section analysis to confirm adequate margins. According to the Enneking System for Staging Benign and Malignant Musculoskeletal Tumors, no stage of proximal humeral bone tumors in these series exceeded ⅡA. A wide resection was preferable, excessive resection of the adjacent soft tissue seemed unnecessary, and the rotator cuff was preserved.
Meantime, one assistant prepared the bone cement prosthesis on another operating table. The inside surface of the bone cement mold was evenly coated with a thin layer of bone wax, which made it easier to remove the bone cement prosthesis after hardening (Fig. 2. b2). After the bone cement was poured into the mold and before it self-setting, 2 shaped Kirschner wires were inserted right in the middle of the bone cement (Fig. 2. b3). The Kirschner wires supported as scaffolds, connecting the proximal bone cement to the medullary canal of the distal humerus. Curling the distal part of the Kirschner wires could improve the anti-movement stability of the bone cement prosthesis and prevent them from prolapse. The overall length of the bone cement prosthesis in this case was 270 mm, and the distance between the apex of the humerus head and the expected osteotomy plane was 140mm as preoperatively designed (Fig. 2. c). Low-viscosity cement was injected into the medullary canal of the distal humerus, and then the stem of the bone cement prosthesis was inserted and adjusted to a proper position (Fig. 2. d). After the bone cement hardened, a final fluoroscopy was conducted to confirm the implant position. The rotator cuff and capsule were repaired as much as possible. Finally, we closed the wound and immobilized the shoulder in an abduction humeral splint.
Section 2: The resection and reconstruction of distal radius
A 57-year-old woman who was diagnosed with an aneurysmal bone cyst at her left distal radius in 2019 underwent intralesional curettage followed by cancellous allografts. The patient had a local recurrence one year later (Fig. 3. a-b). She chose the option of a bone cement prosthesis to reconstruct the defect after resecting the bone tumor.
The details of the operation were as follows:
A dorsal approach was used to expose the distal radius. The extensor retinaculum was divided, and a spacing interval between the third and fourth extensor tendons was developed to enable full exposure of the distal radius. A Gigli saw was used to perform en bloc resection of the distal radius, and the osteotomy plane was at a distance of minimal 2.0 cm from the proximal boundary of the tumor lesion (Fig. 3. c). The mean length of the resected segment of the distal radius was 6.9 cm (range, 6.0 cm to 8.0 cm) in this case series. The fabrication of the bone cement prosthesis and the reconstruction procedure were similar to those used in the reconstruction of the proximal humerus described above (Fig. 3. d-f). Lastly, the dorsal retinaculum was repaired as much as possible.
No significant complications occurred during the intraoperative or immediate postoperative periods. The surgical positions of the patients were required for fixation by plaster for 4–6 weeks. Continuous passive movements started once the plaster was removed. We suggested active rehabilitation exercises but did not recommend full weight-bearing. Patients with high-grade osteosarcoma received adjuvant chemotherapy routinely after the wound had healed.
Patients were followed up every 3 months during the first year and every 6 months thereafter, during which, physical examination, VAS assessment, and X-ray scanning were performed. To rule out pulmonary metastases, a CT scan of the chest was performed every six months in the first year and every year thereafter. The functional outcomes of the reconstructed limb were assessed by a physical therapist based on the Musculoskeletal Tumor Society score (MSTS score) (15).
All statistical analyses were performed by SPSS 22.0 Statistical Package for the Social Sciences (SPSS, Inc., Chicago, IL, USA), and a p-value of less than 0.5 was accepted statistically significant. The data are presented as the mean ± standard deviation. Statistical differences in pre-and post-surgery VAS scores were compared using the paired-sample t-test.