Osteosarcoma of the spine
Experience of the Cooperative Osteosarcoma Study Group
Article first published online: 20 FEB 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 4, pages 1069–1077, 15 February 2002
How to Cite
Ozaki, T., Flege, S., Liljenqvist, U., Hillmann, A., Delling, G., Salzer-Kuntschik, M., Jürgens, H., Kotz, R., Winkelmann, W. and Bielack, S. S. (2002), Osteosarcoma of the spine. Cancer, 94: 1069–1077. doi: 10.1002/cncr.10258
- Issue published online: 20 FEB 2002
- Article first published online: 20 FEB 2002
- Manuscript Accepted: 19 SEP 2001
- Manuscript Revised: 11 SEP 2001
- Manuscript Received: 8 JUL 2001
- Alexander von Humboldt Foundation
- Deutsche Krebshilfe
Due to the low incidence rate, the optimal strategy for the treatment of patients with spinal osteosarcoma is unknown.
Twenty-two patients with osteosarcoma of the spine (15 with tumors of the sacrum and 7 with tumors at other sites) who received chemotherapy according to the Cooperative Osteosarcoma Study Group protocol were analyzed. Six patients presented with metastasis, and 16 patients had no evidence of metastasis at the time of entry into the protocol. Of 12 patients who underwent excision of their tumors, 2 patients underwent wide excision, 3 patients underwent marginal excision, and 7 patients underwent intralesional excision. Eight patients received irradiation: Six patients received conventional radiotherapy only, one patient received neutron beam therapy, and one patient received samarium-153-ethylene diamine tetramethylene phosphonate therapy. Follow-up ranged between 24 months and 105 months (median, 47 months).
The median survival was 23 months, and three patients have survived without disease for > 6 years. Patients with primary metastases (P = 0.004), large tumors (P = 0.010), and sacral tumors (P = 0.048) had lower overall survival compared with patients who had no metastasis, small tumors, and nonsacral tumors, respectively. There was a significant difference in overall survival between 5 patients who underwent either wide or marginal surgery and 17 patients who underwent either intralesional surgery or no surgery (P = 0.033). Among 17 patients who underwent no surgery or intralesional surgery, overall survival tended to be better in 7 patients who received irradiation compared with the overall survival in 10 patients who did not receive irradiation (P = 0.059).
Patients with metastases, a large tumors, and sacral tumors had a poor prognosis in the current study with small numbers of patients. Wide or marginal excision of the tumor improved survival. Patients with osteosarcoma of the spine should be treated with a combination of chemotherapy and at least marginal excision for those with surgically accessible tumors. Postoperative radiotherapy may be beneficial. Cancer 2002;94:1069–77. © 2002 American Cancer Society.
Spinal osteosarcoma accounts for 3.6–14.5% of primary spinal tumors1, 2 and 0.85–3.0% of all osteosarcomas.3 It tends to occur in a slightly older age group, with a mean age of 38 years,4 than osteosarcoma of the extremity.3, 5, 6 It has been found that the prognosis for patients with osteosarcoma of the trunk is worse compared with the prognosis of patients with osteosarcoma of the extremities.3, 5, 7 In recent years, aggressive adjuvant and neoadjuvant chemotherapy has improved the treatment outcome of patients with osteosarcoma.3, 5, 8–11 There have been several articles on the treatment of patients with osteosarcoma of the spine;3, 5, 12 however, those publications included patients who were treated over a 30-year or 40-year period under different concepts of treatment3, 5 or were a type of case report6, 13–16 because of the rarity of spinal or sacral osteosarcomas.
In this study, we analyzed our experience treating patients with osteosarcoma of the spine who received chemotherapy according to a Cooperative Osteosarcoma Study Group (COSS) protocol between 1980 and 1998. To our knowledge, this work represents the largest series of patients with spinal osteosarcoma treated with a single concept and the first report on the outcome of multi-institutional treatment for patients with osteosarcoma of the spine.
MATERIALS AND METHODS
From the end of 1979 until June 1998, 1905 patients with osteosarcoma were registered on the consecutive prospective neoadjuvant studies performed by COSS. The site of origin of the tumor was recorded for all patients. The results obtained in this cohort of patients have been reported to our study center. For this study, we selected all patients with tumors arising in the spine and the sacrum (Table 1). Patients with tumors of the sacroiliac or the costotransversus region of the spine were included only if the major part of the osteosarcoma was located in the sacrum or the thoracic vertebrae, respectively. The diagnosis had to be confirmed histologically as high-grade osteosarcoma.
|No.||Gender||Age (yrs)||Site||Localizationab||Size (cm)||Type||Metastasis||Neurologic gradingc||Chemotherapy||Laminectomy||Tumor resection||Surgical margind||Histologic effect||Radiotherapy (gy)||Recurrence||Prognosis||Follow-up (mos)|
|Local||Metastasis (mos after diagnosis)|
|3||M||19||Sacrum, SI joint||IV||18||Unclassified||+ (multiple)||D3||COSS82||−||—||—||—||—||+||+ (0)||DOD||17|
|5||M||16||L1, L2||6||6||Teleangiectatic||−||A||COSS86||+||Resection and bone graft||Intralesional||?||—||−||−||CDF||105|
|6||F||27||Sacrum SI joint||IV||5||Chondroblastic||+||D3||COSS91||−||Internal hemipelvectomy||Intralesional||4||—||+||+ (0)||DOD||19|
|7||M||14||Sacrum SI joint||IV||?||Chondroblastic||+||B||COSS91||−||—||—||—||—||+||+ (0)||DOD||7|
|9||F||24||T5, T6||6||10||Osteoblastic||−||COSS86C||−||Resection||Intralesional||?||—||+||+ (17)||DOD||22|
|11||M||16||Sacrum, SI joint||IV||8||Osteoblastic||−||COSS86C||−||Resection||Marginal||4||Neutron beam (65)||−||−||CDF||75|
|12||F||34||Sacrum, L5||IV||8||Unclassified||−||COSS86C||−||Resection and posterior fusion||Wide||5||—||−||+ (22)||DOD||33|
|16||M||16||T1, T2||6||12||Small cell||+||D1||COSS86C||−||Resection, spacer implantation, and posterior fusion||Intralesional||?||—||+||+ (0)||DOC||22|
|17||F||16||Sacrum, L5||IV||12||Chondroblastic||+||D3||COSS86C||−||Resection, bone graft, and posterior fusion||Intralesional||?||—||+||+ (0)||DOD||15|
|19||F||28||T8, T9||6||7||Osteoblastic||−||COSS96||−||Resection and bone graft||Wide||?||—||−||−||CDF||32|
|20||F||22||Sacrum, SI joint, L5||IV||17||Chondroblastic||+||COSS96||−||—||—||3||Samarium-153-EDTMP, 60 gy to metastasis||+||+ (0)||DOD||17|
|21||F||5||C4||?||3||Teleangiectatic||−||D3||COSS96||+||Resection, bone graft, and anterior fusion||Intralesional||6||50||−||−||AWD||26|
|22||F||10||Sacrum, SI joint||IV||11||Chondroblastic||−||D2||COSS96||+||—||—||—||56||+||−||AWD||24|
In the COSS protocols COSS 80, COSS 82, COSS 86, COSS 86C, and COSS 917–10, patients were designated to receive high-dose methotrexate with leucovorin rescue and doxorubicin in different combinations with one or several of the following agents: cisplatin, ifosfamide, bleomycin with cyclophosphamide and actinomycin D, or α-interferon. In COSS 96 with risk-stratified therapy, etoposide and carboplatin were added in a group of high-risk patients.17, 18 In COSS 80 through COSS 96, both preoperative and postoperative chemotherapy were scheduled.7–10, 17 All studies were approved by the local ethics committee and/or the protocol review committee of the German Cancer Society or the German Ministry for Science and Technology. Before the initiation of treatment, informed consent was obtained from all patients or their legal guardians.
Preoperative neurologic deficit was classified using the information available at the COSS center according to Bradford and McBride's modification of the Frankel grading system (Table 2) 19: The location and extension of the spinal tumors (not sacral tumors) could be classified into seven types according to the criteria by Tomita et al.20 (Table 3). The location and extension of sacral tumors also were classified into four types according to the criteria by Tomita et al.21 (Table 4, Fig. 1). The surgical margin was classified according to the method of Enneking et al.22 After surgery with preoperative chemotherapy, the surgical specimens were examined histologically and classified into six categories of regression grades according to the criteria published by Salzer-Kuntschik and associates.23 Then, Grades 1–3 were used to classify good responses, and Grades 4–6 were used to classify poor responses.
|A||1||No motor or sensory function|
|B||2||Preserved sensation only|
|C||1||Preserved motor function (nonfunctional)|
|D1||1||Preserved motor function (lowest functional grade) or with bowel or bladder paralysis|
|D2||3||Preserved motor function (midfunctional grade) or neurogenic bowel or bladder|
|D3||4||Preserved motor function (high functional grade) and normal bowel or bladder|
|E||Normal or complete recovery|
|Grade||Patient no.||Compartment||Tumor site|
|Type 1||0||Intracompartmental||Site 1 or 2 or 3 (anterior or posterior lesion in situ)|
|Type 2||0||Intracompartmental||Site (1 + 2) or (3 + 2) (extension to pedicle)|
|Type 3||0||Intracompartmental||Site (1 + 2 + 3) (anterior-posterior development)|
|Type 4||0||Extracompartmental||Any site + 4 (epidural extension)|
|Type 5||0||Extracompartmental||Any site + 5 (paravertebral development)|
|Type 6||5||Extracompartmental||Involvement to adjacent vertebra|
|Type 7||0||Multiple, skip lesion||—|
|Grade||Patient no.||Tumor site||Surgery|
|Type I||0||Localized below S3||Low sacral amputation|
|Type II||0||Localized below S2||High sacral amputation|
|Type III||5||Involvement of S1||Total sacrectomy|
|Type IV||10||Extended into the lumbar spine, the sacroiliac joint, or the intrapelvic organs||Extended sacrectomy|
The Mann–Whitney U test was used to evaluate differences in the mean rank between two groups. The cumulative probability of survival was calculated using the Kaplan–Meier method. Tests of the difference between survival curves were carried out using the log-rank test and the Breslow–Gehan–Wilcoxon test. Statistical software used was the Stat View (version 5.0; StatSoft, Inc., Tulsa, OK) on a Macintosh system (Apple Computers, Cupertino, CA).
Twenty-two of 1905 patients with osteosarcoma (1.2%) had spinal tumors (Table 1). Fifteen tumors were located in the sacrum, 5 tumors were located in the thoracic spine, one tumor was located in the cervical spine, and one tumor was located in the lumbar spine. The age of patients ranged between 5 years and 55 years (median, 17 years). Nine patients were male, and 13 patients were female. One patient (Patient 14) had had a previous malignant disease. This patient had received both chemotherapy as well as radiotherapy of the mediastinum with 44 grays (Gy) 18 years previously for Hodgkin disease, and osteosarcoma developed in the thoracic spine within the field of irradiation. There was no Paget osteosarcoma. Six patients had detectable primary metastases at the time of diagnosis: Five patients had pulmonary metastases, and one patient (Patient 3) had pulmonary, chest wall, and skin metastases. The histologic diagnosis was made by the local pathologist from biopsy and/or resection material in all patients. Moreover, 15 tumor samples were reviewed centrally. In 17 patients, the histologic subtype was specified: There were 6 chondroblastic tumors, 6 osteoblastic tumors, 2 small cell tumors, 2 teleangiectatic tumors, and 1 fibroblastic tumor. Follow-up ranged from 24 months to 105 months (median, 47 months).
The period of symptoms between onset and diagnosis ranged from 2 months to 18 months (median, 5.0 months). The duration of symptoms did not differ significantly between patients who were diagnosed in the 1980s (median, 5.7 months) and patients who were diagnosed in the 1990s (median, 4.9 months). The most common symptom was pain, which was reported in 20 patients. The second most common symptom was neurologic disorder, which was reported for 12 patients: 1 patient with Grade A, 2 patients with Grade B, 1 patient with Grade C, 1 patient with Grade D1, 3 patients with Grade D2, and 4 patients with Grade D3 (Table 2).
In 20 patients, radiologic findings on X-ray could be classified from existing radiology reports. Eight patients had osteoblastic findings (Fig. 2), seven patients had osteolytic findings, four patients had mixed findings, and one patient had no finding on plain X-ray.
Size and Tumor Extension
Tumor size could be estimated in 19 patients by the greatest tumor dimension. Tumor size ranged between 3 cm and 18 cm (median, 8 cm). Eight tumors were large (≥ 10 cm), and 11 tumors were small (< 10 cm). The median size of 15 sacral tumors was 8.0 cm, and the median size of 7 nonsacral tumors was 6.7 cm. There were five type 6 spinal tumors: All tumors showed paravertebral development and involvement of adjacent vertebra (Table 3). There were 5 type III sacral tumors and 10 type IV sacral tumors (Table 4, Fig. 4).
Two patients received chemotherapy according to the COSS 80 protocol, two patients received the COSS 82 protocol, one patient received the COSS 86 protocol, 9 patients received the COSS 86C protocol, 3 patients received the COSS 91 protocol, and 5 patients received the COSS 96 protocol. Information on the histologic effect of preoperative chemotherapy was available for 6 patients: One patient had a good response (Grade 3), and 5 patients had poor responses (two Grade 4 responses, two Grade 5 responses, and one Grade 6 response).
Six patients underwent surgery without knowledge of the correct diagnosis: One patient underwent an initial surgery for osteomyelitis, four patients underwent emergent laminectomy, and one patient underwent simple resection. Finally, tumor resection was attempted in 12 patients and was not planned in 10 patients. Of 12 patients, 2 patients achieved wide margins, and 3 patients achieved marginal margins. Seven patients achieved intralesional margins. Local recurrence developed in neither of the two patients who underwent wide resection, in one of three patients who underwent marginal resection, and 4 of 7 patients who underwent intralesional resection.
Eight patients received irradiation to the primary tumor site. One patient who underwent intralesional surgery received postoperative radiotherapy consisting of 50 Gy. One patient who underwent marginal surgery received neutron beam therapy (total, 65 Gy). Five patients who did not undergo surgical removal of their tumors received 30 Gy, 30 Gy, 45 Gy, 54 Gy, and 56 Gy irradiation as first-line local therapy, and one patient received high-activity Samarium (Sm)-153-ethylene diamine tetramethylene phosphonate (EDTMP; 150 megabecquerels per kg of body weight; total, 8.1 gigabecquerels) with peripheral blood stem cell rescue in addition to 60 Gy irradiation to a pulmonary metastasis.
Nineteen of 22 patients (86%) survived > 1 year. At the time of analysis (August, 2000), three patients have survived in a continuously disease free condition for > 6 years after they were diagnosed with osteosarcoma. Seven patients with nonsacral tumors had better overall survival compared with 15 patients with sacral tumors (P = 0.048) (Table 5). Six patients with primary metastases had significantly worse overall survival compared with 16 patients without metastases (P = 0.004) (Fig. 3). In 19 tumors in which the greatest dimension was known, 11 patients with tumors measuring < 10 cm had better overall survival compared with 8 patients with tumors measuring ≥ 10 cm (P = 0.010) (Fig. 4).
|Factor||Log-rank test (Mantel–Cox test)||Breslow–Gehan–Wilcoxon test|
|Site (sacrum or others)||0.081||0.048|
|Primary metastasis (positive or negative)||0.004||0.015|
|Size (≥ 10 cm or < 10 cm)||0.010||0.015|
|Margin (wide marginal or intralesional)||0.013||0.033|
|RT in patients with intralesional or no surgery (positive or negative)||0.059||0.103|
The four patients who underwent emergent laminectomy survived for 24–105 months (median, 44 months). There was a significant difference in overall survival between 5 patients who underwent wide resection (2 patients) or marginal resection (3 patients) and 17 patients who underwent either intralesional resection or no surgery (P = 0.033) (Fig. 5). Of 17 patients who underwent no surgery or intralesional resection, 7 patients received radiotherapy. Although the local control rate was not different between patients with and without irradiation, there was a tendency for an improved overall survival rate in the 7 patients who received irradiation (P = 0.059) (Fig. 6).
The spine rarely is affected by osteosarcoma.2, 3, 5, 24, 25 Primary osteosarcoma of the spine accounts for 1.2% of all patients in this series, which is very similar to previous reports.3, 4, 24 Twenty of 22 patients had pain, and 12 of 22 patients had a neurologic deficit. This tendency also was similar to the review report by Green et al.4 The osteoblastic shadow is a common radiographic feature of osteosarcoma arising in the spine (Fig. 2A,B); however, osteolytic lesions also occurred frequently in 7 of 20 patients (35%) compared with a rate of 13.7% of all osteosarcomas.26 Conversely, one patient with osteosarcoma in this series showed no findings on plain radiogram, and only computed tomography scans or magnetic resonance imaging could show the tumor. In particular, magnetic resonance imaging is indispensable in the assessment of patients with intraosseous tumor spread, particularly in the sacrum or sacroiliac joint, or in the identification of neural compression.4
The Pediatric Oncology Group has reported a median interval between symptom onset and diagnosis of 56 days in pediatric patients with osteosarcoma at all sites.27 In the current series, however, the median latency was 5 months. Because the early diagnosis of spinal sarcomas is difficult,27 treatment often is delayed. Shives et al.5 reported that only 7 of 26 patients who were treated between 1909 and 1980 survived for > 1 year. The median survival has been reported in the range of 6–10 months for patients with osteosarcoma of the spine.3, 5, 16, 25 Our series suggests that patient outcome can be better with aggressive interdisciplinary treatment. Nineteen of 22 patients (86%) patients survived > 1 year, and 3 patients survived > 6 years.
There are only a few reports on long-term survivors of spinal osteosarcoma after receiving chemotherapy with3 or without radiation therapy.15 In general, however, the effect of chemotherapy is only temporary, and preoperative chemotherapy should be followed by intensive local treatment. In recent years, not only chemotherapy7–10, 17 and imaging techniques28 but also surgical techniques20, 25, 29–31 have been developed in spinal sarcoma surgery. Kawahara et al.14 reported reconstruction after en bloc resection of the affected vertebra by osteosarcoma. Recent progression of surgical techniques may enable not only corporectomy but also total sacrectomy20, 29, 32 to improve the survival of patients. However, the surgical intervention is large, and the complication rate seems to be high in these surgeries. Radical surgery, such as sacrectomy, leads to significant compromise of extremity function and bowel and bladder function. It is true that surgery is challenging and may be attractive to the orthopedic tumor surgeon, but the question remains whether this type of surgery is really beneficial to the affected patients. Our results show that complete tumor resection may improve their prognosis.
In patients with acute spinal paralytic symptoms, laminectomy must be performed as soon as possible. Laminectomy or root decompression or both may result in transient improvement of neurologic signs or symptoms. However, tumor cell contamination in the operative field poses a problem, and resection with adequate margins during definitive surgery becomes difficult. In this series, emergency laminectomy did not show an impact on the overall survival of patients. At the time of laminectomy, it is important that as little as possible dissemination of tumor tissue should be permitted. Also, with regard to biopsy, computed tomography-guided needle biopsy may be optimal to prevent contamination of tumor tissue.33
In the current series, overall survival for patients who underwent either intralesional surgery or no surgery and received adjunctive irradiation was better compared with overall survival for patients without this modality, a phenomenon similar to that reported by Shives et al.5 Sundaresan et al.25 also suggested that the combination of complete tumor resection (spondylectomy) with chemotherapy and radiation may increase survival. In their report, three patients who underwent surgical tumor resection and irradiation with chemotherapy survived longer than 36 months without evidence of disease. Targeted internal radiotherapy with Sm-153-EDTMP may offer an additional treatment option for some patients with inoperable osteosarcomas.34, 35
Primary metastasis, a large tumor, a sacral tumor, and intralesional or no tumor surgery were associated with poor survival in the current study with a small number of patients. Although this technique can be used only for patients with surgically accessible tumors, wide or marginal excision improves the survival of affected patients. Postoperative irradiation influenced the improvement in survival for some patients. Patients with osteosarcoma of the spine should be treated with a combination of chemotherapy and at least marginal surgery if they have surgically accessible tumors. Postoperative radiotherapy may be of benefit in selected patients.
- 11Osteogenic sarcoma of the extremity with detectable lung metastases at presentation. Results of treatment of 23 patients with chemotherapy followed by simultaneous resection of primary and metastatic lesions. Cancer 1997;79: 245–54., , , , , , et al.
- 21Indication and operative method of sacral amputation. In: KanedaK, editor. Orthopaedic surgery now, no. 22. Surgical techniques for disorders of the thoracolumbar, lumbar and lumbosacral spine. Tokyo: Medical View, Inc., 1996: 188–97., , , .
- 24Dahlin's bone tumors general aspects and data on 11,087 cases, 5th ed. Philadelphia: Lippincott-Raven, 1996..