1100
Metastatic Osteosarcoma at Diagnosis Prognostic Factors and Long-Term Outcome—The French Pediatric Experience
Valerie Mialou, M.D.1 Thierry Philip, M.D., Ph.D.1 Chantal Kalifa, M.D.2 David Perol, M.D.3 Jean-Claude Gentet, M.D.4 Perrine Marec-Berard, M.D.1 Helene Pacquement, M.D.5 Pascal Chastagner, M.D., Ph.D.6 Anne-Sophie Defaschelles, M.D.7 Olivier Hartmann, M.D.2 1
Pediatrics Department, Leon Berard Center, Lyon, France.
2
Pediatrics Department, Gustave Roussy Institute, Villejuif, France.
3
Public Health Department (Unite´ de Biostatistique et d’Evaluation des The´rapeutiques [UBET]), Leon Berard Center, Lyon, France.
4
Pediatrics Department, Timone Children’s Hospital, Marseille, France.
5
Pediatrics Department, Curie Institute, Paris, France.
6
Pediatrics Department, Nancy, France.
Children’s
Hsopital,
7
Pediatrics Department, Oscar Lambret Center, Lille, France.
BACKGROUND. The objective of this report was to estimate long-term outcome and prognostic factors in children and adolescents who presented with metastatic osteosarcoma at diagnosis. Patients were treated in six French pediatric oncology centers with surgery and multiagent chemotherapy, mainly with high-dose methotrexate. Their medical records were reviewed retrospectively. METHODS. The medical records of patients who were treated for metastatic osteosarcoma from 1987 to 2000 were reviewed. Patients were treated with the chemotherapy regimens recommended for nonmetastatic disease in children (the French Society of Pediatric Oncology OS 87 and OS 94 protocols) or, in a few patients, with other chemotherapy regimens. Surgical excision of the primary tumor and, when possible, of all metastatic sites was performed based on a personalized assessment of each patient’s situation.
RESULTS. Seventy-eight patients age ⬍ 20 years were treated. Forty-six patients (59%) had only 1 metastatic site (35 to the lungs and 11 to bone). Twenty-eight patients (36%) achieved a complete remission after combination chemotherapy and surgery. The event-free survival and overall survival rates at 5 years were 14% and 19%, respectively. To date, 14 patients (18%) have remained alive with a median follow-up of 112 months. Pretreatment features associated with a shorter event-free survival in the multivariate analysis were metastasis to at least two organs and high alkaline phosphatase level. Patients with at least 1 of these poor prognostic factors had a 2.6% event-free survival rate at 5 years despite treatment. CONCLUSIONS. The survival of patients with metastatic osteosarcoma were treated with conventional chemotherapy and surgery remained very poor. Patients should be classified into different prognostic groups and treated accordingly. New therapeutic approaches are warranted to improve the prognosis for patients with the most severe disease. Cancer 2005;104:1100 –9. © 2005 American Cancer Society.
KEYWORDS: metastatic osteosarcoma, children, prognostic factors, treatment.
O The authors thank L. Saint Ange for editing. Address for reprints: Vale´rie Mialou, M.D., Service d’He´matologie Pe´diatrique, Hoˆpital Debrousse, 29 Rue sœur Bouvier, 69005 Lyon, France; Fax: (011) 33 04 72385503; E-mail:
[email protected] Received November 29, 2004; revision received March 31, 2005; accepted April 11, 2005.
steosarcoma is a malignancy that is observed mainly in teenagers and young adults. The prognosis for patients with this illness in childhood has improved greatly during the past 3 decades with the use of multiagent chemotherapy, with the improvement of diagnostic technologies (such as computed tomography [CT] scans and magnetic resonance imaging [MRI]), and with a more aggressive surgical approach. The 5-year event-free survival rate for patients with localized disease has attained 70%.1 In contrast, the prognosis for patients with metastatic osteosarcoma remains poor, with an event-free survival rate of only approximately 20% at 5 years.2–7 Fifteen to 20% of patients present with metastatic dissemination at diagnosis,7 and until recently, such patients usually have been treated like patients with nonmetastatic lesions using conventional chemotherapy regimens and surgical excision of each tumor site,
© 2005 American Cancer Society DOI 10.1002/cncr.21263 Published online 13 July 2005 in Wiley InterScience (www.interscience.wiley.com).
Prognostic Factors: Metastatic Osteosarcoma/Mialou et al.
whenever possible. Unfortunately, most clinical trials in osteosarcoma exclude patients with metastatic disease at presentation. The therapeutic regimen is chosen by the physicians treating the patient, which signifies that these patients do not usually receive consistent and homogenous treatment. The paucity of published data on advanced osteosarcoma and the small size of each study prompted us to review our experience to evaluate the outcome of this group of patients. The objective of this study was to identify potential prognostic factors in these patients at diagnosis, to study their survival, and to evaluate the results of the treatments used to date.
MATERIALS AND METHODS Patients Between 1987 and 2000, patients with metastatic osteosarcoma (Stage IIIB according to the Enneking staging system8) were not included prospectively in therapeutic protocols in France. However, most of these patients were recorded in protocols as diagnosed with metastatic disease and were treated according to the existing therapeutic recommendations. A list of patients who were diagnosed with metastatic osteosarcoma at initial presentation between January 1, 1987 and January 31, 2000 was generated. All medical and imaging records of patients who were treated in the six biggest centers were reviewed, and data were collected. The date the study began was chosen as the beginning of standard chemotherapy regimens based on high-dose methotrexate and, as far as technically possible, aggressive surgical treatment with en bloc resection of the primary tumor and metastases. A minimal follow-up of 3 years was required for the inclusion of patients in the study.
Diagnosis All patients underwent a biopsy of the tumor site (usually the primary tumor site) to establish the diagnosis. The size and volume of the primary tumor were evaluated on CT scans or MRI studies (volume was assessed by multiplication of width, height, and length). The work-up for metastases included a chest X-ray and CT scans; 99mtechnetium bone scan; and, if necessary, a standard X-ray or MRI of all painful bony locations. Patients with an isolated, small (⬍ 5 mm) lung metastasis that was not proven histologically were classified with nonmetastatic disease and were not included in this study. Patients with only skip metastases also were not included in the study. All pulmonary lesions were discussed for each patient. Only patients who had unequivocally metastatic lesions
1101
were included in the study: At least 1 metastasis had to measure ⬎ 5 mm in greatest dimension.
Treatment Among the 78 patients recorded, 1 patient died of respiratory distress before receiving any chemotherapy. The 77 other patients were treated with a combination of multiagent chemotherapy and surgery.
Chemotherapy Neoadjuvant chemotherapy was used as often as possible (n ⫽ 69 patients) with agents used for nonmetastatic disease. It consisted of combinations of drugs with known effectiveness against osteosarcoma, such as high-dose methotrexate,9 doxorubicin,10 cisplatinum,11 etoposide, and high-dose ifosfamide.12 Twenty-two patients were treated according to the French Society of Pediatric Oncology (SFOP) OS 87 protocol,13 39 patients were treated according to the OS 94 protocol,14 and 8 patients received other regimens. Based on the protocols used for patients with nonmetastatic disease, adjuvant chemotherapy was adapted after surgical excision of the primary tumor according to the histologic response: Drugs were modified if the histologic response was poor (⬍ 90% tumor necrosis obtained), and the same regimen was continued if the histologic response was high (⬎ 90% tumor necrosis). Among the eight patients who first underwent surgical excision of the primary tumor, 7 patients received adjuvant chemotherapy according to the OS 87 protocol.13
Clinical response Response was appraised clinically before each course of chemotherapy and by a complete imaging evaluation before surgery or earlier if there was clinical suspicion of progressive disease (PD). Complete remission (CR) was defined as the total disappearance of all tumor sites, partial remission (PR) was defined as a decrease ⬎ 50% in the volume of each measurable site, no response was defined as a decrease or an increase ⬍ 25% in tumor volume, and progressive disease (PD) was defined as progression of ⱖ 1 tumor sites or a new metastatic site.
Surgery Eight patients first underwent surgery because of tumor size and/or infection or because of severe pain associated with a very huge tumor. This was particularly true for patients arriving from developing countries with a delayed diagnosis. In these patients, primary amputation often was the only appropriate local treatment. Most patients (69 of 77 patients) were treated with
1102
CANCER September 1, 2005 / Volume 104 / Number 5
neoadjuvant chemotherapy. The surgical approach was aggressive and was aimed at removing the primary tumor and all metastatic sites using the recommended surgical procedures for sarcoma.15,16 Limbsalvaging surgery was performed as often as possible. The chronological course of surgery for local and metastatic disease was adapted for each patient according to the clinical evolution of the disease and postsurgical complications, avoiding a long period off chemotherapy whenever possible. In the OS protocol, surgical excision of the primary tumor took place usually 1 week after the completion of neoadjuvant chemotherapy between the Weeks 12 and 14 of treatment.
Histologic Response According to Huvos et al.,15 a good histologic response is defined as ⬍ 10% of detectable viable tumor cells. Patients were divided into 3 groups: good responders (⬎ 90% tumor necrosis), poor responders (⬍ 90% tumor necrosis), and patients with an unevaluable histologic response (those who initially underwent surgery).
TABLE 1 Patient Characteristics (n ⴝ 78 patients)
Characteristic No. of patients Male/female ratio Primary tumor Long bones Flat bones Common histology Bone metastases Isolated Multiple Lung metastases ⱕ8 ⬎8 Unilateral/bilateral Current status CR1 CR2 CR3
No. with lungs metastases
No. with bone metastases
No. with metastases at multiple sites
35 19/16
11 4/7
32 18/14
34 1 25
10 1 8
30 2 21
0 0
3 8
10 16
23 12 11/24
0 0 0
14 18 4/28
6 2 1
3 0 0
1 0 0
CR: complete remission.
Statistical Analysis The endpoints of this study were event-free survival and overall survival. Event-free survival was defined as the period from the start of the treatment date to the date when disease progression (or death) was observed. Overall survival was defined as the period from the start of the treatment date to the date of death. Patients who remained alive at the time of data cut-off were censored at the last date the patient was known to be alive. Survival distributions were estimated using the Kaplan–Meier method.17 Correlations between survival and potential prognostic features were analyzed using the log-rank test.18 All prognostic features were transformed for use in a categorical form. The baseline clinical and biochemical characteristics examined included gender, age at diagnosis (ⱕ 14 yrs vs. ⬎ 14 yrs), site (legs vs. others) and size (ⱕ 200 cm3 vs. ⬎ 200 cm3) of the primary tumor, histologic type (sarcoma with osteoblastic differentiation vs. others), number of metastases (1 site vs. ⬎ 1 site), sites of metastases (lung, bone), and serum alkaline phosphatase (AP) level (ⱕ 500 UI/L vs. ⬎ 500 UI/L; 500 UI/L was the median AP level in patients at diagnosis and was used as the predetermined cut-off level to avoid biasing the results). The tissue AP level was not studied as a prognostic factor, because it was not available at diagnosis. Treatment features examined included: complete resection of metastases (yes vs. no) and histologic response to neoadjuvant chemotherapy. Four baseline clinical and biochemical characteristics were included
in Cox proportional hazards regression models19 on the basis that they were associated significantly with event-free survival and overall survival in a univariate analysis (P ⬍ 0.05). A backward procedure was used to eliminate noninfluential variables (variables were rejected in the model if their P value was ⬎ 0.05). In addition, in the last step of the models, we investigated whether the association of different prognostic factors with survival was independent of complete resection of metastases.
RESULTS Patient Characteristics Between January 1987 and December 2000, 78 patients with metastatic osteosarcoma were treated in 6 French centers (Table 1). Only one patient had a previous underlying disease (bone fibrotic dysplasia). The median patient age was 13.5 years (range, 2–19 yrs). The gender ratio was 1.1 (37 girls and 41 boys). The median Karnofsky index at diagnosis was 80 (range, 30 –100). Symptoms caused by the primary tumor were present for a median of 2.8 months (range, 0.4 –17 mos) before treatment.
Disease Characteristics Table 1 shows that the main primary tumor location was the long bones, and primary tumor site distribution was similar to that of patients without metastatic disease. The median volume of the primary tumor, as
Prognostic Factors: Metastatic Osteosarcoma/Mialou et al. TABLE 2 Metastatic Sites No. of metastatic sites One metastatic site Lung ⱕ 8 metastases ⬎ 8 metastases Bone Isolated/multiple Axial/peripheral bones More than one metastatic site Bone 2 Sites 3 Sites
No. of patients (%)
35 (45) 23 (29) 12 (15) 11 (14) 3 (4)/8 (10) 9 (12)/2 (3) 32 (41) 23 (29) 9 (12)
measured on imaging, was 525 cm3 (range, 75–5355 cm3). The majority of patients had only 1 metastatic site (59%) (Table 2), which affected the lung in most of them (45%); 6 patients had an isolated metastasis alone, and only 1 of them was not proven histologically after surgical biopsy at diagnosis. Twenty-one patients had ⬍ 8 pulmonary metastases. Disease was not proven histologically in four patients, but all four patients died of their disease. Only nine patients had respiratory symptoms (dyspnea, pain, or coughing) related to lung metastases. Two-thirds of the patients had a common histologic type with osteoblastic differentiation (n ⫽ 54 patients). Eight patients had tumors with chondroblastic differentiation, six patients had a mixed histology, five patients had an anaplastic form, and five patients had another histologic type.
Treatment Among the 78 patients, 1 patient died of respiratory distress before any treatment, leaving 77 patients evaluable for response. Eight patients initially underwent amputation, whereas the other 69 patients were able to receive neoadjuvant chemotherapy. Chemotherapy regimens were at the discretion of the physicians responsible for treatment. Most patients (n ⫽ 63 patients) received chemotherapy regimens that contained high-dose methotrexate, such as those recommended in SFOP pediatric protocols (according to Rosen’s T10,20 OS 87,13 and OS 9414) combined with other drugs with known effectiveness against this disease (ciplatinum, doxorubicin, ifosfamide, and etoposide). Five other patients were treated with multiagent chemotherapy regimens without methotrexate and with higher doses of ifosfamide. A clinical response to neoadjuvant chemotherapy was obtained in 37 of 69 evaluable patients (54%).
1103
Twenty-one of the 77 evaluable patients had disease progression despite chemotherapy and were not eligible to undergo local surgery. Fifty-six of 69 patients who did not undergo amputation initially underwent surgical excision of their primary tumor after neoadjuvant chemotherapy, and 38 of those patients underwent limb-salvaging surgery. The surgical procedure resulted in an amputation in 18 patients. Eight of those patients initially underwent surgery because of unbearably painful bulky disease or severe infection that prohibited neoadjuvant chemotherapy. Ten patients underwent radical surgery after chemotherapy, because a limb-salvaging procedure was not possible due to their tumor location. Thirty-five patients underwent surgical excision of metastases (22 patients for lung metastases alone, 3 patients for bony metastases alone, and 10 patients for multiple metastatic sites). Disease was resected completely in 20 patients. Among the 43 patients who had metastases that were not resected, disease progressed or recurred in 36 patients who ultimately died, and 7 patients achieved a CR after chemotherapy, 3 of whom remained alive in first CR.
Histologic Response to Treatment Tumor necrosis was assessable in the primary tumor after neoadjuvant chemotherapy in 48 of 56 patients. Indeed, 30 patients were not evaluable, 8 patients because of surgery first, 14 patients because of disease progression that did not allow local surgery, and 8 patients because of nonavailable data. Twenty of 48 patients (42%) were good responders to neoadjuvant chemotherapy, and 28 patients (58%) were poor responders. It was determined that this did not have prognostic value in univariate analysis (Table 3).
Postoperative Treatment Fifty-five of the 56 patients who underwent resection of their primary tumor received adjuvant chemotherapy (1 patient had no subsequent treatment). Fortyeight patients already had received neoadjuvant chemotherapy: Nine of 20 good responders were treated with the same chemotherapy regimen before and after surgery, and 35 patients received a different chemotherapy protocol. The other two patients in this group were lost to follow-up soon after surgery after a few courses of adjuvant chemotherapy. In seven patients, chemotherapy was administered only postoperatively (primary amputation). Six patients received high-dose thiotepa with autologous stem cell transplantation support at the end of their conventional treatment as part of a Phase II
1104
CANCER September 1, 2005 / Volume 104 / Number 5
TABLE 3 Baseline Patient Characteristics EFS
Characteristics Age ⱕ 14 yrs ⬎ 14 yrs Gender Male Female Site of primary tumor Legs Others Size of primary tumor ⱕ 200 cm3 ⬎ 200 cm3 Histology Osteogenic sarcoma Others histology No. of metastatic sites 1 ⬎1 No. of lung metastases ⱕ8 ⬎8 Bilateral lung metastases Yes No Bone metastases Yes No Complete resection of metastases Yes No Histologic response Yes No Alkaline phosphatase ⱕ 500 UI/L ⬎ 500 UI/L
OS No. of events
Five-yr OS (%)
0.54
31 28
23 14
0.52
13 15
0.73
32 27
17 21
0.75
52 11
15 14
0.83
48 11
20 15
0.77
7 46
4 37
34 16
0.44
4 33
33 24
0.76
53 24
43 20
15 13
0.96
41 18
18 20
0.60
46 31
34 29
21 4
0.002
30 29
30 3
0.0004
37 29
31 24
14 9
0.03
27 24
24 8
0.007
51 15
43 12
11 14
0.09
41 10
15 27
0.055
36 41
32 31
11 16
0.049
32 27
11 26
0.02
35 42
27 35
21 9
0.0003
17 31
39 12
0.0001
20 28
14 21
30 17
0.11
10 18
44 26
0.08
30 30
24 28
14 4
0.002
21 27
24 7
0.002
No. of patients
No. of events
Five-yr EFS (%)
43 34
35 28
13 15
40 37
34 29
64 13
P valuea
P valuea
EFS: event-free survival; OS: overall survival. a Log-rank test.
study. None of them was in CR at the time of intensification. At the end of treatment, 27 patients were in CR, 13 patients were in PR, 34 patients had PD, 2 patients had died during therapy, and 1 patient had died before any treatment of disease progression.
remained alive, including 10 patients in first CR, 2 patients in second CR, 1 patient in third CR, and 1 patient with PD.
Patients Lost to Follow-Up Recurrence and Disease Progression Sixty-two patients (80%) experienced either disease progression (n ⫽ 32) or disease recurrence (n ⫽ 30). Events occurred 0.3–31.0 months after the start of the treatment (median, 6.3 months). Fourteen patients
Four patients who attained a PR were lost to follow-up during or soon after treatment: They often returned to their country of origin and probably died within a few months of treatment but were considered still alive at last follow-up and were censored thereafter.
Prognostic Factors: Metastatic Osteosarcoma/Mialou et al.
Survival and Metastatic Sites Lung metastases Thirty-five patients had lung metastases alone at presentation (Tables 1, 2) with 1–50 tumor nodules (median, 4 tumor nodules). The site of the primary tumor was mainly the femur (22 of 35 patients). The histologic type was osteoblastic in most patients (25 of 35 patients). Twenty-nine of 35 patients underwent surgical excision of the primary tumor, and 22 of 35 patients also underwent resection of metastases, which resulted in a complete excision in 14 patients. Adjuvant chemotherapy was modified in 15 patients. Seven of 14 patients who underwent complete surgical excision of all tumor sites were alive (with 6 patients in first CR and 1 patient in third CR) as of March 2004, whereas only 3 of the other 21 patients were alive (2 patients in second CR and 1 patient with PD). Patients in first CR had a follow-up of 78 –251 months after their diagnosis. Three patients were lost to follow-up while they were in PR, and 22 patients died. The two patients who remained alive in second CR without initial surgery of the metastases were in CR under chemotherapy alone. They both developed recurrences with unique pulmonary metastases and are long-term survivors after intensive chemotherapy and surgery.
Bone metastases Eleven patients had bony metastases alone at presentation (Tables 1, 2). The primary tumor site was mainly long bones, and histology was mainly osteoblastic. Eight of these 11 patients underwent surgical excision of the primary, and 6 of them had local treatment of metastatic lesions (3 patients underwent surgery, and 3 patients received irradiation for unresectable metastatic lesions). Among the six patients who underwent local treatment of the primary and metastatic lesions, three patients remained alive in first CR. This first remission has been sustained for 70 –102 months after the diagnosis. The duration of first CR for the patient who received radiation therapy for unresectable bone metastasis was 102 months, and the 2 others patients who achieved a CR at metastatic sites with chemotherapy alone had a first CR duration of 164 –244 months. All patients who did not receive local therapy for metastatic lesions died of their disease.
Metastases at multiple sites Thirty-two patients had several metastatic sites at presentation. Twenty-two patients had bone metastases associated with lung metastases alone (n ⫽ 17 patients) or with other multiple metastatic sites (n ⫽ 6 patients). The other 10 patients had ⬎ 2 metastatic
1105
sites. The sites of the primary tumors were similar to those reported in other subgroups with metastases (Tables 1, 2). Most patients had bilateral lung metastases (28 of 32 patients), and 50% of patients had multiple bone metastases. Among those 32 patients, 27 patients were treated with neoadjuvant chemotherapy, 4 patients underwent primary surgery, and 1 patient died before any treatment. Only in 19 of 32 patients underwent surgery for the primary tumor, and only 10 of 32 patients underwent surgery for metastatic sites. Only 6 patients achieved a CR at the end of the treatment (3 of 6 patients underwent surgery for metastases), and only 1 patient (3%) remained alive with a sustained first CR at 164 months after the diagnosis. All patients with lymph node metastases (n ⫽ 13 patients) that were local or distant died (n ⫽ 12 patients) or were lost to follow-up (n ⫽ 1 patient).
Survival and Prognostic Factors Fourteen patients remained alive 29 –251 months postdiagnosis (median, 112 months). Ten patients were in first CR, 2 patients were in second CR, 1 patient was in third CR, and 1 patient had PD after a recurrence. Among the 13 disease-free patients, 12 patients were in CR and 1 patient was in PR at the end of treatment. Among the 13 patients who remained alive in CR, 12 patients had only 1 metastatic site, including: 3 patients with metastasis to the skeleton (2 to axial bones and 1 to peripheral bone) and 9 patients with metastasis to the lungs (bilateral in 5 patients and ⬍ 8 nodules in 7 patients). The last patient had both metastatic sites. All of these patients received chemotherapy that contained high-dose methotrexate; all underwent complete resection of the primary tumor; and, in 9 of 13 patients, resection of all metastatic sites was complete (1 patient with unresectable bony metastases received radiotherapy, and the 2 other patients achieved a CR at their metastatic site with chemotherapy alone). The 5-years event-free survival rate for patients with metastases was poor at 14% (median follow-up, 10.6 months) as was overall survival at 19% (median, 17.5 months) (Fig. 1). Baseline characteristics that were associated significantly (P ⬍ 0.05) with a limited event-free and overall survival were the number of metastatic sites (⬎ 1 site), the presence of metastases in bone, the number of lung metastases (⬎ 8 metastases), bilateral lung metastases, and the AP level (⬎ 500 UI/L) (Table 3). These variables were included in the multivariate analysis. Two parameters were identified as independent prognostic factors for event-free survival and overall survival in this series: the AP level (ⱕ 500 IU/L vs. ⬎ 500 IU/L) and the number of metastatic sites (1
1106
CANCER September 1, 2005 / Volume 104 / Number 5
significantly poorer 5-year event-free survival (24% vs. 3%; P ⫽ 0.0005) and overall survival (39% vs. 5%; P ⫽ 0.0001) (Fig. 2A,B).
DISCUSSION
FIGURE 1. Overall and event-free survival (EFS) is illustrated in 77 pediatric patients with metastatic osteosarcoma. TABLE 4 Results of the Multivariate Analysis for Survival EFS Factor
OS
HR 95%CI P valuea HR 95%CI P valuea
Alkaline phosphatase ⬎ 500 UI/L 3.6 1.8–7.1 0.001 Several metastatic sites 3.2 1.6–6.2 ⬍ 10⫺4
2.7 1.4–5.3 0.004 2.7 1.4–5.4 0.004
EFS: event-free survival; OS: overall survival; HR: hazard ratio; 95%CI: confidence interval. a Cox proportional hazards regression model.
TABLE 5 Results of the Multivariate Analysis for Survival after Adjustment for Treatment Factors EFS Factor
OS
HR 95%CI P valuea HR 95%CI P valuea
Alkaline phosphatase ⬎ 500 UI/L 2.7 1.5–4.8 0.001 Several metastatic sites 2.8 1.5–5.2 0.001 Incomplete resection of metastases 2.0 1.1–3.5 0.02
2.2 1.2–4.1 0.01 2.5 1.3–4.7 0.004 2.2 1.2–4.1 0.01
EFS: event-free survival; OS: overall survival; HR: hazard ratio; 95%CI: confidence interval. a Cox proportional hazards regression model.
site vs. ⬎ 1 site) (Table 4). These results remained stable after adjustment for complete resection of metastases (Table 5). Consequently, each patient was assigned to two prognostic groups according to the number of prognostic factors, i.e., no factors versus one or two factors. Patients with 1 or 2 factors had a
Until the 1980s, metastatic osteosarcoma was considered a dismal disease, because was no hope of survival for these patients. In the first trials that used intensive chemotherapy,20 patients with primary or secondary metastatic disease were studied together. It was concluded that only an aggressive regimen could increase survival for these patients. The first publication on patients with metastatic osteosarcoma at presentation6 confirmed that, until 1982, all patients had died. In addition, it was announced that the use of intensive chemotherapy, along with improved diagnostic and therapeutic techniques (CT scan, surgery), resulted in an increase in 2-year survival from 0% to 30%. The main characteristics of the patients in the current study (gender, age, and primary tumor site) were are similar to those found in recently published studies.21,22 The current study included patients who were treated in several of the largest French pediatric oncology centers. The surgical approach probably was more aggressive in those centers than in smaller institutions, and this may have had an impact on the survival of such patients. The current treatment results were similar to those published previously (Table 6), in which 5-year even-free and overall survival rates between 14% and 46% and between 11% and 53%, respectively, were reported (Table 4). In the current study, those rates were approximately 14% and 19%, respectively (Fig. 1). These differences probably reflect dissimilarities between the populations studied: In the current study, 55% of patients had extrapulmonary metastases compared with from 0%6 to 29%21 of such patients in the other studies with better treatment results. In the current study, several prognostic factors were investigated in the univariate analysis. Some of those factors have been described in the literature. Patient age did not appear to have prognostic value, an observation that is at variance with the findings of Meyers et al.4 The number of metastatic sites was of major prognostic significance, as shown previously by Kager et al.23 Patients with metastases in a single organ (mainly lungs or bone) had a better prognosis than patients who had several organs involved. This finding was not linked to the difficulties in achieving a complete excision, because these two factors were independent in the multivariate analysis. It may suggest that the parameter “multiple metastatic sites” is associated with a different biologic profile of tumor cells. Further genomic studies may explore this hy-
Prognostic Factors: Metastatic Osteosarcoma/Mialou et al.
1107
FIGURE 2. Event-free survival (A) and overall survival (B) is illustrated according to the number of prognostic factors at diagnosis (no factors vs. one or two factors). TABLE 6 Published Data on Pediatric Metastatic Osteosarcoma: Event-Free Survival and Overall Survival Study
Study period
No. of patients
Patient age (yrs)
EFS (%) (yrs since DX)
OS (%) (yrs since DX)
Marina et al., 19926 (St. Jude; USA) Meyers et al., 19934 (NY; USA) Pacquement et al., 19965 (France) Harris et al., 19982 (POG; USA) Kaste et al., 19997 (USA) Ferguson et al., 200122 (COG; USA) Goorin et al., 200221 (POG; USA) Kager et al., 200323 (Germany)
1982–1989 1975–1984 1980–1990 1987–1990 1977–1997 1980–1999 1990–2000 1979–1998
18 62 73 30 32 36 41 202
⬍ 21 No limits ⬍ 19 ⬍ 30 ⬍ 25 ⬍ 23 ⬍ 30 2–60
— — — 46.7 (5) 14.0 (5) 24.0 (3) 43.0 (2) —
50.0 (3) 11.0 (5) 15.0 53.3 (5) 29.0 (5) 32.0 (3) 55.0 (2) 31.0 (5)
EFS: event-free survival; OS: overall survival; POG: Pediatric Oncology Group; COG: Children’s Oncology Group.
pothesis. Similar to the findings from other studies,7,6,22 the current results confirmed that complete excision of all metastatic sites carries a better prognosis. The prognostic value of resectability of all metastatic sites probably also is linked to the sensitivity of the disease to chemotherapy. Indeed, patients with resectable lesions were selected during therapy, because only responders to chemotherapy underwent curative surgery. The number of lung metastases also had prognostic value in the univariate analysis: Greater than eight lung metastases signaled a poor outcome, probably because numerous lung metastases were less likely to
be resected completely, signifying a high risk of residual disease after surgery. This was confirmed by the multivariate analysis, in which the prognostic value of the number of lung metastases was cancelled in favor of resectability. The variable “bone metastases” was an adverse prognostic factor in the current study, as also reported by others.4,2,5 In the other series, only patients who were treated recently survived after bone metastases (3 of 12 patients [Goorin et al.21]; 3 of 7 patients [Ferguson et al.22]). Similarly, 3 of our 11 patients with bone metastases alone remained alive at the time of this report: Two of them had resectable bone metas-
1108
CANCER September 1, 2005 / Volume 104 / Number 5
tases, and 1 had radiotherapy (45 Gy) to an inoperable metastatic site. All three of those patients had an isolated bone metastasis. Among the eight patients who died, three patients underwent a complete resection of disease, but all had several bone metastases that could be classified as osteosarcomatosis. Apparently, two different situations may arise24: some patients have osteosarcoma with one or two true bone metastases, and the others have multifocal osteosarcoma that presents as different, synchronous, primary tumor sites sometimes called osteosarcomatosis.25 This disease has never been cured; and, to date, all affected patients have died.26 None of our eight patients with more than two metastatic skeletal sites (without lung metastases) have survived. This is in accordance with previously published data.24 –26 It seems that the only likelihood of improving the prognosis for patients with synchronous, multifocal osteosarcoma may be to use a very aggressive approach combining intensive chemotherapy with surgical resection of all involved bones.22 The histologic response to neoadjuvant chemotherapy is a recognized prognostic factor in patients with nonmetastatic osteosarcoma15 and also in patients with metastatic disease.1 This was not confirmed in the current study, but only 60% of our patients were assessable for histologic response, and that may explain this difference with findings in the literature. Moreover, the current study included more patients with bone metastases than were included in others studies. This very strong prognostic factor may explain why histologic response did not appear to be a prognosis factor. Surgical excision of all tumor sites is a major, independent prognostic factor that has been reported in other studies.2,4 However, 1 of the long-term survivors (102 months) with bone metastases in that study was treated with radiotherapy instead of surgery for inextirpable metastatic localizations. Three of our patients who currently are alive in first CR did not undergo surgical resection of their pulmonary metastases, because they entered CR of the metastases at the end of neoadjuvant chemotherapy. The AP level is an indicator of bone formation and reflects bone tumor volume and growth kinetics. A high AP level signals a worse prognosis because of the existence of a more aggressive and bulky tumor. We found that this marker had a prognostic value, as demonstrated exclusively in patients without metastases in previous reports.1,4 In the current population of patients with metastases, the proportion with bone lesions was greater (55% vs. 22%) compared with the proportion found in other studies, and this fact could may the difference between their results and ours.4
According to our results, the AP level may be useful as a simple prognostic factor. Prospective studies should be performed to confirm this finding. Finally, at the time of diagnosis, two main prognostic factors appear to be relevant: the AP level and the number of metastatic sites. During the course of treatment, data on these factors are completed after surgery, allowing patients to be stratified into prognostic groups so that treatment may be adapted to the severity of the disease. Although the outcome of patients with osteosarcoma has improved with the use of intensive chemotherapy regimens,16 the results obtained with conventional therapies remain poor for patients with metastatic disease. Treatment adapted to prognostic factors could change their outlook. For patients with “low-risk metastatic disease” (an AP level ⬍ 500 UI/L and resectable metastases at a single site), it is possible that the current treatment with a very aggressive chemotherapeutic and surgical approach may lead to acceptable results. Some have suspected that metastases from osteosarcoma may be resistant to high-dose methotrexate27 through an increase in the reduced folate carrier in tumor cells. However, in the current study, the majority of patients received methotrexate as first-line therapy; and all patients, currently alive and disease free, were among them. Thus, the efficacy of high-dose methotrexate appears to be significant, and this drug should remain a major tool in future protocols. High-dose chemotherapy followed by autologous stem cell transplantation is a standard strategy in pediatric tumors such as high-risk neuroblastoma.28 A Phase II study of highdose thiotepa has yielded an encouraging response rate in osteosarcoma.29 Other combinations (highdose etoposide and carboplatin) also have produced promising results.30 However, the place of this approach has not been defined clearly in the therapeutic strategy for patients with metastatic osteosarcoma. For patients with “high-risk metastatic disease” (an AP level ⱖ 500 UI/L and/or ⬎ 1 metastatic site and unresectable metastases), new and innovative therapeutic approaches are needed. Novel approaches with cellular molecular targets are emerging and probably will occupy a major place in treatment for osteosarcoma either alone or in association with chemotherapy. Furthermore, a better understanding of the biology of osteosarcoma31–36 should help describe new prognostic markers in the near future.
REFERENCES 1.
Meyers PA, Heller G, Healey J. et al. Chemotherapy for non metastatic osteogenic sarcoma: the Memorial Sloan-Kettering experience. J Clin Oncol. 1992;10:5–15.
Prognostic Factors: Metastatic Osteosarcoma/Mialou et al. 2.
3.
4.
5.
6.
7.
8. 9.
10.
11.
12.
13.
14.
15.
16.
17. 18.
19. 20.
Harris MB, Gieser P, Goorin AM, et al. Treatment of metastatic osteosarcoma at diagnosis: a Pediatric Oncology Group Study. J Clin Oncol. 1998;16:3641–3648. Bacci G, Ferrari S, Longhi A, et al. High-grade osteosarcoma of the extremity: differences between localized and metastatic tumors at presentation. J Pediatr Hematol Oncol. 2002: 24:27–30. Meyers PA, Heller G, Healey JH, et al. Osteogenic sarcoma with clinically detectable metastasis at initial presentation. J Clin Oncol. 1993;11:449 – 453. Pacquement H, Kalifa C, Fagnou C, et al. Metastatic osteogenic sarcoma at diagnosis. Study of 73 cases from the French Society of Pediatric Oncology between 1980 and 1990 [abstract]. Med Pediatr Oncol. 1996;27:264. Marina NM, Pratt CB, Rao BN, Shema SJ, Meyer WH. Improved prognosis of children with osteosarcoma metastatic to the lung(s) at the time of diagnosis. Cancer. 1992;70:2722– 2727. Kaste SC, Pratt CB, Cain AM, Jones-Wallace DJ, Rao BN. Metastases detected at the time of diagnosis of primary pediatric extremity osteosarcoma at diagnosis: imaging features. Cancer. 1999;86:1602–1608. Enneking WF. A system of staging musculoskeletal neoplasms. Clin Orthop. 1986;204:9 –24. Rosen G, Caparros B, Huvos AG, et al. Preoperative chemotherapy for osteogenic sarcoma: selection of postoperative adjuvant chemotherapy based on the response of the primary tumor to preoperative chemotherapy. Cancer. 1982;49: 1221–1230. Tan C, Etcubanas E, Wollner N, et al. Adriamycin an antitumor antibiotic in the treatment of neoplastic disease. Cancer. 1973;32:9 –17. Ochs JJ, Freeman AI, Douglas HO Jr., et al. Cis-dichlorodiamineplatinum (II) in advanced osteogenic sarcoma. Cancer Treat Rep. 1978; 62:239 –245. Gentet JC, Brunat-Mentigny M, Demaille MC. Ifosfamide and etoposide in childhood osteosarcoma. A Phase II study of the French Society of Paediatric Oncology. Eur J Cancer. 1997;33:232–237. Kalifa C, Brunat-Mentigny M, Pacquement H, et al. Treatment of osteosarcoma with an intensive chemotherapy regimen (OS 87 protocol) [abstract O-159]. Med Pediatr Oncol. 1996;26:227. Kalifa C, Guinebretie´re JM, Gentet JC, et al. Comparison of doxorubicin versus etoposide-ifosfamide in osteosarcoma: a randomised trial by the French Society of Pediatric Oncology [abstract 1634]. J Clin Oncol. 2002;21S:409. Huvos AG, Rosen G, Marcove R. Primary osteogenic sarcoma: pathologic aspects in 20 patients after treatment with chemotherapy en bloc resection and prosthetic bone replacement. Arch Pathol Lab Med. 1977;101:14 –18. Rosen G. Neoadjuvant chemotherapy for osteogenic sarcoma: a model for the treatment of other highly malignant neoplasms. Recent Results Cancer Res. 1986;103:148 –157. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457– 481. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep. 1966;50:163–170. Cox DR, Oakes D. Analysis of survival data. New York: Chapman and Hall, 1990. Rosen G, Huvos GH, Mosende C, et al. Chemotherapy and
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
1109
thoracotomy for metastatic osteogenic sarcoma. Cancer. 1978;41:841– 849. Goorin AM, Harris MB, Bernstein M, et al. Phase II/III trial of etoposide and ifosfamide in newly diagnosed metastatic osteosarcoma: a Pediatric Oncology Group trial. J Clin Oncol. 2002;20:426 – 433. Ferguson WS, Harris MB, Goorin AM, et al. Presurgical window of carboplatin and surgery and multidrug chemotherapy for the treatment of newly diagnosed metastatic or unresectable osteosarcoma: Pediatric Oncology Group trial. J Pediatr Hematol Oncol. 2001;23:340 –348. Kager L, Zoubek A, Potschger U, et al., Cooperative GermanAustrian-Swiss Osteosarcoma Study Group. Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol. 2003;21:2011–2018. Mahoney JP, Spanier SS, Morris JL. Multifocal osteosarcoma: a case report with review of the literature. Cancer. 1979;44:1897–1907. Taccone A, Di Stadio M, Oliveri M, Oddone M, Occhi M. Multifocal synchronous osteosarcoma. Eur J Radiol. 1995; 20:43– 45. Jones RD, Reid R, Balakrishnan G, Barrett A. Multifocal synchronous osteosarcoma: the Scottish Bone Tumour Registry experience. Med Pediatr Oncol. 1993;21:111–116. Guo W, Healey JH, Meyers PA, et al. Mechanisms of methotrexate resistance in osteosarcoma. Clin Cancer Res. 1999; 5:621– 627. Matthay KK, Villablanca JG, Seeger RC, et al. Treatment of high-risk neuroblastoma with intensive chemotherapy, radiotherapy, autologous bone marrow transplantation, and 13-cis-retinoic acid. Children’s Cancer Group. N Engl J Med. 1999;341:1165–1173. Lucidarme N, Valteau-Couanet D, Oberlin O, et al. Phase II study of high-dose thiotepa and hematopoietic stem cell transplantation in children with solid tumors. Bone Marrow Transplant. 1998;22:535–540. Miniero R, Brach Del Prever A, Vassallo E, et al. Feasibility of high-dose chemotherapy and autologous peripheral blood stem cell transplantation in children with high grade osteosarcoma. Bone Marrow Transplant. 1998;22(Suppl 5):37– 40. Kilpatrick S, Geisinger KR, King TS, et al. Clinicopathologic analysis of HER2/neu immunoexpression among various histologic subtypes and grades of osteosarcoma. Mod Pathol. 2001;14:1277–1283. Gorlick R, Huvos AG, Heller G, et al. Expression of HER2/ erbB-2 correlates with survival in osteosarcoma. J Clin Oncol. 1999;17:2781–2788. Zhou H, Randall RL, Brothman AR, et al. Her-2/neu expression in osteosarcoma increases risk of lung metastasis and can be associated with gene amplification. J Pediatr Hematol Oncol. 2003;25:27–32. Maitra A, Roberts H, Weinberg AG, Geradts J. Loss of p16 (INK4a) expression correlates with decrease survival in pediatric osteosarcoma. Int J Cancer. 2001;95:34 –38. Serra M., Scotlandi K, Reverter-Branchat G, et al. Value of P-glycoprotein and clinicopathologic factors as the basis for new treatment strategies in high grade osteosarcoma of the extremities. J Clin Oncol. 2003;21:536 –542. Baldini N, Scotlandi K, Barbanti-Brodano G, et al. Expression of P-glycoprotein in high-grade osteosarcomas in relation to clinical outcome. N Engl J Med. 1995;333:1380 –1385.
