Cancer in Children: Clinical Management, 5th Edition

Chapter 19. Osteosarcoma

Stefan S. Bielack

Mark L. Bernstein


Osteosarcoma is the most frequent primary cancer of bone. The approximate annual incidence is 2–3 per million in the general population; it is <1 per million in children aged <five years, 2 per million at age 5–9 years, 7 per million at age 10–14 years, and peaks at 8–11 per million at age 15–19 years. In the period of peak incidence in adolescents aged 15–19 years, it accounts for more than 5 per cent of all cancers. A second smaller age peak in older patients is due to osteosarcomas arising in abnormal bones, such as those affected by Paget's disease or prior irradiation. This chapter describes osteosarcoma in children and adolescents. Tumours of older individuals will not be discussed further.

Osteosarcoma affects males approximately 1.4 times more often than females (Fig. 19.1). The primary tumour is usually located in the metaphysis of a long bone of an extremity, with the distal femur and the proximal tibia being the most frequent sites of involvement, followed by the proximal humerus and proximal fibula. In sum, two-thirds of all paediatric osteosarcomas arise around the knee (Fig. 19.2). Tumours of the axial skeleton or craniofacial bones occur almost exclusively in older patients.


By definition, osteosarcoma is a mesenchymal malignancy in which the malignant cell population produces osteoid [Fig. 19.3(a)]. The extent of osteoid production can vary considerably. Both abundant production leading to hard and sclerotic tumours and very scanty production are consistent with the diagnosis. Conventional osteosarcoma, a high-grade central malignancy of bone, accounts for 80–90 per cent of all osteosarcomas. Its most frequent subtypes are osteoblastic, chondroblastic, and fibroblastic osteosarcomas. Unusual histologic subtypes, such as sclerosing osteoblastic, osteoblastoma-like, chondromyxoid-fibroma-like, malignantfibrous-histiocytoma-like and chondroblastoma-like osteosarcomas, as well as giant-cell-rich and epithelioid osteosarcomas, are also included among the conventional osteosarcomas.1 Conventional, telangiectatic, high-grade surface and small-cell osteosarcomas all have a very similar clinical course and must be treated by multimodal regimens which include chemotherapy. Low-grade central and parosteal osteosarcomas are treated by surgery only. Craniofacial osteosarcomas, apart from those of the skull, metastasize less frequently than conventional osteosarcomas, as do periosteal osteosarcomas, so that there is no general consensus as to whether they should be treated by surgery alone or by surgery plus chemotherapy. Extraskeletal osteosarcomas, usually high-grade malignancies, are rare and are included among the soft tissue sarcomas.


The aetiology of osteosarcoma remains obscure in most patients. Trauma has often been accused, but little evidence exists to support a causal relationship. The predilection of osteosarcoma for the age of the pubertal growth spurt and the sites of maximum growth suggest a correlation with rapid bone growth. There has been a suggestion of a higher incidence in taller individuals, but epidemiologic studies have not been conclusive. A minority of osteosarcomas are caused by radiation exposure. The risk is related to the dose administered. Exposure to alkylating agents may also contribute to osteosarcoma development. Together, radiation therapy, alkylators, and genetic tumour predisposition, as described below, make osteosarcoma one of the most frequent secondary solid malignancies following therapy for childhood cancer. A viral aetiology has been suggested based on evidence that bone sarcomas can be induced in select animals by viruses and by the presence of SV40-like sequences in some human osteosarcomas. However, no convincing evidence has emerged that viruses are a major aetiologic factor in humans.2

Fig. 19.1 Age and sex distribution of 1791 paediatric patients from the Cooperative Osteosarcoma Study Group (COSS) with primary high-grade central osteosarcoma.

The incidence of osteosarcoma is increased in several well-defined hereditary disorders associated with germ-line alterations of tumour suppressor genes, but even taken together, these account for only a few per cent of all osteosarcomas. Survivors of hereditary retinoblastoma with germ-line mutations of the retinoblastoma gene RB1 on chromosome 13q14 carry a risk which is 500–1000 times greater than that of the general population. The Li–Fraumeni cancer family syndrome, in which germ-line mutations in the p53 gene are found, is associated with a 15-fold increase. Among patients with sporadic osteosarcoma, approximately 3 per cent will have germ-line p53 mutations. Although germ-line mutations of p53 and RB are rare, these genes are altered in many osteosarcoma tumour samples. Consequently, loss of function of the p53 and RB tumour suppressor genes, which regulate cell cycle progression in normal cells, are believed to have an important role in osteosarcoma tumourigenesis. Rothmund-Thomson syndrome and Bloom syndrome, rare conditions caused by mutations in tumour suppressor genes coding for helicases, are also associated with an increase in osteosarcomas, as is Werner syndrome (adult progeria) (see Fletcher et al.1 for reviews of congenital and inherited disorders associated with osteosarcoma development.

Fig. 19.2 Skeletal distribution of osteosarcoma (based on 1791 primary high-grade central osteosarcomas in paediatric COSS patients).

Numerous oncogenes are also altered in osteosarcoma tumour cells. These include amplifications of the product of the murine double minute 2 gene, amplification of cyclin-dependent kinase 4 and overexpression of human epidermal growth factor receptor 2. Although it is clear that alterations in tumour suppressor genes and oncogenes are necessary to produce osteosarcomas, it is not clear which of these events occurs first and why or how it occurs. Moreover, it is not clear which, if any, of the alterations is essential for tumour development and therefore might represent a therapeutic target.

Signs, symptoms, and natural disease course

Patients with osteosarcoma usually do not feel ill until very late in the course of the disease. They typically seek medical attention because of first intermittent and then continuous pain, which is often erroneously attributed to a recent trauma of the involved region, for instance a sports injury. Tumour-related swelling and loss of function of the adjacent joints usually develop later. In approximately 10 per cent, the first sign of disease is a pathologic fracture. Pain at bony sites other than the primary may represent metastatic involvement. However, metastases are most likely to occur in the lungs, and these produce respiratory symptoms only with extensive involvement. Systemic symptoms, such as fever and weight loss, occur rarely in the absence of very advanced disease.

Fig. 19.3 Osteoblastic osteosarcoma (a) before and (b) after preoperative chemotherapy, showing a good response to induction chemotherapy. (Courtesy of H. Bürger.)

The differential diagnosis of osteosarcoma includes traumatic lesions, osteomyelitis, benign bone tumours such as exostosis, fibroma, osteoid-osteoma, chondroma, giant cell tumour of bone, bone cysts, and others, as well as other primary malignant lesions of bone such as Ewing sarcoma or lymphoma, and metastases from malignancies such as neuroblastoma or soft tissue sarcoma.

At diagnosis, even the most accurate staging procedures will reveal metastases in no more than 10–20 per cent of patients. Primary metastases are limited to the lungs in 80 per cent of affected individuals. The remainder have bone metastases with or without additional pulmonary involvement. Skip metastases (isolated tumour foci within the same bone as the primary tumour) occur in a minority of patients. Regional lymph node metastases are rare, and other sites are almost never involved at initial diagnosis.3 If no systemic treatment is given, most patients with seemingly localized disease will go on to develop metachronous metastases within 1–2 years. Lungs and, to a lesser extent, distant bones are again the organs involved. Death is usually due to respiratory failure caused by extensive pulmonary involvement.

Diagnostic evaluation

History and physical examination

The evaluation of suspected osteosarcoma begins with a detailed history, physical examination, and plain radiographs. As stated above, most patients present with a history of pain of the involved region. Physical examination is typically remarkable only for a mass at the primary site. Loss of motion of neighbouring joints, infiltration of the skin, and neurologic deficits may occur, depending on the location and extent of the tumour.

Laboratory studies

There are no known specific laboratory parameters. Increases of alkaline phosphatase or lactic dehydrogenase (LDH) serum levels, which are observed in a considerable number of patients, do not correlate reliably with disease extent but may have negative prognostic significance.


The characteristics and extent of the primary tumour must be evaluated by plain radiographs and cross-sectional imaging techniques. On plain radiographs, which are mainly used to describe the bony compartment, osteosarcoma often presents with lytic or sclerotic changes, or both. Ossification in the soft tissue in a radial or ‘sunburst’ pattern is a typical finding, but neither sensitive nor specific. Periosteal new bone formation with lifting of the cortex leads to the appearance of Codman's triangle [Fig. 19.4(a)]. MRI is the most useful tool to define the intramedullary tumour extent, soft tissue component, and relation of the tumour to vessels and nerves [Fig. 19.4(b)].4

The search for metastases must focus on the two organ systems in which >95 per cent of all osteosarcoma metastases arise: the lungs and the skeleton. Plain radiographs [Fig. 19.5(a)] and a CT-scan of the thorax [Fig. 19.5(b)], preferably high-resolution spiral CT, must be performed to exclude pulmonary metastases. Bone metastases are searched for using a [99mTc]MDP bone scan [Fig. 19.6]. Skip metastases may also be visualized on the bone scan, but MRI of the whole bone is preferable because of its higher sensitivity. There is currently no established role for positron emission tomography (PET), which is inferior to CT for the detection of lung metastases and to bone scintigraphy for the detection of bone metastases. Whole-body MRI may lead to a higher detection rate of bone metastases, but its place in diagnostic evaluation remains to be defined.


While imaging will often result in a high index of suspicion, the diagnosis of osteosarcoma must always be verified histologically. In order to ensure appropriate biopsy techniques and an appropriate evaluation of the material obtained, it is strongly recommended that biopsies should be performed only in specialized centers. Open biopsy may be most suitable to obtain sufficient material for histologic evaluation and ancillary studies. The biopsy specimen should be forwarded to the pathologist without prior fixation.

Staging systems

Previous editions of the TNM staging system were not very well adapted to the necessities of bone sarcomas. Most clinicians, especially tumour surgeons, still prefer the system developed by the Musculoskeletal Tumour Society (MSTS)5 The MSTS system categorizes localized malignant bone tumours by grade (low grade = stage I or high grade = stage II) and by the local anatomic extent (intracompartmental = A or extracompartmental = B). The compartmental status is determined by whether or not the tumour extends through the cortex. It remains to be seen whether the latest edition of the TNM classification,6 which allows a more accurate description of osteosarcomas than previous versions (Table 19.1), will gain wider acceptance than its predecessors.

Fig. 19.4 (a) Conventional radiograph of an osteosarcoma of the distal femur in a 15-year-old girl with extensive intramedullary sclerosis and typical cloudy soft tissue calcifications. (b) MRI of a large osteosarcoma of the proximal humerus.

Treatment strategy

Many patients with osteosarcoma can be cured (Fig. 19.7). Inappropriate use of diagnostic tools and suboptimal initial therapy can irrevocably compromise a patient's chances. Therefore all patients with osteosarcoma should be treated in specialized experienced centers able to provide access to the full diagnostic and therapeutic spectrum. Treatment within prospective clinical trials is considered standard clinical practice in many countries. An intergroup trial, the European and American Osteosarcoma Study (EURAMOS-1), is currently being performed by several European and North American study groups.

Local treatment of osteosarcoma is surgical. However, most patients have already developed micrometastatic disease by the time their osteosarcoma is detected. Prior to the 1970s, when treatment was exclusively surgical, the outcome was extremely poor. Almost 90 per cent of patients who presented with apparently localized disease developed recurrences and died within 1–2 years of diagnosis. This dismal outlook was dramatically improved when multiagent chemotherapy was added to surgery. Many trials have since reported disease-free survival rates in the range of 50–70 per cent. Most investigators believed that the favourable results of single-arm studies of surgery plus combination chemotherapy were sufficient to demonstrate the superiority of the combined modality approach over surgery alone. However, a minority questioned the validity of historical comparisons. Therefore the American Multi-Institution

Osteosarcoma Study (MIOS) randomized patients with localized extremity osteosarcoma between surgery plus adjuvant multi-agent chemotherapy and surgery plus observation. Not surprisingly, this study confirmed the low cure rate (only 11 per cent disease-free survival) for patients who did not receive chemotherapy, while 66 per cent of the patients who received adjuvant chemotherapy were disease-free survivors.7 Ever since, the multimodal approach has remained the undisputed standard of care.

Fig. 19.5 (a) Radiograph of the chest with multiple pulmonary metastases after forequarter amputation for an osteosarcoma of the proximal humerus. (b) CT scan of the chest with pulmonary metastatic osteosarcoma. Note the round metastasis in the dorsal periphery of the left lung.

Effective agents

The majority of current treatment protocols are based on two or more of only four active agents: doxorubicin, cisplatin, high-dose methotrexate, and ifosfamide, although its absolute role has still not been estabished. Even after more than two decades of experience with these agents against osteosarcoma and despite numerous clinical trials, the exact role of each of the agents and the optimal way in which they should be combined and delivered are still being debated, as is the potential benefit of additional drugs.

Fig. 19.6 99mTC bone scan showing increased uptake in multiple foci in the pelvis (and proximal femur) in a 16-year-old patient with multifocal osteosarcoma

Table 19.1. TNM classification (6th edition) for malignant bone tumours and suggested staging system1,6

TNM classification


Primary tumour cannot be assessed


No evidence of primary tumour


Tumour ≤8 cm in greatest dimension


Tumour >8 cm in greatest dimension


Discontinuous tumours in the primary bone site


Regional lymph nodes cannot be assessed


No regional lymph node metastasis


Regional lymph node metastasis


Distant metastases cannot be assessed


No distant metastases


Distant metastases




Other distant sites

Staging system

Stage IA

T1, N0, M0 (low grade)

Stage IB

T2, N0, M0 (low grade)

Stage IIA

T1, N0, M0 (high grade)

Stage IIB

T2, N0, M0 (high grade)

Stage III

T3, N0, M0 (any grade)

Stage IVA

Any T, N0, M1a (any grade)

Stage IVB

Any T, N1, any M (any grade)
Any T, any N, M1b (any grade)

Fig. 19.7 (a) Overall survival with multimodal therapy. (b) Comparison of good responders (solid curve) and poor responders (broken curve). A good response is defined as < 10% viable tumour. Results are from 1152 COSS patients aged < 20 years with localized extremity osteosarcoma.



Doxorubicin was first introduced into osteosarcoma treatment in the 1970s and has remained an integral part ever since. A large meta-analysis of osteosarcoma trials concluded that, of all agents, only the dose intensity of doxorubicin was closely related to efficacy.8 Given the threat of long-term anthracycline cardiotoxicity, doxorubicin administration to young patients with a fairly high cure rate is troublesome. Some osteosarcoma protocols include measures aimed towards reducing doxorubicin cardiotoxicity. The use of cardioprotectants such as dexrazoxane and the reduction of anthracycline peak levels by continuous doxorubicin infusions feature most prominently. Sequential studies by American and European groups suggest that there is no major loss of efficacy, but no controlled studies evaluating whether cardiotoxicity can be reduced without influencing efficacy against osteosarcoma have been reported.


Methotrexate, a folate antagonist, blocks the action of dihydrofolate reductase, the enzyme responsible for reducing folate to its active form, tetrahydrofolic acid. In osteosarcoma, methotrexate is given at supralethal doses, usually in the range 8—12 g/m2. The severe toxicity which follows high-dose methotrexate administration must be antagonized by the antidote leucovorin (activated folate), thereby bypassing the blocked enzyme. The treatment concept of high-dose methotrexate is based on the assumption that normal cells can be rescued more effectively than tumour cells, which may lack active folate transporters. High-dose methotrexate therapy requires meticulous attention and extensive supportive measures, including hydration, alkalinization of the urine, and leucovorin administration adapted to methotrexate serum levels. Inadequate supportive care will result in severely delayed methotrexate clearance and excessive toxicity, including myelosuppression, increases of serum creatinine or frank renal failure, and severe gastrointestinal side effects. Some patients will experience such toxicity despite adequate supportive care; the risk appears to increase with increasing patient age. The enzyme carboxypeptidase G2, which cleaves methotrexate, may be of benefit in selected patients with renal failure and severely delayed methotrexate clearance, but most cases of delayed methotrexate clearance can be handled with high-dose leucovorin as sole therapy.

No other drug used against osteosarcoma has been associated with as much controversy as high-dose methotrexate. There is no doubt that some osteosarcomas show marked responses to this therapy. The superiority of high over conventional doses was clearly demonstrated in the randomized IOR I trial from Italy.9 There is still no consensus as to whether individual pharmacokinetic parameters, such as the peak methotrexate serum level, correlate with efficacy. Some investigators believe that the overall efficacy of any osteosarcoma protocol is strongly related to the amount of methotrexate included and the way in which it is administered. In contrast, the European Osteosarcoma Intergroup has claimed that the incorporation of methotrexate into doxorubicin/cisplatin-based protocols may compromise efficacy.10 The conclusions from their randomized trial EOI 80831 have been questioned, as the overall success rate in both arms was comparatively low, doxorubicin and cisplatin dose intensity was lower in the arm with methotrexate, and the chosen methotrexate dose of 8 g/m2 was relatively low. Given the relative lack of myelotoxicity and the resulting ability to schedule methotrexate at times when other more myelotoxic agents cannot be administered, most groups continue to incorporate high-dose methotrexate into their osteosarcoma protocols.


The efficacy of cisplatin against osteosarcomas was demonstrated in early phase II trials. The agent was subsequently incorporated into most multi-agent regimens. Cisplatin therapy requires supportive hyperhydration. Otoand nephrotoxicity are dose-limiting toxicities. Both can be reduced by administering the drug as a continuous infusion. Intra-arterial cisplatin administration directly into the artery supplying the tumour was investigated in the 1980s, but was later largely abandoned when comparative trials could not demonstrate enhanced antitumour effects compared with intravenous administration.11


Following positive phase II trials, ifosfamide has been part of many osteosarcoma protocols since the mid-1980s. Its efficacy may be related to the dose administered. Supportive measures necessary to prevent the otherwise frequent haemorrhagic uropathy after ifosfamide include hydration and the administration of mesna (Uromitexan). Ifosfamide may also lead to chronic renal tubular toxicity (the Fanconi syndrome) and sterility.

Based on the results of sequential trials, both the Istituto Rizzoli and Cooperative Osteosarcoma Study (COSS) group have reported that the inclusion of ifosfamide into their respective multi agent regimens was associated with improved event-free survival rates. However, a recent randomized trial by the Pediatric Oncology Group (POG) and the Children's Cancer Group (CCG) (intergroup 0133) could not demonstrate that the addition of standard dose ifosfamide to a regimen of high-dose methotrexate, doxorubicin, and cisplatin improved outcome.12 The results of intergroup 0133 must be interpreted with caution, as a second randomization of liposomal muramyl tripeptide (MTP-PE) may have interfered with the ifosfamide question. Also, the ifosfamide arm contained no preoperative cisplatin.

Other agents

No other agents have come even close to replacing the four standard substances described above. A combination of bleomycin, cyclophosphamide, and actinomycin D (BCD) was used in the early days of chemotherapy, but was later largely abandoned because of questionable efficacy. Carboplatin has some activity against osteosarcoma, but seems to be less active than cisplatin. Etoposide is almost inactive when given as a single agent, but may enhance the effect of carboplatin or ifosfamide. Negative phase II studies have been reported for paclitaxel, docetaxel, and topotecan. Gemcitabine seems to be marginally active, with stable disease having been reported for some patients. High-dose chemotherapy with autologous peripheral blood stem cell transplantation was unsuccessful in the few reported series.

Neoadjuvant chemotherapy

Currently, most institutions use an approach consisting of preoperative chemotherapy (also called neoadjuvant chemotherapy or induction chemotherapy), followed by definitive surgery and postoperative adjuvant chemotherapy. The neoadjuvant concept was first introduced into osteosarcoma therapy 25 years ago. The theoretical advantages include early treatment of micrometastatic disease and facilitation of the eventual surgical procedure because of tumour shrinkage and decreased vascularity. A theoretical concern is that delayed removal of the bulk tumour could lead to the emergence of chemotherapy resistance. Only one relatively small randomized trial has prospectively compared patients treated by preand postoperative chemotherapy with patients treated by primary surgery followed by the same chemotherapy. In this POG trial, treatment results did not differ between the two arms.13 Similarly, the COSS Group could not detect a survival difference between the two approaches in a large retrospective comparison of 157 patients with primary surgery and 1451 with primary chemotherapy.14 Given the advantages in facilitating limb salvage procedures and assessing chemotherapy efficacy, the use of induction chemotherapy has become the standard treatment approach. Response of the tumour to induction chemotherapy can be evaluated histologically to assess the effectiveness of therapy [Fig. 19.3(b)]. Most investigators would define a good response as <10 per cent viable tumour.

Prognostic factors

Several prognostic factors have been identified. The largest reported series of 1702 osteosarcoma patients found primary metastases, axial or proximal extremity location, and large tumour size to be of independent negative prognostic value.14 Survival may be better in fibroblastic and telangiectatic tumours than in osteoblastic and chondroblastic tumours.15 Other factors associated with an adverse outcome in some series include very young age or older age, high serum levels of alkaline phosphatase or LDH, and the immunohistochemical detection of p-glycoprotein or HER2erbB2. However, the relative risk associated with any presenting factor is lower than that of two treatment-related variables: Incomplete surgery was the most important negative prognostic indicator in the COSS series, followed by a poor histologic response to induction chemotherapy.14 There is ample evidence that axial, primary metastatic, and even secondary osteosarcomas can be cured if a complete surgical remission of all affected sites can be obtained.

Response and outcome

Response to induction chemotherapy is believed by many to be the most important prognostic factor for resectable osteosarcoma. However, response is not an all-or-nothing effect. Even a very moderate response is associated with a better prognosis than no response at all.14 Response rates may be higher in fibroblastic and telangiectatic tumours and lower in chondroblastic tumours.15 Methods used to predict response preoperatively include serial evaluation by angiography, quantitative bone scans, dynamic MRI, or PET. Early reports from the Memorial Sloan–Kettering Cancer Center suggested that the poor prognosis associated with poor response could be improved by altering postoperative chemotherapy.16 However, the COSS Group's trial COSS-82 failed to detect a benefit of salvage treatment even in poor responders who had received only very low intensity treatment preoperatively and then went on to very intensive doxorubicin/cisplatin salvage therapy.17 A reanalysis of the Sloan–Kettering results failed to confirm the salvage effect reported earlier.18 Several other attempts have been made to improve the prognosis of poor responders by altering postoperative chemotherapy, but none of these was a randomized trial in which a salvage approach was compared with unaltered postoperative chemotherapy. The only trial suggesting that salvage chemotherapy might work was the second study by the Rizzoli Institute, Bologna, where ifosfamide/etoposide was added postoperatively for those patients whose osteosarcomas had responded poorly to ifosfamide-free preoperative regimens.19

Local therapy

Despite the efficacy of chemotherapy against microscopic disease, it cannot reliably control clinically detectable osteosarcoma. Therefore surgery of the primary tumour and, if present, all metastases remains an integral part of successful therapy. Radiotherapy does not reliably sterilize osteosarcomas and is reserved for inoperable sites or those that can only be operated on with inadequate margins.

Osteosarcoma surgery has three aims: First and foremost, the tumour must be removed completely. Secondly, the patient should be left with good extremity function. Thirdly, surgery should, if possible, result in a cosmetically acceptable appearance. Complete tumour removal is of paramount importance, while functional and cosmetic aspects can only be secondary goals.

Definitive surgery must be planned and performed so that the complete tumour, including the biopsy scar and biopsy track, is removed with an unviolated cuff of normal tissue surrounding it. This corresponds to ‘wide’ margins as defined by the Musculoskeletal Tumour Society.5 The MSTS classification distinguishes between radical, wide, marginal, and intralesional resection margins (Table 19.2). Radical and wide margins are considered adequate, while marginal or intralesional margins are associated with an increased risk of local recurrence,20 which in turn carries a very poor prognosis.

The type of surgery needed to achieve wide margins varies according to the location, size, and regional anatomy of the tumour. En bloc resection with limb salvage is possible in many cases, but ablative techniques, such as amputation, will be required in others. The relation of the tumour to nerves and vessels of the popliteal fossa or axilla as well as structures of neighbouring joints must be carefully evaluated before deciding upon the type of surgery.

Particular care must be taken in tumours with insufficient response to preoperative induction chemotherapy, as this is associated with an excessive local failure rate in the case of inadequate margins.20 If pathologic evaluation of the resected tumour specimen reveals inadequate (marginal or intralesional) margins, this is usually an indication for revision surgery, even if this implies severe mutilation. Radiotherapy is indicated in cases where even the most aggressive surgical approaches will not result in wide margins, as is often the case in pelvic tumours.

Table 19.2. Surgical margins in musculoskeletal oncology5




Within the lesion


Through the pseudocapsule or reactive tissue


Lesion (including biopsy scar), pseudocapsule and/or reactive zone, and an unviolated cuff of normal tissue completely surrounding the mass removed as single block


Entire anatomic compartment containing the tumour removed as single block

Fig. 19.8 Changing distribution of definitive surgical procedures for extremity osteosarcoma (based on paediatric and adolescent COSS patients, 1980–2000).14

Only a minority of osteosarcomas are located in expendable bones, such as the proximal fibula, where no reconstruction is required. In all other cases, the choice of reconstruction must be based on the bone, nerves, vessels, soft tissues, and skin remaining after wide resection of the tumour. Many patients have not yet reached their final height at the time of tumour surgery, so that the remaining growth expectation must also be included in the decision-making process. Until ~30 years ago, amputation was the only form of bone sarcoma surgery. This has since changed dramatically, and limb-salvage techniques are now used in the majority of cases (Fig. 19.8). Even today, however, amputation may be the most appropriate type of surgery for selected patients with unfavourable tumour characteristics. Advantages of amputation include oncologic safety, a low complication rate, and a low rate of revision surgery. Disadvantages include mutilation, phantom sensations, and functional deficits. The functional outcome is often rather poor after proximal amputations, but below-knee amputations can lead to a good functional outcome.

Rotation-plasty, used for osteosarcomas of the distal femur and occasionally the proximal tibia, is the classic example of a resection–reimplantation procedure. The knee is removed en bloc together with the tumour; the only structure left in situ is the popliteal nerve bundle. The distal part of the tibia, together with the foot, is then rotated by 180°, the tibia is fused with the femoral stump, and anastomoses are created between the femoral and tibial vessels. The result of this reimplantation is a shortened extremity, in which the rotated foot substitutes for the knee and carries the prosthesis. Advantages of rotation-plasty include oncologic safety even in very large tumours, lack of phantom sensations, and an infrequent need for revision surgery, as well as an extremity function rivalling that of even the most favourable limb-salvage procedures. The highly unusual cosmetic appearance of the rotated foot is the main disadvantage of the procedure.

Today, most osteosarcoma patients are operated on with limb-salvage procedures. Retaining the extremity requires reconstruction of bony and soft tissue defects resulting from tumour resection with wide margins. Because of the predilection of the tumour for the metaphyses of the long bones, this usually implies replacement of one of the major joints. Allografts are used by some surgeons, but modular endoprosthetic systems which can be assembled to fit in the operating room are employed more commonly (Fig. 19.9). Autologous bones, such as a vascularized fibula bridging the defect left by the resection of an osteosarcoma of the femoral diaphysis, can be employed in selected situations. Special expandable endoprostheses are available for growing extremities, but these carry with them a need for multiple revision procedures. Disadvantages of limb salvage include a significant complication risk, to which infections, fractures, and prosthetic wear contribute. Several publications report a local recurrence rate which is approximately three times higher than after amputations or rotation-plasties, pointing to the fact that the resection margins achieved with limb-salvage procedures may sometimes not be as wide as expected.

Treatment of primary metastatic and relapsed disease

Detectable metastases must be removed by surgery if therapy is to be curative. As most metastases develop in the lung, this usually implies thoracotomy. There is evidence that a significant proportion of patients with apparently unilateral lung metastases may, in fact, have bilateral disease. In general, acceptable surgical alternatives include bilateral thoracotomy or median sternotomy. Complete resection of osteosarcoma pulmonary metastases requires palpation of the lung, which is not possible thoracoscopically.

Approximately a quarter of all patients with proven metastatic disease at diagnosis and 40 per cent of those who achieve a complete surgical remission of both the primary and all metastases in the context of an intensive polychemotherapy regimen will go on to become long-term survivors. Patients with solitary primary metastases may have a prognosis similar to those with localized disease.3

Osteosarcoma recurrences also usually involve the lung. Bone metastases and local recurrences are much less common, and other sites are rarely affected. Unfortunately, the survival rates after relapse are low, with <20 per cent of affected patients becoming long-term survivors. A short latency period and more than one or two metastases at relapse are associated with an especially poor outcome. Complete surgical removal of all sites of recurrence is the only therapeutic measure with unequivocally proven impact on survival. It may be prudent to irradiate suitable inoperable lesions in order to slow the progression of disease, but it is unlikely that this will lead to permanent cure. In a recent Italian series21 patients with inoperable osteosarcoma relapse who received chemotherapy survived longer than those who did not. The exact role of adjuvant chemotherapy in the treatment of operable relapsed osteosarcoma is still being debated. So far, success has been limited at best, and there is no accepted standard regimen outside clinical trials.


Remission status

Suggestions for a follow-up programme are given in Table 19.3. The intervals between clinic visits mirror the declining risk of relapse with time. As there have been no prospective trials addressing follow-up, all suggestions must necessarily remain somewhat arbitrary. However, any follow-up for osteosarcoma must include regular assessments of the remission status as well as tests for possible late effects of (successful) treatment. Tumour-directed follow-up should focus on the few organ systems where relapses are likely to occur. Pulmonary metastases, which are part of >80 per cent of recurrences, are potentially curable only as long as they are resectable.21 They will usually not cause symptoms until they have reached a very large size or penetrate the pleura. In order to detect them at an earlier stage, lung metastases must be searched for by appropriate imaging studies, which include serial radiographs and/or CT scans of the thorax. Local recurrences and bone metastases occur in a much lower percentage of patients and are often first detected because they cause symptoms, most noticeably pain. However, most centers will include sequential radiography of the primary site into the follow-up programme. There is no evidence that serial bone scans are beneficial.

Fig. 19.9 Modular endoprosthetic replacement after surgery for a distal femur osteosarcoma: (a) radiograph; (b) extremity function.

Table 19.3. Suggestions for a follow-up schedule after multimodal therapy for osteosarcoma



Late effects


Chest radiograph and CT
Radiograph and CT/MRI of primary site

Echocardiogram, audiogram, liver and kidney
function, hepatitis B/C and HIV serology

Years 1 and 2

Chest radiograph every 6–12 weeks
Radiograph of primary site every 4 months

Echocardiogram every 1–2 years, audiograma,
livera and kidneya function

Years 3 and 4

Chest radiograph every 2–4 months
Radiograph of primary site every 4 months

Echocardiogram every 1–2 years

Years 5–10

Chest radiograph every 6 monthsb
A few relapses reported as late as two decades after treatment
Discuss with patient whether to continue chest radiograph every 6–12 months

Echocardiogram every 2–4 yearsc
Echocardiogram every 2–4 years

Every clinic visit must include detailed history and physical examination. Many institutions will add complete blood counts.
Evaluate any site with unexplained pain or swelling. A chest CT scan is optional, but should always be performed if the chest radiograph shows metastasis or is inconclusive. Add consultation with orthopaedic surgery and physical therapy as indicated. Offer fertility testing for males. Additional investigations may be indicated.

a Need not be repeated if normal at 1 year.
b Some groups recommend continued annual radiographs of the primary site until year 10.
c Longer interval if normal function and post-pubertal at diagnosis.

Late effects

Fortunately, many former osteosarcoma patients will go on to lead relatively normal and productive lives. However, it cannot be expected that intensive therapy for an otherwise deadly disease can remain free of late effects. Late complications may be caused by the tumour itself or by the surgery and chemotherapy needed to control it. Functional and cosmetic consequences for the musculoskeletal system depend on the location and extent of the osteosarcoma as well as the type of surgery employed. Amputations, and even more so rotation-plasties, carry the stigma of mutilation, but may lead to functional results which are similar to those after limbsalvage procedures. Ablative surgery is most often definitive, while revision surgery is frequently needed after limb-salvage procedures where periprosthetic infection, loosening, fractures, and prosthetic wear can occur. The use of expandable prostheses in skeletally immature patients is predictably associated with the need for multiple revision procedures.

Anthracycline cardiomyopathy is a feared complication of osteosarcoma therapy in which most protocols include high cumulative doxorubicin doses. Lifelong cardiac follow-up is required. This is usually accomplished by serial echocardiograms. Renal function may be permanently compromised by cisplatin, where changes usually affect glomerular filtration, or ifosfamide, with proximal tubular damage. If end-of-treatment evaluations do not reveal impaired kidney function, it is unlikely to manifest later. The same holds true for hearing loss due to cisplatin. High-frequency loss is frequent, but only a minority of patients will require hearing aids. Protocols incorporating ifosfamide lead to sterility in most males and in some females. Secondary malignancies will develop in 3 per cent of patients within the first 10 years after treatment. Both the mutagenic side effects of cytostatic treatment and individual predisposition contribute to the development of second cancers.


Osteosarcoma remains a disease that presents challenges to the treating team. It is sensitive to a small number of medications, all of which have significant shortand long-term toxicities. Radiation is of limited effectiveness. Even sensitive tumours are incompletely eliminated by chemotherapy. Therefore surgical resection and reconstruction are necessary, with their attendant morbidity and imposed requirement for intensive rehabilitation. About two-thirds of patients with localized disease at initial diagnosis are cured, whereas only one-third of those with initially metastatic disease achieve cure. Patients whose disease has recurred are difficult to cure, but there is a somewhat better outlook for those with isolated resectable metastases.

New initiatives are required. These include the investigation of newer cytotoxic agents, singly or in combination. An example is the combination of trimetrexate and high-dose methotrexate, designed to circumvent dysfunction of the reduced folate carrier present in many osteosarcoma cells at the time of initial diagnosis. A phase I study to demonstrate tolerability of the combination is being planned at Memorial Sloan–Kettering Cancer Institute (R. Gorlick, personal communication, Spring 2003). Other avenues must be sought as well. An example is the use of trastuzumab in patients who have widely metastatic tumours that express her2 at initial diagnosis, which is currently being studied by the Children's Oncology Group. Another is the use of interferon-α as maintenance therapy after the completion of standard cytotoxic chemotherapy, since it may have both anti-angiogenic and anti-neoplastic activity, as is planned in the upcoming European–North American (EURAMOS) study. Augmentation of the fasfas ligand death pathway through the use of granulocyte-monocyte colony stimulating factor (GM-CSF) as an inhalation in patients with isolated pulmonary metastatic disease will soon be studied by the Children's Oncology Group, and will be followed by a study of the combination of ifosfamide and inhaled GM-CSF. Investigations of the underlying oncogenic defect, possibly a loss of genomic control, are ongoing. In addition, increasing collaborative efforts with veterinary colleagues, to take advantage of the higher rate of osteosarcoma in large dog species, are underway. Initial studies of the role of bisphosphonates in the therapy of osteosarcoma are planned (C. Khanna and J. Hock, personal communication, Spring 2003). Recent reports suggest that in selected cases external beam radiotherapy, which is not reliably effective against inoperable osteosarcoma, may be augmented by targeted internal radiotherapy with high-dose [153Sm]EDTMP plus blood progenitor cell support. It is hoped that exploration of these novel approaches and the increasing cooperation of European and North American investigators will lead to improved outcome for patients with osteosarcoma.


1. Fletcher CDM, Unni K, Mertens K (eds) (2002). WHO Classification of Tumours. Pathology and Genetics of Tumours of Soft Tissue and Bone. Lyon: IARC Press.

2. Fuchs B, Pritchard DJ (2002). Tumor biology: etiology of osteosarcoma. Clin Orthop 397, 40–52.

3. Kager L, Zoubek A, Pötschger U, et al. (2003). Primary metastatic osteosarcoma: presentation and outcome of 202 patients treated on neoadjuvant Cooperative Osteosarcoma Study Group (COSS) protocols. J Clin Oncol 21, 2011–18.

4. Sanders TG, Parsons TW III (2001). Radiographic imaging of musculoskeletal neoplasia. Cancer Control 8, 221–31.

5. Enneking WE, Spanier SS, Goodmann MA (1980). A system for the surgical staging of musculoskeletal tumors. Clin Orthop 153, 106–20.

6. Sobin LH, Wittekind C (2002) UICC TNM Classification of Malignant Tumors (6th edn) New York: Wiley.

7. Link MP, Goorin AM, Miser AW, et al. (1986). The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremities. N Engl J Med 314, 1600–6.

8. Smith MA, Ungerleider RS, Horowitz ME, et al. (1991). Influence of doxorubicin dose intensity on response and outcome for patients with osteogenic sarcoma and Ewing's sarcoma. J Natl Cancer Inst83, 1460–70.

9. Bacci G, Picci P, Ruggeri P, et al. (1990). Primary chemotherapy and delayed surgery (neoadjuvant chemotherapy) for osteosarcoma of the extremities. The Istituto Rizzoli experience in 127 patients treated preoperatively with intravenous methotrexate (high vs. moderate doses) and intraarterial cisplatin. Cancer 65, 2539–53.

10. Bramwell VH, Burgers M, Sneath R, et al. (1992). A comparison of two short intensive adjuvant chemotherapy regimens in operable osteosarcoma of limbs in children and young adults: the first study of the European Osteosarcoma Intergroup. J Clin Oncol 10, 1579–91.

11. Winkler K, Bielack S, Delling G, et al. (1990). Effect of intraarterial versus intravenous cisplatin in addition to systemic doxorubicin, high dose methotrexate, and ifosfamide on histologic tumor response in osteosarcoma (Study COSS-86). Cancer 66, 1703–10.

12. Meyers PA, Schwartz CL, Bernstein M, et al. (2001). Addition of ifosfamide and muramyl tripeptide to cisplatin, doxorubicin and high-dose methotrexate improves event-free survival (EFS) in localized osteosarcoma (OS). Proc Am Soc Clin Oncol 20,1463 (abstract).

13. Goorin AM, Schwartzentruber DJ, Devidas M, et al. (2003). Presurgical chemotherapy compared with immediate surgery and adjuvant chemotherapy for nonmetastatic osteosarcoma: Pediatric Oncology Group Study POG-8651. J Clin Oncol 21, 1574–80.

14. Bielack S, Kempf-Bielack B, Delling G, et al. (2002). Prognostic factors in high-grade osteosarcoma of the extremities or trunk. An analysis of 1702 patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol 20, 776–90.

15. Bacci G, Bertoni F, Longhi A, et al. (2003). Neoadjuvant chemotherapy for high-grade central osteosarcoma of the extremity. Histologic response to preoperative chemotherapy correlates with histologic subtype of the tumor. Cancer 97, 3068–75.

16. Rosen G, Caparros B, Huvos AG, et al. (1982). Preoperative chemotherapy for osteogenic sarcoma: selection of postoperative adjuvant chemotherapy based on the response of the primary tumor to preoperative chemotherapy. Cancer 49, 1221–39.

17. Winkler K, Beron G, Delling G, et al. (1988). Neoadjuvant chemotherapy of osteosarcoma: results of a randomized cooperative trial (COSS-82) with salvage chemotherapy based on histological tumor response. J Clin Oncol 6, 329–37.

18. Meyers PA, Heller G, Healey J, et al. (1992). Chemotherapy for nonmetastatic osteogenic sarcoma: the Memorial Sloan–Kettering experience. J Clin Oncol 10, 5–15.

19. Bacci G, Ferrari S, Bertoni F, et al. (2000). Long-term outcome for patients with nonmetastatic osteosarcoma of the extremity treated at the Istituto Ortopedico Rizzoli according to the Istituto Ortopedico Rizzoli/Osteosarcoma-2 protocol: an updated report. J Clin Oncol 18, 4016–27.

20. Picci P, Sangiorgi L, Rongraff BT, et al. (1994). The relationship of chemotherapyinduced necrosis and surgical margins to local recurrence in osteosarcoma. J Clin Oncol 12, 2699–705.

21. Ferrari S, Briccoli A, Mercuri M, et al. (2003). Postrelapse survival in osteosarcoma of the extremities: prognostic factors for long-term survival. J Clin Oncol 21, 710–15.

If you find an error or have any questions, please email us at Thank you!