Fax: (713) 792-0608
Single and multiple metachronous osteosarcoma tumors after therapy
Article first published online: 3 NOV 2003
Copyright © 2003 American Cancer Society
Volume 98, Issue 11, pages 2457–2466, 1 December 2003
How to Cite
Jaffe, N., Pearson, P., Yasko, A. W., Lin, P., Herzog, C. and Raymond, K. (2003), Single and multiple metachronous osteosarcoma tumors after therapy. Cancer, 98: 2457–2466. doi: 10.1002/cncr.11800
- Issue published online: 17 NOV 2003
- Article first published online: 3 NOV 2003
- Manuscript Accepted: 26 AUG 2003
- Manuscript Revised: 15 AUG 2003
- Manuscript Received: 11 JUN 2003
- metachronous tumors after therapy;
- skeletal metastases;
The objective of the current study was to determine the incidence, clinical and pathologic characteristics, and outcome of patients with conventional osteosarcoma who developed metachronous tumors after treatment for the primary tumor and prevention of pulmonary metastases.
The medical records of 270 pediatric patients (younger than age 18 years) were reviewed. The prevention and absence of pulmonary metastases was confirmed by chest radiographs and computerized scans of the lungs. Radionuclide bone scans were used to confirm the absence of skeletal metastases.
Eleven patients with metachronous tumors were identified. Index primary tumors involved the femur (n = 8), the tibia (n = 2), and the radius (n = 1). Single metachronous tumors developed in the femur (n = 6), in the humerus (n = 1), and multifocal in multiple bones (n = 4). Two patients later developed second metachronous tumors. The interval between identification of the primary tumor to development of the single metachronous tumors varied from 11 months to 78 months and from 12 months to 42 months for synchronous multifocal tumors. Metachronous tumors were treated with single-agent cisplatin or ifosfamide. Only 1 patient experienced > 90% tumor necrosis. Pulmonary metastases were not detected in 10 of 11 patients at the time metachronous tumors were discovered. In the 11th patient, synchronous pulmonary metastasis with the metachronous tumor was noted. Three patients had a prior history of bilateral retinoblastoma. The Li–Fraumeni syndrome may have been present in another patient. Six patients died. Five patients have survived for 20+ to 50+ months after the appearance, treatment, and resection of metachronous tumors.
With improvement in the cure rate, metachronous osteosarcoma should be recognized as an important sequela in long-term survivors. The etiology of this disease is unknown. Speculation rests on a skeletal multicentric origin, which includes an inherited predisposition to develop osteosarcoma in retinoblastoma and in the Li–Fraumeni syndrome. Meticulous follow-up is required to permit early detection and successful therapeutic intervention. Cancer 2003. © 2003 American Cancer Society.
The biologic behavior of osteosarcoma is consistent with the premise that 80% of patients harbor micrometastases in the lungs at the time of diagnosis.1 These metastases are undetected in imaging studies. Untreated, they surface 6–9 months later and usually are responsible for the patient's demise. In a small percentage of patients, metastases appear in the lungs and also in other parts of the body.2 These include skeletal metastases. The synchronous detection of such skeletal lesions with the primary tumor is designated as multifocal sclerosing osteosarcoma.3 In contrast, lesions may appear later, after treatment of the primary tumor. These lesions are designated as metachronous tumors (Greek: meta, “occurring later”; chronos, “time”). In the current article, metachronous refers to the delayed appearance of a single tumor or multiple tumors in the skeleton after successful treatment of the primary tumor in the absence (prevention) of pulmonary metastases. By inference, the skeletal lesions, in the absence of the pulmonary metastases, did not metastasize from the primary tumor or the lungs but developed later and spontaneously in other parts of the skeleton.
MATERIALS AND METHODS
During the past 25 years, 270 patients younger than age 18 years were treated in the Pediatric Department of the University of Texas M. D. Anderson Cancer Center (Houston, Texas) with cisplatin-based therapy (Treatment and Investigation of Osteosarcoma Studies [TIOS] I–IV and the Intergroup Osteosarcoma protocol, Children's Cancer Group 9721).4 The chemotherapeutic agents comprised various combinations of cisplatin, doxorubicin, high-dose methotrexate, cyclophosphamide, and ifosfamide. All patients, including 31 patients who were treated on the TIOS II study who initially were treated with chemotherapy and were not designated to undergo surgical resection of the primary tumor, were included in the analysis. (None of those 31 patients developed metachronous osteosarcoma). The following studies, initiated at diagnosis and during routine surveillance, were used to determine the prevention and continued absence of detectable pulmonary and skeletal metastases: 1) monthly chest radiographs for 2 years—the interval between the radiographic studies then was extended progressively, and ultimately, studies were obtained annually by the fifth year; 2) computer scans of the lungs at 3–6-month intervals for 3.5 years; and 3) radionuclide bone scans at 6-month intervals for 3.5 years.
Eleven of 270 patients were identified with single or multiple posttherapy metachronous osteosarcomas (Table 1). The primary lesions in the 11 patients were located in the following sites: femur (n = 8), tibia (n = 2), and radius (n = 1). The metachronous tumors appeared as single lesions in 7 patients and were located in the femur (n = 6) or the humerus (n = 1); synchronous multifocal lesions were observed in 4 patients (Patients 1, 2, 4, and 8). The interval between the diagnoses of the primary tumor and the single metachronous tumors varied from 11 to 78 months and from 12 to 42 months for the 4 patients with synchronous multifocal tumors. Selected examples of primary and metachronous tumors in individual patients are depicted in Figures 1–5.
|Patient no.||Age at diagnosis of primary tumor (yrs)||Tumor necrosis from chemotherapy (%)||Interval between primary and metachronous tumors (mos)||Pulmonary metastases (interval from detection of metachronous tumor) (mos)||Chemotherapy for primary tumor||Survival from detection of metachronous tumor (mos)||Comments|
|Primary tumor site||Metachronous tumor site|
|Site/histology||Necrosis (%)||Site/histology||Necrosis (%)|
|1||9.0||Distal radius (L), osteoblastic||95||Multifocal, synchronous (osteoblastic by imaging)||—||22||5||Cisplatin, doxorubicin, methotrexate||10||Death due to disseminated tumor|
|2||11.5||Shaft femur (L), chrondoblastic||70||Distal femur (R), telangiectatic||70||56||—||Cisplatin, doxorubicin, methotrexate||2||Death due to doxorubicin-related cardiac toxicity|
|3||15.0||Distal femur (R), osteoblastic||95||Multifocal, occipital bone, proximal femur (L), osteoblastic||—||12||—||Cisplatin, doxorubicin, cyclophosphamide||4||Cranial nerve palsies; death due to cachexia; inability to swallow|
|4||12.0||Distal femur (L), osteoblastic||86||Multifocal, proximal tibia (R) and proximal humerus (R), chondroblastic/ osteoblastic||83||42||12||Cisplatin, doxorubicin, ifosfamide||16||Retinoblastoma; prior treatment with vincristine, doxorubicin, and cyclophosphamide; death due to disseminated tumor|
|5||12.0||Distal tibia (R), osteoblastic||85||Distal femur (L), osteoblastic||99||78||—||Cisplatin, doxorubicin, methotrexate||20+||Metachronous tumor treated with cisplatin, ifosfamide, and methotrexate|
|6||16.25||Distal femur (R), osteoblastic||99||Proximal humerus (R), osteoblastic||19||42||Synchronous with metachronous tumor||Cisplatin, doxorubicin, methotrexate||10||Renal failure after ifosfamide treatment for metachronous tumor; death due to disseminated tumor|
|7||10.0||Proximal tibia (L), periosteal||—a||Distal femur (R), osteoblastic||80||36||4||Cisplatin, doxorubicin, methotrexate||12||Pulmonary and intracranial metastases; death due to disseminated tumor (? Li–Fraumeni syndrome)|
|8||12.0||Distal femur (L), osteoblastic||89||1) Multifocal, calcaneus (L) and distal tibia (L); 2) Proximal tibia (L), osteoblastic/telangiectatic||—||1) 18; 2) 24||—||Cisplatin, doxorubicin, ifosfamide, methotrexate||28+ (16+)b||Retinoblastoma; prior treatment with cyclophosphamide, second metachronous tumor (proximal tibia); first metachronous tumor treated with ifosfamide and methotrexate, second treated with ifosfamide|
|9||10.0||Distal femur (L), chondroblastic||98||1) Distal femur (R), chondroblastic; 2) Rib||20||1) 11; 2) 17||—||Cisplatin, doxorubicin, ifosfamide, methotrexate||13+ (8+)b||Second metachronous tumor (rib); first metachronous tumor treated with cisplatin, second treated with ifosfamide|
|10||18.0||Distal femur (R), osteoblastic||—||Proximal femur (R), osteoblastic||—||23||—||Cisplatin, doxorubicin||10+||—|
|11||11.0||Proximal tibia (R), osteoblastic||—||Distal femur (R), osteoblastic||—||35||—||Cisplatin, doxorubicin, methotrexate||50+||Retinoblastoma|
Two patients (Patients 8 and 9) developed second metachronous tumors after successful sequential treatment of the primary tumor and the first metachronous tumor. The primary tumor in Patient 8 involved the femur. Two metachronous multifocal tumors were discovered simultaneously 18 months later: 1 in the distal tibia and the other in the calcaneus. The second metachronous tumor developed in the proximal tibia 6 months later. The primary and first metachronous tumors in Patient 9 involved the left and right femoral bones, respectively. The first metachronous tumor was discovered 11 months after the appearance of the primary tumor. The second metachronous tumor involved a rib and developed 6 months later.
In all but 1 patient (Patient 6), the lungs were free of pulmonary metastases at the time the metachronous tumors were discovered. Patient 6 presented with synchronous involvement of the lung and a metachronous tumor in the proximal humerus. Before the discovery of the metachronous tumor, the lungs and skeleton had been continually free of disease, and the lesion in the lung was attributed to the metachronous tumor.
The metachronous tumors, except for the one in the calcaneus (Patient 8; see above), presented with characteristics similar to those of the primary tumors: pain, swelling, discomfort, and difficulty with movement. The metachronous tumor in the calcaneus was silent. It was discovered on radionuclide scanning concurrently when the tumor in the distal tibia was identified.
Histologic characteristics of the primary tumors were as follows: osteoblastic (n = 8), chondroblastic (n = 2), and periosteal (n = 1). Ten of 11 metachronous tumors were biopsied for confirmation of the diagnosis. Patient 1, who developed multifocal osteoblastic lesions, did not undergo a biopsy. The histology of the 10 metachronous tumors was as follows: telangiectatic (n = 1), chondroblastic (n = 1), chondroblastic/osteoblastic (n = 1), osteoblastic/telangiectatic (n = 1), and osteoblastic (n = 6). Thus, the histology in 10 metachronous tumors was concordant with 7 of the primary tumors. In the nonconcordant series, the histology was as follows: The primary tumor in Patient 2 was chondroblastic, and the metachronous tumor was telangiectatic. The primary tumor in Patient 4 was osteoblastic, and the metachronous tumor was chondroblastic/osteoblastic. The primary tumor in Patient 8 was osteoblastic, and the first and second metachronous tumors were osteoblastic/telangiectatic.
Three patients had a prior history of bilateral retinoblastoma. The father of one other patient, after discovery of the metachronous tumor, developed a leiomyosarcoma of the arm, an aunt had a sarcoma of the uterus, and an uncle had a brain tumor. This suggested the presence of Li–Fraumeni syndrome.5 There was no other history of sarcoma or malignancy in the family members of the other patients.
The primary tumor in each patient, except in Patient 7 (who had a periosteal osteosarcoma), was treated with preoperative chemotherapy. Chemotherapy consisted primarily of cisplatin and doxorubicin. Additional agents included ifosfamide, cyclophosphamide, and high-dose methotrexate. All primary tumors were resected, and histologic responses in chemotherapy-treated patients varied from 80% to 99% (Table 1). Metachronous tumors were treated initially with single-agent cisplatin or ifosfamide. These single agents, at optimal dose intensity, were selected because the patients all had been exposed to the maximum cumulative dose of doxorubicin, and high-dose methotrexate was not considered as effective as cisplatin and ifosfamide. Only 1 metachronous tumor experienced > 90% tumor destruction.
Six patients died, due to the following causes: pulmonary metastases (n = 4), congestive cardiac failure (n = 1), and cachexia (n = 1). Five patients survived and were free of disease for 20+ to 50+ months after the appearance of the first metachronous tumors. This includes the two patients who had second metachronous tumors that also were treated successfully. Patient 9 also developed pulmonary metastases after the appearance of the second metachronous tumor; these tumors were resected.
The lungs are the most common site for the development of metastases in patients with osteosarcoma.1, 2 Skeletal metastases, although less common, also were reported in early studies,3, 6–10 generally in association with pulmonary metastases. Jeffree et al. reported a 32% clinical incidence of bone involvement within 30 months of initial diagnosis and an overall incidence of 48%.2 Lockshin and Higgins similarly reported a 41% incidence of bone metastases among 22 patients who were dying in hospital.10 The metastases were detected radiologically or at autopsy.
Compared with the detection of pulmonary and skeletal metastases (concurrently or later), the detection of single or multiple skeletal lesions without pulmonary involvement subsequent to the initial presentation of the primary tumor has been relatively uncommon. It has not been determined whether these skeletal lesions represent metachronous tumors or metastases from the primary tumor (with or without overt lung involvement). The skeletal sites generally were similar to the sites affected by primary osteosarcoma.
The subsequent discovery of a single lesion or multiple lesions, after adequate control of the primary tumor and without intervening pulmonary involvement in our series, is designated as metachronous osteosarcoma, as indicated above. This definition and concept also should be distinguished from lesions that may develop later in 15–20% of patients in nonosseous sites, e.g., brain, heart, kidney, skin, and subcutaneous tissue. In our experience, the majority of these lesions (although rare) also occur in the presence of pulmonary metastases.
Metachronous osteosarcoma may occur as a single lesion11, 12 or as multiple lesions.13–29 The clinical and radiographic features generally resemble those of primary ‘classic’ osteosarcoma, although Howat et al. reported a case of the multifocal metachronous periosteal variety.16 The interval between discovery of the primary tumor and the metachronous (first) tumors in our series varied from 11 months to 7.33 years. In a report on a large series by Fitzgerald et al., the interval between discovery of the primary and metachronous lesions varied from 9 months to 14 years.26
The etiology and pathogenesis of metachronous osteosarcoma are unknown. An incidence as great as 10% has been reported.30 Several early reports suggested a multicentric origin. In those reports, most lesions were discovered in the absence of pulmonary metastases.15, 16, 18–29 This finding is in harmony with our definition. In contrast, Morse et al. suggested that in the absence of overt pulmonary disease, skeletal lesions also could constitute metastatic disease.31 In these circumstances, it was postulated that the metastases migrated through the vertebral venous system and bypassed the portal, caval, and pulmonary circulations.32 However, because the blood supply to the periosteum is poor compared with the blood supply to intramedullary bone, Howat et al. considered their case of periosteal osteosarcoma of primary origin.16
Fitzgerald et al. described 12 patients with multiple metachronous osteosarcoma.26 It was not possible to determine whether the metachronous sarcomas represented late metastases or new ‘primary’ tumors. The multiple lesions were considered to be metastatic deposits from a single primary tumor, because there were insufficient criteria to establish a multicenter hypothesis, even in the absence of pulmonary lesions. Four patients in that series developed a third metachronous tumor, and two patients developed a fourth metachronous tumor.
Parham et al., in reviewing clinical and pathologic features of metachronous osteosarcoma, concluded that the absence of pulmonary metastases was not sufficient to prove multiple primary sites of origin.27 Mahoney et al. suggested that the lesions probably represented metastases, because pulmonary involvement often was present.15 However, it was conceded that some cases may represent new primary lesions developing in damaged or dystrophic mesenchymal tissue.
Speculation has been raised that other preexisting conditions, such as Rothmund–Thomson syndrome, Paget disease, and Fanconi anemia, may contribute to the development of metachronous osteosarcoma.11, 33 Three of 11 patients in our series had bilateral retinoblastoma. This disease is known to be associated with the development of osteosarcoma and other malignant tumors, and its occurrence with osteosarcoma has been noted.5, 34–36 In view of the propensity of such patients and those with Li–Fraumeni syndrome (see below) to develop several malignancies, these patients probably should be considered in a different category than patients without any known predisposing factors to develop secondary tumors.
Fitzgerald et al., in discussing etiology, raised the possibility of an abnormality in the immune response of the host.26 It was speculated that tumor-associated neoantigens, which were demonstrated in a variety of experimental and human malignancies, including osteosarcoma, possibly may be contributory. Similarly, Amstutz speculated on a relation between metachronous osteosarcoma and an underlying mesenchymal or humoral disturbance.28
One patient probably had a predisposition to develop osteosarcoma as part of the Li–Fraumeni syndrome.5 Thus, it is possible that inherent genetic factors caused the primary osteosarcoma and, with improved survival, could have been operative in the later development of tumors in the other bones. This is consistent with the premise that cancer may be a ‘two-hit’ phenomenon.37 The role of chemotherapy and environmental factors in producing the ‘second hit’ in these circumstances remains to be determined. Furthermore, with increased survival, second malignant neoplasms other than osteosarcoma also have been reported.38, 39 None was metachronous osteosarcoma.38
The histology in 8 of 11 metachronous tumors was concordant with the primary tumors. However, from a histopathologic standpoint, there was no mechanism to determine whether the new tumors were metastatic. It is noteworthy that the metachronous tumors in both patients had identical histologic characteristics. They appeared in the same bone (Patient 10) and extremity (Patient 11), and the possibility that they may have represented intraosseous and transarticular skip lesions, respectively, cannot be excluded. However, before the appearance of the metachronous tumors, there was no evidence of their presence on bone scintigraphy.
We were able to find two classifications for metachronous osteosarcoma. Amstutz reported on two patients and divided the lesions into three broad categories.28 Type I lesions were defined as synchronous osteosarcomas in childhood and adolescence with symmetric or asymmetric, simultaneously (or nearly simultaneously) appearing lesions; Type II lesions were defined as synchronous osteosarcomas in adults with low-grade, multiple bone lesions (at first examination) confined to the axial and proximal appendicular skeleton and no pulmonary metastases; Type III A tumors were defined as early metachronous osteosarcomas occurring > 5 months and up to 24 months after diagnosis; and Type IIIB tumors were defined as late metachronous lesions appearing after 24 months.
Lowbeer described two main types of metachronous osteosarcomas7: Type I lesions were defined as unicentric osteosarcomas with bone metastases; Type IIA tumors were defined as multiple synchronous osteosarcomas in young patients with the early appearance of metachronous, metaphyseal, symmetric, and usually sclerotic bone lesions followed by rapid demise; and Type IIB tumors were defined as osteosarcomas occurring in older patients with the appearance of metachronous lesions up to 8 years after discovery of the initial lesion.
In one of the largest reported studies of metachronous osteosarcoma, survival varied from 5 months to 11 years.26 In another series, survival was found to be correlated with time to development of the metachronous tumor39: Patients who developed metachronous tumors ≥ 24 months after the initial diagnosis had a 5-year postmetachronous survival rate of 61% after combined modality therapy. In contrast, the 5-year postmetachronous survival rate was 8% for patients who developed metachronous tumors ≤ 24 months from diagnosis of the primary osteosarcoma. Five of 11 patients in the current series achieved survival rates varying from 18+ to 50+ months after detection of the first metachronous tumor. This result was observed despite the finding that the response of the metachronous tumors to chemotherapy was not as encouraging as the response observed in the primary tumor. The current study indicates that lifelong follow-up of surviving patients with osteosarcoma is warranted and that metachronous osteosarcoma, upon discovery, should be treated with curative intent.
- 6The spread of tumours in the human body (2nd edition). London: Butterworth, 1952..
- 7Multifocal osteosarcomatosis: a rare entity. Bull Pathol. 1968; 9: 52–53..
- 9Sarcomata of the osteogenic series (osteosarcoma, fibrosarcoma, chondrosarcoma, paraosteal ostegenic sarcoma and sarcomata arising in abnormal bone). An analysis of 552 cases of bone. J Bone Joint Surg. 1966; 48A: 1–26., , , .
- 12Metachronous osteosarcoma. A case report. J Korean Radiol Soc. 1990; 20: 940–943., , .
- 13Multiple metachronous osteosarcoma with eleven-year survival. A case report. J Bone Joint Surg Am. 1981; 83: 317–322., , .
- 14Multifocal osteogenic sarcoma. Report of a case. J Bone Joint Surg. 1957; 39B: 524–533., .
- 18Multiple osteosarcoma. Arch Pathol. 1936; 21: 88–95..
- 19Multiple localized osteosarcoma. Rev Ortop Traumatol. 1945/1946; 15: 85–96., .
- 22Sclerosing osteogenic sarcomatosis. A radiologic entity. Radiology. 1956; 86: 41–45., .
- 26Multiple metachronous osteogenic sarcoma. J Bone Joint Surg Am. 1973; 53: 595–605., , .
- 29Metachronous skeletal osteosarcoma in patients successfully treated with adjuvant or neoadjuvant chemotherapy for non-metastatic osteosarcoma [abstract]. Proc Am Soc Clin Oncol. 2002; 21: 405a., , , et al.
- 30Osteosarcoma. In UnniKK, editor. Dahlin's bone tumors: general aspects and data on 11,087 cases. Philadelphia: Lippincott-Raven, 1996: 143–183..
- 31Sclerosing osteogenic sarcoma. AJR Am J Roentgenol. 1962; 88: 691–965., , .
- 35A new familial cancer syndrome: a specimen of malignant and benign tumors including retinoblastoma, carcinoma of the bladder and other genitourinary tumors, thyroid adenoma and a probable case of multifocal osteosarcoma. J Natl Cancer Inst. 1977; 58: 205–207., .