INTRODUCTION

Metallic orthopaedic hardware and prosthetic joints have been used for decades to repair skeletal defects and replace poorly functioning or severely painful joints. Over 140,000 total hip arthroplasties, 103,000 revisions, and 30,000 partial hip replacements were performed in the United States in 1996 (0.7% of the U.S. population) (1) and metallic hardware is routinely used for fracture fixation as well as other orthopaedic procedures. The biomaterials used in these devices are generally considered to be nontoxic; however, some of the constituents have been shown to be potentially carcinogenic in laboratory animal studies (2, 3, 4, 5, 6, 7) and malignancies have been associated with orthopaedic implants in pets (8). In humans the development of an orthopaedic implant associated malignancy is a rare, but well recognized complication. Thirty-one case reports of implant-related sarcoma (9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37) and two cases of implant-related lymphoma are present in the English language literature (38, 39). The literature describing this phenomenon has stressed the clinical aspects of the cases or the metallic composition of the implants that have the highest association with malignant transformation; however, none has focused on the pathology of orthopaedic implant-related sarcoma. Herein, we present the clinical aspects and review in detail the pathologic findings of 12 implant-associated sarcomas.

MATERIALS AND METHODS

The patients were treated at the authors institutions or their slides were reviewed in consultation by some of the authors. Between 1 and 45 hematoxylin and eosin stained slides were available for review on each case. Electron microscopy was performed on Case 4, and immunohistochemistry was performed on Cases 4 and 6. Follow-up information was obtained from the patient charts or from the referring physicians. Five of the cases were previously reported (20, 21, 23, 34).

RESULTS

(Table 1.) The patient's ages ranged from 18 to 85 (mean 55) years at the time of diagnosis of the malignancy. Three patients were female, and nine were male. All tumors were associated with implants in the femur; eight in the proximal femur and four in the distal femur. Five patients had undergone total hip arthroplasty for osteoarthritis, four had intramedullary nail placement for fracture fixation (three in the proximal femur and one mid-shaft), one had staples placed after osteotomy for a genu valgum deformity, one had staples placed for an unknown reason, and one had hardware placed for fracture fixation followed 33 years later by a total hip arthroplasty on the same femur. The time interval between hardware placement and sarcoma detection was known in 11 cases and ranged from 2.5 to 33 (mean 11) years. The presenting symptoms were pain, stiffness and swelling. Ten sarcomas arose within the bone, and two in the adjacent soft tissue. Radiographically the tumors were located immediately adjacent to or surrounding the implant (Fig. 1) and demonstrated permeative destruction of the adjacent bone. Four sarcomas demonstrated obvious neoplastic matrix production.

TABLE 1 Clinical Data on Patients with Implant-Related Sarcomas
Full size table
FIGURE 1
figure 1

Radiograph demonstrating the blastic component of an osteosarcoma that surrounds the implant in Case 10.

Full size image

In five patients the initial symptoms of the tumors were attributed to the pre-existing orthopaedic problem, which may have delayed the diagnosis (cases 2, 3, 8 to 10). In Case 2, an 18-year-old man who had a long history of Morquio's syndrome and related orthopaedic deformities recently developed pain but did not seek medical assistance for 2 months, at which time a large mass had become palpable. The patient in Case 3 sought medical advice because of pain and a biopsy was interpreted as infection. The biopsy was repeated one month later because of persistent pain and swelling, and the diagnosis of infection was again rendered. Four months later, when the mass was 25 cm in greatest diameter, the patient was rebiopsied and the diagnosis of inflammatory malignant fibrous histiocytoma was given. The patient in Case 8 had pain at the site of an intramedullary nail. An X-ray was interpreted as negative. After the pain persisted, the nail was removed, but the pain did not subside. A total hip replacement was performed several weeks later because the pain was thought to be secondary to osteoarthritis. Sarcoma was noted at the time of surgery. The young man in Case 9 experienced pain in the leg, which contained an intramedullary nail. An X-ray disclosed calcifications at the nail tip, but was thought to be otherwise normal. A lumbar disk was implicated as the cause of continued pain, and the patient underwent laminectomy. Because of persistent pain, the nail was finally removed at a later date, which led to perfuse bleeding because of an underlying telangiectatic osteosarcoma. In Case 10 the patient waited 3 months before seeking medical attention, at which time a large mass was noted on X-ray. After 4 weeks of pain, the patient in Case 4 sought medical attention at which time a large soft tissue mass was discovered. The remaining three patients in which clinical information is known sought medical assistance soon after the onset of pain. Radiographic studies in Cases 1, 7, and 11 obtained at that time of presentation demonstrated destructive bone lesions (Fig. 3).

FIGURE 3
figure 2

A destructive, lytic malignant fibrous histiocytoma surrounds the intramedullary nail in Case 7.

Full size image

Pathology

(Table 2.) All sarcomas arose in close proximity to the implant (Fig. 3), 10 in bone and 2 in soft tissue, and all were high grade. Seven tumors were osteosarcoma, four malignant fibrous histiocytoma (MFH) and one a malignant peripheral nerve sheath tumor. The size of the tumors ranged from 5 to 25 (mean 11) cm in greatest dimension and they varied from tan-grey to tan-yellow, and usually had areas of hemorrhage and necrosis (Figs. 4 and 5). Seven tumors were centered within bone and demonstrated bone formation by malignant cells. Histopathologically, five of these were osteoblastic osteosarcomas (Fig. 6), displaying plump atypical cells with a high nuclear to cytoplasm ratio and abundant mitotic activity. Prominent in one of these tumors were areas where the malignant cells were spindled and formed fascicles arranged in a storiform pattern (Fig. 7) (mixed osteoblastic and malignant fibrous histiocytoma-like osteosarcoma). Another osteosarcoma displayed large areas of high-grade malignant cartilage admixed with neoplastic bone (Fig. 8) (mixed osteoblastic and chondroblastic osteosarcoma). Lastly, one osteosarcoma was extensively cystic, with multiple blood-filled cavities (Fig. 9). The cyst walls contained malignant cells with hyperchromatic, enlarged nuclei, which produced osteoid (telangiectatic osteosarcoma).

TABLE 2 Pathologic Findings in Patients with Implant-Related Sarcoma
Full size table
FIGURE 4
figure 3

A pathologic fracture is present through the tan and red, focally cystic osteosarcoma in Case 2.

Full size image
FIGURE 5
figure 4

Areas of hemorrhage are present within the tan, fleshy malignant fibrous histiocytoma, which surrounds the femur in Case 3.

Full size image
FIGURE 6
figure 5

Plump tumor cells demonstrate osteoid production in Case 9.

Full size image
FIGURE 7
figure 6

Osteosarcoma, which has a focally fascicular growth pattern (Case 1).

Full size image
FIGURE 8
figure 7

The osteosarcoma in Case 5 has areas of neoplastic cartilage formation.

Full size image
FIGURE 9
figure 8

Cystic, spaces are characteristic of the telangiectatic osteosarcoma in Case 2.

Full size image

Of the four cases of MFH, three were the storiform-pleomorphic type and were located within bone. They were composed of severely atypical and bizarre spindle cells that were arranged in a storiform pattern (Fig. 10). Mitoses were numerous. The remaining MFH (Case 3) originated within the soft tissue adjacent to a hip prosthesis and was composed of atypical epithelioid and spindle shaped cells that were infiltrated by numerous inflammatory cells including eosinophils, neutrophils and lymphocytes and was considered to be an inflammatory MFH (Fig. 11). Immunohistochemistry performed on formalin-fixed paraffin-embedded tissue showed that the tumor cells in Case 6 expressed vimentin, actin, smooth muscle actin, alpha-1-antichymotrypsin, and alpha-1-antitrypsin. They did not express desmin or S-100 protein. This staining pattern supports the histologic impression of MFH with myofibroblastic features. Immunohistochemistry was not performed on the other cases of MFH.

FIGURE 10
figure 9

Whorling arrangement of spindle cells in a malignant fibrous histiocytoma (Case 12).

Full size image
FIGURE 11
figure 10

Acute and chronic inflammatory cells virtually obscure the malignant component of the inflammatory malignant fibrous histiocytoma (Case 3).

Full size image

The malignant peripheral nerve sheath tumor (MPNST) was composed of large, spindle and focally epithelioid cells that had eosinophilic cytoplasm (Fig. 12) and were arranged in small swirls around vessels or other tumor cells. Multinucleate tumor cells were scattered throughout the mass. Mitotic activity was high, with many atypical forms and there was focal necrosis. Immunohistochemistry performed on formalin-fixed, paraffin-embedded tissue showed that the tumor cells expressed vimentin, alpha-1-antitrypsin, alpha-1-antichymotrypsin, and were focally positive for keratin. The tumor cells did not stain for desmin, smooth muscle actin, leu-7, S-100, or KP-1. Ultrastructural analysis revealed long, intertwining cell processes, intracellular junctions, and basal lamina, diagnostic of Schwann cell differentiation.

FIGURE 12
figure 11

This malignant peripheral nerve sheath tumor is composed of epithelioid cells with a high mitotic rate (Case 4).

Full size image

Treatment and Follow-Up

Three patients had pulmonary metastases at the time of diagnosis. The surgical treatment is known for 11 patients. Nine patients underwent wide resection of their tumor, and usually received adjuvant chemotherapy and/or radiation therapy (see Table 2). One patient had surgical debulking combined with radiation therapy and chemotherapy, and the patient with the telangiectatic osteosarcoma refused surgery and had radiation therapy and embolization of the tumor. Follow-up information is available on eight patients. Five patients died of disease 2 months to 8 years (mean 26 months) after diagnosis. Two patients died without evidence of disease 7 and 30 months after diagnosis. One patient is alive without disease 6 years after diagnosis.

DISCUSSION

The use of orthopaedic hardware and prosthetic joints has greatly advanced the treatment of skeletal diseases. Although the materials composing these implants are relatively biologically inert they may induce a variety of tissue responses (40). By far the most serious complication of orthopaedic hardware and prosthetic material is malignancy. Animal studies analyzing the biological effects of the components of prosthetic joints revealed carcinogenic properties of beryllium, cadmium, chromium, cobalt, iron, lead, nickel, selenium, zinc, and titanium (2, 3, 4, 7, 41). In humans, prostheses containing cobalt, chromium, stainless steel, nickel, iron, manganese and silicon have been implicated as being carcinogenic. Other authors state that the placement of an implant may induce osteonecrosis, which in itself is a risk factor for the development of a sarcoma (37, 42). To date only 31 cases of implant-related sarcoma have been reported in the literature since the 1950s (Table 3) (9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37). This number is minuscule when compared with the 270,000 hip surgeries in addition to the hundreds of thousands of other hardware-related procedures performed in the U.S each year (1). Furthermore, a Finnish study of 31,651 patients who received polyethylene and metal hip prosthesis reported no associated sarcomas (43). However, the fact that malignancies have arisen in close proximity to metallic implants is strong evidence that there may be an important relationship. Chromosomal aberrations have been reported in the marrow adjacent to hip arthroplasties (44). In other anatomic sites, such as the lung, metal foreign bodies have also been associated with the development of malignancy (45).

TABLE 3 Prior Reports of Implant-Related Sarcoma
Full size table

An animal model for the development of sarcomas in rats that have been implanted with various biomaterials has shown that up to 67% of implant sites developed either a sarcoma or a reaction that they termed “proliferative, probably preneoplastic” (5). There was no correlation between the morphology of the sarcoma and the type of biomaterial that was implanted. The materials most often associated with these histologic changes were polymethylmethacrylate (67% of sites), polyethylene (59% of sites), silicone (56% of sites), and aliphatic polyurethane (52% of sites). Pure titanium, nickel chromium, cobalt-chromium alloy and aluminum oxide were less often associated with proliferative and malignant changes. Preneoplastic mesenchymal lesions have not been well defined or classified and its exact meaning in this study is unclear.

The majority of our cases of implant-related sarcomas were osteosarcoma, followed by MFH, and a single case of MPNST. Combined with the cases reported in the literature previously, 15 are MFH, 12 osteosarcoma, 2 angiosarcoma, 2 Ewing's sarcoma, 1 each of fibrosarcoma, MPNST, epithelioid sarcoma, synovial sarcoma, epithelioid hemangioendothelioma, chondrosarcoma in a patient with Maffucci's syndrome, and an unclassified sarcoma. The age range of all patients who developed implant-related sarcomas is 11 to 87 (mean 50) years, which is similar to the age range in our series. The time interval from the insertion of the implant to the development of malignancy has ranged from 6 months to 30 years (mean 9) years, which is also similar to our results. Thirty-one sarcomas occurred in the femur, three in the tibia, two in the acetabulum, one in the humerus, and one involved both the femur and acetabulum. The propensity of the sarcomas to arise in the femur is probably related to the fact that it is the most common bone to be treated with an orthopaedic implant.

CONCLUSION

Large, high-grade sarcomas arising in bone or soft tissue have a poor prognosis. When coupled with a delay in the diagnosis of malignancy, prosthetic-related sarcomas have a high mortality. Follow-up information available on 14 of the 26 implant-related sarcomas reported in the literature demonstrates this point; one patient was free of disease 15 months after surgical therapy, 5 patients were dead 1 year after diagnosis, and 8 patients were dead less than 1 year after diagnosis. At least 8 of the 13 patients that died had metastatic disease at the time of death. The clinical outcome of our cases is similar. Of the eight patients with follow-up information, seven died of disease a mean of 26 months after diagnosis, and only one patient is alive and disease free. Clinically it may be difficult to detect an implant associated malignancy as implant-related symptoms caused by other processes are more common and radiographic imaging studies are compromised by the artifact induced by the metal component of the implant. Although the possibility of a sarcoma developing in association with an implant is extremely small, physicians should at least consider this possibility when pain or other new findings develop in relation to metallic orthopaedic hardware.

FIGURE 2
figure 12

A large, tan osteosarcoma is adjacent to the implant in Case 10.

Full size image