81

Soft Tissue Sarcomas (Excluding Retroperitoneum)

Sarcomas are rare malignancies that arise from the connective tissues in any organ or at any anatomic location of the body. This chapter addresses sarcomas that arise in the extraskeletal, nonvisceral connective tissues of adults, excluding the retroperitoneum. Despite the diversity of tissues and locations of origin, these soft tissue sarcomas are grouped together because of overall similarities in natural history and treatment. Retroperitoneal sarcomas, pediatric sarcomas, osteosarcomas, sarcomas arising in visceral organs, Kaposi's sarcoma, and sarcomas arising in the vasculature are discussed elsewhere in this textbook.

Anatomy

The majority of soft tissue sarcomas occur in the muscle groups of the extremities (Table 81.1). These tumors often remain confined to the muscle compartment of origin. The thigh is the most common subsite of origin and is partitioned into three compartments (37). The muscle compartments of the arm, forearm, and leg are similarly defined (4) (Fig. 81.1). Anatomic knowledge is essential for the radiation oncologist because it allows appropriate positioning of the limb to encompass the portion of the compartment at risk, while avoiding compartments that are not involved.

Epidemiology, Genetics, and Risk Factors

Approximately 9,400 cases of soft tissue sarcoma are diagnosed yearly in the United States, accounting for 0.7% of cancers and an estimated 3,500 deaths (68). Men are more frequently affected than women, and rates are higher among African Americans than whites. Most sarcomas arise in a sporadic fashion, without identifiable etiology. Sarcomas do not appear to develop from pre-existing benign lesions. Associated factors can be identified in certain subsets of sarcomas, including predisposing genetic mutations, previous ionizing radiation or chemical exposures, and chronic soft tissue injury or lymphe-dema.

The Li-Fraumeni syndrome is an autosomal dominant familial cancer predisposition syndrome in which the risk of breast and other invasive cancers, including sarcomas, by age 35 years is almost 50% (88). A germline mutation in the p53 tumor suppressor gene is identifiable in most of the affected families (91). The p53 gene is central in modulating a cell's response to DNA damage by arresting the cell cycle and inducing apoptosis (81). Somatic mutations of p53 are among the most common genetic alterations seen in mesenchymal tumors, occurring in nearly 60% of sarcomas (26). The activity of p53 can also be disrupted by amplification of the MDM2 gene, located at chromosome 12q13-q14 and coding for a nuclear phosphoprotein that inactivates wild type p53. MDM2 amplification has been demonstrated in 10% to 30% of sarcomas (43).

Patients with hereditary retinoblastoma inherit a germline mutation in the RB gene, and a “second hit” in the remaining allele results in malignancy. The RB protein regulates the cell cycle, governing the entrance into the DNA synthesis (S) phase of the cell cycle. In addition to malignant retinoblastomas of the eye, these patients are at increased risk of developing osteosarcomas and soft tissue sarcomas later in life, particularly after exposure to therapeutic radiation (145). Genetic disruption of the RB pathway is observed in over 50% of sarcomas (22).

Patients with neurofibromatosis type 1 (NF1, von Recklinghausen's neurofibromatosis) develop multiple neurofibromas and are at increased risk for gliomas and malignant peripheral nerve sheath tumors (MPNSTs) (155). As in heritable retinoblastoma, a germline mutation in NF1, followed by somatic mutation of the remaining allele, results in malignant degeneration (59).

Ionizing radiation exposure produces a small but detectable risk of both bone and soft tissue sarcoma. Radiation-induced sarcomas were first reported in the 1920s among workers painting radium watch dials (48). Sarcomas arising after therapeutic irradiation, reported since the 1930s (16), develop after a latency period (between 2 and 25 years) within the radiation portal and are histologically distinct from the primary malignancy (17). In one review, 3.3% of 1,089 sarcoma patients met these criteria (94). The median latency period was 14 years, and risk was increased after high radiation doses. In a large Finnish cohort study, the absolute risk of postirradiation sarcoma with long-term follow-up was 0.03% (140). The most commonly observed radiation-induced sarcomas arise after radiation therapy for breast cancer. These aggressive malignancies often involve a large portion of the breast (Fig. 81.2). Among 194,798 women diagnosed with localized or regional invasive breast cancer between 1973 and 1995, the relative risks of angiosarcoma and other sarcoma subtypes were 15.9 and 2.2 for irradiated patients compared with unirradiated patients (66). Despite high relative risks, the absolute risk of developing radiation-induced sarcomas is small, 0.28% and 0.48% at 15 years after in two large series (78,116). Angiosarcoma is also observed in patients with chronic lymphedema (Stewart-Treves syndrome), which may or may not be associated with radiation therapy (73).

Epidemiologic studies of industrial chemical exposures and sarcoma risk are limited by the small numbers of individuals exposed to a variety of different agents. Studies have suggested links between vinyl chloride and hepatic angiosarcoma (36), as well as phenoxy herbicides, particularly those contaminated with chlorinated dioxins, and soft tissue sarcomas (38). Studies of United States veterans exposed to Agent Orange, a dioxin-containing herbicide used extensively during the Vietnam War, have shown no evidence of increased sarcoma risk (133).

Natural History

Extremity soft tissue sarcomas spread directly by local extension along the longitudinal axis of muscular compartments. Fascial planes and bone are rarely violated and constitute barriers
P.1809

to local spread. Grossly, lesions appear encapsulated; however, this is a pseudocapsule, representing compressed normal tissue and reactive fibrosis (37). Subclinical disease can infiltrate adjacent tissues, extending 5 to 10 cm beyond the pseudocapsule, “skipping” areas that appear uninvolved. Biopsy procedures can potentially change the pattern of spread if they violate an uninvolved compartment or if an extensive hematoma results (4). In the trunk or head and neck regions, the disease more commonly invades adjacent structures.

High-grade sarcomas have the potential to metastasize. Because lymph nodes are involved in <10% of sarcoma cases, routine lymph node sampling is usually not performed. Clear cell sarcoma, epithelioid sarcoma, angiosarcoma, rhabdomyosarcoma, and synovial cell sarcoma have higher rates of nodal spread (44), and sampling should be considered for these histologies. Hematogenous metastases occur frequently in patients with high-grade sarcomas; most occur in the lungs (111), with less frequent metastasis to other soft tissue sites, bone, liver or skin (139). The median time to metastasis is approximately 1 year (107). However, metastasis >5 years after initial diagnosis is not uncommon (87).

Clinical Presentation

Soft tissue sarcoma classically presents as a growing, painless mass. Several-month delays in presentation to a physician, establishment of a sarcoma diagnosis, and referral to a sarcoma center are not uncommon (21). Numbness, pain, or edema may be caused by tumor-induced neurovascular compromise. Deep tumors may attain an enormous size before coming to clinical attention. Metastases are noted at the time of diagnosis in <10% of patients (118).

Clinical Evaluation

The history should detail family history and previous radiation exposure. The physical examination must detail the size, location, and depth (superficial or deep) of the mass, as well as its proximity to joints. Evidence of neurovascular compromise and fixation to bone should be sought because the ability to perform a limb-sparing procedure in a patient with these findings is greatly decreased. A careful lymph node examination should always be performed.

Obtaining diagnostic imaging before an attempt to obtain tissue diagnosis may be advantageous because it provides images devoid of biopsy-related changes, and may provide guidance for appropriate biopsy technique. Plain radiographs and ultrasound of the affected area are underused, and often provide valuable information including the presence of a solid versus cystic mass, calcification, or bony invasion. Magnetic resonance imaging (MRI) is increasingly preferred as an imaging modality for soft tissue masses and provides excellent soft tissue detail (Fig. 81.3). Computed tomography (CT) supplements MRI and is particularly helpful in identifying bony invasion or destruction. Although certain soft tissue neoplasms may have characteristic radiographic features, no imaging modality has sufficient specificity to specifically distinguish benign from malignant masses (27). Because the lungs are the predominant site of distant metastasis for soft tissue sarcomas, chest imaging by posterior and lateral chest x-ray or CT is appropriate at initial evaluation and subsequent surveillance for distant metastases, particularly for patients with high-grade disease.

The clinical role of positron emission tomography (PET) in the evaluation of soft tissue sarcomas is an active area of investigation. Uptake of [F-18]-fluorodeoxy-D-glucose (FDG) is somewhat variable in soft tissue neoplasms, but is generally increased in malignant compared with benign, and in high-grade compared with low-grade neoplasms (12). FDG-PET activity may have prognostic significance and may have promise for predicting treatment response (121,123).

A biopsy should be performed on any soft tissue mass that persists or grows, with the exception of subcutaneous lesions that have remained unchanged for years. Ideally, biopsy of a soft tissue mass should be performed by an experienced surgeon. Consideration should be given to biopsy technique and selection of biopsy site, given that all potentially contaminated tissue may need to be removed in a subsequent definitive resection and included in radiation therapy target volumes. Excisional biopsies should be avoided in all but the smallest superficial lesions. Fine-needle aspiration is used by experienced groups for the diagnosis of soft tissue tumors (6), but does not allow for the examination of tissue architecture and is not preferred for diagnosis by most pathologists. Fine-needle aspiration may be most useful to diagnose recurrence or metastasis in patients with an established histologic diagnosis (135). Sufficient tissue for diagnosis is more easily obtained by incisional biopsy or core needle (TruCut) biopsy. The incision for biopsy should be oriented along the longitudinal axis of the extremity such that it can be encompassed in a subsequent resection. Core needle biopsy is minimally invasive as well as easier and cheaper to perform. Although less volume of tissue is obtained by core needle biopsy, it has become standard practice and has been proven to be accurate for diagnosis in the majority of cases (62).

Staging

The American Joint Committee on Cancer staging system (2002 edition) emphasizes grade (G) as the most important prognostic factor for soft tissue sarcoma (Table 81.2) (54). A three- or four-tier system of histologic grading may be used. The prognostic importance of tumor size and depth of invasion is incorporated in the primary tumor (T) stage. The presence of either nodal (N) or distant (M) metastases constitutes stage IV disease. The primary anatomic site of disease is not considered. Comparisons with alternate staging systems developed by investigators at Memorial Sloan-Kettering Cancer Center (MSKCC) and other institutions confirm that tumor depth, grade, and size are the most predictive of systemic relapse (147). However, most systems provide little prognostic information relevant to local recurrence.
P.1810


P.1811


Pathologic Classification

Soft tissue sarcomas are classified according to their presumed tissue of origin, using histologic designations such as liposarcoma (adipose tissue), leiomyosarcoma (smooth muscle), or angiosarcoma (vascular tissue) (41). Tumors without identifiable histogenesis are designated according to morphologic appearance or the presumed “line of differentiation” of the tumor cells. Changes in histologic classification have occurred over time. For example, the category of malignant fibrous histiocytoma (MFH) was established in the 1970s, but subsequently became the most common histologic diagnosis for adult soft tissue sarcomas, coinciding with a reduction in the number of cases classified as pleomorphic rhabdomyosarcoma (Table 81.3). The MFH classification is controversial because neither the tissue of origin nor the line of differentiation is clear. In a review, a specific line of differentiation could be identified in the majority of MFH specimens when reanalyzed histologically, immunohisto-chemically, or ultrastructurally (40), suggesting that the MFH designation is overused.

Pathologic grading is subjective, but the significance of grade as a predictor of metastasis has been demonstrated repeatedly (41). No grading system is uniformly accepted, but most assign tumors to one of three or four categories. The grading system developed by the French Federation Nationale des Centres de Lutte Contre le Cancer assigns a tumor to grade 1 through 3 based on differentiation, mitotic rate, and degree of necrosis (24). The National Cancer Institute system uses histologic type, cellularity, nuclear pleomorphism, frequency of mitoses, and degree of necrosis. There was 34.6% discordance in grading between the two systems in a study of 410 nonmetastatic soft tissue sarcoma cases (58). Although both systems were prognostic, the French Federation Nationale des Centres de Lutte Contre le Cancer system yielded the best correlation with distant metastasis and overall survival. In our current practice, every attempt is made to assign a given tumor into either a high- or low-grade category in order to facilitate clinical decision-making.

Two immunohistochemical stains that are useful for distinguishing sarcomas from more common carcinomas are vimentin (positive in almost all sarcomas, and negative in most carcinomas) and cytokeratin (positive in almost all carcinomas, and negative in most sarcomas). S100 and HMB-45 are positive in melanoma, but may also be positive in specific soft tissue sarcomas. Sarcoma subclassification can be aided by desmin or myoD1 (positive in myogenic tumors), vascular markers (positive in angiosarcomas), and MIC2 (positive in peripheral neuroectodermal tumors). These stains may confirm a diagnosis already considered on morphologic grounds or may raise the possibility of a diagnosis not previously considered.

Metaphase cytogenetics or polymerase chain reaction-based molecular testing may be useful for the identification of chromosomal rearrangements and gene fusions specific to particular subtypes of soft tissue sarcomas (153), such as the t(12;16)(q13;p11) translocation that creates a fusion between adipocyte differentiation gene CHOP and nuclear RNA-binding protein TLS (28), t(X;18)(p11.2;q11.2) that results in a fusion between the SYT gene and either the SSX1 or SSX2 genes in synovial cell sarcoma (20), t(12;22)(q13;q12) (ATF1-EWS) in clear cell sarcoma (42), t(11;22)(p13;q12) (WT1-EWS) in desmoplastic small round cell tumor (50), t(9;22)(q22;q12) (CHN-EWS) in
P.1812


P.1813

extraskeletal myxoid chondrosarcoma (64), t(17;22)(q22;q13) (COL1A1-PDGFB) in dermatofibrosarcoma protuberans (105), or t(X;17)(p11;q25) (ASPL-TFE3) in alveolar soft parts sarcoma (70). Chromosomal and molecular analyses such as gene expression profiling can subclassify sarcomas based on molecular biology, providing insight into tumor biology and identifying potential therapeutic targets (60). Molecular classification may supplement or even supplant traditional histologic classification in the future.

Prognostic Factors

The most important prognostic factor for distant metastasis and survival is grade (24,107,152). For low-grade tumors, the risk of distant metastases at 5 years is <10%, compared with almost 50% for high-grade tumors. Tumor size and depth are also prognostic with respect to distant metastasis. Certain histologic subtypes such as MPNST or leiomyosarcoma may be associated with increased distant metastasis and worse survival (24,107). Nomograms predicting disease-specific survival after resection of localized soft tissue sarcoma have been developed (74,93).

Risk factors for local recurrence are distinct from those for distant metastasis and survival. Multiple prospective and retrospective studies have demonstrated that the presence of tumor cells at the surgical margin and inadequate surgical excision are the most important adverse risk factors for local recurrence (23,56,107,115,119,130,134,143,152). Age >50 years, locally recurrent disease, MPNST or fibrosarcoma histology, the presence of symptoms at presentation, deep location, and withholding of radiation therapy have also been associated with increased local recurrence risk.

No conclusive link has been demonstrated between local control and survival in soft tissue sarcoma. Randomized trials have not detected a survival difference between patient groups with disparate local control (108,112,149). However, Gronchi et al. (56) found that the 10-year rates of distant metastasis and cause-specific mortality were higher for patients with positive margins, compared with patients with negative margins. Although these differences were not large, several studies have suggested this trend (23,35,63,134,142,152). Interestingly, Gronchi et al. reported a cause-specific mortality hazard ratio of 0.7 (p = .032) in favor of patients receiving adjunctive radiation therapy. Whether local recurrence seeds distant metastasis to impact survival, or it is simply a reflection of biologically aggressive disease, remains controversial.

A variety of molecular pathologic factors have been evaluated for prognostic significance. Proliferative activity as assessed by Ki-67 (MIB-1) immunohistochemistry has been shown to be prognostic (61). Increased expression of p53 and MDM2 has been associated with a poor prognosis in some studies (148), but not others (61). In synovial sarcomas, the presence of the fusion gene SYT-SSX2 was shown to associate with higher metastasis-free survival than SYT-SSX1 (75). Despite these preliminary data, routine application of molecular prognostic biomarkers awaits prospective validation in larger patient cohorts. These molecular markers may be most useful for selecting high-risk patients for future trials of adjuvant chemotherapy.

General Management

The majority of soft tissue sarcoma patients require multimodality treatment. Treatment is optimally delivered by a multidisciplinary team of dedicated surgical, orthopaedic, medical and radiation oncologists, plastic and reconstructive surgeons, pathologists, and radiologists with specific interest and expertise in mesenchymal malignancies (51). Given the rarity of soft tissue sarcomas, it is understandable that treatment results are optimized at specialized sarcoma centers.

Surgery

Surgical resection is the primary and only potentially curative treatment for soft tissue sarcomas. The primary goals of sarcoma surgery are to achieve optimal oncologic resection while preserving maximal function with minimal morbidity. Surgical specimens, including all surgical margins, should be thoroughly assessed by expert pathologists. For an optimal oncologic resection, negative surgical margins should be obtained if at all feasible, and often may require re-resection. If these goals cannot be anticipated with primary surgery, strong consideration should be given to preoperative treatment with chemotherapy and/or radiation therapy.

Four categories of surgical procedures have been described, based on the surgical plane of dissection (37). An intralesional procedure results in partial tumor removal with violation of the pseudocapsule. Although appropriate for a planned incisional diagnostic biopsy, it is not an appropriate therapeutic procedure. A marginal procedure (simple excision or “shellout”) removes the tumor within the confines of the pseudocapsule with a high likelihood of local recurrence due to residual subclinical disease (156). In wide local excision, the tumor is removed with a margin of normal tissue from within the same muscle compartment without removal of the entire structure of origin. Radical excisions, including compartmental resections and most amputations, remove the entire tumor and the structure of origin (entire anatomic compartment) en bloc. Local recurrence rates after surgery alone range from <10% after radical excision to ≥80% after marginal excision.

Historically, radical resections were performed to maximize local control, but also severely compromised limb function (143). Subsequently, more conservative, limb-sparing surgical procedures have become standard. Surgery alone may be sufficient for selected, small soft tissue sarcomas excised with wide (>1 cm) margins (49,72). However, conservative surgery with adjunctive radiation therapy is required in the majority of cases and results in local control comparable to amputation (89,111,127,144) with superior functional and cosmetic results (114,126). Amputations are now applied to <5% of patients at major sarcoma centers, and are reserved for massive disease in which functional limb-preservation is not feasible. Amputation may also be used for salvage of patients with local recurrence after previous conservative resection and radiation therapy, although limb salvage may still be possible in these cases (18).

Multidisciplinary communication regarding surgical technique can influence the effectiveness of postoperative radiation therapy as well as the incidence of late complications. Surgical scars and drain sites, which are at risk for subclinical disease, should be positioned and oriented such that their inclusion in the radiation treatment portal avoids circumferential (or near-circumferential) limb irradiation. Surgical clip placement at the boundaries of the tumor bed facilitates radiation treatment planning (131). Prophylactic bone stabilization in antici-pation of circumferential bone irradiation may reduce risk of subsequent fracture.

Chemotherapy

The efficacy of chemotherapy for soft tissue sarcoma is difficult to assess because of the heterogeneity of patients and drugs studied and the small sizes of individual trials. There remains no uniform consensus regarding the value of chemotherapy in patients with soft tissue sarcoma. Anthracyclines (doxorubicin and epirubicin) achieve response rates of 15% to 25% in patients with metastatic disease (98). Single-agent ifosfamide achieves similar response rates at conventional doses, and may be more
P.1814

active at higher doses (84). Combination chemotherapy regimens may be more active than single-agent regimens, although with increased toxicity (120). Regimens combining ifosfamide with an anthracycline appear to result in higher response rates than those without ifosfamide (146).

Early randomized trials of adjuvant chemotherapy in patients with localized disease showed no significant benefit for treatment (5); however, a meta-analysis found that patients receiving adjuvant doxorubicin had significantly improved local and distant recurrence-free survival (1). A 4% absolute benefit in overall survival at 10 years was not statistically significant, although for patients with extremity sarcomas, an absolute overall survival benefit of 7% at 10 years was statistically significant. Criticisms have centered on the inclusion of patients with visceral soft tissue sarcoma and the lack of central pathology review in this analysis (138). A recent randomized Italian trial randomized 104 patients with ≥5 cm or locally recurrent high-grade extremity sarcomas to five cycles of adjuvant epirubicin and ifosfamide or observation (47). With an updated median follow-up of 90 months, the 5-year survival for the chemotherapy group was 66% compared with 46% for the control group (p = .04) (46). A retrospective analysis of 245 patients with resected high-grade liposarcomas of the extremity treated with or without adjuvant chemotherapy suggested improved disease-specific survival with ifosfamide-based chemotherapy but not doxorubicin-based chemotherapy (34). In 215 patients with resected synovial cell sarcoma, distant metastasis-free survival was improved in patients receiving adjuvant chemotherapy (39). However, when 674 patients treated with neoadjuvant or adjuvant doxorubicin-containing chemotherapy at the MSKCC and the M.D. Anderson Cancer Center were combined for a retrospective analysis, the benefit of chemotherapy in improved disease-free survival was only sustained for 1 year after treatment (25), emphasizing the importance of sufficient follow-up in clinical studies of adjuvant chemotherapy for soft tissue sarcomas.

Chemotherapy given in the neoadjuvant setting allows investigators to judge clinical and pathologic response to treatment, and may provide a basis for identifying patients for whom additional chemotherapy may provide a benefit (106). A retrospective study of patients treated with surgery alone versus those receiving neoadjuvant doxorubicin and ifosfamide before surgery showed improved disease-specific survival for those receiving neoadjuvant chemotherapy, with the benefit mainly seen in patients with tumors >10 cm (55). However, a prospective randomized phase II trial of surgery alone versus neoadjuvant doxorubicin and ifosfamide followed by surgery failed to demonstrate a survival benefit (53). For many institutions, including our own, clinical trials of neoadjuvant and adjuvant chemotherapy in selected patients with high-grade extremity sarcomas are ongoing.

The next generation of systemic treatment strategies for soft tissue sarcoma may arise from current translational research that has identified specific molecular targets for therapy. The development and use of the imatinib mesylate (STI 571) in patients with gastrointestinal stromal tumor provides proof of principle for this type of approach (69). Molecular agents designed to modulate various receptor tyrosine kinases and their downstream signaling pathways governing growth and differentiation, survival and apoptosis, normal and aberrant transcription, invasion and metastasis, and angiogenesis are all being investigated in soft tissue sarcoma (14).

Radiation Therapy

Radiation therapy plays a central role in the treatment of soft tissue sarcoma. Although historically considered to be “radioresistant,” sarcomas have similar radiosensitivity to epithelial neoplasms (117). Multimodality treatment combining conservative surgery and radiation therapy achieves excellent local control rates while minimizing morbidity and maximizing long-term extremity function in comparison to radical surgery. Radiation therapy may be delivered using external beam, brachytherapy, or intraoperative electron beam techniques, and advancing technologies such as intensity-modulated radiation therapy (IMRT) and proton or other charged particle radiation therapy are also being applied to sarcomas (31,108,149).

Adjunctive radiation therapy may be effectively and safely delivered either before or after surgery. Postoperative radiation therapy allows for examination of resected specimen, including assessment of the surgical margins, to aid in treatment planning. Preoperative radiation therapy may allow for smaller radiation treatment volumes and may reduce the risk of local and distant dissemination at the time of resection. A phase III Canadian trial randomized 190 patients to preoperative (50 Gy preoperative radiation therapy with 16 to 20 Gy postoperative boost for positive margins) versus postoperative radiation therapy (50 Gy to large field and 16 to 20 Gy cone down boost) with a primary end point of acute wound complications (102). At a median follow-up of 3.3 years, there was a 35% incidence of wound complications in patients treated with preoperative radiation therapy, compared with 17% of patients treated with postoperative radiation therapy (p = .001). Interestingly, increased wound complications were only observed for lower extremity tumors. Updated results with a median follow-up of 6.9 years reveal that local and distant control rates as well as survival are equivalent between the two arms; however, a higher rate of late complications including fibrosis was observed with postoperative radiation therapy (29,103).

Preoperative radiation therapy (44 Gy in split-course) interdigitated with MAID chemotherapy (mesna, doxorubicin, ifosfamide, and dacarbazine) was developed as a strategy to enhance local control and limb preservation (30). Of the 66 patients treated with this regimen as part of a Radiation Therapy Oncology Group trial, 83% experienced grade 4 toxicities and there were three treatment-related deaths (79). Only 22% of patients had partial responses; however, 91% of patients were able to have complete tumor resection, and 27% of resected tumors showed no residual viable tumor. Regional delivery of intraarterial doxorubicin with concurrent preoperative radiation therapy has been shown to result in excellent local control rates, but was also associated with a substantial incidence of wound complications (90,141).

Although many sarcoma centers use preoperative radiation as standard treatment, at our own institution we use mainly postoperative radiation therapy because of concerns about acute wound complications and the fact that many patients are treated on clinical trials of neoadjuvant chemotherapy. If attempted resection will clearly result in gross residual disease and limb-sparing treatment is still desired, preoperative radiation therapy should be considered in an attempt to avoid amputation. The current National Comprehensive Cancer Network practice guidelines include each radiation treatment strategy (preoperative external beam, brachytherapy, and postoperative external beam) because all are effective at achieving excellent local control rates, and there are no data to suggest that one approach has greater efficacy (33). The optimal sequencing for surgery, radiation therapy, and chemotherapy remains unknown.

Retrospective series of highly selected soft tissue sarcoma patients treated with wide local excision with generous margins alone have reported high local control rates (3,7). Interestingly, size and depth were not associated with local relapse in these series. The subset of patients that may be adequately treated with radiation therapy alone has not been well defined; however, for lesions that have been properly excised with wide negative margins (all margins >1 cm), it is reasonable to consider observation, particularly if local recurrence in the tumor
P.1815

bed could be re-excised with preservation of function. These criteria are met in fewer than 10% of patients. Current strategies place less emphasis on grade because data from randomized trials of adjuvant brachytherapy (108) and external-beam irradiation (149) show similar incidence of local recurrence for both low- and high-grade tumors. Adjuvant radiation therapy appears to improve local control for both low- and high-grade tumors.

Radiation therapy can also be delivered with radical intent for patients who refuse surgery or have unresectable sarcomas (124,132). However, the local failure rate remains unacceptably high. Several alternative approaches have been investigated for these patients, including preoperative radiation therapy with concurrent chemotherapy (discussed earlier), preoperative radiation therapy with concurrent hyperthermia (113), the use of iododeoxyuridine or other radiosensitizers (125), high linear energy transfer radiation (fast neutrons) (122), and isolated limb perfusion with tumor necrosis factor-α, melphalan, and interferon-γ (86). If these strategies or others could improve local control rates in patients with unresectable disease, consideration could then be given to limb-sparing procedures in the 5% to 10% of patients with extremity sarcomas who otherwise would still require amputation to achieve clear proximal margins.

Radiation Therapy Techniques

Radiation therapy techniques for treatment of sarcomas of the extremity, trunk, and head and neck are described here; retroperitoneal sarcomas are discussed in Chapter 73.

Compartmental Nature of Soft Tissue Sarcomas

Before commencing a course of radiation therapy, the radiation oncologist must evaluate the extent of tumor involvement in the muscle compartment, understand the anatomy of the compartment, and be able to assess the risk of extracompartmental involvement based on MRI and CT imaging. For patients treated in the postoperative setting, attendance in the operating room at the time of resection and surgical clip placement is invaluable in this regard.

Volume at Risk

The radiation target volume is determined based on physical examination, radiologic studies, and knowledge of anatomy and the natural history of sarcomas. Normal structures and organs in proximity to the targeted region must be identified, and appropriate dose constraints for each must be considered. In the postoperative setting, details from the surgeon regarding the extent of dissection or observations from the resection itself must be considered. Some authorities recommend treating the entire compartment (origin to insertion) because hematoma can theoretically track cells to the farthest reaches of a muscle compartment (150). Others recommend margins around the tumor or tumor bed ranging from less than 5 cm up to 15 cm (in the long axis of the extremity), based on the grade and size of the tumor (128). One retrospective analysis of patients treated with postoperative radiation therapy demonstrated that an initial margin of <5 cm was associated with a significantly higher rate of local failure, compared with ≥5 cm (99). Our general practice is to include the resection bed with a 5-cm margin, the incision, and any drain sites in the initial treatment volume. However, the MSKCC brachytherapy experience calls this into question because excellent local control is achieved with a technique that does not cover the surgical scar, the drain sites, or the wide margins discussed previously (108). Clearly, margins should not extend beyond natural barriers of spread (i.e., fascial planes, bone). Regional lymph nodes are rarely at risk in extremity sarcomas, and there are no convincing data that prophylactic lymph node irradiation is beneficial.

Positioning the Extremity

The extremity should be positioned so as to treat the region of the affected compartment with minimal treatment of uninvolved tissue. The anterior compartment of the thigh can be treated in the “frog-leg” position, with external hip rotation, separating the anterior compartment from the medial and posterior compartments. A lateral decubitus position, with the affected thigh closest to the couch with flexion of the uninvolved extremity, allows treatment of the posterior compartment (Fig. 81.4A). The anterior compartment of the arm (biceps) can be treated by having the shoulder abducted approximately 90 degrees and maximally internally rotated (Fig. 81.4B). Positioning of extremities can be difficult for patients because of effects of tumor or surgery. If the extremity is placed at too extreme an angle, CT scanning may be more difficult. It is often necessary to assess multiple limb positions to discover the optimal setup. The body part must be immobilized with a device such as a foam cradle, plaster mold, or thermoplastic cast (Fig. 81.4), with the limb secured above and below the treatment area to reduce the possibility of rotation in the cradle. Cradle material can be removed from the region to be treated, if it will not compromise the immobilization, to reduce skin toxicity due to bolus effect.

P.1816


Treatment Planning

It is common practice to use a “shrinking-field technique” for treatment of sarcomas. For postoperative radiation therapy, the initial treatment fields are designed to encompass the resection bed with generous margins. Subsequently, reduced fields encompassing the preoperative tumor volume can be boosted with smaller margins. For preoperative radiation therapy, the gross tumor itself with a margin is treated. Usually, no field reduction is made prior to surgery, but a postoperative boost can be delivered in case of a positive margin.

For either CT-based, three-dimensional planning, or conventional fluoroscopic simulator-based, two-dimensional planning, it is useful to construct a clinical target volume (CTV) that encompasses any gross disease as well as a volume to account for potential subclinical (microscopic) disease. In the postoperative setting, the initial CTV can be constructed based on the volume of the resection bed (defined by placement of surgical clips and consultation with the surgeon), the preoperative tumor volume (based on preoperative imaging), and additional volume for extension of potential subclinical disease. This initial CTV should be generous, covering the resection bed with a 3- to 6-cm margin, as well as the surgical scar and drain sites. If the scar is being struck tangentially by the irradiation fields, no bolus is necessary. However, if the scar is being irradiated with a direct perpendicular field, bolus should be applied to ensure a brisk skin reaction and full dose over the scar itself. The “boost” CTV usually is limited to the preoperative tumor volume only with smaller 2- to 3-cm margins. In preoperative cases, the CTV can be derived by expansion of the visible gross tumor volume, again with generous 3- to 6-cm margins. The planning target volume should be an expansion of the CTV accounting for potential variation in daily setup, easily 1 cm or more for extremity targets.

A ≥1 cm strip of soft tissue in the circumference of the extremity should be spared to avoid subsequent edema. Attempts should be made to avoid circumferential bone radiation, if possible, to reduce fracture risk, and to minimize joint irradiation, if possible. Three-dimensional conformal radiation therapy and IMRT treatment planning may be useful in achieving the desired dose distribution in selected extremity cases (Fig. 81.5). Although these techniques are useful to spare normal tissues and bone to reduce morbidity (65), extra caution must be used with steep dose gradients to ensure adequate dose coverage of target volumes.

Treatment of thin regions of anatomy (e.g., hand, foot, and forearm) presents additional technical concerns. Skin-sparing by high-energy photon beams can produce underdosed regions inside the tumor volume, and bolus to the entire treatment volume may be necessary. Treatment of the involved region inside a water bath ensures uniform dosage to the affected area, although the complete loss of skin-sparing can produce marked skin reactions. With meticulous technique, limb-sparing surgery with adjuvant radiation therapy can be safely applied (95,129).

Radiation Energy and Dose

Lower energy (6-MV) photons are usually used because higher energies could potentially spare too much superficial tissue. However, higher energy (10- to 16-MV) photons are occasionally
P.1817

required for thigh or buttock lesions to produce reasonable dose homogeneity. Sarcomas are usually treated to high doses, even in the adjuvant setting. In postoperative therapy, the initial volume is usually treated to 45 to 50 Gy, with subsequent cone downs to a final dose of 60 to 66 Gy, using 1.8- or 2.0-Gy daily fractions. For preoperative irradiation, 45 to 50 Gy is often delivered 2 to 4 weeks before resection with an intraoperative or postoperative boost as indicated by the surgical margin. Brachytherapy or intraoperative radiation therapy may be used in combination with either preoperative or postoperative external-beam radiation therapy. Doses of 12 to 25 Gy may be given by intraoperative electron beam, or perioperative low dose rate (LDR) or high dose rate (HDR) afterloading brachytherapy, with 36 to 50 Gy external beam (2,19,32,80). Brachytherapy may be used as the sole radiation therapy mode of treatment, using doses of 42 to 50 Gy. For unresectable sarcomas, doses above 70 Gy are used, limiting the high-dose volume to the tumor plus a minimal margin.

Truncal and Head and Neck Sarcomas

Tumors arising in the trunk are usually more superficial than extremity sarcomas, but have similar clinical behavior (57). Tumors on the chest wall and abdominal wall often can be treated with oblique tangential fields. After 45 Gy of photon irradiation, a direct electron boost to the surgical bed can be use to minimize dose to underlying lung or bowel.

In the head and neck, target volumes and sensitive normal tissues are often in close proximity, and toxicity is a significant concern. Treatment planning techniques with three-dimensional conformal radiation therapy and IMRT are particularly useful in these locations. Treatment planning must account for the different patterns of spread that distinguish sarcomas from squamous cell carcinomas in the head and neck region, including the substantially lower risk of nodal spread for sarcomas. Several techniques for entire scalp irradiation have been described, and may be used for the treatment of angiosarcomas, which are infiltrative and prone to local recurrence after surgery alone (77,96,136).

Interstitial Brachytherapy

Interstitial brachytherapy can be used to deliver all or part of the radiation dose (8). After surgical excision of the tumor, hollow plastic afterloading catheters are inserted using sharp metal trocars in a single plane at approximately 1-cm intervals within the tumor bed (Fig. 81.6). Surgical clips placed at the margin of the tumor bed permit the target volume to be delineated for planning purposes, and the catheters are secured in place. Orthogonal localization films are obtained 2 to 4 days after surgery, and catheter positions may be digitally recorded into a radiation therapy planning system. In contrast to the wide margins typically employed for external-beam irradiation, the MSKCC experience demonstrates that a brachytherapy CTV encompassing only the clipped tumor bed with a 2-cm margin resulted in adequate local control (108). The dose is prescribed to 5 to 10 mm from the implant plane. Catheters are loaded with wired LDR 192Ir seeds or connected for HDR treatments no sooner than the sixth postoperative day to reduce the risk of wound complications. After completion of the treatment, sources are removed and catheters are cut at one end for removal by pulling through the skin.

For LDR implants, a dose of 42 to 45 Gy has been shown to be adequate adjuvant treatment when used alone for high-grade lesions (108). If brachytherapy is to be used in combination with external-beam radiation therapy, a dose of 15 to 25 Gy is used with 45 to 50 Gy external beam (2,32). HDR implants allow for more customization of the treatment plan because the dwell times of the single HDR source at each position can be manipulated. HDR treatments are usually given twice daily at 2 to 5 Gy per fraction to 35 to 50 Gy when used alone, or 15 to 20 Gy when to be used with postoperative external beam (19). HDR treatments can be delivered using conventional interstitial catheters as already described; a technique for intraoperative HDR treatment has also been described (80). A detailed list of recommendations for brachytherapy has recently been published by the American Brachytherapy Society (100).

Results of Standard Treatment

Retrospective reports and a prospective randomized trial have demonstrated conclusively that limb-sparing surgery plus adjunctive radiation therapy produces local control and survival rates similar to those achieved with amputation (89,112,115,127,151). The value of adjuvant radiation therapy after limb-sparing surgery has been demonstrated in randomized trials (108,149). Local control rates for patients with intermediate and high-grade sarcomas treated with surgical resection and adjunctive radiation therapy are generally in the 80% to 90% range, and representative series are presented in Table 81.4. The brachytherapy literature is limited in comparison to the extensive literature supporting the local control benefits of external-beam radiation therapy. Nonetheless, investigators from MSKCC and a limited number of other institutions have demonstrated that brachytherapy achieves comparable local control benefits for intermediate and high-grade disease when used alone or in combination with external irradiation (2,32,108). That these different techniques produce similar and excellent local control further validates the basic concept that radical treatment can be achieved with limb preservation. Even for patients with large high-grade lesions, a local control rate of approximately 85% can be achieved with the use of limb-sparing, wide local excision, and meticulous radiation therapy techniques.

Local control rates for low-grade lesions are also excellent with either postoperative or preoperative external-beam irradiation. In a series of patients treated at the National Cancer Institute, adjuvant external-beam irradiation significantly decreased local recurrence rates, primarily in patients with positive margins (92). Importantly, a randomized trial from MSKCC revealed
P.1818

that brachytherapy does not improve local control compared with surgery alone in low-grade lesions (109). Therefore, external beam is preferred over brachytherapy for treatment of low-grade soft tissue sarcoma. Unlike the high- and inter-mediate-grade lesions, low-grade tumors have almost no metastatic potential. Therefore, local control is tantamount to cure in this group.

Despite the increased technical demands, similar local control rates with limb-sparing procedures have been described for sarcomas of the distal extremities. Local recurrences can be salvaged with additional surgery (amputation) with no apparent decrement in survival (15,95,129). Although the head and neck represents another technically challenging site because of the difficulty obtaining negative margins, local control rates have been reported in the 75% to 90% range when radiation therapy is combined with wide local excision (11,85,104).

Overall survival for patients with soft tissue sarcoma closely relates to the development of distant metastases. This is related to the current American Joint Committee on Cancer stage, as can be seen in Fig. 81.7.

Unresectable Sarcomas

In patients who are not eligible for surgical resection, radiation therapy alone can be considered but results in relatively low rates of durable local control. In one series, local control was related to radiation dose and tumor size (76). Neutron radiation, carbon ion beam or photon radiation in combination with radiosensitizing iododeoxyuridine, or isolated limb perfusion with tumor necrosis factor-α, melphalan, and interferon-γ have also been used (52,71,122,124,125,132). Although the optimal treatment of sarcomas clearly involves complete excision, high-dose irradiation may at least achieve palliative benefits.

Treatment of Metastatic Disease

For patients with metastatic disease and controlled primary tumors, complete surgical resection of pulmonary metastases may be potentially curative and can result in disease-free survival rates of 40% at 3 years (13,137). The role of radiation therapy in treatment of metastatic disease is mainly limited to palliation of sites of disease causing local symptoms, although the possibility of using extracranial stereotactic radiation techniques for patients with solitary or oligometastasic disease in unresectable locations may be an area for future investigation.

Aggressive Fibromatosis and Dermatofibrosarcoma Protuberans

Aggressive fibromatosis (desmoid tumor) and dermatofibrosarcoma protuberans (DFSP) are soft tissue neoplasms that almost never metastasize but can be very invasive locally. Desmoids arise within the muscle or its fascial coverings, and DFSPs arise within the dermis. Microscopically, bundles of spindle-shaped fibroblasts are surrounded by abundant fibrous stoma devoid of mitotic figures. Complete surgical excision alone is usually curative for these tumors, but is not always possible because of their size and location. Local recurrence is common. For desmoid tumors, postoperative radiation therapy improved local control for patients with positive margins or gross residual disease (10,97,101,154). It should be noted that some authorities prefer to observe surgically resected patients with microscopically positive margins if the disease site can be readily followed and a local recurrence could be re-excised with minimal morbidity. Primary radiation therapy achieves high rates of local control when surgery is not feasible. Little evidence exists for a dose–response relation, and doses of 50 to 55 Gy are used for either subclinical or gross disease. Tumor responses are rarely seen in <6 months, but can occur after 1 to 2 years. Nonsteroidal antiinflammatory agents, hormonal agents, cytotoxic chemotherapy, and imatinib have activity against desmoid tumors (67). For
P.1819

DFSP, radiation therapy enhances local control in patients with positive margins after surgery, or as sole treatment (9).

Sequelae of Treatment

The most significant short-term toxicity of radiation therapy for sarcomas is usually moist desquamation in the high-dose volume. This can be very uncomfortable in patients with proximal thigh tumors who receive significant dose to the perineum. Patients treated for truncal and head and neck sarcomas experience toxicity similar to breast cancer and head and neck squamous cell carcinoma patients. Major wound complications (delayed wound healing or need for surgical intervention) occur in approximately 5% to 15% of patients after surgical resection with postoperative irradiation, and perhaps more commonly with preoperative irradiation.

The long-term sequelae after conservative surgery and irradiation for extremity lesions must always be considered because they may significantly limit the function of the preserved limb. They include decrease in range of motion related to fibrosis, contracture of the joint, edema, pain, and bone fracture. In centers treating high volumes of patients with soft tissue sarcoma, the incidence of moderate-to-severe late effects is <10% (110). The risk of these complications may be reduced by sparing a strip of normal tissue (to allow lymphatic drainage from the extremity) and a portion of the circumference of uninvolved bone. If possible, joint spaces should be excluded after a dose of 40 to 45 Gy to avoid fibrotic constriction of joint capsules. Collaboration with physical therapy specialists is essential in minimizing disabilities after treatment of soft tissue sarcomas. Mobility of the extremity should be stressed, and patients should be on an exercise and range-of-motion program early in the course of therapy. In the treatment of patients with truncal sarcomas, it is particularly important to use cone down fields to limit the dose to normal tissues deep to the target volume (e.g., lung and bowel). In contrast to acute wound complications, late limb morbidity may be reduced with preoperative radiation, likely due to the lower doses and smaller volumes used with preoperative treatment (29). With attention to these details, a high local control rate can be achieved with minimum sequelae.

High-dose irradiation does not appear to compromise the viability of skin grafts used to repair defects after sarcoma surgery if adequate time is allotted for healing (at least 3 weeks) (83). Fertility can be preserved in men undergoing irradiation for lower extremity sarcomas through the use of a gonadal shield to decrease testicular dose (45). The risk of a second malignancy associated with adjuvant irradiation must also be considered, particularly in young patients with low-grade tumors in which an otherwise normal life expectancy is anticipated.

Future Directions

The most significant challenge in the management of soft tissue sarcomas is to reduce the mortality related to systemic disease in patients who present with M0 disease. Further investigation into the benefits of conventional cytotoxic chemotherapy, as well as new molecularly targeted agents, will ultimately determine whether mortality rates can be reduced. Molecular characterization of individual patients and tumors will improve patient selection in future trials. Many clinical trials in soft tissue sarcoma currently use neoadjuvant chemotherapy. Earlier systemic treatment may have a greater capacity to influence occult micrometastases, and allows the assessment of in vivo response. Unfortunately, given the rarity of the disease and the small size of many trials, statistical power will continue to be limited in power to resolve the current controversies regarding the value of chemotherapy. Ultimately, these questions must be addressed in multi-institutional cooperative group studies.

The local treatment of soft tissue sarcomas is markedly different today than it was 25 years ago. Most patients with extremity lesions are now treated with limb-preserving methods, achieving local control rates of ≥90%. Although we prefer wide local excision followed by postoperative irradiation for resectable extremity tumors, excellent results can be obtained with preoperative irradiation or brachytherapy. Wide local excision and meticulous shrinking-field radiation therapy given either before or after surgery have improved the local control rates for patients with truncal and head and neck sarcomas almost to that of extremity lesions. Because wide local excision alone would lead to approximately a 50% local failure rate, radiation therapy appears to permit organ preservation without a significant sacrifice in control rates.

Several issues remain important with respect to local control. Although it has become less common, some patients with advanced extremity sarcomas still require amputation to achieve clear proximal margins, or have unresectable tumors. Continued innovations in the use of combined-modality therapy or radiosensitizers may lead to further improvements in this area. The Radiation Therapy Oncology Group is building on its previous study of preoperative chemotherapy and radiation therapy with another phase II trial investigating interdigitated MAID with combined thalidomide and radiation therapy for high-grade disease, and combined thalidomide and radiation therapy for low-grade disease (see http://www.clinicaltrials.gov/ct/show/NCT00089544). The issues of acute and late treatment morbidity also remain active areas of investigation. The results of the Canadian trial comparing preoperative and postoperative irradiation have further emphasized the relation between field size and radiation dose to toxicity. The necessity of large 5- to 7-cm margins compared with the more conservative 2- to 3-cm margins successfully used in brachytherapy may be an area for randomized study. Currently, Canadian investigators are running a prospective trial investigating the potential benefit of IMRT in reducing wound complications (see http://www.clinicaltrials.gov/ct/show/NCT00188175). Finally, it will be important to develop methods to prospectively identify those patients who may be managed adequately with wide local excision alone. Although adjuvant irradiation can often be delivered without significant acute or late toxicity, selective elimination of its use without loss of local control would represent an additional success in the management of these patients.

80

Osteosarcoma

Epidemiology and Risk Factors

Osteosarcoma is the most common malignant bone tumor in childhood, representing approximately 50% of newly diagnosed malignant pediatric bone tumors or 700 new U.S. cases annually (27). The annual incidence is 4.5 per million in girls and 5.5 per million in boys (27). This incidence peaks in those ages 10 to 19 years (40). There does not appear to be a difference in incidence among African Americans and whites.

The etiology of osteosarcoma is unknown in most cases. The incidence does correlate with the growth spurt in teenagers. However, specific pathways associated with this finding are elusive. For a minority of patients, a specific risk factor is identified. These risk factors include, prior radiotherapy (58), and specific genetic syndrome. Survivors of hereditary retinoblastoma carry a risk of osteosarcoma of 6% at 18 years (20), Li-Fraumeni syndrome (6), and in older adults, there is an association between Paget's.

Clinical Presentation

Most patients present with pain in the affected limb or region and soft tissue swelling. In some patients, trauma and a subsequent pathologic fracture brings the individual to medical attention.

Approximately 90% present in the diaphysis of the extremities, with the distal femur and proximal tibia being the most common sites. Other sites such as the pelvis and head and neck represent significant minority of the locations (27).

Diagnostic Evaluation

Radiologic investigation begins with a plain radiograph (Fig. 80.1A). Classic findings include an ill-defined zone of transition, Codman's triangle (defined as osteoid formation under the periosteum), and bone formation in the adjacent soft tissue. The lesion itself may be sclerotic (Fig. 80.2), lytic (Fig. 80.1A), or mixed. Most lesions are subsequently evaluated by magnetic resonance imaging (Fig. 80.1B). This will show the proximal and distal extent of involvement, evaluate any soft tissue component, and establish the proximity of nerves, vessels, and the joint space. Skip metastases are a well-defined but uncommon entity in osteosarcoma. Modern series place the incidence of isolated skip metastases at diagnosis at <5% (33,51).

At diagnosis, approximately 15% of patients have detectable distant metastases. More than 80% of metastases are pulmonary, followed by metastases at bony sites (5). Therefore, chest computed tomography and radionuclide bone scan are needed to complete staging.

Positron emission tomography is being investigated as a part of the initial staging work-up and as a modality to evaluate response to chemotherapy (8). However, to date, it is not a part of the recommended work-up.

Staging Systems

There are two major staging systems for this disease: the Enneking system (22) and the American Joint Committee on Cancer system (1) (Table 80.1). However, most practitioners usually classify the disease state as nonmetastatic or metastatic, based on the presence or absence of distant metastases.

Pathology

The commonly accepted histologic description of osteosarcoma is based on the World Health Organization classification (53). This divides osteosarcomas into intramedullary and surface subtypes. The most commonly encountered subtype is the conventional category of medullary tumors. These are further subclassified into osteoblastic, chondroblastic, fibroblastic, and mixed types based on the pathologist's visualization of the specific elements. Other categories of medullary (or conventional) osteosarcoma are small cell, telangiectatic, and well-differentiated (or low-grade) types.

Surface osteosarcomas are divided into parosteal (juxtacortical), periosteal, and high grade.

P.1802

Overall Management

Chemotherapy is essential for cure. In the nonmetastatic setting, the overall schema consists of chemotherapy followed by resection of the primary tumor, and adjuvant chemotherapy.

In patients with pulmonary metastatic disease, the treatment program is the same as for nonmetastatic disease, with the addition of possible resection of any pulmonary nodules remaining after the completion of chemotherapy.

Low-grade osteosarcomas are usually managed with surgery alone.

Surgical Management

Resection of the primary tumor is part of the standard management. Subsequent to an en bloc resection of the tumor, reconstruction is usually required. The goal of the surgical intervention is to remove the tumor en bloc and achieve adequate negative margins.

Presurgical planning includes careful evaluation of the pre- and postneoadjuvant chemotherapy imaging and determination of the anticipated reconstruction. For extremity tumors, imaging will often show a decrease in the soft tissue component of the tumor and allows visualization of the neurovascular structure, muscle groups, and fascial planes; the relationship of the tumor to the epiphysis and articular surface; and provides an estimate of the length of bone to be removed.

There are several options for surgery. Amputation should be recommended if the patient will be left with a nonfunctioning limb (39). Most individuals will undergo some type of limb-sparing procedure. Reconstruction options include autologous bone grafts, allografts, and endoprosthetics. Less commonly, rotationplasty or arthrodesis is employed. In the current era, 80% to 90% of patients will undergo a limb salvage (39).

Reconstructions can suffer infections, nonunion, and fracture, depending on the technique. Endoprosthetics are prone to infection. Allografts can fracture up to 20% of the time (37). However, the functional outcome of various reconstructive techniques can be good in 60% to 90% of cases (24,37).

Traditionally, pelvic osteosarcomas present a challenge to the orthopaedic oncologist. Small tumors may be adequately resected with or without reconstruction. Resection of large tumors may mean not only the loss of the ipsilateral lower extremity, but compromising of bowel and bladder function.

Chemotherapy

Systemic chemotherapy is standard of care for all patients who are able to tolerate the intensive regimens.

Two randomized studies demonstrated the efficacy of adjuvant chemotherapy (21,29). Table 80.2 shows various randomized trials of adjuvant chemotherapy. Notably, an early randomized trial had negative findings (36). The standard agents used are methotrexate, cisplatin, and doxorubicin, all of these with or without ifosfamide.

P.1803

Subsequent studies investigated neoadjuvant chemotherapy as a way to evaluate tumor response. A randomized Pediatric Oncology Group study showed no difference in outcome whether preoperative or postoperative chemotherapy was administered (26) Advantages of neoadjuvant chemotherapy include the determination of the pathologic response, early treatment of micrometastatic disease, and allowing adequate time for surgical planning and ordering of a custom prosthesis.

The percent necrosis after neoadjuvant chemotherapy is a prognostic factor (5,32). The classification scheme is according to the Huvos grade. The overall survival of patients with nonmetastatic disease with >90% necrosis is near 70%, compared with 50% in those with <90% necrosis (31). Therefore, the next therapeutic question was whether the survival of poor responders could be improved by altering and/or intensifying chemotherapy administered after surgery. Several studies have investigated this, but no improvement in survival has been demonstrated (3,61). Likewise, attempts to intensify the chemotherapy regimen delivered preoperatively have failed to show an increase in survival despite a small increase in the percentage of good responders (41,48).

Radiotherapy

Historically, radiotherapy has been used in the treatment of osteosarcoma. Prior to effective chemotherapy, Cade (13) pioneered a technique of radiotherapy with delayed amputation in those who did not develop distant metastases. Subsequently, others questioned the need for amputation. The radiotherapy doses employed were 5,000 to 8,000 R. Of note, in these series many patients did have resolution of their symptoms (pain and swelling) soon after starting radiation. Beck et al. (4) report only 1 of 21 survivors in a group treated with definitive radiotherapy. However, prior to death, three patients had local recurrences. deMoor (18) describes a cohort treated with “radical radiotherapy.” Of the 27 initial patients, 9 had survived at least 5 years and 3 had local recurrences.

P.1804

With the advent of chemotherapy, Caceres et al. (12) reported on a group of 16 patients who were treated with chemotherapy and definitive radiotherapy. Tumors and surrounding tissue received 6,000 rad after one cycle of chemotherapy. Chemotherapy was then continued for 1 year. Biopsies were performed at the primary site in 15 of 16 patients every 3 months after the initiation of treatment. Results of this study showed that 80% of patients had a complete pathologic response. Complications included soft tissue fibrosis in nine patients, fracture in four, infection in two, and necrosis in two.

The role of radiotherapy in osteosarcoma therapy in the 21st century is now limited to select situations. Specifically, irradiation is considered in patients who refuse surgery, those with positive margins after resection, those with sites that are not amenable to resection and reconstruction, and palliation.

Modern External-beam Radiotherapy

Table 80.3 gives an overview of modern radiotherapy treatment. In recent years, contemporary chemotherapy and definitive radiotherapy in those refusing amputation has been reported by Machak et al. (38). A median of 60 Gy was given using conventional fractionation. The 5-year local progression-free survival was 56%, with an overall survival of 61%. Those with a good response to neoadjuvant chemotherapy had an overall survival of 90%, compared with 35% in those who were poor responders. This phenomenon was also paralleled in local control. There were no local failures in good responders, but nearly one third of poor responders failed locally.

Conversely, Delaney et al. (17) reported only a 22% local control rate in patients who were treated with chemotherapy and local radiotherapy. For the group of patients receiving radiotherapy adjuvantly after surgery, the local control rate was 74%, with a gross total resection or a subtotal resection.

Dincbas et al. (19) recently reported preoperative radiotherapy integrated in the usual osteosarcoma treatment protocol. Local control was excellent at 97% with good limb salvage. However, this is similar to what would be expected in the cooperative group trials. Therefore, it is not clear that radiotherapy added to the overall outcome.

Extracorporeal and Definitive Intraoperative Radiotherapy

The techniques of extracorporeal and definitive intraoperative radiotherapy (IORT) have been investigated in bone tumors (Table 80.4) (9,15,28,44,58,59). The extracorporeal technique includes en bloc resection of the tumor and surrounding soft tissues, irradiation of the specimen, and reimplantation, often with the aid of prostheses. With definitive IORT, the operative field is exposed and radiotherapy is administered. No resection of the tumor is performed.

Extracorporeal irradiation is associated with a low rate of local recurrence (<5%). Chen et al. (15) noted a higher rate of complications (62%) in their initial series. The events included fractures, nonunions, wound infections, and loss of cartilage. Subsequently, they incorporated the use of prostheses placed at the time of reimplantation. Their local recurrence rate

P.1805

continued to be low and there was only one complication (a nerve palsy) in their series of 14 patients (15).

The reported local control rate for definitive IORT is 20% to 25% (44,57). The complication rate is >50% as reported by Tsuboyama et al. (57), but minimal in the hands of Oya et al. (44).

Particle Therapy

Because of the difficulty of achieving adequate local control with photons, neutrons and protons have been employed in the treatment of osteosarcoma (Table 80.5). Neutrons are thought to have a higher relative biologic effectiveness and oxygen-enhancement ratio, making them radiobiologically more effective against osteosarcomas. The advantage of protons is in the physical properties of the Bragg peak, which falls off rapidly and spares adjacent tissue.

The earliest studies of particle therapy are with neutrons in the 1970s and 1980s, prior to optimal chemotherapy and surgical reconstruction. The review of the early data by Laramore et al. (35) shows an overall local control rate of 55% in 73 patients pooled from seven institutions worldwide.

In a more recent review of head and neck sarcomas, Oda et al. (43) report local control in a patient treated with chemotherapy, surgery, and neutron irradiation. One other patient who received only surgery and neutron therapy had local failure. Carrie et al. (14) describe local control in 4/4 pelvic osteosarcomas treated with modern chemotherapy and a combination of photons and neutrons.

The major complications surrounding neutron therapy are severe fibrosis and scarring of the soft tissues and adjacent organs (35).

The largest proton experience is at the Massachusetts General Hospital (30). Fifteen patients with osteosarcoma of the base of skull or vertebra were treated by this form of therapy. The 5-year local control is reported at 59%.

Whole-lung Irradiation

Prophylactic lung irradiation has been investigated in osteosarcoma. Three randomized trials were conducted in the 1970s and 1980s (Table 80.6) (9,11,49). The Mayo Clinic and first European Organisation for Research and Treatment of Cancer studies were conducted prior to the routine use of chemotherapy (9,49). They both showed trends toward improved survival with whole-lung irradiation. However, a three-arm EORTC/SIOP study that compared chemotherapy, whole-lung irradiation, or a combination of both, showed the same disease-free survival and overall survival in both arms (43% and 24%) (11). Therefore, with the recognition of the other advantages of systemic therapy, prophylactic lung irradiation has fallen out of favor (60).

Radionuclide Therapy

Several investigators have used radionuclides in the treatment of bony metastatic osteosarcoma (Table 80.7). There are case reports of the use of rhenium (52), strontium (25), and samarium (10). The major toxicity is decreased in the platelet and white blood cell counts.

Anderson et al. (2) conducted a phase I dose-escalation study of samarium-153 in metastatic osteosarcoma. The goal was to evaluate the toxicity of increasing doses of radionuclide using hematopoietic stem cells to decrease the bone marrow toxicity. Bone marrow toxicity and transient hypocalcemia were seen at the highest dose level. The authors report good pain relief.

Results of Radiotherapy in Specific Disease Sites

Pelvis

The management of large pelvic osteosarcomas continues to present a challenge. Definitive surgery often includes a hemipelvectomy. Despite being the most common nonextremity site of osteosarcomas, the percentage is <10%. The overall local failure rate in the 22 patients with spinal primaries was 70% in the Cooperative Osteosarcoma Study Group (45). Eleven of 67 patients received radiotherapy. Seven patients were treated definitively and four were treated in a postoperative fashion. The definitive dose was 56 to 68 Gy and the postoperative dose was 45 to 51 Gy. The majority of those patients receiving radiotherapy failed locally (6/7 treated definitively, and three of four treated after an intralesional surgery).

In the St. Jude Children's Research Hospital experience, local control was achieved in three-fourths of the patients using 50 to

P.1806

98 Gy and modern chemotherapy (50). Promising local control was achieved in the University of South Florida series of five patients treated with intra-arterial cisplatin and radiotherapy (23).

Spine

Fewer than 2% of patients present with spinal primaries. Within the Cooperative Osteosarcoma Study (COSS) studies, overall survival for patients with spinal primaries is <2 years and the local failure rate was near 70% in the 22 patients studied (46). Seven of the 17 patients who underwent an intralesional procedure or biopsy received radiotherapy only as part of their care. Radiotherapy doses ranged from 20 to 60 Gy. Five of seven patients had local recur-rences.

When the group from Memorial Sloan-Kettering Cancer Center analyzed their series, 5 of 11 patients in the cohort treated with resection, external-beam radiotherapy, and chemotherapy were long-term survivors (56).

Head and Neck

Most head and neck osteosarcomas present in the mandible or maxilla. The age of presentation tends to be somewhat older than that of patients with extremity lesions (55). The review by Kassir et al. (34) finds an overall local control rate of 50% at these sites. Almost 40% of the patients received radiotherapy (external-beam or brachytherapy), but no comment is made on the effect of irradiation on local control. Those receiving radiotherapy did have a lower survival rate than those treated with surgery and chemotherapy.

St. Jude Children's Research Hospital researchers reported on four children who received 31 to 74 Gy postoperatively (21). The two who received 31 Gy and 40 Gy both had local failure. In the University of Washington experience, five patients received postoperative radiotherapy (49). The three who received chemotherapy have maintained local control. However, the two who did not receive chemotherapy died, but no comment was made regarding the status of the primary site.

Late Effects

Late complications are largely related to chemotherapy and surgical interventions. Doxorubicin can cause cardiomyopathy (47) and cisplatin results in high-frequency hearing loss in about half of patients (54). Some patients will exhibit transient changes in renal function, but late complications are unusual. Second malignancies, with a minimum 5-year follow-up, were reported in 7% (42).

Nicholson et al. (42) report long-term survivors having more difficulty climbing stairs; the patients had similar employment and marital status as sibling controls.

With respect to radiotherapy, the data are limited. This is largely because this modality is used in patients with unfavorable prognoses, with a low chance of long-term survival. Laramore et al. (35) report a 25% to 40% complication rate of study results gleaned from reviewing the literature for neutron therapy, which is often related to dense fibrotic reactions. Delaney et al. (17) report a 24% complication rate in a proton/photon cohort. In a definitive external-beam radiotherapy series, Machak et al. (38) describe three-quarters of the patients as having good limb function. Three of the 31 patients had pathologic fractures and 1 had skin necrosis.

79

Chapter 79

Plasma Cell Myeloma and Plasmacytoma

David C. Hodgson

Joseph Mikhael

Richard W. Tsang

Epidemiology and Etiology

Plasma cell neoplasms account for 22% of all mature B-cell neoplasms in the Surveillance, Epidemiology, and End Results (SEER) program in the United States (85). The majority of plasma cell neoplasms are multiple myeloma, with solitary plasmacytoma accounting for ≤6% of cases and rarely plasma cell leukemia. Although the incidence of multiple myeloma has gradually increased in the 1970s through the 1990s (100), recently there has been a downward trend, with an annual decrease of 0.3% from 1992 to 2002 (101). Data from SEER indicate an incidence rate of 5.5/100,000 per year and mortality rate of 3.8/100,000 per year during the period 1998 to 2002 (101). For 2007, it is estimated that there will be 19,900 new cases and 10,790 deaths due to multiple myeloma in the United States (5). The incidence rate exceeds that of Hodgkin's lymphoma and is about one-quarter that of non-Hodgkin's lymphoma. The incidence rate rises with advancing age, with a median age at diagnosis of 70 years (101), and <1% of cases are diagnosed in persons <35. Nonregistry studies usually report a lower median age ranging from 60 to 66 years (50,68). There is a slight male predominance, and for black Americans, the incidence and mortality rates are approximately double that of whites. The 5-year relative survival rates have increased, from 26% in 1975 to 1977, to 33% in 1996 to 2002 (p <.05) (5). The etiology of multiple myeloma is not known, but there are studies reporting association with prior exposure to radiation (e.g., atomic bomb survivors in Hiroshima) (88), certain chemicals such as petroleum products (18,29), and monoclonal gammopathy of unknown significance (MGUS) (69). The previously reported association with herpes simplex virus type 8 does not appear to be causally related (113).

Natural History, Genetic Mechanisms

Multiple myeloma is a disease with a wide clinical spectrum, ranging from the condition known as MGUS to the most aggressive form, plasma cell leukemia (Table 79.1). In all cases, a plasma cell clone exists but to varying degrees. The secretion of a monoclonal protein by these plasma cells, along with their interaction with the bone marrow environment, is the source of organ damage in patients with this illness (54). These concepts have become particularly important as the molecular mechanisms by which the disease progresses through these “stages” provide essential information that may help to better understand the disease and its potential therapies.

Monoclonal Gammopathy of Unknown Significance

MGUS has traditionally been considered a benign or a premalignant condition in which only a small proportion of patients will progress to multiple myeloma or related diseases (see Table 79.1). In MGUS, the monoclonal protein is ≤3 g/dL and the bone marrow clonal plasma cells are <10% with no related organ damage. This condition is likely much more common than initially thought, as it has been documented in 3% of the overall population and 5% in those over the age of 70 (70). The risk of transformation to myeloma and related diseases (such as amyloidosis or Waldenstrom's macroglobulinemia) has been estimated at 1% per year, based on a 30-year follow-up of 1,384 patients at the Mayo Clinic (69).

Asymptomatic Multiple Myeloma (Smoldering Myeloma)

Asymptomatic multiple myeloma represents an intermediate form of myeloma whereby patients meet serological monoclonal protein and bone marrow criteria for the diagnosis of myeloma (in excess of 10% clonal plasmacytosis) but have yet to develop evidence of end organ damage (see Table 79.1). These patients are not significantly anemic, do not have renal insufficiency, and do not have bony disease. Although the risk of transformation to multiple myeloma is much higher than in MGUS (5–10% per year), some patients may “smolder” for many years. These patients generally do not require therapy but should be followed closely to monitor for progression.

Pathophysiology

Multiple myeloma arises from malignant transformation of a late-state B cell. Although the seminal event has yet to be defined, one of the earliest genetic events is the illegitimate switch recombination of partner oncogenes into the immunoglobulin heavy (IgH) chain (65). Other events may occur such as cytogenetic hyperploidy and upregulation of cell cycle control genes. The result of these genetic abnormalities is the development and propagation of a clonal population of B cells within the bone marrow; this, however, is common and can be seen in up to 5% of the general population over the age of 70 (70). Most of these will not go on to develop myeloma, so there must be additional events to create the malignant phenotype of multiple myeloma. These secondary events may include mutations of kinases, deletions of chromosomes, and up-regulation of enzymes such as c-myc (43). Having sustained a secondary event, the malignant plasma cells begin to proliferate in the bone marrow microenvironment, producing monoclonal proteins and causing osteolytic bone disease. The slow accumulation of these malignant cells gradually results in the characteristic clinical features of myeloma of anemia, bone resorption, hypercalcemia, renal failure, and immunodeficiency. Established myeloma is dependent and sustained on a number of microenvironment features, including the bone marrow stroma itself and the cytokines interleukin-6 and insulinlike growth factor-1 (54). The bone disease that arises in myeloma appears to be mediated in part by RANK ligand/osteoprotegerin and the Wnt-signaling antagonist dickkopf 1 (DKK1) (114).

P.1791

Clinical Presentation

Solitary Plasmacytomas

The median age at diagnosis of solitary plasmacytoma (SP) is 55 to 65 years, on average about 10 years younger than patients with multiple myeloma (90,110,117). Males are affected predominately (male:female ratio 2:1) (90). A diagnosis of SP is made if all the following criteria are satisfied at presentation: a histologically confirmed single lesion with negative skeletal imaging outside the primary site, normal bone marrow biopsy (<10% monoclonal plasma cells), and no myeloma-related organ dysfunction (37). A monoclonal protein is present in 30% to 75% of cases (particularly for an osseous presentation), the level is usually minimally elevated (IgG <3.5 g/dL, IgA <2.0 g/dL, and urine monoclonal kappa or lambda <1.0 g per 24 hours) (37,121).

The disease more commonly presents in bone (80%). Such cases are considered stage I multiple myeloma according to the Durie and Salmon (38) staging system. The most common location is the vertebra (90). Patients with bone involvement often present with pain, neurologic compromise, and occasionally pathologic fracture. A lytic lesion is typical, with or without adjacent soft tissue mass. Less commonly SP presents in an extramedullary site (20%), usually as a mass in the upper aerorespiratory passages that produces local compressive symptoms (4,90,110,116). The histologic diagnosis of extramedullary plasmacytoma (EMP) can be difficult, with the main differential diagnosis being extranodal marginal zone lymphoma (MALT type), where there can be extensive infiltration by plasmacytoid cells (4,58).

Multiple Myeloma

Bone pain and symptoms due to anemia, such as easy fatigability, are the most common (68). Because of the myriad effects of the disease, other insidious symptoms can result from a combination of hypercalcemia, renal impairment, infection, neurologic compression, and occasionally, hyperviscosity. Bone disease manifesting as generalized osteopenia and multiple lytic bone lesions can frequently lead to pathologic fractures. In the vertebral column, this often results in a diminished height. Sclerotic lesions at presentation are rare.

Laboratory evaluation generally confirms anemia, high erythrocyte sedimentation rate, and a variable degree of granulocytopenia and thrombocytopenia. An abnormal monoclonal immunoglobulin (M-protein) in the blood and/or urine is characteristic (68), most commonly immunoglobulin G (IgG) or immunoglobulin A (IgA). Biclonal disease is also recognized, and, rarely, nonsecretary disease. Occasionally only monoclonal light chains are detected. It is important to assess for hypercalcemia, renal dysfunction, and integrity of the skeleton because these complications require appropriate management. A constellation of polyneuropathy, organomegaly, endocrinopathy, M-protein, and skin changes characterize a rare plasma cell dyscrasia known as polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy, and skin changes (POEMS) syndrome (33,89).

Plasma Cell Leukemia

Plasma cell leukemia is a very rare variant of multiple myeloma, where the proliferation of plasma cells is not confined to the bone marrow but may be detected in the peripheral blood. It carries a very poor prognosis with median survival of only 3 to 6 months (123). There is currently no standard therapy for this condition, but patients are usually treated with high-dose, multiagent chemotherapeutic regimens or with experimental therapies.

Diagnostic Work-Up and Staging

The recommended tests for the diagnosis of plasma cell neoplasms are outlined in Table 79.2. The most important components relate to the measurement and quantification of the M-protein, bone marrow examination with ancillary

P.1792

studies, serum β2 microglobulin and albumin, and diagnostic imaging of involved bony sites. The M-protein should be measured with serum protein electrophoresis (SPEP). Quantification of the monoclonal immunoglobin with immunofixation techniques is also acceptable and especially useful if the M component is at a low level. If no M-protein is detectable, assays for free light chains should be performed in the serum and in the urine (Bence-Jones proteinuria). The standard imaging is the skeletal survey, as radionuclide bone scan usually does not detect lytic disease and has limited value (37). For localized areas of concern, both computed tomography (CT) or magnetic resonance imaging (MRI) should be liberally utilized. MRI is preferred to assess the extent of vertebral disease and the presence of spinal cord or nerve root compression. With advances in diagnostic imaging, it is likely that “stage migration” has occurred (41). It has been documented that some patients with presumed solitary plasmacytoma of bone will be upstaged following the detection of multiple vertebral lesions or bone marrow disease by MRI (74,76,118) or by 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) (105). The staging criteria for the widely used Durie and Salmon staging system are detailed in Table 79.3 (38). The newer International Myeloma staging system is simple, validated, and of importance particularly for present and future clinical trials (see Table 79.3) (50). Criteria for the diagnosis of MGUS and asymptomatic (smoldering) myeloma are also well established (37,51).

Prognostic Factors

Solitary Plasmacytoma

Age is a factor affecting the risk of progression to myeloma in some series (14,24,117) but not in others (20,56,77,90,106). A bony presentation has been consistently demonstrated to have a significantly higher risk of subsequent development of myeloma with a 10-year rate of 76%, compared with an extramedullary presentation where the 10-year rate was 36% (Fig. 79.1) (90). Subclinical bone disease, either detected as generalized osteopenia (45) or through abnormal MRI scan of the spine (76,86,118), predicts for rapid progression to symptomatic multiple myeloma. A suppression of the normal immunoglobulin classes, also known as immunoparesis, has been shown to correlate with a higher risk of progressing to myeloma (45,59). Where there was an elevation of M-protein pretreatment, the persistent of the M-protein following radiation therapy (RT) predicts for progression to myeloma (74,121). Many of these factors reflect the presence of occult myeloma. Therefore, it is not surprising that generalized disease becomes manifest once the local disease is controlled. Pathologic factors have been examined in some studies, with the finding that anaplastic plasmacytomas (those with a higher histologic grade) (112), and those tumors expressing a high level of angiogenesis (67) are associated with a poor outcome. Anaplastic plasmacytomas share some common pathologic and clinical features with aggressive B-cell lymphomas (plasmablastic type) and can arise in the context of immunosuppression and Epstein-Barr virus infection (28,42).

With respect to local control, tumor bulk appears to be an important unfavorable factor. Tumors <5 cm achieved a high level of local control with 35 Gy, whereas those >5 cm had a local failure rate of 58% (7/12 patients, total dose range 25 to 50 Gy) (117). The importance of tumor bulk is also supported by other studies (56,77,90).

Multiple Myeloma

Univariate analysis of over 1,000 patients evaluated at the Mayo Clinic revealed the following adverse prognostic risk factors: Eastern Cooperative Oncology Group performance status 3 or 4, serum albumin <3 g/dL, serum creatinine ≥2 mg/dL, platelet count <150,000/µL, age ≥70 years, β2 microglobulin >4 mg/L,

P.1793

plasma cell labeling index ≥1%, serum calcium ≥11 mg/dL, hemoglobin <10 g/dL, and bone marrow plasma cell ≥50% (68).

A new International Staging System has been validated to assist in prognostication (50). Over 10,000 patients were evaluated, and the three-stage system was developed based on two variables: serum albumin and β2 microglobulin (see Table 79.3). In addition to stage, the other area emerging as important to prognosis is cytogenetics. Much like acute leukemia, cytogenetic and molecular features are influencing treatment options. Some abnormalities demonstrated to carry a poorer prognosis include: deletion of chromosome 13 (39), presence of the t(4;14) translocation (25), and p53 deletion (92). It is expected that additional cytogenetic and molecular features of prognostic significance will be identified, especially with enhanced techniques such as fluorescence in situ hybridization and gene microarray analysis.

Management of Solitary Plasmacytoma

RT is the standard treatment for solitary plasmacytoma. Surgery should be considered for structural instability of bone or rapidly progressive neurologic compromise such as spinal cord compression (37,109,111). For patients treated with gross tumor excision, RT is still indicated due to a high likelihood of microscopic residual disease. Surgery alone without RT leads to an unacceptably high local recurrence rate (90). A review of the literature for solitary bone plasmacytoma in Table 79.4 indicates a high local control rate with RT (79% to 95%), yet a modest overall survival of approximately 50% at 10 years. This is due to a high rate of progression to multiple myeloma in the bone plasmacytomas, a finding consistently reported from all series (see Fig. 79.1) (14,24,44,45,56,59,63,90,117,121). As shown in Table 79.4, over 60% of patients with solitary bone tumor progressed to myeloma, at a median of 2 to 3 years after treatment. When actuarial methods were not used, the progression rate is slightly lower (crude rates ranges 53% to 54%) (44,56). Therefore, solitary plasmacytoma of the bone appears to be an early form of multiple myeloma. Studies have documented about 29% to 50% of patients with apparent solitary plasmacytoma will have multiple asymptomatic lesions detected in the spine on MRI (76,87,118). Provided that all the other diagnostic criteria for solitary plasmacytoma are satisfied, it is still appropriate to treat with local RT to the presenting site (109). For these patients the risk of developing symptomatic myeloma in a short time is high (76,86,118). Chemotherapy can be started at the time of symptomatic progression. The presence of low level M-protein preradiation is extremely common and is not associated with a higher risk of progression to multiple myeloma. However, its persistence following radiation is highly predictive of subsequent systemic failure (31,45,74,121), attesting to the importance of monitoring this as part of posttreatment follow-up.

The addition of adjuvant chemotherapy is theoretically attractive, both in enhancing local control and eradicating subclinical disease to prevent the development of myeloma. One randomized trial suggested a benefit with adjuvant melphalan and prednisone given for 3 years after RT (10). With a median follow up of 8.9 years, those treated with chemotherapy had a myeloma progression rate of 12%, whereas with RT alone it was 54% (10). However, this was a small study and the concerns regarding prolonged use of alkylating agents on the bone marrow (negative effect on stem cell reserve and risk of leukemia) do not justify its routine use.

It has been observed that some patients recur with plasmacytoma(s) of bone or soft tissues, without bone marrow involvement (14,56,64). This is infrequent and the subsequent development of multiple myeloma is high, 75% in one series (14).

In the management of EMP, while complete surgical excision may be curative for small lesions, most patients with larger lesions or those with tumor location not amenable to complete excision should receive local RT. Postoperative RT is indicated for incompletely excised lesions. In contrast to bone plasmacytoma, EMPs are frequently controlled with local radiation (Table 79.5), with a lower rate of progression to myeloma, ranging from 8% to 44% (22,27,46,61,64,75,90,108,110,112,116,122), indicating a significant proportion of patients are cured of their disease. Although the 10-year survival varies widely in the reported literature (range 31% to 90%), the two largest series report 10-year survival rates of 72% (90) and 78% (46). The issue of dose will be discussed later.

P.1794

Management of Multiple Myeloma

A description of therapy of myeloma would not be complete without addressing the need to treat not only the disease itself, but the complications of this disease. Patients often present with both bony disease and anemia—both of these complications are treatable, allowing an improved quality of life. Erythropoietic agents have become the mainstay of anemia management, as have bisphosphonates (15,16,17,35) and local RT for bony disease. Newer surgical techniques such as vertebroplasty and kyphoplasty are also being used to improve back pain and spinal symptoms. Other supportive care interventions being addressed include diet, exercise, and patient support groups.

Initial Treatment of Symptomatic Multiple Myeloma

Patients who have symptomatic multiple myeloma require treatment of the malignant plasma cell clone. Once the decision is made to treat, however, the first step is to determine candidacy for autologous stem cell transplantation (ASCT) (Fig. 79.2). As this modality has become the standard of care for eligible patients, it is necessary to stratify patients initially so that the ability to collect stem cells is not compromised by induction therapy (49).

Patients Eligible for Autologous Stem Cell Transplantation

In patients who are candidates for ASCT, various regimens can be used to induce response prior to stem cell collection. Historically most regimens are high dose and steroid based, either with high-dose dexamethasone alone (3) or with vincristine, adriamycin, and dexamethasone (VAD) (104). An alternative induction is the combination of thalidomide and dexamethasone. Rajkumar et al. (95) demonstrated in 50 patients that this combination yields a response rate of 64% (similar to VAD), without compromising the ability to collect stem cells, but with a rate of deep vein thrombosis of 12% (95). Newer agents that have been validated in the relapse setting, such as bortezomib and lenalidomide, are now being tested as initial therapy with impressive results. Early studies with bortezomib, a first-in-class proteosome inhibitor, have produced response rates of 75% to 100%, with complete remission rates of 20% to 30% (97). Lenalidomide, a derivative of thalidomide, has also been tested in newly diagnosed patients. Rajkumar et al. (96) evaluated the combination of lenalidomide with dexamethasone in 34 patients, with a response rate of 91%.

Patients Not Eligible for Autologous Stem Cell Transplantation

In patients who will not be undergoing a transplant, there are various options available for initial therapy. Most will receive alkylator-based therapy, commonly melphalan and prednisone (MP) (94). This regimen yields partial remissions in approximately 55% of patients, with the occasional complete response. Recent trials have evaluated the addition of thalidomide to melphalan and prednisone (MPT). For patients aged 60 to 85, Palumbo et al. (91) demonstrated a 76% response rate with MPT, superior to the 48% in the MP arm; however, thromboses were more common with thalidomide with an incidence of 12% (vs. 2% in the MP group). Newer agents are now being incorporated into trials in this population as well, including bortezomib and lenalidomide. The precise role of these agents in older patients has yet to be determined.

Autologous Stem Cell Transplantation

ASCT has become the standard of care for eligible patients, as it has been demonstrated in multiple trials to improve the likelihood of complete response, prolong disease-free survival, and

P.1795

extend overall survival (9,19,66). Treatment-related mortality rates are now <2%, and often the transplant can be performed entirely as an outpatient. Melphalan 200 mg/m2 is the most commonly used conditioning regimen, although it may be reduced in elderly patients or patients with renal insufficiency.

Tandem Transplantation

Tandem or double transplantation refers to a planned second ASCT after the patient has recovered from the first. A phase III trial in France evaluated tandem transplant versus single ASCT and demonstrated superior overall survival in the tandem group (8); however, when further analyzed, the patients who benefited most from the second transplant were those who did not achieve a 90% reduction in their disease after the first ASCT. Therefore, it may be more prudent to consider tandem transplantation only in patients whose response to the first ASCT is suboptimal.

Allogeneic Stem Cell Transplantation

Myeloablative stem cell transplant is perhaps the only current potential cure for patients with myeloma, as the graft is not contaminated with tumor cells and may produce a profound graft versus myeloma effect (79). However, its use is very limited due to the lack of donors, age restriction, high treatment-related mortality, and graft versus host disease.

The general approach to myeloma is to provide sequential therapies to patients, knowing each will not be curative but will prolong the period of disease control. The goal is to convert the disease into a chronic illness. Whereas there used to be very limited treatment options, the armamentarium available has grown considerably over the past few years. This has contributed to a prolongation of the median survival of patients with myeloma. Patients will relapse after a median of 2 years after first ASCT (81), and several options may be pursued for treatment (Table 79.6). The most exciting is the development and availability of novel, biological agents. Bortezomib is the first proteosome inhibitor to be used in clinical trials. Various phase I and II studies have been completed, and a large multicenter phase III trial compared bortezomib to high-dose dexamethasone (99). The updated results of 669 patients revealed time to progression of 6.2 months in the bortezomib arm and 3.5 months in the dexamethasone arm, along with a superior 1-year overall survival (80% vs. 67%) favoring bortezomib (98). Side effects included peripheral neuropathy, cyclical thrombocytopenia, and diarrhea.

Lenalidomide is an immunomodulatory drug derived from thalidomide and is currently undergoing extensive clinical investigation worldwide. Its role has been established for relapsed disease based on a large phase III trial comparing the combination of lenalidomide and dexamethasone with dexamethasone alone (32). The trial was stopped prematurely due to a large difference between the treatment groups favoring the lenalidomide combination arm. Time to progression was 4.7 months with dexamethasone alone and 11.3 months in the combination arm. Overall response rates were 24% (4% complete response [CR] and 20% partial response [PR]) with dexamethasone alone and 59% (17% CR and 42% PR) in the combination group. There was also an overall survival advantage, with median overall survival of 104 weeks in the dexamethasone alone arm, but this end point had not been reached yet in the lenalidomide arm (32). Side effects included neutropenia, thrombocytopenia, and constipation.

Radiation Therapy of Multiple Myeloma

Total Body Irradiation

Some high dose chemotherapy protocols for multiple myeloma incorporate total body irradiation (TBI) into the conditioning regimen. Because of toxicity concerns (mucosal and

P.1796

hematologic) with TBI, many programs use chemotherapy alone, most commonly melphalan. A phase III French study (Intergroupe Francophone du Myelome [IFM] trial 9502) examined melphalan, 200 mg/m2 alone (M200) versus melphalan 140 mg/m2 with TBI, 8 Gy in four fractions (M140/TBI) (84) and found that patients in the TBI-containing arm suffered more grade 3 or 4 mucosal toxicity, heavier transfusion requirement, and longer hospitalization stay. There was a higher toxic death rate in the M140/TBI arm (3.6% vs. 0% for the M200 arm). The event-free survival was no different between the two treatments, but the 45-month overall survival favored the M200 arm (M200: 65.8%, M140/TBI: 45.5%; p = .05) (84).

Similarly, another IFM protocol tested TBI in the tandem transplant setting by intensifying the conditioning regimen for the second transplant to melphalan 200 mg/m2 without TBI and comparing with the standard tandem regimen (M140 for the first, M140/TBI for the second). There was no benefit with TBI, and increased toxicity was again observed. Therefore, all subsequent IFM trials abandoned the use of TBI (52). Another study from the Spanish bone marrow transplant registry compared M140/TBI with three other chemotherapy conditioning regimens (71). There were no significant differences in the hospitalization duration, hematologic recovery, event-free survival, and overall survival among the four regimens. The authors concluded that no one regimen was clearly superior to another.

The Toronto protocol with intensification of the conditioning regimen to melphalan 140 mg/m2, etoposide 60 mg/kg, and fractionated TBI (12 Gy in six fractions over 3 days, with a high dose rate) was used in 100 patients. The main toxicity was interstitial pneumonitis (28% of patients) of whom seven died (1), leading to discontinuation of the TBI in the subsequent protocol. Presently, the use of TBI is based on institutional experience and the specific drug regimen used for conditioning. The tandem transplant program at the University of Arkansas (11,12,30) and the Memorial Sloan-Kettering Cancer Center (57) continue to use TBI in their ASCT programs for specific indica-tions.

Hemibody Radiation

Diffuse bone pain involving wide areas of the skeleton can be effectively palliated by half body radiation with single doses of 5 to 8 Gy (21,78,115), although this is rarely used now. The bone marrow in the unirradiated half body serves as a stem cell reserve and will slowly repopulate the irradiated marrow after treatment. The dose for upper half body should not exceed 8 Gy due to lung tolerance (119). The main toxicity is myelosuppression. The use of hemibody radiation must be carefully considered in patients heavily pretreated with chemotherapy. Growth factor support may be helpful, while transfusions of blood products should be given as needed. The sequential hemibody radiation technique has been used in phase II (102,107) and phase III trials as “systemic” treatment to control myeloma, in patients with or without skeletal pain. A phase III trial by the Southwest Oncology Group (SWOG) included newly diagnosed patients treated initially with chemotherapy, with complete responders randomized to sequential hemibody radiation (7.5 Gy in five fractions, upper hemibody, followed 6 weeks later by lower hemibody) or further chemotherapy. Survival was significantly poorer with radiation compared with chemotherapy (103). At present, there is no standard role for sequential hemibody radiation as systemic treatment for myeloma outside of a clinical trial, although it may remain useful for palliation of advanced disease in chemotherapy-refractory patients.

Local External Beam for Palliation

The most common use of RT in the management of plasma cell tumors is for palliative treatment of bony disease (2,21,73), relief of compression of spinal cord (6,93,120), cranial nerves, or peripheral nerves. It has been estimated that approximately 40% of patients with multiple myeloma will require palliative RT for bone pain at some time during the course of their disease (40). In practice the actual proportion is lower than estimated varying from 24% to 34%, leading investigators in Australia to suggest that this potentially useful modality of treatment has been underutilized, even taking into account the beneficial effect of bisphosphonates, particularly for the elderly (40). Palliative RT to the spine reduces the incidence of future vertebral fractures or appearance of new lesions (72). However, the role of RT in preventing impending pathologic fracture is unclear. In general, lesions at high risk for pathologic fracture should be referred for surgical stabilization, and RT can be administered after surgery for control of residual disease at the local site.

When RT is given for pain due to disease involving a long bone, a local field suffices. It is unnecessary to treat the entire bone (23). Doses of 10 to 20 Gy (in five to 10 fractions) are effective, although the pain relief is often partial (82). With an average dose of 25 Gy given to 306 sites in 101 patients, Leigh et al. (73) found a symptomatic response rate of 97% (complete pain relief in 26%, and partial relief in 71%). There was no dose-response relationship above 10 Gy. Recurrence of symptoms requiring further treatment was seen in 6% of sites after a median of 16 months.

It is not clear if pain relief is better if RT is given concurrently with chemotherapy. A study by Adamietz et al. (2) reported complete pain relief in 80% of patients receiving RT with chemotherapy, compared with 40% among those receiving RT alone. In contrast, Leigh et al. (73) found no significant difference in pain relief when RT was given with or without concurrent chemotherapy. For spinal cord compression, motor improvement is expected in approximately 50% of irradiated patients. A multicenter study suggested that a longer fractionated regimen (30 Gy in 10 fractions or higher) was associated with better neurologic recovery than 20 Gy in five fractions or a single 8 Gy fraction (93). With the availability of newer drugs, the advantage of radiation sensitizing efforts with drug-radiation combinations requires continued investigation, both in terms of enhancing local control (48) and possible toxicity. Bortezomib and spinal radiation given concurrently was reported to result in severe enteritis (83). The use of bisphosphonates (e.g., pamidronate) has been shown to reduce skeletal complications and pain (15,16,17,35) with a reduction of the use of RT from 50% to 34% in one study (16).

Radioimmunotherapy Approaches

Bone seeking radiopharmaceuticals targeting the bone marrow have been studied as an alternative to TBI. Typically a β-emitting isotope is conjugated to a phosphonate complex, such as 153Samarium-ethylene diamine tetramethylene phosphonate (153Sm-EDTMP). The isotope also emits a γ-ray permitting scanning to locate areas of uptake. This agent has been used for palliation of bone metastasis (7,13). The feasibility of this approach in a small number of myeloma patients has been reported for stem cell transplantation both in the autologous (34,55) and allogeneic settings (62). Another bone-seeking pharmaceutical is 166Holmium-DOTMP (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetramethylene-phosphonic acid), with a higher energy β-emission (maximum energy 1.85 MeV) than 153Sm, and a shorter T1/2 of 26.8 hours. It also has a γ-emission (81 KeV) suitable for imaging. A phase I/II study incorporating 166Holmium-DOTMP into a transplant regimen has been performed at the M.D. Anderson Cancer Center with encouraging results (47). With the ability to deliver much higher doses to the bone marrow than TBI, in the range of 30 to 60 Gy, yet sparing the dose-limiting normal tissues such as lung, mucosa, and kidneys, the concept of targeted radiation therapy is tantalizing.

P.1797

However, there remains a problem of heterogeneity of uptake in the skeleton, and the dosimetric variation may be even larger at a microscopic level due to the limited range of the β particle. Whether this approach will have a more favorable therapeutic ratio than standard conditioning regimens in the transplant setting awaits larger scale phase II and phase III trials.

Radiation Therapy Techniques

Radical Radiation Therapy for Local Control of Solitary Plasmacytoma

Accurate evaluation of tumor extent is an important feature of radical RT for solitary plasmacytoma. MRI is useful to evaluate the extent of disease both within and beyond bone. This is particularly true for the paranasal sinuses, where inflammatory changes may be difficult to distinguish from tumor on CT imaging. Currently, the accuracy of FDG-PET in the evaluation of tumor extent is uncertain.

There are few data to support specific guidelines regarding RT treatment volumes. CT and MRI imaging should be used to determine gross tumor volumes (GTV). Clinical target volumes (CTV) should encompass probable routes of microscopic spread, recognizing that barriers to the extension of local disease will vary according to anatomic location, as will the morbidity of treating adjacent normal tissues (Fig. 79.3). For the spine, inclusion of two vertebral bodies above and below the grossly involved vertebra(e) is a common practice. As this is based on relapse patterns seen following RT for spinal metastases for solid tumors rather than plasmacytomas, it may not be directly applicable to solitary plasmacytoma.

For RT of long bone lesions, while coverage of the entire involved bone has been recommended by some authors, a study of palliative RT to only the symptomatic area for multiple myeloma found that recurrence in the untreated portion of the involved bone was rare (23), and similarly, no marginal recurrences were seen among 30 patients with solitary plasmacytoma treated with RT that encompassed only the tumor with a margin (60). Prophylactic regional nodal coverage is not necessary in solitary plasmacytoma of bone as multiple studies have found a very low risk of regional nodal failure after involved-field radiation without intentional coverage of adjacent nodes (i.e., 0% to 4%) (60,75,112,117). For extramedullary plasmacytoma, nodal involvement at presentation is observed in 10% to 20%, and occasional nodal failure in the literature led to a common practice of extending the RT coverage to the draining lymph node region (20,57,110). Some authors specifically recommend this practice if the primary disease involves a lymphatic structure (e.g., lymph nodes, or Waldeyer's ring) (53,64,117). However, this is controversial as some series reported a low incidence of regional nodal failure without routine prophylactic nodal irradiation (53,64,77), leading to variation in practice between centers (110). After reviewing their own series of 26 patients with EMP and contrasting the results with the literature, Strojan et al. (110) concluded that prophylactic nodal radiation is probably unnecessary.

Planning target volumes (PTV) should account for day-to-day setup variation and will typically add 5 to 10 mm around CTV volumes depending on the immobilization technique employed (see Fig. 79.3). Overall, RT field edges are typically 2 to 3 cm from gross tumor seen on imaging. Although parallel-opposed fields are commonly adequate to encompass disease without significant irradiation of normal tissues, CT-based planning, and the use of conformal techniques, including intensity modulated radiation therapy, should be employed when needed to treat the PTV adjacent to critical structures. This can be particularly important in extramedullary disease involving the paranasal sinuses, where avoidance of the optic structures and salivary glands is desirable.

Radiation Therapy Dose

Studies evaluating RT dose response in plasmacytoma have produced differing results. Most studies have found response rates >85%

P.1798

among patients treated with ≥35 Gy; some investigators have found better local control following doses ≥45 Gy (44,116), while others have found no indication of improved outcome with higher doses (60,90). Based on a dose–response analysis of 81 patients by Mendenhall et al. (80) reported in 1980, a minimum dose of 40 Gy was recommended, including osseous and extramedullary lesions. A total dose of 40 Gy and above resulted in a local failure rate of 6% versus 31% for lower doses. Therefore, the usual practice is to administer a dose of 40 to 45 Gy or even higher for bulky tumors. However, in the largest of these studies (n = 258), there was no evidence of improved local control with RT doses ranging from 30 to 50 Gy, including a subset of patients with tumors >4 cm (90). In fact there was a worse local control rate for the group receiving total dose ≥50 Gy, although not statistically significant (90). It should be noted, however, that retrospective studies of dose response are typically confounded by selection bias, as higher doses are prescribed to larger tumors with worse prognosis. Several studies have demonstrated durable local control in >85% of tumors <5 cm with 35 to 40 Gy, and there is little evidence that higher doses are necessary for small tumors, regardless of bone or EMP locations. In contrast, plasmacytomas >5 cm have worse local control (90,117), and doses of 45 to 50 Gy are recommended in these bulkier tumors, which also tend to be EMPs. However, one should be aware that the quality of evidence supporting the use of higher RT doses is limited, and local failures are occasionally observed even after doses exceeding 50 Gy (80,90,117).

Assessment of Response and Follow-Up

Reimaging is of greatest value in the response assessment of extramedullary plasmacytoma. Repeat imaging, preferably MRI, should be done approximately 6 to 8 weeks following completion of treatment. It is rare to have symptoms suggestive of local progression that necessitate reimaging prior to this. It is common for a residual soft-tissue abnormality to persist on follow-up imaging, and periodic reimaging may be required every 4 to 6 months until any residual mass disappears or remains stable on consecutive scans. It is generally not beneficial to continue to reimage a stable abnormality.

Bone destruction caused by tumor can produce persistent abnormalities on imaging following RT for painful bone metastases or isolated plasmacytoma of bone. Consequently, repeat imaging is of less value in establishing response in such cases.

With a high risk of recurrence of disease as multiple myel-oma, the occurrence of new bone pain requires further investigations, including imaging as appropriate. Repeat measurement of the M-protein often detects the onset of systemic disease prior to the development of symptoms and can be used as an indicator of disease burden (26,121). Complete blood counts should be taken periodically to evaluate bone marrow function. A team of international investigators have recently developed recommendations for uniform response criteria for assessing the treatment of multiple myeloma (36).

RT doses used for myeloma are rarely associated with significant delayed side effects. Treatment of significant volumes of the parotid or submandibular glands may result in prolonged xerostomia and should be avoided. As noted previously, TBI has been associated with significant toxicity and is not widely used. Evaluation of renal function should be undertaken prior to initiating RT that may include the kidneys, and blood counts should be evaluated prior to treating a large volume of bone marrow in the spine or pelvis. Reirradiation of vertebral metastases is possible, but careful evaluation of all prior RT records is required to ensure that the tolerance of the spinal cord is not excee-ded.