| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Orthopedic Oncology, Department of Orthopedics, Massachusetts General Hospital, Yawkey 7B-7926, 55 Fruit Street, Boston, Massachusetts 02114
2 Department of Radiation Oncology, Massachusetts General Hospital, Yawkey 7B-7926, 55 Fruit Street, Boston, Massachusetts 02114
3 Hematology/Oncology, Department of Medicine, Massachusetts General Hospital, Yawkey 7B-7926, 55 Fruit Street, Boston, Massachusetts 02114
4 Department of Pathology, Massachusetts General Hospital, Yawkey 7B-7926, 55 Fruit Street, Boston, Massachusetts 02114
5 Surgical Oncology, Department of Surgery, Massachusetts General Hospital, Yawkey 7B-7926, 55 Fruit Street, Boston, Massachusetts 02114
Correspondence: Address correspondence and reprint requests to: Sam S. Yoon, MD; E-mail: syoon{at}partners.org
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Patients with a diagnosis of angiosarcoma treated at our institution between 1980 and 2006 were analyzed for patient demographics, tumor characteristics, multimodality treatment, and outcomes.
Results: A total of 82 patients were divided into those with primary and advanced disease. Overall, the median age was 65 (range, 22–91) years, and 44% of patients were women. Median size of tumors was 3.8 cm, and 76% of tumors were intermediate or high grade. Tumors were located throughout the body: 32 cutaneous, 22 deep soft tissues or organs, 10 radiation or lymphedema field, 8 bone, and 7 nonirradiated breast. Of 46 patients with primary disease, all patients underwent surgical resection, 67% received radiotherapy, and 27% received chemotherapy. Five-year disease-specific survival was 60%, and negative prognostic factors included intermediate or high grade, and tumors arising in a radiated or lymphedema field. Of 36 patients with advanced disease, 36% underwent a palliative operation, 78% received radiation, and 58% received chemotherapy. Median survival was just 7.3 months, and cutaneous tumors predicted a better prognosis compared with other sites.
Conclusions: Primary angiosarcomas treated with aggressive surgical resection and the addition of radiation for close margins or worrisome pathologic features can result in long-term survival in most patients. The role of adjuvant chemotherapy is unclear. Patients with advanced disease have a poor prognosis, but there can be dramatic responses to chemotherapy in a minority of patients.
Key Words: Sarcoma • Angiosarcoma • Outcome • Radiotherapy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Angiosarcomas are a relatively rare histological subtype of sarcomas and represent <1% of all sarcomas. Despite their rarity, angiosarcomas display remarkable clinical heterogeneity. These tumors can occur in any location in the body, and one review found one-third to be cutaneous (primarily in head and neck), one-fourth to be in soft tissue, and the remainder at miscellaneous sites such as the breast, liver, and bone.3 Most angiosarcomas present as intermediate- or high-grade lesions. Angiosarcomas also occur in special clinical circumstances including developing in a previously irradiated field4 and in a chronically lymphedematous extremity (Stewart-Treves syndrome).5 Other characteristics that are somewhat unique to angiosarcomas include a propensity for multifocal disease6 and an ability to spread to lymph nodes and distant soft tissues.7
There are few contemporary large clinical series of patients with angiosarcomas. In this study, we sought to examine the clinical presentation, tumor characteristics, treatment, and outcome of 82 patients with angiosarcoma treated at a single institution over 25 years.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All patients were evaluated by a complete history and physical examination. Radiological imaging included a computed tomographic scan or magnetic resonance imaging of the primary site along with a chest computed tomographic scan. Data of the cases, including radiological imaging and pathology slides, were reviewed at the weekly Massachusetts General Hospital multidisciplinary sarcoma conference, which was attended by orthopedic and surgical oncologists, radiation oncologists, medical oncologists, pathologists, and radiologists, all with a clinical interest in sarcomas. Treatment plan was individualized for each patient on the basis of a consensus reached after open-format discussion.
Clinical information regarding patients was obtained from the institution’s medical records and the Massachusetts General Hospital prospective sarcoma databases. Recorded patient data included age, sex, symptoms, and risk factors. Characteristics of the primary tumor were recorded and included size, site, grade, and depth. The maximum single dimension of a tumor was used to determine size. Tumors originating in the dermis were considered cutaneous, and tumors below the dermis were considered deep tumors. Breast tumors, bone tumors, and tumors originating in a previously irradiated or lymphedematous field were analyzed separately. A tumor arising in a previously irradiated breast was considered in the irradiated or lymphedematous field category. All biopsy and surgical specimens were analyzed by experienced sarcoma pathologists (A.P.R, G.P.N, V.D.), who determined grade (I = low, II = intermediate, III = high) and margin of resection (microscopically negative, microscopically positive, or grossly positive). Grossly positive margin was defined as a resection in which the surgeon thought that tumor cells were left unresected. Microscopically positive was defined as the presence of tumor cells at the inked specimen margin. Information regarding radiotherapy was analyzed for all patients. Chemotherapy information was available for most patients.
In general, patients were examined at follow-up at least every 3 months for the first 2 years and every 6 months from year 2 through year 5. Follow-up beyond year 5 was performed at the discretion of the treating physician. Imaging studies of the primary site and possible metastatic sites were obtained at the discretion of the treating physician. The location and time of initial local and distant recurrences were recorded. Cutaneous recurrence near the site of the primary tumor was considered a local recurrence, even if it was not contiguous with the primary tumor site. Lymph node metastases were considered distant recurrences.
Statistical Analysis
Local recurrence-free survival, distant recurrence-free survival, and disease-specific survival were determined for patients with primary disease, and disease specific survival was determined for patients with advanced disease. Survival curves were estimated by the Kaplan-Meier method.8 Clinical variables were associated with survival by the univariate Cox proportional hazard model.9
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Forty-four percent of patients were female, and 56% were male. Nine patients (11%) had their angiosarcoma arise in a previously irradiated field or in a lymphedematous extremity.
|
). Twenty patients (61%) presented with an asymptomatic mass or lesion. Thirteen patients (39%) presented with symptoms: 10 patients had pain or discomfort associated with a mass or lesion, and one patient each presented with a neuropathy, anemia and weight loss, and a pathologic fracture. Most patients with a cutaneous angiosarcoma presented with a painless discolored skin lesion or superficial nodular lesion, but a few had pain, bleeding, or ulceration.
Tumor Characteristics
Table 2 lists the tumor characteristics for primary and patients with advanced disease. The median size of all primary tumors was 3.8 cm. Overall, half of tumors were <5 cm in size, with nearly two-thirds of tumors in patients with primary disease being <5 cm and only a third of tumors in patients with advanced disease being <5 cm. Ten percent of tumors were >10 cm in size
|
). Two patients presented with widespread metastatic disease and had no identifiable primary tumor. Of the 33 cutaneous lesions, 20 were on the scalp, 7 were on the face, 4 were on the extremities, and 2 were on the trunk. Of the 22 deep soft tissue or organ sarcomas, nearly half were located in the extremity (n = 10); other locations included the trunk (n = 3), heart (n = 3), head/neck (n = 2), liver (n = 1), spleen (n = 1), and adrenal gland (n = 1). Ten angiosarcomas occurred in a radiated or lymphedema field. Six of these lesions occurred as breast or chest wall lesions 5–11 years after surgery and radiation for breast cancer, and one occurred in the pelvis 22 years after radiation for a reticulum cell sarcoma. Three patients developed an angiosarcoma in a chronically lymphedematous arm 9, 18, and 30 years after treatment of breast cancer. There were eight primary bone tumors and seven primary breast tumors. Two patients developed angiosarcomas in preexisting benign lesions. One patient developed an adrenal angiosarcoma in the region of a long-standing lymphangioma, and one patient developed an axillary angiosarcoma in the region of a brachial plexus neuroma. There were eight primary bone angiosarcomas and six primary breast angiosarcomas. Three-quarters of angiosarcomas were intermediate or high grade; only 11 patients had low-grade lesions.
Treatment of Patients with Primary Disease
All 46 patients with primary disease or isolated local recurrence underwent complete surgical resection (Table 3). Margins were negative in 34 patients (74%), microscopically positive in 5 patients (11%), and grossly positive in 6 patients (13%). Two-thirds of patients received adjuvant radiation. Data on chemotherapy were available in 37 patients. Twenty-seven percent of patients received adjuvant chemotherapy, and 73% of patients did not. Chemotherapy regimens included primarily doxorubicin-based regimens or vinorelbine.
|
Only one patient had a local recurrence, and this patient also had distant metastases and died 4 months after surgery. Three patients had distant recurrence only and died 6 months, 8 months, and 30 months after surgery. Three patients with no recurrence died of other causes 5 months, 29 months, and 5 years after surgery. One patient had no evidence of disease at 1 year.
For the seven nonscalp cutaneous lesions, five were on the face and two were on the buttock. Sizes ranged from 1.9 to 10 cm. All were intermediate or high grade except for one low-grade lesion on the eyelid. All lesions were excised with negative microscopic margins. Adjuvant radiation was delivered to five patients. Two patients had a local recurrence 14 and 48 months after resection; they underwent additional salvage treatment. One patient had a distant recurrence 3 months after resection and died of disease. At last follow-up, three patients had no evidence of disease, and three patients had died of other causes.
Angiosarcoma of Deep Soft Tissues or Organs
Twelve patients with deep soft tissue or organ angiosarcoma were treated with resection; 10 patients had negative margins and 2 patients had microscopically positive margins. Radiation was delivered to eight patients. Five patients were alive without disease 30–87 months after surgery. Two patients had a local recurrence 8 and 33 months after surgery, and five patients had a distant recurrence 1–55 months after surgery. Five patients died of disease at 3–64 months. One patient was alive with disease at 18 months, and one patient died of other causes at 108 months.
Cutaneous Angiosarcoma Associated with Radiation or Lymphedema
Six patients had tumors arising in a previously irradiated field and two in a lymphedematous extremity. Tumors were between 1.5 and 11 cm in size, and all were intermediate or high grade. All patients were treated with aggressive surgical resection. Five patients who developed angiosarcomas after breast surgery and radiation underwent resection of the skin and underlying subcutaneous tissue and muscle. Chest wall resection and reconstruction was performed if the tumor extended to the chest wall. One patient with a radiation-induced sarcoma of the pelvis underwent hemipelvectomy. Of the two patients who developed angiosarcoma in a lymphedematous extremity, one had a wide local excision and one had a forequarter amputation. Overall, margins were microscopically negative in six of eight patients.
Four patients were able to receive additional Radiotherapy. Repeat radiotherapy was considered in patients with a long interval since receipt of previous radiation, if tumor pathology showed worrisome pathologic features such as multiple tumors or satellite nodules distant from the dominant lesion, if wide surgical margins could not be achieved, or if previously unirradiated tissue could be brought in for flap coverage after resection. In previously irradiated patients, consideration was given to delivering radiation with specialized techniques such as intraoperative radiation, brachytherapy, or electrons to allow localized additional dose delivery.
Overall prognosis in these patients was poor. Six patients had local recurrence between 1 and 63 months, and six patients had distant recurrence between 1 and 77 months. At last follow-up, five patients had died of disease 6 months to 7 years after surgery, two patients were alive with disease 4 months and 4 years after surgery, and only one patient was alive without disease 5 years after surgery.
Bone Angiosarcoma
Five patients had bone angiosarcomas that were resected. Two of these were located in the humerus, and there was one each in the clavicle, spine, and pelvis. Three lesions were high grade, two were low grade, and one tumor grade was not reported. Surgical margins were negative in three patient and microscopically or grossly positive in three patients. Radiation was delivered to all three patients with positive margins. No patient had a local recurrence, but one patient had a distant recurrence and died of disease 5 months after surgery.
Breast Angiosarcoma
Six patients had primary breast angiosarcoma, and all underwent mastectomy with negative margins. Four patients received adjuvant radiotherapy. One patient had a local recurrence, and three patients had a distant recurrence 2–45 months after surgery. Two patients died of disease 11 months and 5 years after surgery, one patients died of other causes 14 years after surgery, and four patients were alive without disease 5, 23, and 25 years after surgery.
Recurrence and Survival in Patients with Primary Disease
Local recurrence occurred in 12 patients (26%) at a median interval of 11 (range, 1–63) months (Fig. 1A). Median local recurrence-free survival was not reached. Univariate analysis of patient, tumor, and treatment data was performed (Table 4), and site of tumor in a radiation or lymphedema field was statistically significantly associated with local recurrence (Fig. 1B). Patients with tumors arising in a radiated or lymphedema field had a median local recurrence-free survival of 18.1 months; the median local recurrence-free survival was not reached for other sites. Intermediate- and high-grade tumors had a higher rate of local recurrence than low-grade tumors, but this did not reach statistical significance (Fig. 1C). Interestingly, size of tumor and margin status were not associated with local recurrence.
|
|
). Median distant-recurrence free survival was 55 months. Univariate analysis of factors associated with distant recurrence was performed (Table 4), and positive resection margin was associated with distant recurrence (Fig. 2B). Other studies have also found that locally infiltrative sarcomas are associated with decreased survival.11 Patients with a positive microscopic or gross margin had a median distant disease-free survival of 8 months; the median distant disease-free survival was not reached in patients with negative microscopic margins. Patients with intermediate- or high-grade tumors or tumors arising in a radiation or lymphedema field tended to have a higher risk of metastasis, but this was not statistically significant (Fig. 2C, D).
|
). Univariate analysis (Table 4) found that intermediate- and high-grade tumor (Fig. 3B) and location in a radiation or lymphedema field (Fig. 3C) were associated with worse disease-specific survival. Size and site of tumor were prognostic of disease-specific survival.
|
Thirteen patients underwent palliative operations, which included operation at the primary site in five patients, operation at a distant site in four patients, and operation at both primary and distant sites in four patients (Table 5). Of the nine patients who presented with only advanced tumors at the primary site and no metastatic disease, palliative surgical resection was performed in four patients (two scalp, one face, one brain). All these resections were considered noncurative. The five other patients who did not undergo surgical resection had either extensive scalp lesions (n = 4) or a cardiac tumor (n = 1). All nine of these patients received radiotherapy.
|
Of the 22 patients who presented with both local and metastatic disease, nine palliative operations were performed for disease at the primary site: seven were resections at the primary site, one was a laminectomy for spinal tumor, and one was an intramedullary nailing of a femur lesion with pathologic fracture. Seven palliative procedures were performed for distant disease: three brain metastasectomies, two spine decompressions and/or stabilizations, one pericardial resection, and one soft tissue resection. Seventeen of these 22 patients received radiotherapy at some point. Nine patients received radiation to the primary tumor site, three patients received radiation to metastatic sites (two to the spine and one to the brain), and five patients received radiation to both the primary tumor site and metastatic sites (two for brain; and one each for soft tissue, bone, and lung).
Information regarding chemotherapy was available on 31 of the 36 patients with advanced disease, and 21 patients received some form of chemotherapy. Patients declined chemotherapy primarily because of older age or complicating medical conditions. The initial chemotherapy treatment was doxorubicin-based chemotherapy in 12 patients. MAID (mesna, doxorubicin, ifosfamide, dacarbazine) was the preferred regimen and used in six patients. Doxorubicin was used alone in three patients and combined with dacarbazine in two patients. CYVADIC (cyclophosphamide, vincristine, doxorubicin, and dacarbazine) was used in one patient. Other chemotherapy agents included vinorelbine, which was used as an initial regimen in three patients. Palliative responses were common, with 64% of patients having at least a partial response to one or more chemotherapy regimens, but responses were generally short-lived.
Survival in Patients with Advanced Disease
At time of last follow-up, 32 of 36 patients had died of disease, 1 patient had died of non–sarcoma-related causes, and 3 patients were alive with no evidence of disease. The median survival for these patients was 7.3 months (Fig. 4A). Univariate analysis (Table 6) found that only cutaneous tumor location was associated with improved survival (Fig. 4B). Chemotherapy was not associated with improved survival in this retrospective analysis, but the only long-term survivor did receive chemotherapy. This single patient presented with a multifocal scalp angiosarcoma along with lymph node and brain metastases. She initially underwent resection of her symptomatic brain metastases, followed by whole-brain radiation. She then underwent six cycles of MAID chemotherapy followed by radiation to the metastatic neck lymph nodes and primary scalp lesions. She currently has no evidence of disease 10 years after diagnosis. The other two patients alive without disease underwent operative resection of all gross disease and had short follow-up times (<6 months).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Patients with low-grade tumors had better survival. Thirty-six patients presented with advanced disease. One-third of these patients had palliative operations, three-quarters had radiotherapy, and over half had chemotherapy. Despite multimodal treatment in many patients with advanced disease, median survival was only 7.3 months.
|
Chronic lymphedema and chronic inflammation may lead to the development of angiosarcomas. Classically, this has been seen patient with a history of breast cancer after mastectomy in a lymphedematous arm (Stewart-Treves syndrome).15 This series had three patients with this syndrome. The development of angiosarcoma has also been described after chronic lymphedema to be due to congenital defects, trauma, and infections, including filarial infection.16 Several case reports also describe angiosarcoma developing near defunctionalized arteriovenous fistulas in patients with renal transplants17 as well as adjacent to foreign-body material.18
Angiosarcomas have been described to occur in preexisting benign vascular lesions (portwine stain and lymphangioma) as well as benign and malignant nerve sheath tumors, neurofibroma (in the setting of neurofibromatosis), leiomyoma, spindle cell hemangioma (in the setting of Maffucci syndrome), retinoblastoma (in the setting of Rb1 deletion), Klippel-Trenaunay syndrome, xeroderma pigmentosum, malignant germ cell tumor, herpes zoster lesion, and Aicardi syndrome.19 This series had two cases of angiosarcomas arising from a preexisting benign lesion (lymphangioma and neuroma). Several environmental exposures are specifically associated with angiosarcomas. Thorium dioxide (Thorotrast) was used in the past for cerebral angiography, vinyl chloride was used in the production of synthetic rubber, and arsenic was used in pesticides; all are associated with hepatic angiosarcomas.20,21 The one patient with liver angiosarcoma in this study had no history of such exposures.
Diagnosis
The histopathological features angiosarcomas can be quite diverse, ranging from well-differentiated to poorly differentiated lesions.19 Low-grade, well-differentiated lesions are usually composed of vascular channels that had irregular contours and grow with an infiltrative fashion. They frequently use preexisting collagen as scaffolding as they dissect through tissue. The neoplastic endothelial cells were plump and had hyperchromatic nuclei; mitoses were infrequent. High-grade or poorly differentiated tumors are generally densely cellular neoplasms with only focal vascular channel formation. The neoplastic cells range from spindled or epithelioid and have large, irregular nuclei with coarse chromatin. The cytoplasm is eosinophilic. These tumors have irregular margins and frequently destroy the tissue within which they arise. Mitoses and necrosis are commonplace.
In particularly poorly differentiated tumors, the diagnosis can be confirmed with immunohistochemistry and electron microscopy. By immunohistochemistry, these tumors are usually positive for factor VIII-related antigen, CD34, and CD31. Ultrastructural features indicative of endothelial differentiation include a partial investiture of basal lamina along antiluminal borders, tight junctions between cells, pinocytotic vesicles, a close relation between tumor cells and erythrocytes, and ramifying clefts. Weibel Palade bodies, when present, are virtually diagnostic of endothelial differentiation.
Although angiosarcomas likely arise from either blood or lymphatic vessels, it is generally not possible to determine a blood vessel angiosarcoma (aka hemangiosarcoma) from a lymphatic vessel angiosarcoma (aka lymphangiosarcoma) origin.19 The term lymphangiosarcoma is often used when an angiosarcoma arises in the setting of lymphedema, but most angiosarcomas arising in this setting are indistinguishable from other angiosarcomas.
Presentation
A comparison of this series to other relatively large, recent series is shown in Table 7. Review of these series demonstrates that the median age ranges from 60 to 71 years, and the sex distribution is roughly equal. Symptoms and signs at presentation depend on the location of the tumor. Cutaneous lesions often present ill-defined, bruiselike lesions that can progress to nodular and ulcerated lesions.19 Noncutaneous tumors can present as an asymptomatic mass. The most common presenting symptom is pain or discomfort. Median size ranges 3–6 cm, and more than three-quarters of tumors are intermediate or high grade. Metastases are most common to the lung and liver, with other less common sites including lymph nodes, soft tissues, and bone.
Cutaneous Angiosarcomas
Cutaneous angiosarcoma are notorious for ill-defined margins and multifocality, making complete resection difficult. They primarily affect elderly persons and are usually located on the head and neck, especially the scalp. In the data from the Armed Forces Institute of Pathology of 101 cutaneous angiosarcomas, 52% occurred in the head and neck and 21% occurred in an extremity.19 In a series of 72 patients from the United Kingdom, median survival was only 15 months, and only 12% of patients survived beyond 5 years.22 Size <10 cm was associated with better prognosis, and radiation seemed to help with local control. Pawlik et al.10 reported on the University of Michigan experience with 29 scalp angiosarcomas over a 27-year period. Nearly all patients underwent surgical excision, but negative margins were achieved in only 21% of patients. Median survival was 28 months, and younger age, tumor size <5 cm, and radiotherapy were associated with improved survival. The M. D. Anderson Cancer Center reported their series of 14 head and neck cutaneous angiosarcomas over 19 years. Eleven patients had multifocal disease. The use of radiation combined with surgery seemed to improve local control, but 63% of patients developed distant metastases.
In our series, there were eight resected scalp angiosarcomas. Negative margins were achieved in only three patients, and all patients received adjuvant radiation. Only one patient experienced local recurrence, but four patients experienced distant recurrence.
Angiosarcoma of Deep Soft Tissues or Organs
Angiosarcomas arising in deep soft tissues or organs may have distinct differences compared with cutaneous angiosarcoma, with a wider age distribution. In a review of 80 patients with soft tissue angiosarcoma by Meis-Kindblom and Kindblom,7 these lesions occurred most often in the seventh decade of life and primarily in the extremities (54%) and trunk (31%). Median size was 5 cm, and there was a wide morphologic spectrum. Fifty-three percent of patients died at a median of 11 months, and 31% were alive without disease at a median of 46 months. Local recurrence occurred in 20% and distant recurrence in 49%, most commonly to the lungs followed by lymph nodes, soft tissues, bone, and liver. Poor prognostic factors for survival included older age, retroperitoneal location, and large tumor size.
Of the nine patients with resected soft tissue angiosarcomas in this series, surgical specimens showed negative margins in seven cases. Seven patients had adjuvant radiation, and there were two local recurrences. Four patients had distant recurrences. At last follow-up, four patients had died of disease.
There are numerous case reports and small series of angiosarcomas arising in solid organs including the heart,23 liver,24 spleen,25 and adrenal gland.26 Angiosarcomas arising in solid organs often have a poor prognosis. In one series of 28 splenic angiosarcomas, only one patient survived without disease for 10 years.25 The one patient with splenic angiosarcoma in this series had metastatic disease at presentation. The one patient with liver angiosarcoma had diffuse, multifocal bilateral liver lesions. The one patient with an adrenal angiosarcoma died 14 months after resection. Cardiac angiosarcomas often present difficulties in terms of local resection and adjuvant radiation.27 There were three patients in this series with cardiac angiosarcomas. One patient underwent piecemeal resection and is currently receiving chemotherapy 18 months after surgery. Another patient presented with bilateral lung metastases and died 2 months after diagnosis. The third patient was thought to have an unresectable primary tumor and underwent proton beam radiation and chemotherapy. He was alive at 5 months after diagnosis.
Cutaneous Angiosarcoma Associated with Radiation or Lymphedema
Monroe et al.14 found approximately 100 cases of angiosarcoma after breast-conservation therapy in their literature review. Fifty-five (73%) of 75 patients recurred after surgery, usually within 1 year. Local recurrences occurred in the tumor bed or along the mastectomy scar, and very few reports of surgical salvage treatment after local tumor recurrence were reported. Distant metastases usually developed at the time of or after local recurrence, and the authors thought that survival rates may be improved if local control could be attained. Adjuvant radiation may also reduce the risk of local recurrence. The authors recommended that these patients undergo preoperative radiation followed by aggressive surgical resection with the placement of flaps as needed.
There have been approximately 300 cases of Steward-Treves syndrome reported worldwide.28 These patients usually present with purplish, multi-centric lesions in a lymphedematous extremity, often have aggressive disease, and are thought to have a poor prognosis. Others have reported some long-term survival in small series.28
We found that patients who developed angiosarcomas in radiation or lymphedema fields were far more prone to local and distant recurrence and death from disease. Of the eight patients who had primary disease, six patients had local recurrence between 1 and 63 months, and six patients had distant recurrence between 1 and 77 months. At last follow-up, five patients died of disease 6 months to 7 years after surgery, two patients were alive with disease 4 months and 4 years after surgery, and only one patient was alive without disease 5 years after surgery.
Angiosarcoma of Bone
Angiosarcoma of bone is rare and accounts for <1% of primary skeletal malignancies; it has few reports in the literature.29 There can be two different patterns of presentation.29,30 One pattern is a tumor presenting as multiple lesions in a single bone, or multiple lesions presenting in adjacent bones or even all the bones of a limb. Tumors with this presentation tend to have a more indolent course. The second pattern is a rapidly progressive solitary lesion that metastasizes quickly to the lung or to other distant bones.
Radiographically, there is no specific finding associated with angiosarcoma of bone as opposed to other malignant bone tumors.31 They may be solitary or multiple, which may be related to the vascular ancestry of the cells. The lesions are usually highly destructive and may grow too fast to invoke a periosteal reaction. They may be eccentric and may have a purely lytic or mixed lytic-sclerotic pattern. Complete destruction of the cortex with extension into soft tissues may be present in high-grade lesions. Some lesions, particularly of the multicentric type, may have a soap-bubble appearance, extending up and down the cortex of a long bone. Involvement in the spine may affect several adjacent vertebrae.
In the largest series of 112 cases, angiosarcoma of bone occurred slightly more commonly in men than women, had a wide age range, and tended to involve the long tubular bones.32 However, any bone can be affected. Multifocal involvement was common, occurring in 25 of 112 cases in one series, but it was not associated with worse survival.
Epithelioid angiosarcoma refers to a variant of angiosarcoma composed of neoplastic cells that have an epithelioid appearance. These tumors are usually poorly differentiated and biologically aggressive.30
This series includes eight patients with angiosarcoma of bone, five of whom presented with primary disease and underwent surgical resection. Radiation was delivered to three patients (all of whom had positive margins), and there were no local recurrences. Only one patient died of disease.
Breast Angiosarcoma
Primary breast sarcomas are rare, and primary breast angiosarcoma (in the absence of prior radiation) is very rare. In the review from the Mayo Clinic of 25 primary breast sarcomas, 6 were angiosarcomas.33 In one series, of 83 women with primary breast sarcomas, only 8 of these were angiosarcomas.34 Angiosarcoma histology was associated with far worse overall survival compared with other histologic subtypes. This is in contrast to another study of 60 breast sarcomas, where angiosarcoma histology was associated with better survival.35
Treatment recommendations for primary breast angiosarcoma generally include simple mastectomy and often radiotherapy, although the recommendations of others have ranged from wide local excision to radical mastectomy. The role of axillary dissection is also unclear. In the series from the M. D. Anderson Cancer Center of 59 patients with any breast sarcoma including 17 angiosarcomas, 59% of patients had axillary dissections, and none of the >300 lymph nodes examined had evidence of tumor. In this study, local recurrence occurred in 34% of patients after mastectomy alone compared with 13% after mastectomy and radiation, although this difference was not statistically significant.
Given the propensity of angiosarcomas to have ill-defined and multifocal disease, we generally recommend simple mastectomy for patients with primary breast angiosarcomas or more extensive resections when there is involvement of adjacent tissues such as the chest wall. Even with negative surgical margins, radiotherapy is also often recommended. Of the six patients in our series, all underwent mastectomy and four underwent radiotherapy; local recurrence occurred in one patient. One patient with breast angiosarcoma who presented to us with advanced disease had been initially treated at an outside institution with wide local excision and radiation, and disease subsequently recurred in bilateral breasts.
Surgical Resection and Radiotherapy
General recommendations regarding the surgical resection and delivery of radiotherapy for angiosarcomas are difficult given the heterogeneous nature of this disease. Surgeons should attempt to widely excise these tumors whenever feasible given the frequency of infiltrative margins. Not uncommonly, these tumors can have discontiguous microscopic deposits at some distance from the tumor edge. Cutaneous lesions often require a skin graft or tissue flap to cover the operative bed.10 Angiosarcomas of deep soft tissues are treated similarly to other sarcoma subtypes arising in soft tissues. Cutaneous angiosarcomas arising in previously irradiated fields often require aggressive resection of adjacent structures such as the chest wall and frequent use of flap reconstruction. Amputations are sometimes necessary for extensive extremity tumors. For primary breast angiosarcomas, we generally favor mastectomy over wide local excision; the recent use of magnetic resonance imaging of breasts at our institution has shown that these tumors are often more extensive than they initially seem.
In this study, a variety of adjuvant radiation approaches were used that were based on the extent and timing of operation. Doses similar to those used for other sarcomas were used.36 Radiation fields tended to be very generous when possible because of the infiltrative nature of these tumors. Patients with scalp lesions underwent radiation to the whole scalp where feasible, and patients with lesions on the breast or chest wall received irradiation to whole breast and chest wall, usually with tangential fields. For head and neck lesions, prophylactic nodal irradiation was given to some patients on the basis of the size of the primary tumor and feasibility of including the draining lymph node stations in a tolerable radiation port. Some patients with large lesions where there was concern about the tumor proximity to critical structures (i.e., neurovascular bundle) did receive neoadjuvant radiotherapy. Irradiation was not provided to some patients who had previously received radiotherapy or to patients with very small superficial lesions that were widely excised (margins >1 cm) or with an intervening fascial barrier.
Prognostic Factors
Various negative prognostic factors for overall or disease-specific survival include size >5 cm, high grade, older age, retroperitoneal location, positive margin, and radiation or lymphedema field location (Table 7). Our 5-year survival of 60% compares favorably with other series, and these results may reflect a smaller median tumor size, a high proportion of cutaneous tumors, and an aggressive treatment strategy of surgical resection with wide margins, liberal use of adjuvant radiotherapy, and selected use of chemotherapy.
One unique finding of this study is the poor outcome of patients who presented with angiosarcomas arising in irradiated or lymphedematous areas after surgical resection. Possibly this is because studies have not compared the outcome of these patients to those of other angiosarcoma patients. One may argue that grouping angiosarcomas arising in a previously radiated field and angiosarcomas arising in a lymphedematous extremity is an artificial combination, even though both are field defects. We only had two patients with primary disease with angiosarcomas arising in a lymphedematous extremity. Reanalysis of six patients with primary disease with radiation-induced angiosarcomas and exclusion of the two patients with Stewart-Treves syndrome found that sarcomas arising in a radiated field remained a statistically significant prognostic factor for local recurrence-free survival and disease-specific survival (P < .05). Other studies have found high grade to be prognostic of poor outcome.6 The only prognostic factor found for patients with advanced disease was a cutaneous primary site, which had a better prognosis than noncutaneous primary sites.
The patients in this series were treated over 26 years, and advances in treatment may have occurred during this time period. Of the 46 patients with primary disease, 13 were diagnosed between 1980 and 1986, 19 were diagnosed between 1987 and 1996, and 14 were diagnosed between 1997 and 2006. All patients underwent surgery, and differences in operative results are difficult to summarize for these tumors located throughout the body. However, our general impression is that operative results have improved with time with better tumor imaging, surgical advances, and perioperative care. For the three time periods, radiation was delivered to 71% of patients between 1980 and 1986, 47% of patients between 1987 and 1996, and 92% of patients between 1997 and 2006. We generally recommended radiotherapy for most angiosarcomas. The varying rates in delivery of radiation by time period reflects varying clinical situations. For example, there were increased patients in the 1987–1996 period who underwent amputation as primary treatment, had bone sarcomas that were widely excised, or had deep solid organ angiosarcomas. Chemotherapy was delivered to approximately 30% of patients in all three time periods (range, 23%–38%). Univariate analysis was repeated for local recurrence-free survival, distant recurrence-free survival, and disease-specific survival, and time period of diagnosis was not a statistically significant prognostic factor for any of the three analyses.
Advanced Disease
Patients with advanced angiosarcomas in general fare poorly. In this study, the median survival of patients with advanced disease was 7.3 months. Several chemotherapeutic regimens have been investigated, with some recent optimism found for paclitaxel. A retrospective review of patient receiving paclitaxel for angiosarcoma of the head and neck at the Memorial Sloan-Kettering Cancer Center revealed that eight of nine patients had a major response, and one patient had a minor response with a median duration of 5 months.37 In a subsequent report from the same institution, 52 patients received chemotherapy for unresectable disease. Paclitaxel was provided as first-line therapy in 26 patients, and median progression-free survival was 4.0 months. For 19 patients with scalp angiosarcoma, median progression-free survival was 4.6 months, and efficacy may have been better with weekly dosing compared with dosing every 3 weeks. Mesna, doxorubicin, and ifosfamide administration resulted in a median progression-free survival of 5.4 months.
Doxorubicin-based chemotherapy regimens, which are commonly used for other sarcoma subtypes, have also been used for angiosarcomas. Skubitz and Haddad38 reported their experience with 13 patients initially treated with either paclitaxel or pegylated-liposomal doxorubicin. Five of eight patients treated with paclitaxel had either partial or complete responses, and three of six patients treated with pegylated-liposomal doxorubicin had partial responses. In our study, patients were generally treated with doxorubicin-based regimens or vinorelbine, and 64% of patients had at least a partial response to at least one regimen. Prolonged responses, however, were uncommon, and better systemic agents are clearly needed.
In summary, the treatment of angiosarcomas is often quite challenging even compared with other sarcomas subtypes, and there have been some advances in treatment over the past decades. Staging has become more accurate with improved imaging modalities. Surgery has been associated with less morbidity and mortality with improvements in perioperative care and in surgical techniques, including free-tissue transfer. The ability to deliver focused radiation has improved with the development of three-dimensional conformal radiotherapy, intensity-modulated radiotherapy, and proton radiotherapy. Some major challenges in the treatment of angiosarcomas include characteristics such as ill-defined borders, long distances of covert extension, frequency of multifocal disease, and propensity for metastatic disease. In addition to surgery, there may be important roles for radiotherapy and possibly chemotherapy in the management of these patients, and thus, optimal results require a multidisciplinary approach to the overall management of these patients.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 26, 2006. Accepted for publication December 7, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. R. Kim, S. Y. Rha, S. H. Cheon, J. K. Roh, Y. N. Park, and N. C. Yoo Clinical features and treatment outcomes of advanced stage primary hepatic angiosarcoma Ann. Onc., April 1, 2009; 20(4): 780 - 787. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
