Angiosarcoma after breast-conserving therapy


  • Alan T. Monroe M.D.,

    1. Department of Radiation Oncology, University of Florida College of Medicine, Gainesville, Florida
    Search for more papers by this author
  • Steven J. Feigenberg M.D.,

    1. Department of Radiation Oncology, University of Florida College of Medicine, Gainesville, Florida
    Search for more papers by this author
  • Nancy Price Mendenhall M.D.

    Corresponding author
    1. Department of Radiation Oncology, University of Florida College of Medicine, Gainesville, Florida
    • Department of Radiation Oncology; University of Florida College of Medicine, PO Box 100385; Gainesville FL 32610-0385
    Search for more papers by this author
    • Fax: (352) 265-0759



Angiosarcoma arising in the irradiated breast after breast-conserving therapy is being reported with increasing frequency. As more women undergo breast-conserving therapy, the incidence can be expected to increase. Surgeons, medical oncologists, and radiation oncologists will be faced with difficult management decisions for this aggressive disease.


A comprehensive review of all English-language reports of angiosarcomas after breast-conserving therapy was performed. Approximately 100 cases were reviewed for treatment details and outcome analysis was performed.


Surgical excision is associated with very high rates of disease recurrence (55 of 75 patients with at least 1 year of follow-up; 73%). Local disease recurrences in the tumor bed or along the mastectomy scar are a component of almost all recurrences (96%). Distant metastases develop simultaneously or shortly after local recurrences. Hyperfractionated radiotherapy has successfully prevented local disease recurrences in a limited number of patients.


Angiosarcoma after breast-conserving therapy is increasingly diagnosed in a small but significant portion of breast carcinoma survivors. The aggressive nature of this disease demands further investigation of adjuvant therapy to prevent recurrence of disease after surgery. Cancer 2003;97:1832–40. © 2003 American Cancer Society.

DOI 10.1002/cncr.11277

Angiosarcomas are malignant tumors with morphologic properties similar to the vascular endothelium.1 They may arise in almost any site but most commonly affect the skin of the face and/or scalp in elderly patients. Overall, less than 1% of all soft tissue sarcomas are angiosarcomas2 and primary angiosarcoma of the breast represents only 1 in 1700–2000 primary breast cancers.1

Over the past 20 years, breast-conserving therapy has replaced modified radical mastectomy as the standard of care for patients with early-stage breast carcinoma. Angiosarcoma arising in the irradiated breast after breast-conserving therapy has been reported with increasing frequency over a similar time period.3–51 Although angiosarcoma after breast-conserving therapy is uncommon (approximately 100 cases are reported in the literature), the incidence has increased as more women are treated with segmental mastectomy and radiotherapy. The purpose of this study is to review the medical literature and determine optimal management strategies for patients who develop secondary angiosarcoma of the breast.



Low-grade, well differentiated angiosarcomas resemble normal endothelium both morphologically and functionally. Ultrastructural findings may include Weibel-Palade bodies (tubular structures found in normal endothelium) and pinocytic vesicles. The majority of angiosarcomas are high-grade tumors with varying degrees of nuclear atypia, hyperchromatic nuclei, large nucleoli, and frequent mitoses. “Blood lakes,” which represent hemorrhage into surrounding stroma, are common features of high-grade angiosarcoma.

Immunohistochemical verification of the diagnosis is frequently attempted but is not always successful given the limited sensitivity and specificity of current markers. Factor VIII-associated antigen and CD34 represent the most frequently employed markers. CD31, a platelet-endothelial cell adhesion molecule, is a more sensitive and specific antigen for endothelial differentiation.52 The limited data on the cytogenetics of angiosarcoma show these tumors to have a complex karyotype involving a number of chromosomal deletions and structural rearrangements.18 BRCA1, BRCA2, p53, and the ataxia telangiectasia gene are potential genetic mutations.53

Primary Versus Secondary Angiosarcoma

Primary angiosarcoma of the breast arises without a recognized associated factor. Despite its overall rarity, it is one of the most common breast sarcomas.54, 55 Primary breast angiosarcoma usually affects women in their 20s or 30s. Axillary lymph nodes are infrequently involved, but the prognosis is poor for women with primary breast angiosarcoma compared with women with invasive ductal breast carcinoma. Chen et al.56 reported a 3-year disease-free survival rate of 14% and Buatti et al.10 reported a mean overall survival period of approximately 2 years.

Secondary angiosarcomas have been distinguished from primary angiosarcomas by their association with a number of presumed etiologic conditions. Secondary angiosarcoma was first reported in association with long-standing extremity edema as the Stewart–Treves syndrome.57 This syndrome was described initially in six patients with chronic lymphedema of the upper extremity after radical mastectomy and complete axillary clearance. Chronic lymphedema due to a variety of causes predisposes patients to angiosarcoma. Congenital,58 idiopathic, traumatic, and infectious causes have been implicated.1

Despite the distinction from primary angiosarcoma, the natural history of secondary angiosarcoma is remarkably similar, with a median survival period of 14.5–34 months and a 5-year survival rate of approximately 15%.25, 59, 60 The distinction between primary and secondary angiosarcoma may prove to be less important as more is learned about the etiology of this rare disease.


The association between radiation and breast angiosarcoma has been established by a number of large cohort population studies. Cozen et al.61 reviewed the incidence of angiosarcoma in a case–control study of women in Los Angeles County with a history of invasive breast carcinoma. The odds ratio was 59 for developing an upper extremity angiosarcoma and 11.6 for developing either a chest wall or breast angiosarcoma. The Surveillance, Epidemiology, and End Results program data compiled by the National Cancer Institute included more than 194,000 women who were treated for breast carcinoma. Among patients in the radiotherapy cohort, the relative risk (RR) of developing angiosarcoma was 15.9.62 Even among patients not receiving radiotherapy, a doubling of the RR was appreciated. This may reflect the role of breast edema as an etiologic factor.

Marchal et al.25 surveyed the French Comprehensive Cancer Center experience with angiosarcoma after breast-conserving therapy. Nine cases were reported after more than 18,000 women were treated for an approximate prevalence of 5 per 10,000. As the number of patients treated with breast-conserving therapy increases, the prevalence of secondary angiosarcoma will likely increase proportionately.


Despite the apparent correlation between reports of secondary breast angiosarcoma and the increasing use of breast-conserving therapy, a true etiologic effect from radiation has been difficult to establish. Classically defined radiation-induced sarcomas63 included lesions of a different histology arising within a previous radiotherapy field, usually after a latency period of more than 10 years.64, 65 In contrast, most recent reports of secondary breast angiosarcoma have been observed with a shorter latency period of approximately 4–7 years.16, 17, 47, 66 The shorter latency period may reflect a different natural history of angiosarcoma after breast-conserving therapy.

Although the preserved breast is the most common site of soft tissue sarcoma in patients receiving radiotherapy for breast carcinoma,29 secondary angiosarcomas may also occur out-of-field. An example is the Stewart–Treves syndrome, which occurs in areas of chronic edema outside the irradiation fields. Radiation exposure alone is unlikely to complete the cause/effect equation. More likely, secondary angiosarcomas have a multifactorial etiology that interacts at the molecular level to produce malignant transformation.

Lymphedema has been implicated as a potential causative factor in the development of angiosarcoma. One of the first reported cases of secondary breast angiosarcoma after breast conservation occurred in a patient with an edematous breast 4 years after segmental mastectomy without radiotherapy.6 Molecular mechanisms linking lymphedema and secondary angiosarcoma are not clearly defined. Chronically obstructed lymphatics may induce the growth of collateral vessels, which could cause the transformation to malignant tumors under the influence of vascular growth factors. An alternative mechanism is the notion that chronic lymphedema may interfere with the host's repair of genetic mutations, rendering an edematous area an “immunologically privileged site.”67 Irradiation may result in a necessary genetic insult or it may simply be a confounding factor that enhances the potential for lymphedema.

Persistent edema is reported in approximately 4% of all patients receiving breast-conserving therapy.68 Edema associated with secondary breast angiosarcoma has been reported by 11–30% of reviews in the literature.25, 43 However, these numbers likely underestimate the incidence of breast edema. Essentially, all patients treated with breast-conserving therapy have some degree of edema, perhaps transient and subclinical, related to the primary surgical and axillary incisions. The breast, being a dependent organ, frequently harbors significant edema, which is not always apparent because of the decrease in overall breast size related to tumor removal. Although most overt edema resolves within a few months of surgery, some degree of clinically apparent edema frequently lingers for several years. The probability of clinically overt edema is much less in patients receiving breast-conserving therapy for noninvasive disease than for patients with invasive disease. One reason for this may be the effect of axillary surgery on lymphatic drainage.

There is no objective measure for breast edema. Therefore, clinicians may often overlook or underestimate its presence. There may be no threshold level of edema that correlates with the development of angiosarcoma. Consequently, careful observation will be important in studying the connection between edema and angiosarcoma.


Approximately 100 cases of angiosarcoma arising in the irradiated breast after breast conservation are described in the English literature. Clinical presentation may vary significantly among patients. As a result, a delay in diagnosis is, unfortunately, common. Patients typically present with erythematous or violaceous skin lesions that are painless and may be multifocal. Both cutaneous and noncutaneous lesions exist,23 with the former being more common. Often, a significant portion of the breast is involved with swelling and/or discoloration at the time of diagnosis. Because of their ecchymotic appearance, lesions may be attributed to trauma and may not brought to the attention of the treating physician.

Radiographic assistance in making the diagnosis has been limited. Mammography may reveal skin thickening and/or an ill-defined superficial mass, but findings are often nonspecific. Nearly 33% of patients with breast angiosarcomas have negative mammograms.69 Magnetic resonance imaging (MRI) scans play an increasing role in breast imaging and has been regarded as a promising modality with which to verify clinically suspicious lesions. Lesions are markedly enhancing70 with low-signal T1 weighting and high-signal T2 weighting. Although experience is limited, excellent correlation between MRI scan findings and mastectomy specimens has been demonstrated. Imaging may serve a confirmatory role. However, the ultimate diagnostic tools for a patient previously treated with breast-conserving therapy are a physical examination followed by the performance of a biopsy of the suspicious lesion.

The clinical course of angiosarcoma is highly dependent on tumor grade. A low-grade angiosarcoma exhibits an indolent, smoldering clinical course similar to that of benign atypical vascular lesions.17 However, local recurrences of low-grade angiosarcoma have been reported.27 Conversely, high-grade angiosarcomas are highly aggressive tumors characterized by explosive growth and relentless local recurrences within months of radical resection. A poorly differentiated histology is a negative prognostic indicator for patients with angiosarcoma.71 Low-grade angiosarcomas may transform into a high-grade angiosarcomas.



Despite the high probability of local disease recurrence, the mainstay of treatment has been a simple mastectomy or a wide local excision. A review of all reported cases in the English literature with at least 1 year of follow-up reveals that 55 of 75 patients (73%) developed tumor recurrences after surgery (Table 1). The majority of tumor recurrences occurred within 1 year (84%) of surgery. Only two patients experienced a recurrence more than 2 years after surgery. Forty-three of 45 patients (96%) had local recurrences in the tumor bed or along the mastectomy scar and for 37 of 45 patients (82%), this was the only site of failure.

Table 1. Review of the Literature
ReferencesPatient's age (yrs)Time to AS (mos)Treatment of AS (adjuvant)MarginsGradeTime to first recurrence (mos)Location of first recurrenceTreatment of first recurrenceDMFollow-up (mos)
  • AS: angiosarcoma; NR: not reported; SM: simple mastectomy; RM: radical mastectomy; MRM: modified radical mastectomy; CW: chest wall; WLE: wide local excision; c/l breast: contralateral breast; LRR: locoregional recurrence; hfx: hyperfractionated; SCF: supraclavicular fossa; b/l lung: bilateral lung; (scar): recurred in surgical scar; mast scar: mastectomy scar; DM: distant metastasis; NED: no evidence of disease; RT: radiotherapy; int: intermediate; CTX: chemotherapy; b.i.d.: twice per day; c/l axilla: contralateral axilla.

  • a

    Four cycles of chemotherapy: doxorubicin, ifosfamide, and dacarbazin.

Autio and Kariniemi, 199937954SMNegativeHighNoneNED (6)
Badwe et al., 199155478SMNegativeNoneNED (6)
Benda et al., 198769548SMCloseNR4Local (scar) + DMAbdomen and thigh
Bolin and Lukas, 1996871120SMNegativeLowNoneNED (12)
Bonetta et al., 199595462SMNRInt12Local (scar)CW resectionNED (26 mos)
Buatti et al., 1994107272SMNegativeHigh6Local (scar)RT and hyperthermiaNED (36 mos after salvage)
Chahin et al., 2001127611SMNegativeHigh3Local and c/l breastNoneDied 2 mos later
Cwikiel et al.,7369483Site unknownDead 9 mos later
 4877NoneNED (14)
Del Mastro et al., 1994136942
Deutsch and Rosenstein, 19981471111SMNegativeHighNoneDead intercurrent disease (23)
Edeiken et al., 1992157952SM8C/l breast and axilla
Feigenberg et al., 2002167366SMCloseHigh1Local scarRT hfx + WLENED (38 after RT)
Feigenberg et al., 2002167277SMNegativeLow2Local scarRT (50 Gy, 1.5 Gy b.i.d.)NED (39 after RT)
Feigenberg et al., 2002167656SMPositiveLow1.5Local scar and flapRTNED (22 after RT)
Fineberg and Rosen, 1994178063RM (doxorubicin)NRHigh<9LocalDied with local disease only
Fineberg and Rosen, 1994177446SMNRHighNED (2)
Fineberg and Rosen, 1994177342NRHighDied 2 weeks after diagnosis
Gil-Benso et al., 1994186660SMNRInt5Local (tumor bed)SMNo follow-up
Givens et al., 19891962150MRMHigh
Hildebrandt et al., 2001207966SMNegativeLow4Local (CW) × 3WLESCF at 3rd LRRAlive with disease (23)
Joshi et al., 19952184SMNegativeHigh3C/l breastSM (+margin)No follow-up
Layfield and Dodd, 1997228084SMLowNo follow-up
Majeski et al., 2000237363SMNegativeNR26Local (CW)WLENED
Marchal et al., 19992550108Extended RM (CTX)??Dead (5.5)
Marchal et al., 1999256374RM (RT)11?Dead (11.5)
Marchal et al., 1999258067RM (RT)3.7?Dead (15.5)
Marchal et al., 1999258077RM (RT)8.6?Dead (15.5)
Marchal et al., 1999256876RM (RT)16?Dead (17)
Marchal et al., 1999255357RM (RT and chemotherapy)7.3?Dead (23)
Marchal et al., 1999257757RM7.6?Dead (25.5)
Marchal et al., 1999257175RM??Dead (5.3)
 7879RM2.9LocalWLENED (32)
Mills et al., 2002517796SMNegativeHighNoneNED (14)
Mills et al., 2002517186SMNegativeHighNoneNED (15)
Molitor et al., 1997265978SMNRHigh9Local (CW) and DMSpine and lungDead (14 after SM)
Molitor et al., 1997266648SMNRIntNoneNED (18)
Molitor et al., 1997264954SMNRLow4LocalChemotherapyDead of regional disease (19 after SM)
Moskhaluk et al., 1992276288WLELow24LocalWLEAt 6 mos LRR treated with SM; NED (96)
Parham and Fisher, 19972879120SMNegativeHighLocal vs. DMNED (13)
Parham and Fisher, 1997286860SMNegativeLowNone
Pendlebury et al., 1995296385Liver metastasis at diagnosisDead 2 mos later
Perin et al., 1997304660SM with removal of prosthesisNRHigh1Local4 cycles of chemotherapya2nd recurrence 8 mos later (CW and c/l breast); treated surgically. 3rd recurrence 1 mo later (LR and R pleural effusion); treated with high-dose ifosfamide; Died 6 mo later.
Polgár et al., 2001317172WLENegativeInt3LocalRMNED (36)
Provencio et al., 1995327975SM (neoadjuvant chemotherapy)NRInt6Local (scar)CW resectionYes sites unknownDied 11 mos after LRR with local and distant disease
Roukema et al., 1991347160SMInt
 6784SMNegativeNoneNED (12)
Rubin et al., 1990356784RMLowNoneNED (84)
Sener et al., 20013673132SMNegativeHigh3Local (CW and c/l breast)RTDead (9) LRR and pleural effusion
Sener et al., 2001367296WLENRHigh2LocalRM followed by paclitaxel × 3NED (14 after RM)
Sener et al., 2001368079MRM (RT, 70 Gy)PositiveHighLocalThalidomide
Sener et al., 2001367254SMPositiveNR3Local × 2C/l axilla at 2nd recurrence
Sessions and Smink, 1992377654SM
Shaikh et al., 1988387748WLEIntSCF and pleural effusion
Slotman et al., 1994397952SMNRNR2Local (scar)RT and hyperthermiaDeveloped c/l axilla 6 mos later
Slotman et al., 1994397256SMNegativeNoneNED (24)
Solin et al., 20014075SMNoneNED (86)
Stokkel and Peterse, 1992417583SMNRHighNoneNED (21)
Stokkel and Peterse, 1992415329SMHigh6Local (mast scar)CW resectionDead 2 weeks postoperatively
Stokkel and Peterse, 1992415769SMNRLowNoneNED (82)
Stotter et al., 198942144
Strobbe et al., 19984376SMNegativeHigh6LocalDead (8)
 78SMPositiveHigh18LocalDead (24)
Strobbe et al., 19984329SMPositiveHigh6LocalDead (6)
Strobbe et al., 199843121SMPositiveHigh12LocalNED (13)
Strobbe et al., 19984398SMPositiveHigh12LocalNED (14)
Strobbe et al., 19984365SMPositiveHigh5LocalNED (6)
Strobbe et al., 19984372SMNegativeInt6LocalDead (35)
Strobbe et al., 19984384SMNegativeInt6LocalDead (8)
Strobbe et al., 19984336SMNegativeInt3LocalNED (15)
Strobbe et al., 199843106SMNegativeIntNoneNED (50)
Strobbe et al., 19984360SMNegativeIntNoneNED (50)
Strobbe et al., 19984355SMNegativeLow3LocalNED (36)
Strobbe et al., 19984368SMNegativeLow78LocalNED (82)
Strobbe et al., 19984384SMNegativeLowNoneNED (12)
Strobbe et al., 19984399SMNegativeLowNoneNED (26)
Strobbe et al., 19984374SMNegativeLowNoneNED (47)
Strobbe et al., 19984352SMNegativeLowNoneNED (5)
Strobbe et al., 19984370SMNegativeLow/int16LocalDead (38)
Strobbe et al., 19984389SMNegativeLow/intNoneNED (48)
Strobbe et al., 19984381SMNegativeLow/intNoneNED (66)
Strobbe et al., 19984358SMPositiveLow/int2LocalDead (8)
Taat et al., 1992447066SMPositive5Local (mast scar)RT and hyperthermiaDead (1 after RT)
Turner and Greenall, 19914566108WLE2LocalMRM and doxorubicin Died secondary to local disease
        and ifosfamide  
Veosoulis and Cunliffe, 2000464596SMNo follow-up
Wijnmaalen et al., 1993485766SMNegativeLow/int16Local (mastectomy scar)CW resection2nd and 3rd LRR treated with RT and hyperthermia. Dead (local disease)
Wijnmaalen et al., 1993487277SMNegativeLow/intNoneNED (30)
Wijnmaalen et al., 1993485984SMNegativeLow/intNoneNED (7)
Williams et al., 1999496044SMNegativeLowNoneNED (5)
Zucali et al., 1994506441SMNRInt13Local (CW)WLE and RT (60 Gy)NED (13 mo after salvage)
Zucali et al., 1994507259SMNRInt4Local (scar) and b/l lungChemotherapyDied shortly after salvage chemotherapy

Very few cases of surgical salvage after local tumor recurrence have been reported in the literature. Even after obtaining negative margins through radical salvage surgery by simple mastectomy, radical mastectomy, or chest wall resections, additional local tumor recurrences are common (14 of 29 patients [48%]). This is important because distant metastases rarely developed as a site of first failure and there is a window of opportunity for controlling disease locally before dissemination occurs. Distant metastases rarely were present at initial presentation (1 of 96 patients [1%]). Instead, they develop simultaneously or shortly after the first local tumor recurrence (6 of 45 patients [13%]) or, occasionally, after multiple local recurrences (three patients). If local control can be attained, the incidence of distant metastasis may be reduced and survival rates may improve.


Given the difficulty in achieving local disease control with surgery alone, adjuvant radiotherapy seems logical. Several investigators used adjuvant radiation in standard regimens in an attempt to improve local control. The explosive growth suggests a high growth faction and a potential role for hyperfractionated radiotherapy. At the University of Florida, results with accelerated hyperfractionated postoperative radiotherapy are encouraging.16

Radiation portals should encompass the entire chest wall or breast with a generous margin. Because the disease occurs in the dermal lymphatics, which have no valves, margins must be even more generous than margins for carcinomas or other sarcomas. Typically, fields include a minimum of 8–10-cm margins around macroscopic disease. The regional lymph nodes are uncommon sites of tumor recurrence, but are at risk in patients with extensive disease. Only one incidence of tumor recurrence in the ipsilateral supraclavicle has been documented in the literature (Table 1). In addition, one treatment failure in the untreated axillary lymph nodes has been observed in our institution. If disease is extensive and is in close proximity to draining lymphatics, consideration should be given to elective treatment of the lymph nodes.

Because radiation is often delivered to previously irradiated surrounding tissue, we recommend preoperative radiation followed by surgical resection. This allows removal of tissue at high risk for late complications and placement of autologous tissue flaps as needed. Preoperative radiation also avoids the potential negative effect of local hypoxia within the tumor bed, minimizes the potential for seeding at the time of surgery, and may allow lower radiation doses.


The most common site of “distant” failure is the contralateral breast, which in most cases is probably local extension through dermal lymphatics. The role of chemotherapy in angiosarcoma patients has not been established, as usually only short-lived responses are reported in the literature.16 Paclitaxel has shown substantial activity in angiosarcoma patients with involved sites other than the breast72 and warrants further investigation. Vascular targeting agents are intriguing new agents that selectively destroy existing tumor vasculature, thereby inducing rapid tumor necrosis. Currently, a Phase I/II clinical trial is being planned at the University of Florida to assess the efficacy of vascular-targeting agents in conjunction with accelerated hyperfractionated radiotherapy in treating angiosarcoma.


Angiosarcoma is increasingly diagnosed in a small but significant portion of breast carcinoma survivors. As more women are treated with breast-conserving therapy for early-stage breast carcinoma, the incidence of angiosarcoma of the breast is expected to rise. Early clinical detection will require a high index of suspicion by health-care providers. Patients should be advised to alert their physician of skin changes and biopsies of suspicious lesions should be performed promptly. Avoiding diagnostic delay is important because of the potentially aggressive kinetics of this disease.

A combination of hyperfractionated radiotherapy and surgery provides the most promising results for patients with high-grade angiosarcoma. The role of systemic therapy is unclear at this time.

The limited prevalence of disease will probably preclude randomized clinical trials. Therefore, it is important to report small successful experiences to aid the clinician in managing these infrequent problems.