A, Recurrence-free survival. B, Overall survival. The dashed line is the position for 50% survival rate.
Locoregional control failure in patients treated with surgery, radiation therapy, and/or chemotherapy increased steadily (70%-80%) over the first 2 to 3 years after diagnosis before leveling off. The dashed line is the position for 50% control failure rate.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Patel SH, Hayden RE, Hinni ML, et al. Angiosarcoma of the Scalp and Face: The Mayo Clinic Experience. JAMA Otolaryngol Head Neck Surg. 2015;141(4):335–340. doi:10.1001/jamaoto.2014.3584
The etiology and optimal treatment are unknown for angiosarcoma, an aggressive malignant tumor that affects vascular endothelial cells and can be mistaken for benign lesions such as hemangioma.
To determine the treatment outcomes of patients with angiosarcoma of the face or scalp treated with a combination of surgery, radiation therapy, and/or chemotherapy.
Design, Setting, and Participants
Retrospective study of 55 patients with angiosarcoma of the face or scalp treated between January 1, 1973, and December 31, 2012, at a tertiary-care academic medical institution.
Surgery, radiation therapy, and/or chemotherapy.
Main Outcomes and Measures
Locoregional control (LRC), recurrence-free survival (RFS), and overall survival (OS).
Fifty-five patients had angiosarcoma localized to the face or scalp. Forty of these patients (73%) received a combination of surgery, radiation therapy, and/or chemotherapy. Eight patients (15%) were treated with surgery alone, 1 (2%) with radiation alone, 5 (9%) with chemotherapy alone, and 1 (2%) with observation alone. Median (range) follow-up for surviving patients was 25.2 (4.7-227.1) months. Five-year LRC, RFS, and OS (95% CI) were 18% (7%-32%), 16% (6%-31%), and 38% (21%-54%), respectively. Of 36 patients with failed treatment, 34 had failure in a local and/or regional site. On univariate analysis, the use of multimodality therapy (vs no multimodality therapy) was associated with higher 5-year LRC (95% CI) (20% [3%-37%] vs 11% [0%-29%]; P = .04), higher RFS (19% [2%-36%] vs 10% [0%-27%]; P = .02), and higher OS (46% [26%-66%] vs 16% [0%-43%]; P = .04). Age 70 years or older (vs <70 years) was associated with lower 5-year LRC (95% CI) (5% [0%-14%] vs 48% [23%-74%]; P = .02) and lower RFS (5% [0%-13%] vs 49% [24%-75%]; P = .04). Radiation therapy (vs no radiation therapy) was associated with higher 5-year LRC (95% CI) (20% [3%-36%] vs 12% [0%-32%]; P = .02) and higher RFS (19% [2%-35%] vs 12% [0%-31%]; P = .004). On multivariable analysis, age younger than 70 years (vs ≥70 years) was associated with improved 5-year LRC (95% CI) (48% [23%-74%] vs 5% [0%-14%]; P = .03) and RFS (49% [24%-75%] vs 49% [24%-75%]; P = .04).
Conclusions and Relevance
Multimodality therapy for angiosarcoma is associated with improved LRC, RFS, and OS. Younger patients with resectable disease undergoing multimodality therapy for angiosarcoma had the best clinical outcomes.
Angiosarcoma is an aggressive malignant tumor of the vascular endothelial cells. In most cases, its etiology is unknown. However, some factors have been associated with an increased risk for angiosarcoma, including prior radiation exposure, chronic lymphedema (Stewart-Treves syndrome), and exposure to chemicals such as polyvinyl chloride, thorium dioxide, arsenic, and radium. Angiosarcoma can involve any anatomic site, with the skin being the most common. It accounts for approximately 5% of cutaneous soft-tissue sarcomas.1 Cutaneous angiosarcomas are more common in the head and neck region and account for more than 60% of cases.2 The diagnosis of cutaneous angiosarcoma can be challenging because it often presents insidiously as a bruiselike lesion or a purplish papule that may be mistaken for a benign lesion such as hemangioma.
Cutaneous angiosarcoma of the head and neck region may involve the skin extensively with multifocal lesions, making it difficult to obtain a wide negative margin despite the attempt for radical resection. It also has a tendency to spread hematogenously. The 5-year survival rate for patients with cutaneous angiosarcoma of the head and neck region ranges from 10% to 54%.3-8 In the early experience at Mayo Clinic, of 13 patients treated between 1920 and 1970, only 2 patients lived longer than 5 years.3 Because of the rarity of the disease, most published studies are single-institution series with a relatively small number of patients. Whenever possible, surgical resection is a critical component of treatment. Adjuvant radiation therapy after maximal surgical resection is commonly administered to improve local control. Chemotherapeutic agents with activity against angiosarcoma include anthracyclines and taxanes. However, the role of adjuvant chemotherapy is not well defined. No prospective randomized studies have been conducted to provide treatment guidelines.
To add to the knowledge of this uncommon disease, we conducted a retrospective study of patients with cutaneous angiosarcoma localized to the face and scalp who were evaluated and treated at Mayo Clinic. Our goals were to assess treatment outcomes and patterns of failure, as well as to identify factors associated with prognosis. This analysis is meant to update an earlier reported analysis of 24 patients who underwent primary therapy between 1974 and 1992.9
This retrospective study was approved by the Mayo Clinic Institutional Review Board. Between January 1, 1973, and December 31, 2012, a total of 56 patients with angiosarcoma of the scalp and/or face were treated at Mayo Clinic. One patient was excluded from analysis because of a diagnosis of congenital angiosarcoma at the age of less than 1 year, for a total of 55 patients whose medical records were reviewed for this analysis.
Patients had no signs of distant metastatic disease outside the head and neck. Primary therapy in the form of surgery, radiation therapy, chemotherapy, or combination therapy was delivered at Mayo Clinic (Scottsdale, Arizona; Jacksonville, Florida; or Rochester, Minnesota). All patients underwent routine history and physical examinations, blood tests, radiologic imaging, and confirmation of diagnosis of angiosarcoma by a Mayo Clinic pathologist.
The medical record of each patient was reviewed to abstract data from operative notes, radiation records, chemotherapy records, subsequent follow-up notes, imaging reports, and surgical pathology reports. The institutional tumor registry was queried to obtain data for events such as local failure, regional failure, distant failure, and vital status. Retrieval of the following data was attempted: age at diagnosis, sex, tumor grade, tumor location (scalp or face), tumor size (≤5 cm or >5 cm), presence of satellite lesions, nodal disease at diagnosis, date of last follow-up, and pattern of failure (local, regional, distant, or combination).
Forty-four patients underwent surgery, which consisted of wide local excision with or without reconstruction. Reconstruction typically consisted of placement of a split-thickness skin graft or rotation flap. Despite attempts to achieve negative margins, including re-resection of positive margins, 12 patients were found to have a positive margin on final pathologic testing. Thirty-seven patients received radiation therapy, which consisted of electron beam or megavoltage photon beam irradiation. Doses ranged from 4000 cGy to 7000 cGy in 10 to 35 fractions. The median dose was 6000 cGy. In the adjuvant setting for negative margins, doses ranged from 4500 cGy to 6000 cGy in 25 to 30 fractions. For positive margins, doses ranged from 6000 cGy to 6600 cGy in 30 to 33 fractions. In the definitive setting, doses typically ranged from 6600 cGy to 7000 cGy in 33 to 35 fractions, with 1 patient receiving 4000 cGy in 10 fractions. Twenty-seven patients received chemotherapy, and chemotherapy details were available for 25 patients, 11 of whom received adjuvant chemotherapy alone (6 paclitaxel based; 1 bevacizumab alone; 2 mitomycin, doxorubicin, and cisplatin; and 2 doxorubicin and ifosfamide). Seven other patients received chemotherapy concurrently with adjuvant radiation therapy (6 paclitaxel and 1 carboplatin). Two patients underwent neoadjuvant chemotherapy, with 1 receiving gemcitabine and docetaxel and 1 receiving paclitaxel and bevacizumab. Five patients underwent chemotherapy alone (4 paclitaxel based and 1 mitomycin, doxorubicin, and cisplatin).
Frequencies and percentages were used to describe characteristics of patients, disease, and treatment. Overall survival (OS) was calculated as the time from cancer diagnosis to date of death due to any cause. Recurrence-free survival (RFS) was calculated as the time from cancer diagnosis to earliest date of recurrence or death due to any cause. Locoregional control (LRC) failure was calculated as the time from cancer diagnosis to the date of local or regional failure. Multimodality therapy was defined as any 2 or more modalities (surgery, radiation, and chemotherapy).
Kaplan-Meier estimates and plots were used to describe time to event end points overall and by group. Five-year OS, RFS, and LRC failure rates were calculated using the Kaplan-Meier method, and the difference between patient groups was tested using Wilcoxon tests. Cox regression analysis was used for multivariable analysis. P < .05 was considered statistically significant.
Patient characteristics are reported in Table 1. Median (range) follow-up for surviving patients was 25.2 (4.7-227.1) months. Five-year OS was 38% (95% CI, 21%-54%), with a median survival of 25.2 months. Five-year RFS was 16% (95% CI, 6%-31%) and 5-year LRC was 18% (95% CI, 7%-32%). Five-year local control was 38% (95% CI, 22%-54%). Figure 1 summarizes RFS and OS.
On univariate analysis, improved OS was associated with the use of combined modality therapy (46% [95% CI, 26%-66%]; P = .04). Age younger than 70 years (49% [95% CI, 24%-75%]; P = .04), use of radiation therapy (19% [95% CI, 2%-35%]; P = .004), and use of multimodality therapy (19% [95% CI, 2%-36%]; P = .01) were associated with improved RFS. Improved LRC was associated with the same factors of age younger than 70 years (48% [23%-74%]; P = .02), use of radiation therapy (20% [95% CI, 3%-36%]; P = .02), and use of multimodality therapy (20% [95% CI, 3%-37%]; P = .04). Table 2 summarizes the univariate analyses for OS, RFS, and LRC.
On multivariable analysis, age younger than 70 years was associated with both improved RFS (49% [95% CI, 24%-75%]; P = .04) and LRC (48% [95% CI, 23%-74%]; P = .03). Table 3 summarizes the multivariable analyses for RFS and LRC.
Of the 37 patients 70 years or older, 25 (68%) received multimodality therapy. Twelve patients 70 years or older received single modality therapy, and 5 of those 12 (42%) did not receive multimodality therapy because of comorbidities.
Of the 55 patients, 36 had failure after initial therapy. Thirty-four failures were initially locoregional, with the majority (n = 26) representing local failures. The majority of initial locoregional failures occurred within 36 months (Figure 2). The 2 initial distant failures were in the abdominal wall and lung. Eventual distant metastasis was documented in 5 additional patients. Two patients developed lung metastases, and 1 patient developed lung and bone metastases. One patient developed liver and bone metastases, and 1 patient developed bone metastases only.
As noted in the earlier Mayo Clinic experience, cutaneous angiosarcoma of the head and neck is a difficult disease to control, with most patients not living longer than 5 years.3,9 Our current series further confirms the poor clinical outcomes of patients with cutaneous angiosarcoma of the head and neck. Furthermore, this disease primarily affects fair-skinned elderly men. Our 5-year OS of 38% represents a modest improvement over previous series that reported a 5-year OS ranging from 10% to 30%.10 In our series, 1 patient elected to have no therapy, and 5 patients were treated with chemotherapy alone because of locally extensive disease, which potentially affected the reported OS for this cohort.
Our results are comparable to those in more recent publications that suggest a 5-year OS ranging from 33% to 54%.2,7,8 In 2011, Albores-Saavedra et al2 reported on a SEER (Surveillance, Epidemiology, and End Results) analysis of cutaneous angiosarcoma that included 135 cases of the scalp and neck, showing a 5-year and 10-year OS of 33.6% and 13.8%, respectively. This improvement in patient outcomes found in more recent studies may be attributed to the increased use of multimodality therapy. In our series, findings from univariate analysis showed multimodality therapy to be associated with significantly improved OS, RFS, and LRC. The 5-year OS of patients undergoing multimodality therapy was 45.9%. Multiple institutional reports have corroborated the benefit of multimodality therapy, particularly the combination of surgery and radiation therapy.5,7,11-13 The use of adjuvant radiation therapy had a favorable impact on RFS and LRC. In addition, the association between the use of surgery and improvement in OS had a P value of .09.
In our study, a few patients treated with nonsurgical multimodality therapy had relatively promising survival. Two patients were not deemed surgical candidates because of the clinically significant extent of their disease. One patient with extensive involvement of the scalp, nose, and left cheek underwent a combination of gemcitabine-docetaxel chemotherapy and radiation therapy and had a dramatic response, then eventually had a regional recurrence 3 years later but is still alive and is currently undergoing salvage therapy. The second patient, who had involvement of the scalp, bilateral cheeks, and right upper eyelid, lived more than 3 years after undergoing paclitaxel-bevacizumab chemotherapy, followed by maintenance bevacizumab chemotherapy and radiation therapy.
Our findings are similar to those in several reports on the combination use of taxane-based chemotherapy and radiation therapy, which suggests a role for nonsurgical multimodality therapy for select patients. Fujisawa et al14 reported on a multicenter retrospective study in which 16 patients underwent taxane-based chemotherapy; 14 patients had a complete response and 1 patient had a partial response. Five-year OS was reported to be 55.8%. Another study by Miki et al15 found a 2-year in-field control rate of 67% in patients treated with docetaxel and radiation therapy. Further studies are needed to fully delineate optimal regimens for patients who are not candidates for surgical therapy, especially for patients with clinically significant disease extent. Tumor size was not significantly associated with treatment outcomes in our study. Other series have shown that tumors smaller than 5 cm or limited disease extent correlates with improved prognosis.2,5,7,9,13,16 The insidious nature of cutaneous angiosarcoma of the head and neck often makes it difficult to ascertain actual tumor size and/or disease extent, which may be one of the reasons our series did not show prognostic significance for size. Whenever feasible, mapping biopsies are attempted to better appreciate tumor extent.
The high tendency for this disease process to recur is evidenced by our findings of a 5-year RFS of 16% and a 5-year LRC of 18%. Mark et al17 found that 81% of initial recurrences were within 24 months, with the majority representing local recurrences, which is similar to our findings. We found age younger than 70 years to be a significant prognostic factor on both univariate and multivariable analysis, with younger patients having a 5-year RFS and LRC close to 50% compared with approximately 5% for older patients. Younger patients have improved survival, as has been shown by multiple authors.2,5,8 Albores-Saavedra et al2 examined 434 cases of cutaneous angiosarcoma in the SEER database and found that an age younger than 50 years correlated with improved survival. Younger patients can likely tolerate multiple modalities of therapy, which therefore translates into improved clinical outcomes for this age group. As noted in our series, more than 40% of elderly patients did not receive multimodality therapy secondary to comorbidities. Therefore, age could be a surrogate marker for poor comorbidity status, having an impact on ability to receive multimodality therapy. Although in our series the association between the use of surgery and improved OS had a P value of .09, it is not clear when negative margins are necessary. On subgroup analyses, our series of patients did not show improved clinical outcomes with a negative margin, a notion supported by multiple series.5,7,18 Dettenborn et al8 found worsened survival with incomplete resection, although it should be noted that more than 50% of their study patients had an R1 (gross complete resection with microscopically positive margins) resection. While surgery is a critical component of therapy, a radical excision does not appear to be necessary. The majority of reports in the medical literature support a more limited resection, thereby limiting the risk of complications and not delaying adjuvant therapy. The 12 patients with a positive margin in our series still had a 5-year OS of 41.1%, which reflects the importance of radiation therapy.
Cutaneous angiosarcoma is an uncommon disease. While our series is one of the larger ones on outcomes of cutaneous angiosarcoma of the head and neck, the number of patients is still small. Our study is retrospective in regard to data collection and subject to the inherent biases of such a study design. Details of subsequent failure patterns for some patients were not available because the patients chose to have follow-up elsewhere rather than within our medical system. Selection bias plays a role in treatment approach, as noted by the heterogeneous strategies used. Patients with more clinically significant disease were more likely to undergo a multimodality approach. Moreover, this study spans a period of 4 decades, with accompanying changes and advances in all the treatment modalities over time, resulting in heterogeneous treatment.
Younger age, the use of radiation therapy, and the use of multimodality therapy are independent prognostic factors for improved LRC and RFS. The use of multimodality therapy is the sole independent prognostic factor associated with improved OS. The value of nonsurgical multimodality therapy requires further investigation, particularly for patients with clinically significant disease extent. Cutaneous angiosarcoma of the head and neck continues to be a difficult tumor to control and is therefore associated with poor outcomes. Further investigation is needed of novel treatment strategies, particularly effective systemic agents.
Submitted for Publication: August 25, 2014; final revision received November 17, 2014; accepted November 25, 2014.
Corresponding Author: Samir H. Patel, MD, Department of Radiation Oncology, Mayo Clinic, 13400 E Shea Blvd, Scottsdale, AZ 85259 (firstname.lastname@example.org).
Published Online: January 29, 2015. doi:10.1001/jamaoto.2014.3584.
Author Contributions: Drs Patel and Wu had full access to all data and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Patel, Hinni, Wong, Foote, Ko, Halyard.
Acquisition, analysis, or interpretation of data: Patel, Hayden, Milani, Wu.
Drafting of the manuscript: Patel, Wong, Foote, Milani.
Critical revision of the manuscript for important intellectual content: Patel, Hayden, Hinni, Wong, Foote, Wu, Ko, Halyard.
Statistical analysis: Patel, Wong, Foote, Wu.
Administrative, technical, or material support: Patel, Hayden, Hinni.
Study supervision: Patel, Hayden, Hinni, Foote.
Conflict of Interest Disclosures: None reported.