Clinical Investigation
A Comparison of Outcomes and Prognostic Features for Radiation-Associated Angiosarcoma of the Breast and Other Radiation-Associated Sarcomas

https://doi.org/10.1016/j.ijrobp.2019.01.082Get rights and content

Purpose

Radiation-associated sarcomas (RAS) are considered to have a poor prognosis. Although the incidence is anticipated to rise, contemporary data regarding predictors of outcomes are few. We performed a retrospective analysis to identify RAS prognostic factors and subset analyses for radiation-associated angiosarcoma arising after treatment for breast cancer (RAAB) and other RAS subtypes (other-RAS).

Methods and Materials

Patients with localized RAS evaluated at an institutional multidisciplinary sarcoma clinic were identified. Clinical and histologic review was performed, and outcomes were assessed to identify prognostic features. A subset of cases underwent molecular analysis by next-generation sequencing.

Results

Among 176 patients, histologic subtypes of RAS included angiosarcoma (41%), undifferentiated/unclassified sarcoma (40%), leiomyosarcoma (8%), malignant peripheral nerve sheath tumor (6%), and osteosarcoma (2%). Sixty-seven patients (38%) had RAAB, and 109 (62%) had other-RAS. RAAB had significantly shorter latency from time of initial radiation compared with other-RAS (8 vs. 15 years; P < .001). Treatment approaches included surgery (91%), chemotherapy (44%), and radiation therapy (27%). Median follow-up was 3.2 years; 3-year overall survival (OS) was 74%. On multivariate analysis, positive margins (P < .0001), deep tumor location (intrathoracic/intra-abdominal, P = .002), and high grade (P < .0001) were associated with worse OS. In particular, 3-year OS with negative versus positive margins was 90% versus 66%. Patients with RAAB versus other-RAS showed a trend for higher 3-year OS (84% vs 68%; P = .09), significantly higher 3-year metastasis-free survival (82% vs 67%; P = .001), but similar 3-year local recurrence-free survival (54% vs 61%; P = .28). Next-generation sequencing identified overall low tumor mutational burden, recurrent MYC amplification in RAAB, and few clinically actionable mutations.

Conclusions

Margin negative excision, superficial tumor location, and low tumor grade are determinants of improved OS for RAS, suggesting that complete surgical excision, when possible, is an optimal component of treatment. RAAB is a clinicopathologically distinct type of RAS with shorter latency from initial RT, different recurrence patterns, and when aggressively managed has potentially better outcomes compared with other-RAS.

Introduction

More than half of all patients with cancer will receive radiation therapy (RT) during the course of their disease.1 Although RT is effective at preventing recurrence, it is associated with significant morbidity in a subset of patients, including the development of radiation-associated sarcomas (RAS). RAS occurs in <1% of treated patients but accounts for up to 5% of all sarcomas.2, 3, 4, 5, 6, 7, 8, 9, 10 As the number of cancer survivors grows, the number of patients developing RAS is expected to rise.11 This is particularly concerning given that multiple studies suggest worse outcomes for RAS compared with non–radiation-associated sarcomas.4, 5, 10, 12

Radiation oncologists are accustomed to having radiation-associated cancers listed on routine treatment consent forms, with RAS being a canonical example of this malignancy, and a deep analysis of the presentation, treatment, and outcomes provides an opportunity to understand the implications of this diagnosis. One area where conflicting data on RAS has been prominent is the influence of histologic subtype on RAS outcome.5, 6, 12 The current World Health Organization (WHO) classification scheme no longer recognizes several histotypes used in these prior studies, such as “malignant fibrous histiocytoma,” and new ancillary techniques have allowed more accurate tumor classification. Therefore, the application of contemporary classification and diagnostic tools can further refine the risks associated with specific subtypes of RAS.

Radiation-associated angiosarcoma of the breast occurring after treatment for breast carcinoma (RAAB) is a clinicopathologic subtype of RAS that constitutes approximately 40% of breast angiosarcomas.5, 13 Given the increasing numbers of breast cancer survivors treated with RT, the prevalence of RAAB is also expected to increase in the coming years.11

Although several reports have examined outcomes of RAS in small cohorts, there are limited data regarding prognostic features, particularly for histotypes other than radiation-associated angiosarcoma of the breast (other-RAS). Furthermore, with recent refinements in sarcoma classification and changing treatment approaches, the generalizability of some of these older data are limited. This study aimed to characterize histologic types of RAS based on the current WHO Classification of Tumours of Soft Tissue and Bone,14 apply uniform grading, evaluate clinical and histologic features, and correlate these data with patient outcomes. Because there are limited data on the genomic landscape of radiation-associated malignancies,15 next-generation sequencing (NGS) data on a subset of cases were analyzed to identify potentially actionable genomic abnormalities.

Section snippets

Identification of patient cohort

This study was conducted with approval from the Brigham and Women's Hospital Institutional Review Board (2016P000965). The pathology laboratory information system of the Brigham and Women's Hospital was searched for sarcomas diagnosed between 1994 and 2016 where the word radiation was used in the report. Pathology reports and medical records were reviewed to identify “true” RAS using criteria similar to those defined by Cahan et al: (1) histologic confirmation of a sarcoma, distinct from the

Patient and primary malignancy characteristics

Initial search identified 1346 potential RAS, among which review of pathology reports and medical records identified 200 RAS. More than 3 months of follow-up was available for 188 patients, of whom 176 had localized disease at diagnosis and served as the final cohort.

Of 176 patients with localized RAS, most were female (79%), including all with RAAB and 66% with other-RAS (P < .001). Median age at primary malignancy diagnosis was 51 years; patients with RAAB were slightly older at diagnosis

Discussion

This study provides a detailed analysis of clinical, histologic, and treatment characteristics and outcomes for a large cohort of patients with localized RAS, applying contemporary classification schemes and including patients treated with modern surgical techniques. The prognosis of RAS has generally been considered to be poor, but our data suggest that clinical outcomes might be better than previously recognized. In our cohort of 176 patients, 3-year OS was 74%, which is substantially better

References (31)

  • W.B. Laskin et al.

    Postradiation soft tissue sarcomas. An analysis of 53 cases

    Cancer

    (1988)
  • A.F. Mavrogenis et al.

    Post-radiation sarcomas. Clinical outcome of 52 Patients

    J Surg Oncol

    (2012)
  • S. Riad et al.

    The clinical and functional outcome for patients with radiation-induced soft tissue sarcoma

    Cancer

    (2012)
  • A.K. Bryant et al.

    Trends in radiation therapy among cancer survivors in the United States, 2000-2030

    Cancer Epidemiol Biomarkers Prev

    (2017)
  • Y.L. Chen et al.

    Outcomes of a large single institutional series of radiation associated sarcomas

    Int J Radiat Oncol Biol Phys

    (2018)
  • Cited by (28)

    • Neoadjuvant Reirradiation for Radiation Therapy–Associated Angiosarcoma of the Breast

      2024, International Journal of Radiation Oncology Biology Physics
    • Postoperative Complications from Breast and Axillary Surgery

      2023, Surgical Clinics of North America
      Citation Excerpt :

      However, patients with certain heritable mutation syndromes who are at higher risk for radiation-induced tumors (eg, retinoblastoma) should be counseled regarding this risk of radiation and tailor the patient’s treatment plan accordingly.117 Radiation-associated angiosarcomas occur at a median 10 years after breast radiation, although cases have been reported as early as 14 months and as late as 54 years.118–120 Cases usually present as skin changes that can be confused for infection, moderate or severe radiation skin changes, or recurrence of the primary breast cancer and any suspicion should prompt biopsy to differentiate the diagnosis.

    • Response Evaluation Following Radiation Therapy With <sup>18</sup>F-FDG PET/CT: Common Variants of Radiation-Induced Changes and Potential Pitfalls

      2022, Seminars in Nuclear Medicine
      Citation Excerpt :

      RAS carries a poor prognosis and has a median latency period of 11 years (range 2-60,6 years). Female breast carcinoma accounts for 41% of all RAS, with cutaneous angiosarcoma as the most frequent manifestation with a mean latency period of 8 years (range 3-25 year).54 Although, conformal RT delivery with IMRT has reduced toxicity in OARs, concerns regarding a risk of RTIM have been expressed.

    View all citing articles on Scopus

    Jeffrey K. Mito and Devarati Mitra made equal contributions to this study. Elizabeth H. Baldini and Leona A. Doyle are co-senior authors.

    This study was partially supported by the Eleanor and Miles Shore Fellowship Program of Harvard Medical School, Boston, Massachusetts.

    Conflict of interest: none.

    View full text