We are grateful to Paul Zum Vörde sive Vörding and Jacques Venselaar, without whom hyperthermia at our departments would not be possible, and to Otto Visser from the Comprehensive Cancer Center Amsterdam for providing the Netherlands Cancer Registry data.
The objective of this study was to evaluate the role of reirradiation and hyperthermia in the treatment of radiation-associated sarcoma (RAS) in the thoracic region, which is an increasing, yet extremely rare condition with a poor prognosis.
Between 1979 and 2009, 16 patients with RAS in the thoracic region were treated in the Academic Medical Center and the Institute Verbeeten with reirradiation and hyperthermia. In 13 patients, this treatment was given for unresectable disease and 3 times after resection as adjuvant treatment. The median latency period between the original malignancy diagnosis and the RAS diagnosis was 86 months (range 19-212 months). Histology was angiosarcoma in 11 patients (69%). The literature on reirradiation with or without hyperthermia for RAS was reviewed.
The median survival was 15.5 months (range, 3-204 months). Four patients were not evaluable for response. The response rate for the remaining 12 patients was 75% (7 complete responses and 2 partial responses). Six patients remained free of local failure until death (5 months and 7 months) or last follow-up (8 months, 11 months, 39 months, and 68 months).
Radiation-associatedsarcoma (RAS) is defined as the appearance of a sarcoma inside an area previously irradiated with a dose of 25 to 80 gray (Gy), a latency interval of at least 3 years, and a histology clearly distinct from the initial neoplasm for which the radiotherapy was applied (according to the criteria of Cahan et al).1 RAS is a very rare disease with a poor prognosis. Between 1989 and 2008, a total of 321 patients with RAS were registered in the Netherlands Cancer Registry (approximately 15 per year).2 Of these, 181 tumors (56%) were located in the chest region, mainly after irradiation for breast cancer. More than half of these 181 tumors (56%) were angiosarcomas. In the current report, we focus on RAS in the thoracic region. The incidence of sarcoma in general appears to increase by a factor 1.5 for patients who previously received irradiation.3-5 The reported median latency period is 6 to 7 years after radiotherapy for breast cancer (range, 1-29 years). The prognosis for patients with RAS is poor, and the reported 5-year survival rate is from 27% to 38%.3-5
Reports in the literature on the role of reirradiation for RAS are extremely rare. To our knowledge, no randomized studies have been performed. Most articles recommend radical surgery as standard treatment if possible.6-8 Adjuvant chemotherapy is advocated because of the poor prognosis.7 Because RAS arises by definition in previously irradiated areas, options for full-dose primary or postoperative reirradiation are limited.
The Academic Medical Center (AMC) and the Institute Verbeeten (BVI) have extensive experience with reirradiation combined with hyperthermia (RT-HT). It is known that hyperthermia enhances the effect of radiation and/or chemotherapy, particularly in hypoxic areas.9-11 Hyperthermia is used frequently to boost the relatively low maximally permitted radiotherapy dose in the treatment of recurrences of breast cancer in previously irradiated areas.12-14
In total, 16 patients with RAS received RT-HT at our institutes. Two of these 16 patients were reported previously in a case report and a letter.15, 16 To investigate the role of combined RT-HT for RAS of the thoracic region, the treatment effects in these 16 patients were analyzed. In addition, the literature on reirradiation with or without hyperthermia for RAS was reviewed. Nineteen patients with reirradiated RAS were identified in the literature in addition to our series of 16 patients.
MATERIALS AND METHODS
From 1979 to 2009, 16 patients (15 women and 1 man) with RAS were referred to the Department of Radiation Oncology and Hyperthermia of the AMC (n = 13) and to the BVI (n = 3). Most of these patients were referred for palliation of unresectable RAS after other treatments had failed.
The mean patient age at the time of referral was 68.5 years (range, 48-91 years). One patient (the only man) had been treated for Hodgkin disease 9.5 years before the RAS occurred. All other patients had been treated for breast cancer either by modified radical mastectomy (MRM) (3 patients) or breast conservation (10 patients). In 1 patient, the primary treatment for breast cancer was unknown. One patient had inoperable breast cancer and received primary radiotherapy (70 Gy) only. One patient underwent an MRM followed by 35 Gy to the internal mammary chain region only. All other patients had received at least 50 Gy locally and/or regionally (range, 50-66 Gy). The interval between radiation of the original tumor and the diagnosis of RAS ranged between 19 months and 212 months (median, 86 months).
Histology of the RAS varied. The majority of the tumors (11 of 16 tumors; 69%) were diagnosed as (lymph)angiosarcoma (Tables 1 and 2). Initial treatment of the RAS consisted mainly of mastectomy (9 patients) and, otherwise, attempted radical resection (3 patients). Four tumors were considered unresectable at first diagnosis. Two of these patients received chemotherapy before referral for RT-HT, and both had progressive disease during chemotherapy. One of these patients (Patient 9) had metastatic disease at the time of RT-HT. Table 1 lists the 4 patients who received RT-HT for the treatment of newly diagnosed RAS, including either RT-HT alone for unresectable RAS or adjuvant RT-HT after surgery. The table also lists the interval between radiation of the original tumor and the diagnosis of the RAS, tumor type, and tumor grade (if available). Table 2 lists the 12 patients who were referred after failure of their initial treatment. Along with interval, tumor type, and tumor grade, Table 2 also lists the type of and response to the initial treatment of RAS. Tumor grade often was not provided in pathology reports. The patients are numbered according to their date of referral to AMC or BVI for RAS.
Table 1. Patients Referred for (Part of) Initial Treatment of Radiation-Associated Sarcoma According to Tumor Type and Grade
Interval CA/RAS, mo
Initial Treatment of RAS
Abbreviations: CA, carcinoma; HT, hyperthermia; NOS, not otherwise specified; RAS, radiation-associated sarcoma; RT, radiotherapy; RT-HT, radiotherapy combined with hyperthermia; UK, unknown.
Adjuvant RT-HT after resection
Unresectable, primary RT-HT
Unresectable, primary RT-HT
Adjuvant RT-HT after resection
Table 2. Patients Referred After Failure of Initial Treatment of Radiation-Associated Sarcoma According to Tumor Type and Grade, Initial Treatment, and Outcome of Initial Treatment
Interval CA/RAS, mo
Initial Treatment of RAS
Local Failure (mo)
Abbreviations: CA, carcinoma; MFH, malignant fibrous histiocytoma; NOS, not otherwise specified; RAS, radiation-associated sarcoma; UK, unknown.
After undergoing macroscopically complete resection of RAS without evident tumor, 3 patients (Patients 5, 6, and 16) were referred for adjuvant RT-HT to obtain optimal long-term locoregional control. One of these patients had recurrent RAS after mastectomy, for which she underwent a second radical resection followed directly by RT-HT. The 13 other patients had macroscopic tumors at referral. In 2 patients, only minor disease (1 nodule measuring 1 cm and 3 nodules measuring 1 cm, respectively) was palpable. In all other patients, there was an area of multiple nodules or patchy discolorations, or a larger tumor mass was palpable, usually on the chest wall or in the shoulder girdle. The tumor dimensions at referral for RT-HT were 0 in the 3 patients referred for adjuvant RT-HT (Table 3); tumor dimensions are listed in Table 4. Radiotherapy at the AMC typically consisted of 32 Gy in 8 fractions given twice per week over a period of 4 weeks; at the BVI, patients received 36 Gy in 12 fractions given 4 times per week. The radiation techniques applied were similar to the standard techniques used for the treatment of recurrent breast cancer, usually photons (6 MV or 10 MV) for the lateral chest wall, axillary, and periclavicular area and electrons (8-15 MeV) for the anterior chest wall. Hyperthermia was given once per week within 1 hour of radiotherapy at AMC or twice weekly at BVI (Tables 3 and 4). Heat was induced electromagnetically by using externally applied, contact flexible microstrip applicators operating at 434 MHz (Fig. 1). For all patients, temperatures were measured on the skin. The target temperature was between 41°C and 43°C for 1 hour. Treatment fields covered at least the macroscopic tumor or the area of surgery with a wide margin. This area was too large to be treated in a single session for 3 patients at AMC; in these patients, 2 hyperthermia sessions per week were scheduled with an overlap of adjacent treatment fields.
Table 3. Adjuvant Radiotherapy Combined With Hyperthermia: Treatment Schedule and Outcome
Survival After RT-HT (mo)
Abbreviations: Gy, grays; HT, hyperthermia; RAS, radiation-associated sarcoma; RT-HT, radiotherapy combined with hyperthermia; UK, unknown; WLE, wide local excision.
WLE of recurrent RAS
Resection of RAS, axilla
Resection of RAS, chest wall
8×4 Gy+4× HT
Table 4. Radiotherapy Combined With Hyperthermia for Macroscopic Disease: Tumor Dimensions, Treatment Schedule, and Outcome
The maximum response during or after RT-HT was reported as a complete response if all target lesions had disappeared and a partial response if there was a decrease >30% in the sum of greatest dimensions of the target lesions (according to Response Evaluation Criteria in Solid Tumors [RECIST] criteria).17 Response and local control were evaluated by clinical examination of the patients. Actuarial local control was calculated from the start of RT-HT to the date of local failure, and patients were censored at the date of death or last follow-up. Actuarial survival was calculated from the start of RT-HT to the date of death, and patients were censored at the date of last follow-up. For comparison with other series, actuarial survival from the date of diagnosis of RAS also was calculated. Toxicity was considered acute when it occurred within 3 months of the start of RT-HT and late when it occurred >3 months after the start of RT-HT.
The majority of patients finished treatment according to the protocol. The first 3 patients at AMC had different treatment schedules. The first patient with RAS (in 1984) was offered a scheme of 3 fractions of 6 Gy plus 3 × hyperthermia, but she refused further treatment after the first RT-HT session. The second patient with RAS (in 1985) received 6 fractions of 2.5 Gy, and each fraction was combined with HT. The patient with RAS who received treatment in 1992 received 8 fractions of 4 Gy combined with 6 × hyperthermia.
The median follow-up of all patients who remained alive was 12 months. The median survival from the start of RT-HT for all 16 patients was 9.5 months (range, 0-68 months) (Tables 3 and 4). Overall, 5 patients were alive after RT-HT at 8 months, 11 months, 12 months, 39 months, and 68 months, respectively, including 4 patients without signs of recurrence. The median survival after the diagnosis of RAS was 15.5 months (range, 3-204, months), and the actuarial 3-year survival rate was 31%.
The treatment schedules, response to RT-HT, time to local failure, and time to death for all 16 patients are listed in Tables 3 and 4. Response could not be evaluated for 4 patients: One patient refused further treatment after 1 session and died within 2 weeks of an unrelated cause (Patient 1, Table 4), and 3 patients received adjuvant treatment after undergoing a macroscopically complete resection (Table 3). For the remaining 12 patients, the response rate was 75% (7 complete responses and 2 partial responses), and 3 patients (25%) had stable disease.
Two of the patients who received adjuvant treatment after a macroscopically complete resection remained tumor free until last follow-up (at 8 months and 68 months). One patient who received adjuvant treatment after resection of recurrent RAS developed metastatic disease shortly after RT-HT and died 10 months later.
Four of 13 patients who were treated for macroscopic disease (31%) had durable local control until either death or last follow-up (at 5 months, 7 months, 11 months, and 39 months). Figure 2 shows Patient 14 before treatment (Fig. 2a) and with complete response after 11 months (Fig. 2b).
Because the skin is the target area for all patients, episodes of acute toxicity from reirradiation and hyperthermia consisted of erythema (7 patients; 44%) and moist desquamation (5 patients; 31%). Other complaints during treatment were nausea (4 patients; 25%), dysphagia (1 patient), and pruritus (1 patient).
Late toxicity occurred in 7 of the 16 patients and consisted of hyperpigmentation (3 patients; 19%), telangiectasia (2 patients), and fibrosis (2 patients). One patient developed severe late toxicity (grade 4), which consisted of ischemia of the arm on the treated side, resulting in forearm amputation 5 years after RT-HT (Patient 6).
Radiation-associated sarcoma is a very rare disease, although an increase in the incidence of RAS of the breast and chest wall is expected as a result of a rising prevalence of breast cancer and a higher percentage of patients who receive irradiation as part of multidisciplinary treatment. The postradiotherapy 10-year cumulative risk for developing RAS is estimated at 0.03% to 0.8%. Some large series have been published that included 32 to 948 patients identified with RAS.4, 18-23 Most of those reports dealt with epidemiology rather than treatment options, and others focused on specific tumor types. The prognosis for patients with RAS is poor and is comparable to the prognosis for patients with sporadic soft tissue sarcoma according to most authors, although Gladdy et al reported a worse prognosis for 130 patients who had RAS compared with their entire cohort of >7000 patients who had sporadic soft tissue sarcomas.3, 4, 8 This may be because of a higher rate of incomplete resections caused by multifocal disease and the less frequent use of adjuvant radiotherapy for patients with RAS in their series.8
The median latency period ranges from 59 to 103 months in the literature, which is in line with our data.8, 24-28 RAS may be of any type, but angiosarcoma was the most common in our patient group and in other series of thoracic RAS.2-4, 19, 22 Angiosarcoma may be associated with chronic lymphedema of a limb or of the chest wall (Stewart-Treves [ST] syndrome), which is diagnosed most frequently after axillary lymph node dissection as a part of breast cancer treatment and can occur in the lymphedematous arm itself or in the shoulder, axillary, or chest wall region. The latency period of ST syndrome appears to be 10 to 12 years, slightly longer than that for RAS.6, 29, 30 Several patients in our series also had angiosarcoma that may have been associated with chronic lymphedema and, thus, was hard to distinguish from ST syndrome.
Evidence for optimal treatment is scarce because of the extreme rarity of the disease. Generally recommended treatment for RAS is radical surgery, the objective of which is microscopically complete (R0) resection. If RAS develops in the breast after breast conservation, then a simple mastectomy may be the surgery of choice. When the initial breast cancer is treated with mastectomy and the sarcoma arises in the mastectomy scar, a major soft tissue resection or full thickness chest wall resection with plastic surgical reconstruction may be required. Radical surgery (R0 resection) is possible in 50% to 75% of patients who have RAS located in any part of the body and results in 5-year local relapse-free survival rates of 34% to 46% and 5-year survival rates of 27% to 44%.7, 17, 31, 32 A review of all 92 patients of angiosarcoma after breast-conserving therapy reported in the English literature revealed that 55 of 75 patients (73%) with at least 1 year of follow-up developed a local recurrence after mastectomy (n = 72) or wide local excision (n = 3).33 In that review, all patients who received adjuvant radiotherapy (n = 1) and/or chemotherapy (n = 3) experienced a local recurrence. Most recurrences (84%) developed within 1 year after surgery.33 Very few cases of surgical salvage after a local RAS recurrence have been reported. In 12 of our 16 patients (75%), radical surgery initially was attempted.
The role of chemotherapy in RAS has not been established. Sher et al retrospectively examined the response of metastatic angiosarcoma of the breast (primary and radiation-associated) to first-line cytotoxic chemotherapy and reported a 48% overall response rate to anthracycline-ifosfamide or gemcitabine-taxane chemotherapy.34 One case report describes a response to paclitaxel in a radiation-associated breast angiosarcoma with local control that lasted for 5 months.35 Docetaxel and thalidomide also are used as palliative treatment that occasionally results in a good response.36, 37
In most cases reported in the literature, patients did not receive reirradiation because of prior radiation therapy and concerns about toxicity. Two cases from the AMC have been described before (Patients 2 and 3).15, 16 Apart from those 2 patients, only 11 patients who received (neo)adjuvant reirradiation could be identified in the literature, and 8 patients receiving reirradiation as salvage treatment for unresectable, macroscopic disease. Three of the latter patients received RT-HT (Table 5). In addition, 4 patients reported by Abraham et al received reirradiation, but the report had insufficient detail to be included in Table 5.38 Feigenberg et al reported 3 patients who received hyperfractionated reirradiation (1-1.5 Gy 2-3 times daily up to a total dose of 50-60 Gy), which was followed by extensive resection and reconstruction in 2 of those 3 patients. The result was no evidence of recurrent disease after 18 months, 38 months, and 39 months (Table 5).33, 39
Table 5. Adjuvant Radiotherapy or Salvage Treatment With Radiotherapy With or Without Hyperthermia in the Literature
Initial Treatment for RAS
Local Failure (mo)
Local Failure (mo)
Overall Survival (mo)
Abbreviations: CT, chemotherapy; CR, complete response; CWR, chest wall resection; HT, hyperthermia; ±HT, with or without hyperthermia; RT, radiotherapy; RT-HT, radiotherapy combined with hyperthermia; NA, not applicable; pCR, pathologic complete response; PD, progressive disease; PR, partial response; RAS, radiation-associated sarcoma; SD, stable disease; SM, simple mastectomy; UK, unknown; WLE, wide local excision.
Combining our series of 16 patients (Tables 3 and 4) with 19 patients identified in the literature (Table 5) yielded the following results of reirradiation with or without hyperthermia. Response to treatment was reported for 23 patients. The complete response rate was 52% (12 of 23 patients). In 21 patients, reirradiation (with or without hyperthermia) was the only treatment applied for macroscopic disease. The local relapse-free survival rate was 29% (6 of 21 patients), and 4 of those patients remained alive at 11 months, 22 months, 36 months, and 39 months. In 14 patients, reirradiation with or without hyperthermia was used in combination with attempted radical surgery as (neo)adjuvant treatment, and 7 of 14 patients (50%) remained without local recurrence until death or last follow-up for a median of 38 months (range, 8-174 months).
Although this is a small retrospective series with various treatment schedules used and various tumor histologies, the response and local control rates are promising and appear to contradict the common assumption that reirradiation is not useful in patients with RAS. In patients with unresectable RAS, reirradiation with or without hyperthermia can offer good palliation and long-term local control. Given the high local recurrence rate after attempted radical resection commonly reported, (neo)adjuvant reirradiation may be considered. Our series and the cases from the literature do not allow comparison of the effectiveness of various schedules of reirradiation or comparison of reirradiation alone with reirradiation plus hyperthermia.
In conclusion, despite the disadvantages of an observational study with a small number of patients, differences in treatment schedules, and variability in tumor histology, the current results suggest that reirradiation combined with hyperthermia is a feasible treatment for RAS. The 75% response rate for the 12 patients who had unresectable, macroscopic disease and the durable local control achieved in 4 of those patients and in 2 of 3 patients who received adjuvant treatment after microscopically incomplete resection are in line with the scarce data about reirradiation in the literature. These response and local control rates suggest that reirradiation plus hyperthermia for RAS is promising.