Pulsed reduced dose-rate radiotherapy: a novel locoregional retreatment strategy for breast cancer recurrence in the previously irradiated chest wall, axilla, or supraclavicular region
Purpose Reirradiation of breast cancer locoregional recurrence (LRR) in the setting of prior post-mastectomy radiation poses a significant clinical challenge due to the high risk for severe toxicity. In an attempt to reduce these toxicities, we have developed pulsed reduced dose-rate radiotherapy (PRDR), a reirradiation technique in which a series of 0.2 Gy pulses separated by 3-min time intervals is delivered, creating an apparent dose rate of 0.0667 Gy/min. Here we describe our early experience with PRDR. Patients and methods We reirradiated 17 patients with LRR breast cancer to the chest wall, axilla, or supraclavicular region using PRDR. The median prior radiation dose was 60 Gy. We delivered a median PRDR dose of 54 Gy (range 40–66 Gy) in 1.8–2.0 Gy per fraction. Eight patients received concomitant low dose capecitabine for radiosensitization. The median treatment volume was 2,084 cm3 (range 843–7,881 cm3). Results At a median follow-up of 18 months (range 4–75 months) only 2 patients have had tumor failure in the treatment region. Estimated 2-year local control rate is 92%. Treatment was well tolerated with 4 patients experiencing grade 3 acute skin toxicity. Despite a median cumulative dose of 110 Gy (range 80–236 Gy), there has been only one grade 3 and one grade 4 late toxicity. Conclusions With a median follow-up of 18 months, PRDR appears to be an effective method to reirradiate large volumes of previously irradiated tissue in selected patients with locoregional chest wall, axilla, and supraclavicular recurrences.
KeywordsBreast cancer Reirradiation Reduced dose-rate Recurrence Radiotherapy
Locoregional recurrence (LRR) in the previously irradiated post-mastectomy breast cancer patient poses a significant clinical challenge. While the LRR risk is low in this patient population, it is not negligible. LRR rates are reported to be 6% at 5 years  and 14% at 18 years  in women with node positive breast cancer who have undergone comprehensive multimodality therapy including mastectomy, axillary surgery, chemotherapy, and post-mastectomy radiation therapy (PMRT). Radiation therapy (RT) is a common LRR treatment option in those who have not undergone PMRT and recent analysis of LRR treatment in the Danish Breast Cancer Cooperative Group (DBCG) 82b&c trials revealed inferior persistent local control at ≥2 years when omitting RT from salvage treatment in previously unirradiated patients .
However, reirradiation of patients previously treated with PMRT is vastly more challenging, particularly when attempting to deliver potentially curative doses of ≥50 Gy. The cumulative RT dose delivered to the surrounding tissues place them at high risk for severe toxicity. These tissues include the brachial plexus, chest wall skin and subcutaneous tissue, uninvolved lymphatics, shoulder joint, and underlying pulmonary and cardiac tissue. Late reirradiation sequelae potentially include painful brachial plexopathy, nonhealing chest wall ulceration, lymphedema, shoulder joint dysfunction, pneumonitis, and cardiac toxicity. These iatrogenic sequelae have the potential of becoming the patient’s chief complaint until their death.
RT delivered below standard dose-rates can preferentially protect normal tissue while producing almost identical tumor cell kill. This can occur due to the superior repair capacity of late responding normal tissues compared to tumor cells. In conventional RT a dose of 2 Gy is delivered at a dose-rate of 4–6 Gy/min. By reducing the apparent dose-rate and increasing the treatment time, cellular repair processes can occur during irradiation. The dose-rate effect, caused by repair of sublethal damage, is most dramatic between 0.01 and 1 Gy/min . A reduced dose-rate can be obtained by dividing a standard treatment fraction into a number of equal sub-fractions that are delivered in a pulsed manner separated by a fixed time interval, thus allowing for repair during each sub-fraction . Early clinical studies investigating fractionated reduced dose-rate external beam RT in the curative setting have demonstrated an improved therapeutic ratio in oropharynx cancer [6, 7] and breast cancer [8, 9].
A second intriguing phenomenon is low dose hyper-radiosensitivity (LDHRS), which is increased radiosensitivity to doses <0.3–0.5 Gy. There are data available on the low dose response of many human cell lines , the majority of which demonstrate LDHRS [11, 12]. Furthermore, analysis of LDHRS in metastatic breast cancer tumors has revealed a potential advantage to low dose ultrafractionated therapy compared to more standard fraction sizes .
We have developed a reirradiation technique where we deliver a series of 0.2 Gy pulses separated by 3-min time intervals, creating an apparent dose-rate of 0.0667 Gy/min . This treatment strategy theoretically takes advantage of LDHRS in tumor cells and increased intrafraction sublethal damage repair in normal tissues. We have termed this reirradiation technique pulsed reduced dose-rate radiotherapy (PRDR) . In this report, we describe our early experience with PRDR in previously irradiated breast cancer patients with LRR.
Patients and methods
Between November 2000 and April 2007, 17 patients with LRR breast cancer were treated with PRDR at the University of Wisconsin. Informed consent was obtained for all patients. The median age at the time of LRR was 57 years (range 38–79). All patients had been previously treated with surgery, RT, and chemotherapy. Prior initial surgery included mastectomy in 13 patients and lumpectomy in 4 patients; all 4 lumpectomy patients recurred in the regional lymphatics without local recurrence in the breast. Prior initial RT was delivered to a median dose of 60 Gy (range 40–182 Gy); 3 of the patients had been reirradiated at standard dose-rate for a previous LRR and had received cumulative doses of 100, 108, and 182 Gy prior to PRDR. All patients received multiple cycles of chemotherapy in the treatment of their initial and/or recurrent disease; some patients were also treated with tamoxifen or an aromatase inhibitor, and/or trastuzumab. The median time from the date of initial diagnosis until diagnosis of the PRDR treated LRR was 58 months (range 19–213 months). In 10 of the patients, PRDR was delivered for local control only, including 8 patients who had a low burden of distant metastatic disease at the time of PRDR treatment. The remaining 7 patients had LRR only and were treated with curative intent.
PRDR was administered broadly to the region of recurrence, including the chest wall in 6 patients, axilla in 14 patients, and the supraclavicular region in 7 patients. Four of 17 patients underwent surgical resection of the LRR recurrence prior to PRDR; including axillary dissection in 3 and resection of the LRR from the mastectomy scar in one. The remaining 13 patients were treated for gross disease ranging from diffuse skin recurrence to multiple axillary masses measuring up to 4.5 cm.
We delivered a median dose of 54 Gy (range 40–66 Gy) in 1.8–2.0 Gy per fraction, resulting in a median cumulative dose of 110 Gy (range 80–236 Gy). All fractions were delivered with 0.2 Gy pulses, each separated by 3 min, thus delivering the radiation at an apparent dose-rate of 0.0667 Gy/min. The treatment volumes were large with the median volume of tissue irradiated to ≥50% of the prescription dose at 2,084 cm3 (range 843–7,881 cm3). Twelve patients were treated with PRDR using photons alone, 2 were treated with PRDR using electrons alone, and 3 were treated using both photons and electrons.
Eight patients received chemotherapy concomitantly with PRDR. Seven were treated with capecitabine at a radiosensitizing dose of 1,000–1,500 mg divided into 2 doses daily. One patient was undergoing treatment with trastuzumab and albumin-bound paclitaxel for lung and bone metastases when she developed PET positive left supraclavicular adenopathy. Since the remainder of her disease had remained stable, the systemic therapy was continued during the delivery of PRDR to maintain systemic disease control.
Clinical response was measured by physical examination and radiographic studies including CT scan and PET scans where applicable. Acute and late toxicity were graded using the NCI CTC toxicity scale. All data were collected in a retrospective fashion. The collection and analysis of data was performed under an IRB approved protocol. Kaplan-Meier estimations and plots were generated using MedCalc v18.104.22.168.
Treatment was well tolerated, with 13 patients experiencing ≤grade 2 acute skin toxicity and only 4 patients experiencing grade 3 acute skin toxicity. Despite a median cumulative dose of 110 Gy (range 80–236 Gy), there has been only one grade 3 and one grade 4 late toxicity. One patient developed a nonhealing chest wall ulcer requiring surgical repair with a latissimus flap and skin graft (grade 4). The second patient required multiple surgical revisions for a non-healing chest wall wound prior to PRDR. Following PRDR she developed a non-healing ulcer and is currently pursuing hyperbaric oxygen therapy (grade 3). Of note, the patient treated with concomitant albumin bound paclitaxel and trastuzumab had a complete response of her left supraclavicular adenopathy to PRDR. She recently relapsed in the lung 18 months post-PRDR, but her PRDR treated disease remains controlled. She experienced only grade 2 acute skin toxicity with no significant late toxicity. There has been no evidence of brachial plexopathy in any patient despite targeting the axilla and/or supraclavicular area in 15 of the 17 patients. Two patients have mild shoulder joint dysfunction with discomfort on movement that is successfully managed with analgesics on an as needed basis.
LRR presents a major clinical challenge in breast cancer patients previously treated with mastectomy, chemotherapy, and comprehensive PMRT. While the rate of LRR in this scenario is relatively low, it is not negligible, with a probability of 6% at 5 years  and 14% at 18 years . Recent analysis of the patterns of LRR in 1,538 patients treated with PMRT on DBCG 82b&c revealed 79 patients with LRR alone (5%) and 73 patients with LRR and simultaneous DM (5%) . Although the 5% presenting with LRR alone are potentially curable, further analysis reveals a 50% probability of DM in this cohort at 2.2 years and an 80% probability at 10 years .
Treatment options for LRR in the postmastectomy setting depend on the initial therapy regimen as well as the location of the LRR, and typically include RT in previously unirradiated patients. A second course of RT is frequently omitted from previously irradiated patients due to toxicity concerns. However recent analysis of the DBCG 82b&c trials revealed that RT provides improved locoregional control after LRR . PMRT patients did not receive RT as part of their LRR therapy, while 64% of unirradiated patients were treated with salvage RT. Looking at the complete remission and persistent locoregional control rate following LRR, it is evident that the use of RT improved persistent locoregional control at ≥2 year after LRR: 45% for surgery + RT, 38% for RT alone, 22% for surgery alone and only 11% for systemic treatment alone . Other studies confirm significantly inferior outcomes in patients with chest wall recurrences who are not treated with RT . Therefore, the use of RT may improve locoregional control compared to chemotherapy alone in PMRT patients, if the reirradiation toxicity can be minimized.
In this same context, it has been our experience that systemic therapy in previously irradiated sites may fail to control LRR while controlling distant disease. This may reflect a compromise of the microvasculature in the previously irradiated site, thus inhibiting drug penetration in these areas. One patient in this series may have reflected this phenomenon. She had lung and bone metastases that remained controlled with systemic therapy for 20 months prior to developing 2 PET positive lymph nodes in the left supraclavicular region. We subsequently treated the left supraclavicular region with 54 Gy of PRDR (cumulative dose of 110 Gy) with a complete response. She remained in systemic remission for a total of 38 months, at which time a lung lesion began to grow. Her PRDR treated supraclavicular region continues to be without evidence of disease progression.
Comparison of reirradiation studies
Number of reirradiated patients
Median FU (months)
Median prior RT dose (Gy)
Median reirradiation dose (Gy)
Local control (%)
Chronic ulceration, necrosis or delayed wound healing
Ability to treat deep axillary or SC tissue?
16% Grade 3
8% Grade 3
van der Zee 
34% 2–3° burns
27% ≤Grade 3
23% Grade 3
Our PRDR technique compares favorably to other reirradiation data. We have achieved a 92% 2-year locoregional control rate while keeping toxicities low. Two patients developed late chest wall ulceration, one grade 3 and one grade 4. The treatment course of the patient with the grade 4 toxicity was complicated by immediate post-mastectomy TRAM flap reconstruction that developed post-operative grade 3 wound complications needing surgical revision, potentially placing this tissue at higher risk for scarring and necrosis. She was subsequently treated to a PRDR dose of 50 Gy (cumulative dose of 99 Gy) with concomitant capecitabine at 1,500 mg/day. The second patient, who had a left chest wall skin recurrence, also underwent multiple surgical resections prior to PRDR due to grade 3 surgical wound complications. She was treated to a PRDR dose of 66 Gy (cumulative dose of 126 Gy) using 9 MeV electrons prescribed to 80% depth-dose with daily 0.5 cm tissue-equivalent bolus. Therefore, it is possible that reirradiation in the setting of complicated post-surgical healing places these patients at an increased risk of developing non-healing ulceration at the surgical site, particularly when delivering ≥60 Gy of PRDR or using radiosensitizing capecitabine concomitantly.
Another advantage of PRDR is that it can be delivered using photons, electrons, or a combination of both, thus giving flexibility to treat disease at any depth. We can treat superficial chest wall recurrences with PRDR using either tangential photons or electrons, maximizing dose to the superficial tissue while minimizing the dose to the underlying lung and heart tissue. Alternatively, deeper axillary or supraclavicular recurrences can be treated with PRDR using high-energy photons, sparing skin and subcutaneous tissue. Historically, these deeper situated LRR are challenging to treat, are often treated with systemic therapy alone in the PMRT setting, and have worse outcomes [3, 21, 22]. These advantages are in contrast to salvage brachytherapy or RT with concomitant hyperthermia, both of which are limited to smaller volume recurrences of the skin and superficial subcutaneous tissues with small additional margins.
An additional benefit of PRDR is the ability to create a treatment plan with standard commercial planning software and to deliver PRDR with a standard linear accelerator, therefore making it a more practical technique for the typical community radiation oncology practice. Brachytherapy and hyperthermia are much more complex techniques, require specialized equipment and treatment planning software, as well as increased physics and nursing support, all of which are barriers that limit their availability.
We hypothesize that PRDR takes advantage of both improved sublethal damage repair of normal tissue and LDHRS of the tumor cells. It is well known from irradiated cell survival data that as the radiation dose-rate is reduced, the survival curve becomes shallower due to the repair of sublethal damage, and that this effect is most dramatic between 0.01 and 1 Gy/min . A lower apparent dose-rate can have profoundly different effects depending on the inherent biology of the irradiated tissue. This stems in part from the fact that typically only the β component of radiation damage is repairable. When the α/β ratio is large, as is typical of most tumors, little impact on repair is expected from lower dose-rates. However, when the α/β is small, as is typical of late responding normal tissues, greater repair occurs with lower dose-rates. Therefore, reduced dose-rates should theoretically reduce normal tissue complications. PRDR potentially exploits the therapeutic gain from improved sublethal damage repair of normal tissue over tumor tissue by delivering the radiation dose at an apparent dose-rate of 0.0667 Gy/min.
PRDR also potentially exploits LDHRS, a phenomenon that results in considerably lower cell survival following low doses of radiation (<0.3–0.5 Gy). LDHRS has been characterized in over 40 human cell lines . Lin and Wu demonstrated a statistically significant decrease in cell survival when delivering 2 Gy in 10 subfractions of 0.2 Gy compared to a single 2 Gy fraction in HT-29 adenocarcinoma cells . Furthermore, potential LDHRS was demonstrated in patients with multiple skin metastases, including metastatic breast cancer, when significant tumor growth delay was noted with an ultrafractionated regimen of 0.5 Gy three times daily compared to standard 1.5 Gy once daily (P < 0.05) .
Clearly, to exploit LDHRS, normal tissue should either not exhibit LDHRS or should have transition doses (the dose threshold below which LDHRS occurs) that are much lower than that of the tumor as well as the pulse dose to be used. Tissue with minimal proliferation, such as the brachial plexus or previous irradiated connective tissue would not be expected to exhibit LDHRS. Of note, the total time to deliver the 10 0.2 Gy subfractions in Lin and Wu’s study was 4 min  (an apparent dose-rate of 0.5 Gy/min), in contrast to our PRDR pulsing strategy which takes approximately 30–45 min to deliver. The longer delivery time of PRDR may increase normal tissue sublethal damage repair, thus providing a potential normal tissue toxicity advantage along with the LDHRS benefit demonstrated above.
There is no consensus standard of care to treat locoregionally recurrent breast cancer in patients who have previously been irradiated. These patients have historically done poorly, particularly when they recur in the deep axilla or supraclavicular region. PRDR is a model-based external beam radiation methodology that takes advantage of: (1) an apparent reduced dose-rate of 0.0667 Gy/min to exploit the improved sublethal damage repair of normal tissue; as well as (2) low subfraction doses of 0.2 Gy to exploit the LDHRS of the tumor cells. We recognize that this report is preliminary and any conclusions drawn must be tempered by its retrospective nature, the low patient number, and the limited median follow-up. However, PRDR was well tolerated in this initial cohort of patients, with minimal acute and late toxicities, and an overall local control that is comparable to other reirradiation series. With a median follow up of 18 months, PRDR appears to be an effective method to reirradiate large volumes of previously irradiated tissue in select patients with locoregional chest wall, axilla, and supraclavicular recurrences.
- 2.Nielsen HM, Overgaard M, Grau C et al (2006) Study of failure pattern among high-risk breast cancer patients with or without postmastectomy radiotherapy in addition to adjuvant systemic therapy: long-term results from the Danish Breast Cancer Cooperative Group DBCG 82 b and c randomized studies. J Clin Oncol 24(15):2268–2275PubMedCrossRefGoogle Scholar
- 4.Hall EJ, Giaccia AJ (2006) Radiobiology for the radiologist, 6th edn. Lippincott Williams & Wilkins, PhiladelphiaGoogle Scholar
- 18.Niehoff P, Dietrich J, Ostertag H et al (2006) High-dose-rate (HDR) or pulsed-dose-rate (PDR) perioperative interstitial intensity-modulated brachytherapy (IMBT) for local recurrences of previously irradiated breast or thoracic wall following breast cancer. Strahlenther Onkol 182(2):102–107PubMedCrossRefGoogle Scholar