Outcome after ipsilateral breast tumor recurrence in patients who receive accelerated partial breast irradiation




The objective of this study was to examine clinical outcomes and patterns of failure in patients with early stage breast cancer who developed an ipsilateral breast tumor recurrence (IBTR) after breast-conserving therapy (BCT) using accelerated partial breast irradiation (APBI).


In total, 1440 patients (1449 tumors) with early stage breast cancer who underwent BCT were treated with the MammoSite device to deliver APBI (34 Gray [Gy] in 3.4-Gy fractions). One thousand two hundred fifty-five patients (87%) had invasive breast cancer (IBC) (median tumor size, 10 mm), and 194 patients (13%) had ductal carcinoma in situ (DCIS) (median tumor size, 8 mm). The median follow-up was 60 months.


Fifty patients (3.5%) developed an IBTR for a 5-year actuarial rate of 3.61% (3.65% for IBC and 3.36% for DCIS). It was determined that 36 recurrences (72%) represented new primary cancers, and 14 recurrences (28%) represented recurrences of the index lesion. Of the 32 recurrences with known histology, 78% were IBC, and 22% were DCIS. After IBTR, 28 of 38 patients (74%) underwent salvage mastectomy, and 9 of 38 patients (26%) had a second attempt at BCT. Adjuvant therapies included tamoxifen in 8 patients (16%) and systemic chemotherapy in 6 patients (12%). The 3-year rates of disease-free survival, cause-specific survival, and overall survival after IBTR were 58.7%, 92.1%, and 80.5%, respectively.


With 5 years of follow-up, APBI produced clinical outcomes and patterns of failure comparable to those achieved with whole breast irradiation. Patients who developed an IBTR after APBI had excellent 3-year survival outcomes after salvage treatments. Cancer 2012. © 2012 American Cancer Society.


With more than 25 years of follow-up, it has been demonstrated that breast-conserving therapy (BCT) provides long-term outcomes equivalent to the outcomes achieved with mastectomy in the management of early stage breast cancer. Although whole-breast irradiation (WBI) has been considered the standard adjuvant radiation therapy (RT) technique in these patients, 1 concern that remains is the 6 to 7 weeks required to deliver RT. Some studies indicate that up to 20% of patients do not receive adjuvant RT, and protracted treatment times and distance from an RT facility have been identified as contributing factors.1 Accelerated partial breast irradiation (APBI) is a relatively new technique that allows for the delivery of adjuvant RT while shortening the treatment course to 1 week or less. Currently, APBI can be delivered using multiple techniques, including interstitial catheters, a balloon-based catheter, or external-beam RT. With 5 to 10 years of follow-up, it has been demonstrated that APBI is associated with excellent clinical outcomes.2, 3

One concern with BCT is how to address recurrent local disease within the previously irradiated breast. One strategy to help stratify these recurrences is to define them either as a true recurrence of the treated malignancy or as a de novo primary carcinoma.4 The basis for this stratification is that new primary carcinomas within the previously treated breast are associated with a more favorable prognosis compared with neoplasms believed to represent true recurrences.5

Salvage mastectomy has been the most common treatment approach for these ipsilateral breast tumor recurrences (IBTRs) after BCT and produces local control rates that range between 50% and 90%.6-8 However, if a patient develops an IBTR after APBI that is believed to represent a new primary cancer, then some data suggest they may be candidates for a second attempt at BCT (primarily because of a smaller volume of initially irradiated breast tissue). Currently, limited data are available on repeat BCT in patients who previously underwent WBI, and the techniques used are quite hetereogenous.9-11 In a similar fashion, even more limited data exist on outcomes after IBTR in patients who received APBI. The purpose of the current analysis was to examine outcomes after IBTR in a large cohort of patients who received APBI and to evaluate patterns of failure after APBI to determine the feasibility of repeat BCT.



Between May 2002 and July 2004, 97 institutions participated in a registry trial designed to collect data on the use of the MammoSite device (Hologic Inc., Bedford, Mass) as a modality for the delivery of APBI. The trial was initiated concurrently by the manufacturer with the clearance of the device for clinical use in May 2002. The goals and objectives of the trial were to provide a method to prospectively, objectively, and systematically collect data on the clinical use of the applicator. Additional data on the technical reproducibility of the device on a large scale, acute and chronic toxicity, cosmesis, efficacy, and adherence to placement in appropriate patients also were sought. On November 17, 2003, the American Society of Breast Surgeons (ASBS) assumed primary management of the trial. Enrollment was closed in July 2004. Patients are followed by either their radiation oncologist and/or their surgeon, and the data collected include cosmetic evaluation, receipt of adjuvant therapy, imaging assessment, recurrence and treatment of recurrence, survival status, radiation recall, and toxicities.

Information on patient enrollment, eligibility criteria, data collection, and management has been published.3 It is noteworthy that the ASBS has continuously collected information in addition to that originally intended at the start of the trial, recognizing the importance of these data in helping to establish the efficacy of APBI before the availability of data from prospective, randomized trials. Since June 2006, a new, independent, full-service contract research organization (BioStat International, Inc. [BSI], Tampa, Fla) assumed full responsibility of the independent management and statistical analysis of the trial data for the ASBS. Since then, an extensive review and updating of database records (for further data accuracy verification) have been performed with emphasis on obtaining accurate cosmesis, recurrence, survival, and toxicity information. Definitions of recurrence and toxicity categories and follow-up visit windows were provided by the ASBS and its independent scientific advisory committee to BSI. Management and analysis of the data at BSI occurs only through in-depth discussions between statisticians at BSI and the ASBS. Note that topics for presentation and/or publication and rules for authorship and data analysis (including the entire preparation of this article) remain under the complete control of the scientific advisory committee of the ASBS. Because data entry and processing are continuous processes for this program, a data cutoff date was chosen for this article to allow for auditing and analysis (May 15, 2009). In total, 1449 breasts treated with APBI (1440 treated patients) are the subject of this analysis (no boost patients are included). Informed consent was obtained from all patients and/or their guardians before they received treatment.

Local, Regional, and Distant Disease Recurrence

For the purposes of the current analysis, an IBTR was defined as the reappearance of cancer in the treated breast (before the development of distant metastases) and was confirmed pathologically. Investigators were asked to classify IBTRs by their clinical location in relation to the initial lumpectomy cavity according to the criteria described by Recht et al.4 A true recurrence/marginal miss (TR/MM) was defined as a recurrence of the treated cancer within or immediately adjacent to the primary tumor site. A failure elsewhere (E failure) was defined as an IBTR several centimeters from the primary site and generally was believed to be a new primary cancer. It is important to note that these designations were made based on clinical judgment after physical examination and review of appropriate imaging studies without loss-of-heterozygosity studies.

All time intervals were calculated from the date of MammoSite removal. For evaluations and summaries of disease characteristics, MammoSite treatment parameters, and local recurrences, the unit of interest was a breast. For demographic information, adjuvant therapy, distant disease, and survival, the unit of interest was a patient. Nonparametric estimates of survival, or recurrence-free distributions, or recurrence (failure) distribution were obtained using life table methods. Descriptive statistics consist of numbers and percentages of responses in each category for discrete measures and of medians, minimum values, and maximum values for continuous measures. The SAS statistical software package (version 8.0 or higher; SAS Institute, Inc., Cary, NC) was used to provide all statistical analyses.


In total, 50 patients (3.5%) developed an IBTR as some component of their initial failure for a 5-year actuarial rate of 3.61% (3.65% for invasive breast cancer [IBC] and 3.36% for ductal carcinoma in situ [DCIS]) (Table 1). The 5-year actuarial rate of TR/MM failure was 1.1%, and the 5-year actuarial rate of E failure was 2.6%. Similar findings were noted when examining invasive and in situ cancers, respectively (Table 1). The median time to IBTR was 42.6 months (range, 7.3-93.7 months). The median time to IBTR was 42.0 months for patients with a TR/MM and 39.1 months for patients with an E failure. The median time to contralateral failure was 44.8 months, and the 5-year actuarial rate was 1.9%.

Table 1. Ipsilateral Breast Tumor Recurrences
 All Patients, N = 1449Patients With Invasive Disease, N = 1255Patients With DCIS, N = 194
Type of RecurrenceNo. (%)5-Year Actuarial Rate, %No. (%)5-Year Actuarial Rate, %No. (%)5-Year Actuarial Rate, %
  1. Abbreviation: DCIS, ductal carcinoma in situ.

All breast failures50 (3.5)3.642 (3.3)3.78 (4.1)3.4
True recurrence/marginal miss14 (1)1.111 (0.9)1.13 (1.5)1.2
Failure elsewhere36 (2.5)2.631 (2.5)2.65 (2.6)2.1
P .41 .66 .27

Characteristics of the patients who developed an IBTR are presented in Table 2. The median age was 66.7 years old, and 86.1% of patients were diagnosed mammographically. The median size of all tumor recurrences was 1.6 cm (range, 0.3-11.0 cm). Eighty-two percent of IBTRs were estrogen receptor positive. Overall, 13.2% of IBTRs were grade 1, 21.1% were grade 2, and 31.6% were grade 3 (Table 2). Of the 50 patients who developed an IBTR, 42 presented initially with IBC, and 8 presented initially with DCIS. The histology of the IBTR was known in 32 patients (27 with IBC, 5 with DCIS). In the patients who presented initially with IBC, 24 (89%) had IBC, and 3 (11%) had DCIS. In the 5 patients who presented initially with DCIS, 1 IBTR (20%), was IBC, and 4 IBTRs (80%) were DCIS.

Table 2. Characteristics of Patients Who Experienced an Ipsilateral Breast Tumor Recurrence, n = 50
 No. of Patients (%)
CharacteristicAll IBTR, N = 50True Recurrence/ Marginal Miss, N = 14Recurrence Elsewhere, N = 36
  1. Abbreviations: CT, computed tomography; ER, estrogen receptor; IBTR, ipsilateral breast tumor recurrence; MRI, magnetic resonance imaging.

Age: Median [range], y66.7 [44.0-86.9]63.1 [45.7-86.3]67.2 [44.0-86.9]
Method of detection   
 Mammography31 (86.1)5 (71.4)26 (89.7)
 Palpation 2 (5.6)1 (14.3) 1 (3.4)
 Ultrasound 2 (5.6)1 (14.3) 1 (3.4)
 MRI 1 (2.8)0 (0) 1 (3.4)
 CT 0 (0)0 (0) 0 (0)
 Other 0 (0)0 (0) 0 (0)
 Unknown147 7
Tumor size   
 No. of patients  evaluated24618
 Median [range], cm1.6 [0.3-11.0]1.4 [0.3-3.9]1.6 [0.3-11.0]
Tumor grade   
 1 5 (13.2)0 (0) 5 (17.9)
 2 8 (21.1)0 (0) 8 (28.6)
 312 (31.6)4 (40) 8 (28.6)
 4 1 (2.6)0 (0) 1 (3.6)
ER status   
 Positive23 (82.1)5 (83.3)18 (81.8)
 Negative 5 (17.9)1 (16.7) 4 (18.2)
Local treatment   
 Mastectomy28 (73.7)6 (75)22 (73.3)
 Breast conservation therapy 9 (23.7)2 (25) 7 (23.3)
 Unknown136 7
Systemic treatment   
 Chemotherapy 6 (12)3 (21.4) 3 (8.3)
 Hormone therapy 8 (16)2 (14.3) 6 (16.7)
 Chemotherapy or hormones only 1 (2.6)0 (0) 1 (3.3)
 Unknown 7 (14)4 (28.6) 3 (8.3)

Outcomes After Ipsilateral Breast Tumor Recurrence

Of the 50 patients who developed an IBTR, 38 patients had information available concerning treatment of the recurrence (Table 2). After they developed an IBTR, 28 patients (73.7%) underwent salvage mastectomy, and 9 patients (23.7%) underwent a second attempt at BCT. Adjuvant therapies included tamoxifen in 8 patients (16%) and systemic chemotherapy in 6 patients (12%). The 3-year rates of disease-free survival (DFS), cause-specific survival (CSS), and overall survival (OS) after IBTR were 58.7%, 92.1%, and 80.5%, respectively, for all IBTRs. With regard to the 9 patients who underwent repeat BCT, the 2-year DFS, CSS, and OS rates were 86.7%, 100%, and 100%, respectively. When a cutoff point of 48 months was used to evaluate survival, the patients who developed an IBTR within 48 months of treatment (n = 33) had 3-year DFS, CSS, and OS rates of 53.1%, 95%, and 80.5%, respectively, whereas patients who developed an IBTR ≥48 months after treatment had 3-year DFS, CSS, and OS rates of 80% each.

Although the analyses were limited because of the focus on clinical assessment, IBTRs were classified as representing either a recurrence of the primary cancer (TR/MM) or a new primary breast cancer unrelated to the index lesion (E failure). In 36 patients (72% of all failures), the IBTR was considered an E failure; and, in 14 patients (28% of all failures), it was considered a TR/MM. The 3-year rates of DFS, CSS, and OS were 41.7%, 87.5%, and 85.7%, respectively, for TR/MM and 65.8%, 94.4%, and 78.1%, respectively, for E failures.


To our knowledge, this is the largest database to date of patients who developed an IBTR after APBI. The results indicate that patients who received APBI had outcomes comparable to the outcomes of patients who received up-front WBI after appropriate salvage treatment.9-11 Similar rates of local control, DFS, CSS, and OS were observed compared with the rates in published historic controls (see Historic Perspective, below). However, results from large phase 3 trials with extended follow-up will be needed to help differentiate the subtle differences that may exist between IBTRs that develop after APBI versus those that develop after WBI, which potentially may have an impact on treatment decisions. Furthermore, our analysis supports the use of repeat BCT, because nearly 75% of recurrences were categorized as new primaries and, thus, amenable to repeat breast-conserving surgery with repeat APBI. Although the number of events was small, patients who had E failures had improved DFS and CSS after IBTR compared with patients who had TR/MM. Also, when evaluating treatment options, it may be important to consider the interval between initial treatment and recurrence, because our data indicated that patients who had later failures had improved DFS at 3 years, although the analysis included a limited number of patients. This is similar to data presented by Haffty et al, who observed increased 5-year survival rates in patients who developed IBTRs >4 years after initial treatment.12

Historic Perspective

Approximately 10% of patients who receive BCT will develop a subsequent IBTR. Currently, the standard treatment remains salvage mastectomy, because repeat attempts at BCT potentially may result in an unacceptable cosmetic outcome. Furthermore, normal tissue toxicity concerns regarding reirradiation exist, potentially limiting the use of a second attempt at BCT in these patients.13 Reported outcomes after salvage mastectomy for IBTRs have indicated chest wall recurrence rates from 7% to 25%, highlighting the persistent risk of local recurrence even with this approach.6-8 Data from Yale University evaluating 116 patients who underwent salvage mastectomy and had 166 months of follow-up indicated a 6.9% rate of local recurrence and a 10-year survival rate of 65.7%.6 Furthermore, an analysis by Voogd et al of 208 patients who underwent salvage mastectomy indicated a chest wall recurrence rate of approximately 25% with nearly 5 years of follow-up.8 Although our current analysis did not examine rates of local recurrence after salvage treatment, our rates of DFS and CSS were comparable to the rates reported in those series.

More recently, repeat breast conservation has been used after IBTR both with excision alone and with various techniques of reirradiation. With regard to excision alone, outcomes have varied significantly with local recurrence rates of 7% to 50%.6-8 In a series from Yale University, the local recurrence rate was 7%; whereas, in 1 of the largest series from Japan, the recurrence rate was nearly 50%. These results highlight the variability in outcomes after excision alone.6, 14 In reviewing the studies described above, survival does not appear to be adversely impacted after a second attempt at BCT (despite the highly variable rates of local recurrence). It is important to realize that patients who undergo local excision alone after IBTR have rates of local recurrence comparable to the rates reported in patients who undergo breast-conserving surgery without the use of postoperative RT de novo. Currently, information about the outcomes of patients who develop an IBTR and receive reirradiation using APBI or alternative techniques are limited; however, reirradiation potentially may reduce the local recurrence rates produced by excision alone (Table 3).9-11, 15-18 Although our analyses were limited by a small number of patients (n = 9), the patients in our series who underwent repeat BCT had excellent outcomes at 2 years.

Table 3. Studies With Repeat Breast-Conservation Therapy Using Accelerated Partial Breast Irradiation
Institution (Reference)No. of PatientsTechniqueFollow-Up After Second BCT, moOutcomes
  1. Abbreviations: 3D-CRT, 3-dimensional conformal radiation therapy; APBI, accelerated partial breast irradiation; BCT, breast-conserving therapy; EBRT, external beam radiation therapy; Gy, grays; HDR, high dose rate; IB, interstitial brachytherapy; IBTR, ipsilateral breast tumor recurrence; LC, local control; LDR, low dose rate; PBI, partial breast irradiation; WBI, whole breast irradiation.

University of Pittsburgh (Deutsch 20029)39EBRT WBI, 50 Gy + 10 Gy51.5LC rate, 80% at 5 y
Beth Israel Medical Center (Chada 200910)15LDR, 30-45 Gy36LC rate, 89% at 3 y
University of Nice (Hannoun-Levi 201015)42HDR IB21LC rate, 97% at 2 y
Drexel University (Trombetta 201111)3621 Patients, LDR IB PBI; 11 patients, MammoSite; 4 patients, 3D-CRT PBI3735 of 36 Patients disease free at 3 y
Barcelona Medical Institute for Radiotherapy and Oncology (Guix 201016)36HDR IB, 30 Gy × 10120LC rate, 89.4% at 10 y
University of Wisconsin (Adkison 201017)11HDR, 3.4 Gy × 1053.7100% disease free at 4 y
University of Paris (Maulard 199518)15LDR, 30 Gy4079% LC at 4 y

One of the largest series reported to date on the use of repeat BCT comes from a retrospective study that was performed at the University of Pittsburgh. In that study, 39 patients who developed IBTRs after WBI received repeat full-dose, external-beam irradiation (50 Gray [Gy] in 25 fractions). With more than 4 years of follow-up, a second IBTR developed in 20.5% of patients. The 5-year OS and DFS rates were 77.9% and 68.5%, respectively. Just as critical, no radiation-induced necrosis developed, and cosmesis was reported as good/excellent in 69% of patients based on physician assessment.9

Chadha et al evaluated the role of repeat BCT with local surgery and brachytherapy for patients who presented with a local recurrence/new primary after standard WBI. Fifteen patients in that analysis received low-dose-rate brachytherapy after undergoing lumpectomy. At a median follow-of 36 months (in the group that received 30 Gy), 1 patient had a documented in-breast failure for an overall in-breast control rate of 89%. In terms of toxicities, no episodes of grade 3 or 4 fibrosis, telangiectasia, ulcers, or necrosis were reported.10

There are limitations to our current analysis. Although data were collected prospectively through the ASBS registry, this was an unplanned retrospective analysis. Furthermore, because of the small number of failures experienced, the power to detect significant differences is limited. Also, no data were available on local control after salvage therapy, and information about outcome was unavailable for 12 of the 50 patients who developed an IBTR. Although the patients were separated into a TR/MM group and an E failure group, a limitation of this analysis was the lack of truly objective criteria, because the division of IBTRs was based strongly on clinical assessment. In the future, objective standards, including loss-of-heterozygosity testing, may be used to better delineate new primaries. Finally, limited numbers of patients underwent repeat BCT; therefore, the outcomes presented require further data to validate our findings. However, despite the small numbers, to our knowledge, the current study represents the largest series in the literature to date.

Finally, only data from ongoing phase 3 trials comparing APBI with WBI can objectively differentiate potential small differences in outcome after IBTR that may exist based on the initial RT technique used, as noted previously. Nonetheless, the results from this study, along with the previous reports discussed above, suggest that clinical outcomes after IBTR with APBI are comparable to the outcomes achieved with WBI at 5 years of follow-up. The Radiation Therapy Oncology Group (RTOG) recently started enrolling patients on protocol RTOG 1014 to prospectively examine the feasibility and acute toxicity of repeat BCT after IBTR using an external-beam, partial breast irradiation technique. In that trial, RT is delivered in 150-centigray fractions twice daily up to a total dose of 45 Gy. It is hoped that, with the added benefits of decreased treatment time and reduced normal tissue toxicity, the use of repeat BCT will continue to increase in well selected patients with early stage breast cancer.

In conclusion, with 5 years of follow-up in a large group of selected patients, APBI produced clinical outcomes and patterns of failure comparable to those reported with WBI. Just as critical, those patients who developed an IBTR after APBI had excellent 3-year survival outcomes after salvage treatment, which, again, was similar to the outcomes of those who developed an IBTR after WBI.


The MammoSite Breast Brachytherapy Registry Trial and this analysis was supported in part by a grant from Hologic, Inc. to the American Society for Breast Surgeons and BioStat International, Inc.


Martin Keisch, M.D.: Consultant/Advisory Board, Hologic, Inc., Bedford, Massachusetts.