758

Three-Year Analysis of Treatment Efficacy, Cosmesis,

and Toxicity by the American Society of Breast
Surgeons MammoSite Breast Brachytherapy Registry
Trial in Patients Treated With Accelerated Partial
Breast Irradiation (APBI)

Frank Vicini, MD1 BACKGROUND. This report presents 3 years of data on treatment efficacy,
Peter D. Beitsch, MD, FACS2 cosmetic results, and toxicities for patients enrolled on the American Society
Coral A. Quiet, MD3 of Breast Surgeons MammoSite (Cytyc, Bedford, Mass) Breast Brachytherapy
Angela J. Keleher4 Registry Trial.
Delia Garcia, MD5 METHODS. A total of 1440 patients (1449 cases) with early stage breast cancer
Howard C. Snider Jr 6 who were undergoing breast-conserving therapy were treated with the Mammo-
Mark A. Gittleman, MD7 Site device to deliver accelerated partial breast irradiation (APBI) (34 Gy in 3.4 Gy
Victor J. Zannis, MD8 fractions). Of these, 1255 (87%) cases had invasive breast cancer (IBC; median
Henry M. Kuerer, MD, PhD9 size 5 10 mm), and 194 (13%) cases had ductal carcinoma in situ (DCIS; median
Maureen Lyden10 size 5 8 mm). Median follow-up was 30.1 months.
RESULTS. Twenty-three (1.6%) cases developed an ipsilateral breast tumor recur-
1
Department of Radiation Oncology, William rence (IBTR) for a 2-year actuarial rate of 1.04% (1.11% for IBC and 0.59% for
Beaumont Hospital, Royal Oak, Michigan. DCIS). No variables were associated with IBTR. Six (0.4%) patients developed an
2
Department of Surgery, Dallas Breast Center, axillary failure. The percentages of breasts with good to excellent cosmetic results
Dallas, Texas. at 12 (n 5 980), 24 (n 5 752), 36 (n 5 403), and 48 months (n 5 67 cases) were
3 95%, 94%, 93%, and 93%, respectively. Breast seromas were reported in 23.9% of
Foundation for Cancer Research and Education,
Arizona Oncology Services, Scottsdale, Arizona. cases (30% in open-cavity implants and 19% in closed-cavity implants). Sympto-
matic seromas occurred in 10.6% of cases, and 1.5% of cases developed fat ne-
4
Department of Surgery, Western Pennsylvania
crosis. A subset analysis of the first 400 consecutive cases enrolled was
Hospital, Pittsburgh, Pennsylvania.
performed (352 with IBC, 48 DCIS). With a median follow-up of 37.5 months, the
5
Department of Radiation Oncology, St. Louis 3-year actuarial rate of IBTR was 1.79%.
Cancer & Breast Center, St. Louis, Missouri.
CONCLUSIONS. Treatment efficacy, cosmesis, and toxicity 3 years after treatment
6
Department of Surgery, Alabama Breast Center, with APBI using the MammoSite device are good and similar to those reported
Montgomery, Alabama. with other forms of APBI with similar follow-up. Cancer 2008;112:758–66.
7
Department of Surgery, Sacred Heart Hospital,
2008 American Cancer Society.
Allentown, Pennsylvania.
8
Department of Surgery, Breast Care Center of KEYWORDS: breast-conserving therapy, brachytherapy, radiation, partial breast
the Southwest, Phoenix, Arizona. irradiation, MammoSite.
9
Department of Surgery, M. D. Anderson Cancer
Victor J. Zannis has received honoraria from Address for reprints: Frank A. Vicini, MD, Depart-
Center, Houston, Texas.
Cytyc. ment of Radiation Oncology, Beaumont Cancer
10
BioStat International Inc, Tampa, Florida. Institute, William Beaumont Hospital, 3601 W. 13
Maureen Lyden receives support from and is the Mile Road, Royal Oak, MI 48072; Fax: (248) 551-
lead researcher for MammoSite, manufactured 0089; E-mail: fvicini@beaumont.edu
by Cytyc.
The MammoSite Breast Brachytherapy Registry Received July 31, 2007; revision received August
Trial was supported in part by a grant from Cytyc Presented at the American Society for Therapeutic 31, 2007; accepted September 6, 2007.
Corporation to the American Society of Breast Radiology and Oncology 49th Annual Meeting, Octo-
Surgeons and BioStat International Inc. ber 28-November 1, 2007, Los Angeles, California.

ª 2008 American Cancer Society

DOI 10.1002/cncr.23227
Published online 7 January 2008 in Wiley InterScience (www.interscience.wiley.com).

A ccelerated partial breast irradiation (APBI) is cur-
rently being explored as an alternative option to
deliver adjuvant radiation therapy (RT) after lum-
pectomy in selected patients with early stage breast
cancer treated with breast conserving therapy (BCT).
Although there are several different types of RT that
can be used to deliver APBI, techniques using
brachytherapy have been the most frequently used
modality.1 The reported 5-year and 10-year rates of
local tumor control have been excellent, but intersti-
tial brachytherapy can be a difficult technique to
teach and learn.2 In addition, because interstitial
brachytherapy requires the use of multiple catheters,
widespread patient acceptance has been limited.3 In
recognition of these problems, a logistically simpler,
technically more reproducible, and patient ‘‘friendly’’
device (the MammoSite breast brachytherapy cathe-
ter manufactured by Cytyc, Bedford, Mass) was
developed to deliver APBI.4 After clearance of the de-
vice by the US Food and Drug Administration (FDA)
for clinical use in May 2002, a registry trial was
initiated concurrently by the manufacturer. The goals
and objectives of the trial were to provide a method
to prospectively, objectively, and systematically col-
lect data on the clinical use of the brachytherapy ap-
plicator.5,6 In November of 2003, the American
Society of Breast Surgeons assumed primary man-
agement of the trial. This article presents data on
treatment efficacy, cosmesis, and toxicity at the
3-year follow-up for patients treated on the trial.
MATERIALS AND METHODS
Between May 4, 2002 and July 30, 2004, 97 institu-
tions participated in a Registry Trial designed to col-
lect data on the clinical use of the MammoSite
device as a modality for the delivery of APBI or as
boost irradiation. The trial was initiated concurrently
by the manufacturer with the clearance of the
brachytherapy device by the FDA 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
brachytherapy applicator. Data on the technical
reproducibility of the device on a large scale, acute
toxicity, cosmesis, efficacy, and adherence to place-
ment in appropriate patients were sought. A previous
trial in 43 patients used to obtain FDA clearance and
to establish the safety of the device as a breast
brachytherapy catheter has been previously re-
ported.4 On November 17, 2003, the American Soci-
ety of Breast Surgeons (ASBS) assumed primary
management of the trial. A total of 951 patients had
been enrolled on the trial when the ASBS assumed
Three-Year Analysis of APBI/Vicini et al. 759
its management. Since assumption of the trial, an
additional 714 patients have been enrolled by the
ASBS.
Since assuming management of the Registry trial,
the ASBS has attempted to collect additional infor-
mation than originally intended at the start of the
trial. Recognizing the importance of these data in
helping to establish the efficacy of APBI, the ASBS
has continuously expanded the type, quality, and na-
ture of the data collected.
Synergos Inc (The Woodlands, Tex), an inde-
pendent full-service Contract Research Organization
(CRO) not affiliated with the ASBS, the manufacturer,
or any of the institutions participating in this trial,
was initially hired to provide services to collect,
manage, and analyze data for the ASBS. This CRO
provides regulatory, clinical trial management, moni-
toring, data management, statistical analysis, and
report-writing services for numerous pharmaceutical
and medical device companies. After assuming man-
agement of the Registry Trial, the ASBS asked Syner-
gos to initiate an additional follow-up protocol to
verify and collect more complete information about
cosmetic results, adjuvant therapy, adverse events,
radiation recall reactions, disease recurrence, and
patient survival. The additional data collection began
in July 2004 for the 1051 patients with at least 1 fol-
low-up visit as of that time and has been completed
for 881 (84%) of these patients.
In June 2006, a new independent full-service
CRO (BioStat International Inc [BSI] Tampa, Fla)
assumed responsibility of the independent manage-
ment and statistical analysis of the Registry Trial data
for the ASBS. Since that time, extensive review and
cleaning of database records have been performed
with emphasis on obtaining clean and correct cosm-
esis, recurrence, survival, and toxicity information.
All paper records were verified to be entered in the
database, site verification of recurrence information
was obtained, and a cleaning of adverse event
records for terminology and missing descriptive in-
formation such as grading and timing of onset
was completed. The consistent and ongoing practice
of timely data cleaning and site communication
has reduced incomplete and missing information
and provided quality data records for summary and
analysis.
Registry Trial Design/Patient Enrollment
All centers that were trained and were using the
MammoSite device clinically were offered and
encouraged to participate in the Registry Trial.
The Registry Trial Program was incorporated into
the regional training programs for the device. The
760 CANCER February 15, 2008 / Volume 112 / Number 4
recruitment goals for the program were to incorpo-
rate as many institutions as feasible to provide for a
large database of patients in various clinical settings
(ie, academic, private practice, and hospital). No site
with adequate resources to complete the required
data forms was denied participation in the study. It
should be noted that patients could be enrolled in
the trial at anytime (before, during, and after treat-
ment) although enrollment before treatment was
strongly encouraged.
Because data entry and processing were a con-
tinuous process for this program, a data cutoff date
was chosen for the current article to allow for audit-
ing and analysis (May 15, 2007). A total of 1449
breasts treated with APBI (1440 treated subjects)
were submitted for analysis from a total of 97 partici-
pating institutions. These 1449 patients treated with
APBI are the subjects of this analysis; no boost
patients are included.
Patient/Technical Eligibility Criteria
Recommended criteria for patient enrollment in the
protocol using the device were based upon previous
publications on the use of APBI by the American
Brachytherapy Society (ABS).7,8 Recommended tech-
nical guidelines were established in the protocol to
exclude the treatment of patients with inadequate
balloon-to-skin distances, excessive cavity size, or
poor balloon-cavity conformance.
Data Collection/Quality Assurance
Information on patient demographics, technical
reproducibility, cosmesis, toxicity, and overall efficacy
were collected on the Registry Trial data forms sup-
plied to investigators. After the American Society of
Breast Surgeons assumed management of the trial,
additional data were collected. This consisted of
additional follow-up to verify and collect more com-
plete information about cosmetic results, adjuvant
therapy, adverse events, radiation recall reactions,
disease recurrence, and patient survival.
All data forms were collected and reviewed for
inaccuracies, omissions, and conflicting information
by Synergos. Upon receipt of the data forms, a medical
review was performed to ensure that no pertinent data
were omitted or contained conflicting information.
Patients were not included in the Registry Trial until
all of the information that was deemed critical was
provided. In addition, an audit (random) of 10% of
these patients was performed (at the request of the
American Society of Breast Surgeons) to verify the
accuracy of the data that was collected. During the
course of the audit, source documents (when avail-
able) were used to verify as much of the data as possi-
ble. (Note: The accuracy of data entered into the
registry database was verified through the use of actual
source documents in >10% of cases).
In addition, the new CRO (BSI) has performed
checking of the database for completeness against
paper records and systematically reviewed 100% of
the critical database elements. BSI identified and cor-
rected information not entered or entered inconsis-
tently and reviewed all adverse events for duplications
and incomplete descriptive information. Changes were
made in adverse-event recording practices, such as the
discontinuation of over-reporting of ongoing events, to
minimize errors and to obtain consistent reporting
across sites. Before data summary, definitions of recur-
rence and toxicity categories, and follow-up visit
windows, were provided by the American Society of
Breast Surgeons to BSI, thus enabling BSI to perform
consistent reporting. As the data continues to be col-
lected, review rules remain consistent and querying of
sites for additional or missing information is timely so
the database can be current. The ongoing emphasis on
completeness and consistency in the management of
the data at BSI and in-depth discussions between sta-
tisticians at BSI and the American Society of Breast
Surgeons has greatly improved the quality and usability
of the Registry Trial database.
Institutional review board (IRB) approval was not
required for participation in the Registry Trial but
was recommended by the sponsor. Eighty percent of
the clinical sites were affiliated with an IRB and
obtained IRB approval to participate in the study. All
patients enrolled in the study were required to sign
an informed consent, and patients who were treated
on or after April 14, 2003 were required to sign a
Notice of Privacy Practices in accordance with the
Health Insurance Portability and Accountability Act
of 1996 (HIPAA agreement) allowing the release of
their data. All clinical sites were provided with a
sample informed consent and HIPAA agreement that
contained all of the required elements necessary for
informed consent. Patient data submitted without an
informed consent and/or HIPAA, if applicable, are
not part of this analysis.
Enrollment was closed in July of 2004. However,
as required by the protocol, all treated patients are
required to return for an annual follow-up visit for 7
years. The American Society of Breast Surgeons and
BSI continue to query sites for follow-up information
on all treated patients. Patients are followed by either
their radiation oncologist or surgeon and data col-
lected include cosmetic evaluation (Harvard Scale),
use of adjuvant therapy, imaging assessment, recur-
rence and treatment of recurrence, survival status,
radiation recall, and toxicities.

Local, Regional, and Distant Disease Recurrence
For the purposes of this analysis, an ipsilateral breast
tumor recurrence (IBTR) was defined as the reappear-
ance of cancer in the treated breast (before the devel-
opment of distant metastases) and was confirmed
pathologically. Investigators were asked to classify
IBTRs by their clinical location in relation to the lum-
pectomy cavity according to the criteria described by
Recht et al.9 A true recurrence/marginal miss (TR/
MM) was defined as a recurrence of the treated can-
cer within or immediately adjacent to the primary tu-
mor site. An elsewhere failure (E failure) was defined
as an IBTR several centimeters from the primary site
and was generally felt to be a new primary cancer.
Investigators were also asked to classify regional fail-
ures as an axillary, supraclavicular, or internal mam-
mary lymph-node recurrence. For the purposes of
this analysis, overall survival reflected all deaths, can-
cer-related or otherwise. Cause-specific survival was
based upon deaths attributed to breast cancer.
Cosmesis/Toxicities
Investigators (radiation oncologist or surgeon) were
asked to evaluate cosmesis at each follow-up visit
using the Harvard criteria (provided in the Registry
Trial data forms).10 An excellent cosmetic result score
was assigned when the treated breast looked essen-
tially the same as the contralateral breast (as it
relates to radiation effects). A good cosmetic score
was assigned for minimal but identifiable radiation
effects of the treated breast. A fair score meant sig-
nificant radiation effects were readily observable. A
poor score was used for severe sequelae of breast tis-
sue secondary to radiation effects.
Investigators were also asked to report the pre-
sence or absence of any seromas and fat necrosis at
all time points after treatment (Note: No specific cri-
teria were given to sites to define these toxicities.) In
addition, whether or not these toxicities produced
symptoms or required some type of intervention was
also reported.
Statistical Methods
All time intervals were calculated from the date of
MammoSite removal. For the evaluations and sum-
marizations of disease characteristics, which were
MammoSite treatment parameters, cosmesis, and
local recurrence, 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 the survival or
recurrence-free distributions or recurrence (failure)
distribution were obtained by life table methods.
Associations between clinical, pathologic, and treat-
Three-Year Analysis of APBI/Vicini et al. 761
ment-related variables and recurrence or survival
event rates were analyzed by fitting parametric mod-
els to the failure-time data and examining the signifi-
cance of the parameter estimates. Associations
between dichotomous cosmetic outcomes (Excellent/
Good vs Fair/Poor) and treatment-related variables
were analyzed by Fisher exact tests for categorical
measurements and Wilcoxon rank sum tests for con-
tinuous measurements. All tests were declared statis-
tically significant if the calculated P-value was
.05.
All tests appear as 2-sided P-values. Descriptive sta-
tistics consist of numbers and percentages of
responses in each category for discrete measures and
of medians, minimum, and maximum values for
continuous measures. Version 8.0 or higher of the
SAS statistical software package (SAS, Cary, NC) was
used to provide all statistical analyses.
RESULTS
Study Populations
Table 1 presents selected clinical, pathologic, and
treatment-related characteristics of the study popula-
tion. In 1249 patients, 1255 (87%) cases had invasive
breast cancer (IBC; 93% T1, 3% N1, median size 5
10 mm), and 194 (13%) cases had DCIS (median size
8 mm). For follow-up, 1236 (85%) breasts have been
followed 12 months, 964 (67%) 24 months, 490
(34%) 36 months, and 79 (6%) 48 months. Median
follow-up for surviving patients was 30.1 months
(range, 0 to 58.6 months).
A subset analysis of the first 400 consecutive
cases enrolled on the study was performed. A total of
352 (88%) of these cases in 351 patients had IBC
(median tumor size 5 10 mm), and 48 (12%) cases
had DCIS (median size 5 9 mm). Median follow-up
for surviving patients was 37.5 months (range, 0.2 to
58.6 months) in this subset of patients. The fre-
quency of all clinical, pathologic, and treatment-
related characteristics of these 400 cases was similar
to the overall population.
Failure Patterns
Ipsilateral breast tumor recurrences (IBTRs); all patients
(N 5 1449)
A total of 23 (1.6 %) cases developed an IBTR as
some component of their initial failure (before dis-
tant metastases) for a 2-year actuarial rate of 1.04%
(1.11% for IBC and 0.59% for DCIS; Table 2). The me-
dian time to IBTR was 23.5 months (range, 4.9–36.8
months). In 17 (74%) breasts, the IBTR was felt to be
a new primary breast cancer unrelated to the index
lesion (E failure) by the investigator and in 6 (26%)
breasts, it was thought to represent a recurrence of
762 CANCER February 15, 2008 / Volume 112 / Number 4

TABLE 1 tal component (EIC), age at diagnosis (<45 vs 45

Patient, Tumor and Treatment-related Characteristics (All Cases) years), bra size, method of placement (open vs closed),
tamoxifen use, or chemotherapy use. The only variable
All cases Invasive DCIS
Characteristic No. (%) No. (%) No. (%) associated with the development of an IBTR included
an extensive intraductal carcinoma (EIC) of marginal
Patients 1440* 1249 194 significance (P 5 .0844) in IBC cases.
Breasts 1449 1255 194
Age
Subset analysis (first 400 cases enrolled)
Median [range], y 65.5 [31.8–93.5] 65.9 [31.8–93.5] 62.1 [40.7–88.0]
>60 929 (64.5) 815 (65.3) 117 (60.3) A total of 6 (1.5%) cases developed an IBTR as some
50–60 380 (26.4) 324 (25.9) 56 (28.9) component of their initial failure (before distant me-
40–50 126 (8.8) 105 (8.4) 21 (10.8) tastases) for a 3-year actuarial rate of 1.79% (2.04%
<40 5 (0.3) 5 (0.4) 0 (0.0) for IBC, 0.0% for DCIS; Table 3). The median time to
Tumor size, mm
IBTR was 23.6 months (range, 12.7–36.8 months). In
Median [range] 10.0 [1.0–45.0] 10.0 [1.0–42.0] 8.0 [1.0–45.0]
<5 130 (9.0) 89 (7.1) 41 (21.1) 5 (83%) breasts, the IBTR was felt to be a new pri-
5 to <10 482 (33.3) 431 (34.3) 51 (26.3) mary breast cancer unrelated to the index lesion (E
10 to
20 705 (48.7) 652 (52.0) 53 (27.3) failure), and in 1 (17%) breast, it was thought to
>20 91 (6.3) 79 (6.3) 12 (6.2) represent a recurrence of the primary cancer (TR/
Unknown 41 (2.8) 4 (0.3) 37 (19.1)
MM failure). The 3-year actuarial rate of TR/MM fail-
T-classification
T1A 160 (11.0) 160 (12.7) N/A ure was 0.34%, and the 3-year actuarial rate of E fail-
T1B 505 (34.9) 505 (40.2) N/A ure was 1.45%.
T1C 506 (34.9) 506 (40.3) N/A Several variables were analyzed for an associa-
T2 84 (5.8) 84 (6.7) N/A tion with the development of an IBTR in all patients
Nodal status
and in those treated for IBC or DCIS. The only vari-
N0 1205 (83.2) 1152 (91.8) 53 (27.3)
NX 206 (14.2) 65 (5.2) 141 (72.7) able associated with the development of an IBTR
N(1) 38 (2.6) 38 (3.0) N/A included an extensive intraductal component (mar-
Margins ginal significance of P 5 .0549 in IBC cases).
Positive 13 (0.9) 11 (0.9) 2 (1.0)
Negative 1326 (91.5) 1155 (92.0) 171 (88.1)
Cosmetic Results
Closey 110 (7.6) 89 (7.1) 21 (10.8)
Systemic therapy The percentage of breasts with good and/or excellent
Chemotherapy alone 87 (6.0) 86 (6.9) 1 (0.5) cosmetic results at 12, 24, 36, and 48 months were
Hormonal therapy alone 752 (52.2) 654 (52.4) 98 (50.5) as follows: 95% (n 5 980), 94% (n 5 752), 93%
Both chemotherapy 80 (5.6) 80 (6.4) N/A (n 5 403), and 93% (n 5 67), respectively (Table 4).
and hormonal therapy
Follow-up,{ mo
Median 30.1 30.3 28.6 Variables associated with optimal cosmetic results
Mean 28.2 28.3 27.4 Multiple variables were analyzed for their association
[Range] [0.0–58.6] [0.0–58.6] [0.2–54.0] with the development of a good and/or excellent
1 y (% of total cases) 1184 (81.7) 1021 (81.4) 163 (84.0) cosmetic result at 2 years and at 3 years. At 24
2 y (% of total cases) 931 (64.3) 805 (64.1) 126 (64.9)
months (n 5 708 evaluable breasts), factors asso-
3 y (% of total cases) 476 (32.9) 417 (33.2) 59 (30.4)
4 y (% of total cases) 79 (5.5) 72 (5.7) 7 (3.6) ciated with good and/or excellent cosmetic results
included increasing balloon-to-skin distance (as a
DCIS indicates ductal carcinoma in situ; N/A, not applicable. continuous variable, P 5 .0007), increasing skin spa-
* Nine patients were treated for bilateral disease. cing as a categorical variable (95.4% vs 83.5%;
y
Close
2 mm.
{
P 5 .0003), no infection (94.8% vs 77.8%; P 5 .0033),
Surviving patients.
and no systemic chemotherapy treatment (94.9% vs
88.4%; P 5 .0249). At 3 years, only breast infection
occurrence had a statistically significant association
the primary cancer (TR/MM failure). The 2-year with cosmetic outcome (P 5 .0063).
actuarial rate of TR/MM failure was 0.33 % and, the
2-year actuarial rate of E failure was 0.71%. Toxicity
Several variables were analyzed for an association Table 5 displays the rates of seroma reporting overall
with the development of an IBTR in all patients and in and at 12, 24, and 36 months after treatment. Most
those treated for IBC or DCIS. These variables seromas were reported in the first year after Mam-
included margin status (positive vs negative), nodal moSite removal. In addition, seromas were reported
status, tumor size and location, an extensive intraduc- more frequently in open versus closed implants
 Three-Year Analysis of APBI/Vicini et al. 763

TABLE 2
Patterns of Failure (All Cases)

All cases, N 5 1449 Invasive cases, n 5 1255 DCIS cases, n 5 194

2-Year 2-Year 2-Year

Event No. (%) actuarial rate No. (%) actuarial rate No. (%) actuarial rate

Breast only failure (IBTR) 21 (1.4) 0.86% 19 (1.5) 0.99% 2 (1.0) 0.00%
Breast & regional failure 2 (0.1) 0.18% 1 (0.1) 0.12% 1 (0.5) 0.59%
All breast failures 23 (1.6) 1.04% 20 (1.6) 1.11% 3 (1.5) 0.59%
TR/MM failure 6 (0.4) 0.33% 4 (0.3) 0.29% 2 (1.0) 0.59%
E failure 17 (1.2) 0.71% 16 (1.3) 0.82% 1 (0.5) 0.00%
Regional failure only* 5 (0.3) 0.28% 5 (0.4) 0.32% 0 (0.0) 0.00%
Axillary only failure 4 (0.3) 0.28% 4 (0.3) 0.32% 0 (0.0) 0.00%
All axillary failures 6 (0.4) 0.46% 5 (0.4) 0.44% 1 (0.5) 0.59%
Distant failure 13 (0.9) 0.42% 12 (1.0) 0.39% 1 (0.5) 0.60%
Disease-free survival — 95.80% — 95.40% — 98.10%
Overall survival — 97.30% — 97.10% — 98.70%
Cause-specific survival — 99.50% — 99.50% — 99.40%
Contralateral breast cancer 9 (0.6) 0.24% 7 (0.6) 0.18% 2 (1.0) 0.60%

Contralateral breast failure, distant failure, disease-free, overall, and cause specific survival are based upon patients treated.
DCIS indicates ductal carcinoma in situ; IBTR, ipsilateral breast tumor recurrence; TR/MM, true recurrence/marginal miss; E, elsewhere failure.
* Axilla and supraclavicular fossa.

TABLE 3
Patterns of Failure (First 400 Consecutive Cases Enrolled)

All cases, N 5 400 Invasive cases, n 5 352 DCIS cases, n 5 48

3-Year 3-Year 3-Year

Event No. (%) actuarial rate No. (%) actuarial rate No. (%) actuarial rate

Breast only failure (IBTR) 5 (1.3) 1.45% 5 (1.4) 1.65% 0 (0.0) 0.00%
Breast & regional failure 1 (0.3) 0.34% 1 (0.3) 0.39% 0 (0.0) 0.00%
All breast failures 6 (1.5) 1.79% 6 (1.7) 2.04% 0 (0.0) 0.00%
TR/MM failure 1 (0.3) 0.34% 1 (0.3) 0.39% 0 (0.0) 0.00%
E failure 5 (1.3) 1.45% 5 (1.4) 1.65% 0 (0.0) 0.00%
Regional failure only* 1 (0.3) 0.27% 1 (0.3) 0.31% 0 (0.0) 0.00%
Axillary only failures 1 (0.3) 0.27% 1 (0.3) 0.31% 0 (0.0) 0.00%
All axillary failures 2 (0.5) 0.61% 2 (0.6) 0.70% 0 (0.0) 0.00%
Distant failure 6 (1.5) 1.30% 6 (1.7) 1.48% 0 (0.0) 0.00%
Disease-free survival — 93.40% — 92.50% — 100.00%
Overall survival — 95.80% — 95.20% — 100.00%
Cause-specific survival — 100.00% — 100.00% — 100.00%
Contralateral breast cancer 3 (0.8) 0.00% 2 (0.6) 0.00% 1 (2.1) 0.00%

Contralateral breast failure, distant failure, disease-free, overall, and cause specific survival are based upon patients treated.
DCIS indicates ductal carcinoma in situ; IBTR, ipsilateral breast tumor recurrence; TR/MM, true recurrence/marginal miss; E, elsewhere failure.
* Axilla/supraclavicular fossa.

(29.6% vs 19.2%) and with the use of larger (5–6 cm lected in a registry trial managed by the American
vs 4–5 cm) brachytherapy balloons (30.3% vs 23.2%; Society of Breast Surgeons to determine 2-year and
data not presented). 3-year cosmetic results and toxicities associated with
its use and short-term efficacy. With a median fol-
low-up of 37.5 months in the first 400 enrolled cases,
DISCUSSION 6 (1.5%) breasts developed an IBTR for a 3-year
Data on the clinical use of the MammoSite breast actuarial rate of 1.79% (2.04% for IBC, 0.0% for
brachytherapy applicator to deliver APBI were col- DCIS). Two patients developed an axillary failure
764 CANCER February 15, 2008 / Volume 112 / Number 4

TABLE 4 Ipsilateral Breast Tumor Recurrence

Cosmetic Results Versus Length of Follow-up One of the potential advantages of APBI is the
reduced time required for the delivery of adjuvant
All cases* First 400 cases*
No. of RT. In order for this treatment approach to be con-
follow-up Excellent/good Excellent/good sidered an acceptable alternative to whole-breast RT,
months No. cosmesis No. cosmesis treatment efficacy (local tumor control) must be
shown to be equivalent. Although 3-year results are
12 980 927 (94.6) 270 254 (94.1)
insufficient to establish long-term efficacy, they do
24 752 708 (94.1) 214 198 (92.5)
36 403 375 (93.1) 178 159 (89.3) provide some indication of the adequacy of this
48 67 62 (92.5) 67 62 (92.5) treatment approach to control the index lesion in
comparison to other forms of APBI with similar
* Treated breasts. follow-up. The reported 3-year actuarial rate of local
failure (1.79%) is comparable to failure rates ob-
served with interstitial brachytherapy to deliver APBI
TABLE 5 in the RTOG 95–17 trial and in multiple single-insti-
Seroma Formation tution experiences from several other centers. These
Time of onset
trials demonstrate 3-year to 5-year rates of IBTR ran-
ging from 0% to 6%. Certainly, longer follow-up will
At any 0–12 >12–18 >18–24 >24 be needed to confirm the stability of these early con-
time months months months months trol rates. Furthermore, additional follow-up will be
No. (%) No. (%) No. (%) No. (%) No. (%) useful in helping to establish clinical, pathologic, and
Cases at risk*
technical factors associated with optimal control
All cases 1449 1449 1217 1088 930 rates and patients most suitable for the use of APBI
Open cavity 653 653 549 490 424 in general, and the MammoSite in particular. None-
Closed cavity 796 796 668 598 506 theless, these early findings are encouraging, and
All seromas patients will continue to be closely followed to fur-
All cases 346 (23.9) 269 (18.6) 33 (2.7) 19 (1.7) 25 (2.7)
Open cavity 193 (29.6) 150 (23.0) 17 (3.1) 10 (2.0) 16 (3.8)
ther document treatment efficacy.
Closed cavity 153 (19.2) 119 (14.9) 16 (2.4) 9 (1.5) 9 (1.8) Regional Failure
Symptomatic seromas
All cases 153 (10.6) 133 (9.2) 13 (1.1) 3 (0.3) 4 (0.4) One of the additional concerns expressed with the
Open cavity 98 (15.0) 84 (12.9) 9 (1.6) 3 (0.6) 2 (0.5) use of APBI is the possibility of higher rates of re-
Closed cavity 55 (6.9) 49 (6.2) 4 (0.6) 0 (0.0) 2 (0.4) gional failure (primarily axillary failure). This concern
Seromas drained is based upon the observation that the lower axilla is
All cases 136 (9.4) 119 (8.2) 11 (0.9) 3 (0.3) 3 (0.3)
not generally included in the radiotherapy target vol-
Open cavity 87 (13.3) 75 (11.5) 8 (1.5) 3 (0.6) 1 (0.2)
Closed cavity 49 (6.2) 44 (5.5) 3 (0.4) 0 (0.0) 2 (0.4) ume with APBI as it many times can be with whole
breast RT.11 However, no recent APBI study has yet
* Treated breasts. to report a significantly higher rate of regional nodal
recurrence, and the low axillary failure rate observed
in this trial (3-year actuarial rate of 0.61%) does not
suggest this should become a significant problem
(AF) for a 3-year actuarial rate of 0.61% (0.3% for a
with longer follow-up. Only additional follow-up
3-year actuarial rate of isolated AF). The percentages
(and mature phase 3 trial data; see below) will con-
of patients with good and/or excellent cosmetic
clusively establish the impact of APBI or whole-
results at 24 (n 5 214), 36, (n 5 178), and 48 months
breast RT on the rate of AF.
(n 5 67 cases) were 93%, 89%, and 93%, respectively.
Although no variable was associated with the de-
Finally, breast seromas were reported in 23.9% of
velopment of an AF (data not presented), the num-
patients overall (29.6% in open-cavity and 19.2% in
ber of events was quite small. Further follow-up and
closed-cavity implants). A total of 10.6% of cases
additional failures will be needed to determine
developed symptomatic seromas. Fat necrosis was
whether any clinical, pathologic, or treatment-related
observed in 1.5% of all patients. These results
factor should be avoided for a patient not to be con-
demonstrate that treatment efficacy, cosmesis, and
sidered an optimal candidate for APBI.
toxicities 2-years and 3-years after treatment with
APBI using the MammoSite device in this registry Cosmetic Results
trial are similar to those reported with other forms of The results of this trial confirm prior observations by
APBI with similar follow-up. Keisch et al. that early cosmesis is strongly related to

balloon-to-skin spacing.4 At 24 months (n 5 752
breasts), increasing skin spacing was associated with
good and/or excellent cosmetic results both as a
numeric (P 5 .0007) and as a categorical variable
(83.5% vs 95.4%, P 5 .0003). As observed in numer-
ous studies of standard whole-breast irradiation, no
infection (94.8% vs 77.8%; P 5 .0033) and no sys-
temic chemotherapy treatment (94.9% vs 88.4%;
P 5 .0249) were also associated with good and/or
excellent cosmetic results with the use of the device.
Certainly, additional follow-up will be needed to con-
firm these observations (further stability of cosmesis
over time) because cosmetic results after standard
BCT generally stabilize after 3 years. However, data
from Benitez et al. and Chen et al. show that cos-
metic results actually improved over time (after 2
years) in patients treated with similar fractionation
schedules using interstitial brachytherapy to deliver
APBI.12,13 This occurred despite the finding that
other endpoints, such as telangiectasia and fat necro-
sis, increased with longer follow-up. As these data
continue to mature, further analyses by the American
Society of Breast Surgeons will be conducted to help
establish the stability of these observations and the
factors associated with optimal long-term cosmetic
results.
Toxicities
The monitoring of specific toxicities (seromas and fat
necrosis) potentially related to the use of this form of
APBI is important to assess its overall efficacy. Breast
seromas were reported in 23.9% of patients overall
(10.6% of cases had symptomatic seromas). Although
this rate appears to be high, it should be noted that
this is very similar to the overall rates reported after
standard BCT (range, 10% to 30%).14 In addition, the
rate of symptomatic seroma formation is quite low
and, again, similar to rates observed with standard
BCT. Unfortunately, specific criteria that have been
used to consistently define the presence and/or se-
verity of seromas after BCT are lacking, making it dif-
ficult to objectively compare rates between various
types of treatment. Fat necrosis was observed in
1.5% of all patients. These toxicities will continue to
be monitored carefully to determine any changes in
their incidence over time and factors associated with
their development.
Treatment of Ductal Carcinoma In Situ With APBI
The treatment of DCIS has evolved significantly over
the past decade. Recommendations have ranged
from simple mastectomy to lumpectomy alone or
lumpectomy followed by whole-breast RT.15 Because
most patients with DCIS now present with small,
Three-Year Analysis of APBI/Vicini et al. 765
mammographically detected lesions, many investiga-
tors have attempted to define subsets of patients
who are adequately treated with excision alone.
Unfortunately, results with this approach have been
inconsistent, and many institutions continue to
recommend postlumpectomy RT for most patients.
Accelerated partial breast irradiation is now
being explored as a possible treatment alternative in
these low-risk patients with the goal of providing
similar local control to whole breast RT but with
reduced morbidity. To our knowledge, the current
study represents the largest compilation of patients
with DCIS treated with APBI (Note: These patients
were enrolled on the Registry trial despite the finding
that it was designed for patients with invasive can-
cer.) These early results (3-year actuarial local control
rate of 100% in the first 48 cases enrolled) are
encouraging and support the continued exploration
of this approach for these local-risk patients.
Study Limitations and Future Considerations
It is important to note that the data presented in this
article are based upon results compiled in a volun-
tary registry study and, therefore, subject to the lim-
itations of this type of investigation. Because this
was not a prospective trial and patients could be en-
rolled at various times, results should be interpreted
with caution. Selection bias in patient enrollment
cannot be ruled out, and, as a result, rates of cosm-
esis, some toxicities, and local recurrence must be
viewed with this in mind. However, previous subset
analyses of the 51% of patients enrolled in the trial
before treatment demonstrated no statistically signif-
icant differences in any of these measures of efficacy
(local control and cosmesis). Also, the growing body
of published data on the clinical use of the device
parallels the early observations in this trial.
It also must be pointed out that much of this in-
formation was collected retrospectively and that sites
were not asked to record some toxicity information
at the initiation of the trial. As a result, under-report-
ing of some toxicities must be seriously considered,
and the data must be viewed cautiously (as with any
type of retrospective analysis). Data from other, more
controlled, prospective trials with longer follow-up
and better defined criteria for objectively assessing
many of these endpoints must be obtained to vali-
date these preliminary observations.
Although these early results on outcome using
the MammoSite to deliver APBI are good, continued
follow-up of these patients and results from other
prospective studies and randomized trials will be
needed to help validate the long-term efficacy of
APBI and the applicability of each APBI technique
766 CANCER February 15, 2008 / Volume 112 / Number 4
for certain clinical settings. Fortunately, the device is
being studied as 1 of 3 possible APBI techniques in
the multi-institutional, prospective, randomized trial
of the National Surgical Adjuvant Breast and Bowel
Project (NSABP) B-39/Radition Therapy Oncology
Group (RTOG) 0413. Enrollment of patients (using
the MammoSite) in this trial is strongly encouraged
to objectively address the critical endpoints of treat-
ment efficacy, cosmesis, and toxicity.
REFERENCES
1. Arthur DW, Vicini FA. Accelerated partial breast irradiation
as a part of breast conservation therapy. J Clin Oncol.
2005;23:1726–1735.
2. Vicini FA, Antonucci JV, Wallace M, et al. Long-term effi-
cacy and patterns of failure after accelerated partial breast
irradiation: a molecular assay-based clonality evaluation.
Int J Radiat Oncol Biol Phys. 2007;68:341–346.
3. Vicini FA, Arthur DW. Breast brachytherapy: North Ameri-
can experience. Semin Radiat Oncol. 2005;15:108–115.
4. Keisch M, Vicini F, Kuske RR, et al. Initial clinical experi-
ence with the MammoSite breast brachytherapy applicator
in women with early-stage breast cancer treated with
breast-conserving therapy. Int J Radiat Oncol Biol Phys.
2003;55:289–293.
5. Vicini FA, Beitsch PD, Quiet CA, et al. First analysis of
patient demographics, technical reproducibility, cosmesis,
and early toxicity: results of the American Society of Breast
Surgeons MammoSite breast brachytherapy trial. Cancer.
2005;104:1138–1148.
6. Zannis V, Beitsch P, Vicini F, et al. Descriptions and out-
comes of insertion techniques of a breast brachytherapy
balloon catheter in 1403 patients enrolled in the American
Society of Breast Surgeons MammoSite breast brachyther-
apy registry trial. Am J Surg. 2005;190:530–538.
7. Arthur DW, Vicini FA, Kuske RR, Wazer DE, Nag S;American
Brachytherapy Society. Accelerated partial breast irradia-
tion: an updated report from the American Brachytherapy
Society. Brachytherapy. 2002;1:184–190. Corrected and
republished in: Brachytherapy. 2003;2:124–130.
8. Vicini F, Baglan K, Kestin L, Chen P, Edmundson G, Marti-
nez A. The emerging role of brachytherapy in the manage-
ment of patients with breast cancer. Semin Radiat Oncol.
2002;12:31–39.
9. Recht A, Silver B, Schnitt S, Connolly J, Hellman S, Harris
JR. Breast relapse following primary radiation therapy for
early breast cancer. I. Classification, frequency and salvage.
Int J Radiat Oncol Biol Phys. 1985;11:1271–1276.
10. Rose MA, Olivotto I, Cady B, et al. Conservative surgery
and radiation therapy for early breast cancer. Long-term
cosmetic results. Arch Surg. 1989;124:153–157.
11. Wong JS, Recht A, Beard CJ, et al. Treatment outcome after
tangential radiation therapy without axillary dissection in
patients with early-stage breast cancer and clinically nega-
tive axillary nodes. Int J Radiat Oncol Biol Phys. 1997;39:
915–920.
12. Benitez PR, Chen PY, Vicini FA, et al. Partial breast irradia-
tion in breast conserving therapy by way of interstitial
brachytherapy. Am J Surg. 2004;188:355–364.
13. Chen PY, Vicini FA, Benitez P, et al. Long-term cosmetic
results and toxicity after accelerated partial-breast irradia-
tion: a method of radiation delivery by interstitial brachy-
therapy for the treatment of early-stage breast carcinoma.
Cancer. 2006;106:991–999.
14. Woodworth PA, McBoyle MF, Helmer SD, Beamer RL. Ser-
oma formation after breast cancer surgery: incidence and
predicting factors. Am Surg. 2000;66:444–450.
15. Silverstein MJ, Lagios MD, Recht A, et al. Image-detected
breast cancer: state of the art diagnosis and treatment.
J Am Coll Surg. 2005;201:586–597.