Breast Cancer

DOI 10.1007/s12282-012-0358-x

ORIGINAL ARTICLE

Prognostic factors for survival after first recurrence in breast

cancer: a retrospective analysis of 252 recurrent cases at a single
institution
Wakako Tsuji Satoshi Teramukai
Masaya Ueno Masakazu Toi Takashi Inamoto

Received: 13 December 2011 / Accepted: 7 March 2012

The Japanese Breast Cancer Society 2012

Abstract
Introduction Previous studies have shown that primary
breast cancer patients with estrogen receptor (ER)-positive
status have better outcomes in terms of both overall survival and disease-free intervals (DFI). However, 25.2 % of
our ER-positive patients experienced recurrence. This
study aimed to define factors potentially predicting survival
after first recurrence in surgically treated patients with
stage IIII breast cancer.
Methods We retrospectively analyzed 252 females with
recurrent breast cancer who had undergone surgery and
been followed at Kyoto University Hospital in Japan. Age,
clinical stage, pathological stage, axillary lymph node
involvement, ER status at the time of diagnosis, progesterone receptor status, human epidermal growth factor
receptor 2 status, operative method, adjuvant chemotherapy, adjuvant endocrine therapy, use of trastuzumab after
recurrence, site of recurrence, DFI, and time of recurrence
were examined for possible influences on survival after the
first recurrence.

W. Tsuji (&)  M. Toi

Division of Breast Surgery, Department of Surgery,
Graduate School of Medicine, Kyoto University,
54 Kawaracho, Shogoin, Sakyo-ku, Kyoto, Japan
e-mail: w-sato@kuhp.kyoto-u.ac.jp
S. Teramukai
Department of Clinical Trial Design and Management,
Translational Research Center, Kyoto University Hospital,
Kyoto, Japan

Results Positive ER status and positive PR status at the

time of diagnosis were significantly favorable factors of
survival after first recurrence for patients with recurrence,
p \ 0.001 and p = 0.021, respectively. More than two
sites of recurrence (p \ 0.001) were associated with
shorter survival time after the first recurrence on multivariate analysis. Survival of patients with recurrent breast
cancer steadily improved from 19801994 to 19952008,
significantly in ER-negative subgroups.
Conclusions Positive ER status at the time of diagnosis is a
powerful predictor for favorable survival after first recurrence. Survival time after first recurrence of breast cancer has
steadily increased in recent decades. Advances in treatments
and attitudes about breast cancer have contributed to this
improvement in survival after first recurrence.
Keywords Estrogen receptor  Survival after first
recurrence  Prognostic factor  Chronological
improvement of survival  Single institutional study
Abbreviations
ER
Estrogen receptor
PR
Progesterone receptor
HER2 Human epidermal growth factor receptor 2
pN
Axillary lymph node involvement
TNM
Clinical stage
pTNM Pathological stage
OS
Overall survival
DFI
Disease-free interval

M. Ueno
Medical Ueno Clinic, Osaka, Japan

Introduction

T. Inamoto
Department of Breast Surgery, Tenri Hospital, Nara, Japan

Breast cancer patients follow a variety of clinical courses,

depending on tumor characteristics. Some relapse a few

123

Breast Cancer

months after operation, while others may relapse many

years later. Recurrent breast cancer patients usually die
from their disease, but the time from recurrence to death
differs markedly among these patients.
Prognostic factors are important for estimating outcomes and determining the optimal form of treatment.
When a patient is considered to be at high risk for recurrence, physicians must plan follow-up and treatments
carefully. Prognostic factors related to overall survival
(OS) and disease-free interval (DFI) have been extensively
evaluated. In addition, we previously reported that estrogen
receptor (ER)-positive patients had better OS- and DFIrelated prognostic factors [1, 2].
Breast cancer treatment is usually combined with surgery, chemotherapy, endocrine therapy, molecular-targeted
therapy and/or radiation therapy. Recent progress in breast
cancer treatments has benefited some patient populations.
There are several reports documenting improved survival
of breast cancer patients [3]. In our previous study, we
described prolongation of OS over time [2]. However,
25.2 % of ER-positive patients in our prior study experienced recurrence.
We retrospectively analyzed data from 252 recurrent
breast cancer patients to determine factors possibly influencing outcomes and changes in the duration of survival over
time after first recurrence. We emphasize that this study was
based on results from a single institution. Such a single
institutional study has the advantage that patient backgrounds
are homogenous and follow-up methods are consistent.

Table 1 Characteristics of the stage IIII patients with recurrence

and without recurrence
Patients with
recurrence

Patients without
recurrence

Age at primary breast operation

\35

11

4.4 %

31

4.6 %

131
110

52.0 %
43.7 %

292
347

43.6 %
51.8 %

Stage I

37

14.7 %

217

32.4 %

Stage II

155

61.5 %

397

59.3 %

Stage III

60

23.8 %

56

8.4 %

Stage I

28

11.1 %

223

32.8 %

Stage II

111

44.0 %

374

55.0 %

Stage III

113

44.8 %

83

12.2 %

pN0

88

34.9 %

452

67.5 %

pN1

64

25.4 %

160

23.9 %

pN2

67

26.6 %

38

5.7 %

pN3

33

13.1 %

20

3.0 %

ER status
Negative

93

36.9 %

197

29.4 %

Positive

114

45.2 %

338

50.4 %

45

17.9 %

135

20.1 %

Negative

84

33.3 %

163

24.3 %

Positive

64

25.4 %

218

32.5 %

104

41.3 %

289

43.1 %

Negative

73

29.0 %

151

22.5 %

Positive

30

11.9 %

52

7.8 %

149

59.1 %

467

69.7 %
38.4 %

3554
[54
TNM

pTNM

pN

Unknown
PR status

Unknown
HER2 status

Methods
Patients

Unknown
Breast surgery

This study was approved by the Kyoto University ethics

committee. From 1980 to 2005, 922 female patients with
stage IIII breast cancer underwent breast surgery with
axillary lymph node dissection at Kyoto University Hospital. Out of 922 total patients, 252 (27.3 %) breast cancer
patients had recurrence before 31 December 2008. The
characteristics of 252 patients of stage IIII with recurrence
and 670 patients without recurrence are shown in Table 1.
All patients were female, and 43.7 % of the patients with
recurrence were older than 55 years old and were postmenopausal. Staging of clinical stage (TNM) and pathological stage (pTNM) were determined by the UICC stage
classification system. Breast-conserving surgery has been
performed since 1990. As a rule, all of the patients who
underwent breast-conserving surgery received radiation
therapy to the conserved breast.
Most of the patients were followed at Kyoto University
Hospital, and some of them were followed at Kitano

123

Breast-conserving surgery
Modified radical mastectomy
Radical mastectomy

73

29.0 %

257

169

67.1 %

408

10

4.0 %

60.9 %
0.7 %

Hospital, Osaka, Japan, which is an affiliate hospital of

Kyoto University Hospital. More than 80 % of patients had
a single recurrent organ at first recurrence, and 17.5 % of
patients had more than two recurrent organs at first recurrence (Table 2). Forty percent of patients had soft-tissue
recurrence, 19.4 % of patients had bone metastasis, 17.5 %
of patients had lung or pleura metastasis, and 5.2 % of
patients had liver metastasis.
Pathological data
Pathological data, including the number of axillary lymph
node metastasis (pN), ER status, progesterone receptor

Breast Cancer
Table 2 Characteristics of recurrence
ER-positive

ER-negative

ER-unknown

Total

Site of recurrence
Soft tissues
Liver

40

35.1 %

41

44.1 %

21

46.7 %

102

40.5 %

5.3 %

4.3 %

6.7 %

13

5.2 %

Lung/pleura

23

20.2 %

15

16.1 %

13.3 %

44

17.5 %

Bone

26

22.8 %

17

18.3 %

13.3 %

49

19.4 %

More than 2 sites

19

16.7 %

16

17.2 %

20.0 %

44

17.5 %

\1year

14

12.3 %

22

23.7 %

12

26.7 %

48

19.0 %

13 year

45

39.5 %

45

48.4 %

16

35.6 %

106

42.1 %

[3year

55

48.2 %

26

28.0 %

17

37.8 %

98

38.9 %

29

25.4 %

22

23.7 %

36

80.0 %

87

34.5 %

85

74.6 %

71

76.3 %

20.0 %

165

65.5 %

Disease-free interval

Time of recurrence
19801994
19952008

(PR) status, and human epidermal growth factor receptor 2

(HER2) status, were evaluated at Department of Pathology
in Kyoto University Hospital. The method for determining
ER and PR status varied during the study period. ER and
PR status have been determined immunohistochemically
(IHC) since 2000. Before 2000, we evaluated ER and PR
status with the dextran charcoal-coating method or by
enzyme immunoassay (EIA). HER2 status has been evaluated since 2000.
Enzyme immunoassay
SRL Inc. (Tokyo, Japan) undertook the measurement.
Briefly, 0.5 g of breast tumor tissue was placed on dry
ice immediately after excision and stored at -80 C until
processing for routine biochemical measurement. Before
cytosol preparation, fat, blood, and necrotic tissue were
eliminated from each sample and discarded, leaving a
piece of tumor tissue between 0.1 and 0.2 g. This tissue
was homogenized in buffer solution and centrifuged at
4 C for 10 min at 8009g. Cytosol was obtained from
the supernatant by centrifugation at 4 C for 60 min at
105,0009g. One hundred microliters of cytosol was
diluted with 100 ll of specimen diluent. The estrogen
receptor concentrations in the diluted cytosol specimens
were determined using the ER-EIA kit from Abbott
Japan Co., Ltd. (Tokyo, Japan), according to the manufacturers instructions. ER values C13 fmol/mg protein
and PR values C10 fmol/mg protein were regarded as
positive.
Immunohistochemistry
For IHC, we used the Ventana HX BenchMark system
(Ventana Medical Systems, Inc., Tucson, AZ, USA),

according to the manufacturers instructions. Briefly, for

antigen retrieval, a retrieval solution (Ventana Medical
Systems, Inc.) was automatically poured onto the sections,
and they were heated on a slide heater to 100 C for
60 min. Endogenous peroxidase activity was quenched by
immersion in 3 % hydrogen peroxide for 4 min. Tissue
sections were incubated with a primary antibody; a
monoclonal antibody to ER (clone SP1, Ventana Medical
Systems, Inc.), PR (clone SP1, Ventana Medical Systems,
Inc.), or HER2 (clone 4B5, Ventana Medical Systems, Inc.)
for 32 min at 42C. Immunoperoxidase staining was performed using the I-VIEW DAB Universal Kit (Ventana
Medical Systems, Inc.), and sections were counterstained
with hematoxylin. ER and PR were considered to be
positive if more than 10 % of cells showed positivity.
However, indications for endocrine therapy were considered in patients with less than 10 % positive cells. HER2
status was defined as positive when IHC score was 3? or a
HER2/centromere probe of chromosome (CEP) 17 ratios
was more than 2.2 in fluorescent in situ hybridization.
Postsurgical therapy
Patients underwent clinical and blood examination every 3
months until 10 years after surgery at both Kyoto University Hospital and Kitano Hospital. Chest X-ray examination, abdominal ultrasonography, and bone scintigraphy
were performed annually before 2001. After 2002, chest
computed tomography replaced chest X-ray examination.
Hormone receptor-positive patients usually receive endocrine therapy for at least 5 years. Tamoxifen was used for
pre- and post-menopausal patients, toremifene for postmenopausal patients. Tamoxifen and toremifene have been
approved since 1981 and 1995, respectively. Medroxyprogesterone acetate was used in some patients.

123

Breast Cancer

Luteinizing hormone-releasing hormone (LHRH) analog

was used for premenopausal patients after 1992. The third
generation aromatase inhibitors have been administered to
postmenopausal patients since 2001. For the patients with
node-positive or high nuclear grades, chemotherapy was
administered. Oral chemotherapy, such as tegafur (FT),
carmofur (HCFU), tegafururasil (UFT), or doxifluridine
(50 DFUR), was administered in the 1980s. In the 1990s,
anthracycline monotherapy was administered. A taxane or
anthracycline-containing regimen was used after 2000.
Trastuzumab was added for patients with a HER2-positive
tumor after 2008.
Treatment after recurrence
Treatment after recurrence varied for each patient. The
goals of treatment for recurrent patients were prolonging
life and maintaining quality of life. Considering patients
general condition, treatments were started based on the
phenotype of the recurrent tumor. For hormone receptorpositive patients, endocrine therapy was applied. Drugs
were chosen among aromatase inhibitors, tamoxifen,
LHRH analog, high dose toremifene, or medroxyprogesterone acetate. For triple-negative tumors, chemotherapy
was given. First-line palliative chemotherapy included an
anthracycline or taxane. Second-line chemotherapy was
mostly based on non-anthracycline non-taxane. For HER2overexpressing tumors, trastuzumab combined with chemotherapy was given after 2000. For bone metastasis,
bisphosphonates were used. If necessary, radiotherapy was
added.
Statistical analysis
We statistically analyzed prognostic factors related to
survival time from first recurrence to death. Survival curves
were estimated with the Kaplan-Meier method. The univariate analysis was performed using the log-rank test and
the univariate Cox regression analysis for estimating hazard ratios. The multivariate Cox regression analysis with
the backward elimination method was used to estimate
hazard ratios and to identify independent prognostic factors. The univariate logistic regression analysis was used to
investigate the independence between recurrent date (period I: 19801994; period II: 19952008) and each factor:
age at primary breast surgery, TNM, pTNM, pN, ER status,
PR status, HER2 status, breast surgery, adjuvant endocrine
therapy, adjuvant chemotherapy, site of first recurrence,
use of trastuzumab after recurrence, and DFI. All reported
p values are two-sided, and a value below 0.05 was considered statistically significant. All statistical analyses were
performed using SAS version 9.2 (SAS Institute, Inc.,
Cary, NC, USA).

123

Results
Factors related to survival after first recurrence
With the univariate analysis, pTNM (p = 0.009), pN
(p = 0.024), ER status (p \ 0.001), PR status (p = 0.005),
adjuvant chemotherapy (p = 0.038), the site of first
recurrence (p = 0.001), DFI (p \ 0.001), and recurrent
date (p \ 0.001) were related to survival after first recurrence. Meanwhile, age at the time of primary breast surgery, TNM, HER2 status, breast surgery, adjuvant
endocrine therapy, and use of trastuzumab after recurrence
had no relationship with survival after first recurrence
(Table 3).
Estrogen receptor status and PR status were strongly
related to survival time after recurrence. The Kaplan-Meier
analysis revealed that ER-positive patients showed more
favorable survival rates throughout the 10 years after the
first recurrence than ER-negative patients (Fig. 1a). ER
positivity was a strong predictor both in pre- and postmenopausal patients (p \ 0.001).
The site of first recurrence was significantly associated
with survival after first recurrence. Patients with liver
metastasis and metastases of more than two organs showed
poorer survival rates than those with single metastasis of
soft tissues, lung/pleura, or bone (Fig. 1b). ER-positive
patients tend to have liver, lung/pleura, or bone metastases.
Otherwise, ER-negative patients tend to have shorter DFI
(Table 2). There was no patient whose first recurrent site
was the brain.
The patients with a longer DFI had better survival rates
after first recurrence than those with a shorter DFI
(Fig. 1c).
Independent prognostic factors after first recurrence
With the multivariate analysis, ER status (p \ 0.001), the
site of first recurrence (p \ 0.001), time of first recurrence
(p \ 0.001), PR status (p = 0.031), and DFI (p = 0.033)
were significant and independent factors influencing the
survival after first recurrence (Table 4). In particular, the
hazard ratio in the ER-positive patients was 0.351 [95 %
confidence interval (CI) 0.2280.540] compared with that
in ER-negative patients after adjustment for the site of
recurrence, time of recurrence, PR status, and DFI. Time of
first recurrence was still significant on the multivariate
analysis.
Chronological changes in survival after first recurrence
curve
The number of patients who were recurrent in period I
(19801994) and in period II (19952008) was 87 and 165

Breast Cancer
Table 3 Univariate analysis for survival after first recurrence
Factors

Postrecurrent survival rate

5-year (%)

Hazard ratio

95 % CI

Log-rank test p value

95 % CI

Age at primary breast surgery

\35

11

4.4

35.8

8.864.8

1.000

3554

131

52.0

46.5

36.855.6

0.953

0.4142.194

NS

[54

110

43.7

38.1

27.448.8

1.172

0.5072.709

TNM
Stage I

37

14.7

54.6

35.270.5

1.000

155
60

61.5
23.8

41.5
38.2

32.550.2
24.551.8

1.417
1.776

Stage I

28

11.1

56.6

30.075.2

1.000

Stage II

111

44.0

46.4

35.256.8

1.349

0.7072.573

Stage III

113

44.8

35.5

25.745.0

2.078

1.1043.910

pN0

88

34.9

48.6

35.560.4

1.000

pN1

64

25.4

47.0

33.059.8

0.958

0.6071.510

pN2

67

26.6

38.0

25.650.4

1.555

1.0272.356

pN3

33

13.1

27.1

11.745.2

1.744

1.0502.895

Negative

93

36.9

26.6

17.336.7

1.000

Positive

114

45.2

53.9

42.164.2

0.442

45

17.9

49.1

32.663.7

Stage II
Stage III

NS
0.8432.382
1.0043.143

pTNM
0.009

pN
0.024

ER status

Unknown
PR status
Negative

84

33.3

30.3

18.842.6

1.000

Positive

64

25.4

52.6

36.766.2

0.530

104

41.3

45.8

35.455.6

Unknown

\0.001
0.3050.641

0.005
0.3380.837

HER2 status
Negative

73

29.0

48.8

33.762.3

1.000

Positive

30

11.9

35.3

16.055.3

1.008

149

59.1

41.4

32.749.7

Unknown

NS
0.5831.807

Breast surgery
Breast-conserving surgery
Modified radical mastectomy
Radical mastectomy

73

29.0

47.9

33.660.9

1.000

169

67.1

39.8

31.547.9

1.404

0.9482.081

NS

10

4.0

53.3

17.779.6

1.223

0.5122.922

43

17.1

61.0

38.777.3

1.000

120

47.6

43.4

33.852.7

1.953

Adjuvant chemotherapy
None
Oral monotherapy

0.038
1.0833.523

Oral multitherapy

15

6.0

44.4

18.567.7

2.348

1.0625.194

Intravenous
Unknown

59
15

23.4
6.0

28.8
32.1

15.543.7
10.255.9

2.488

1.3144.711

None

56

22.2

43.3

28.857.0

1.000

SERM

95

37.7

39.2

37.759.7

0.717

Adjuvant endocrine therapy

NS
0.4571.124

Aromatase inhibitor

18

7.1

70.2

35.788.5

0.510

0.1801.448

LHRH analog ? SERM

16

6.3

45.4

12.574.1

0.909

0.4191.973

2.0

26.7

1.068.6

2.110

0.7445.978

62

24.6

30.9

19.443.3

Medroxyprogesterone acetate
Unknown

123

Breast Cancer
Table 3 continued
Factors

Postrecurrent survival rate

5-year (%)

95 % CI

Hazard ratio

95 % CI

Log-rank test p value

Trastuzumab after recurrence

None

222

88.1

42.9

35.450.1

1.000

Used

30

11.9

38.9

18.459.0

0.873

NS
0.5171.471

Site of recurrence
Soft tissue

102

40.5

48.1

36.858.5

1.000

Liver

13

5.2

13.3

0.843.0

2.214

1.0924.489

Lung/pleura

44

17.5

46.1

27.962.6

0.945

0.5761.552

Bone

49

19.4

55.4

38.569.4

0.936

0.5801.512

44
262

17.5
100

23.8

11.938.0

2.037
0.872

1.3263.129
0.8110.936

\0.001

0.3740.718

\0.001

More than 2 sites

Disease-free interval (year)

0.001

Time of recurrence
19801994

87

34.5

29.2

19.539.7

1.000

19952008

165

65.5

50.8

41.659.3

0.518

(Table 5), respectively. In period I, the percentage of

patients with pathological stage I, II, and III was 6.9, 36.8,
and 56.3 %, respectively. Meanwhile, in period II, the
percentage of patients with pathological stage I and II
increased to 13.3 and 47.9 %, respectively. However, the
percentage of patients with pathological stage III decreased
to 38.8 % in period II. The percentage of patients who
underwent breast-conserving surgery in period I and period
II was 1.1 and 43.6 %, respectively. In period I, no patients
received aromatase inhibitor as adjuvant endocrine therapy
and trastuzumab after recurrence. With the multivariate
analysis, the recurrence date was related to survival time
after the first recurrence (hazard ratio 0.503, p \ 0.001,
Table 4). Survival after first recurrence was significantly
improved from period I to period II (p \ 0.001, Fig. 2a).
From period I to period II, the median survival time was
967 days (2.65 years) and 1831 days (5.02 years), respectively. Survival after first recurrence was improved both in
the ER-positive and ER-negative subgroups (Fig. 2b, c). In
the ER-positive subgroup, the median survival time was
1520 days (4.16 years) and 2569 days (7.04 years) for
period I and period II, respectively (p = 0.053). In the ERnegative subgroup, the median survival time was 660 days
(1.808 years) and 879 days (2.41 years) for period I and
period II, respectively (p = 0.022).
Factors contributing to chronological change
of survival after first recurrence
In order to determine the factors contributing to the chronological change in survival, we performed a univariate
logistic regression analysis. Pathological TNM (p =
0.024), breast surgery (p \ 0.001), adjuvant chemotherapy
(p = 0.014), and DFI (p \ 0.001) had a significant

123

relationship with chronological improvement in survival

(Table 6). The patients in period II had an earlier stage,
longer DFI, and received more breast-conserving surgery
and more intravenous adjuvant chemotherapy than the
patients in period I.

Discussion
Twenty to 30 % of early breast cancer patients experience
recurrence [4]. The course of treatment for these patients is
determined based on the Hortobagyis decision tree [5].
When a patient is diagnosed with recurrent breast cancer,
hormone-receptor status, DFI, age, and menopausal status
are initially assessed. When the disease is hormoneresponsive and non-life-threatening, the patient receives
first-line endocrine therapy. Thus, ER status is the first
important decision-making factor.
We previously reported that ER status was a favorable
prognostic factor of OS and DFI [1, 2]. In this study, we
showed that ER status and PR status at the time of the
operation were independent predictors of survival after first
recurrence (Table 4). Although the assessment process has
changed, ER positivity remained a strong predictor
throughout our study period. The 5-year survival rate after
the first recurrence in ER-negative patients was 27 %
(95 % CI 1737 %), while that in ER-positive patients was
54 % (95 % CI 4264 %). This result confirmed several
reports describing ER status as the best predictor of survival after first breast cancer recurrence [69]. In our
present study, HER2-positive status did not indicate a poor
prognosis with statistical significance. This was due to the
small number of patients with ascertained HER2 status,
because HER2 assessment began in 2000.

Breast Cancer

Survival probability

(A)

1.0
0.8
0.6
0.4
0.2
0.0
0

10

12

14

16

18

Years after first recurrence

(B)

1.0

Survival probability

0.8
0.6
0.4
0.2
0.0

Survival probability

(C)

6
8
10
12
14
Years after first recurrence

16

18

1.0
0.8
0.6
0.4
0.2
0.0
6
8
10
12
Years after recurrence

14

16

18

Fig. 1 Kaplan-Meier estimates for survival after first recurrence

according to a estrogen receptor (ER) status, b site of recurrence, and
c disease-free interval (DFI)

With the univariate analysis, axillary lymph node status

was related to survival after first recurrence (p = 0.024).
However, based on multivariate analysis, axillary lymph
node status was not an independent factor determining
survival after first recurrence. There were several reports
that the multivariate analysis showed axillary lymph node
status to be independently associated with survival after
first recurrence [6, 8, 9], while another report showed that
axillary node status indicated the probability of relapse but

did not influence the length of survival after relapse [7]. In

this study, DFI was also an independent prognostic factor
of survival after first recurrence. Clark et al. [6] and Insa
et al. [8] reported that DFI provided independent information for predicting patient survival after recurrence,
while Koenders et al. [9] observed no association between
survival from the detection of first metastasis and DFI.
Moreover, Howell et al. [7] reported that there was no
significant difference between patients with receptor-positive and receptor-negative tumors in the relapse-free
interval, but survival from first relapse was longer in
patients with receptor-positive tumors. One of the reasons
for this difference might be due to chronological changes
of adjuvant treatments.
The site of initial recurrence is an important determinant
for predicting survival from the time of initial recurrence.
There are several reports indicating ER-positive status to
be significantly associated with a higher rate of bone
metastasis [6, 10, 11]. In this study, 53.1 % of patients with
bone metastasis were ER-positive, and 34.7 % of patients
with bone metastasis were ER-negative (Table 2). The
5-year post-recurrent survival rate was 55 % (95 % CI
3969 %) in patients whose site of first recurrence was
bone, but was only 13 % (95 % CI 143 %) in those with
the liver as first site of recurrence (Fig. 1b), and the multivariate analysis showed the first dominant site of metastasis to be independently associated with survival after
recurrence as showed by others [6, 8, 9]. In our present
study, more than two metastatic sites predicted a poor
outcome. Multiple metastases mean that cancer cells have
disseminated extensively. Some patients develop limited
numbers and sites of metastatic disease. Oligometastasis is
the term describing restricted tumor metastatic capacity
and is associated with a better prognosis [10]. Patients with
oligometastases are sometimes curable following surgical
treatment. Patients with more than two sites of organ
involvement mostly deviate from this concept and are less
curable. Liver metastasis is also associated with a poor
prognosis, because it also means disseminated cancer cells
and leads to early mortality. Even among patients with
liver metastasis, the existence of multiple metastases represents a poorer prognosis after liver resection [11].
Time of the first recurrence is one of the prognostic
factors of survival after the first recurrence (Table 4). The
survival curves after the first recurrence improved from
19801994 to 19952008 (Fig. 2). We previously reported
that OS of primary breast cancer patients in our institute
improved significantly from 19821989 to 19902003 [2].
The improvement of OS is suggested to reflect improvement in survival after the first recurrence.
On logistic regression analysis, breast surgery and
adjuvant chemotherapy were treatment-related factors
contributing to improved survival over time (Table 6). In

123

Breast Cancer
Table 4 Multivariate Cox regression analysis for survival after first
recurrence, n = 252
Factors

Postrecurrent survival
Hazard radio

95 % CI

Positive
PR status

19952008
(n = 165)

Age at primary breast surgery

\35

1.000
0.351

Negative

1.000

Positive

0.583

0.2280.540

\0.001

19801994

1.000

19952008

0.503

0.3570.950

0.3430.738

0.031

\0.001
\0.001*

Site of recurrence
Soft tissue

1.000

Liver

2.988

1.4346.227

0.004

Lung/pleura

1.562

0.9222.646

NS

Bone

1.019

0.6211.671

NS

More than 2 sites

2.653

1.6914.164

\0.001

Disease-free interval (year)

0.921

1.1 %

10

6.1 %

52
34

59.8 %
39.1 %

79
76

47.9 %
46.1 %

Stage I

11

12.6 %

26

15.8 %

Stage II

54

62.1 %

101

61.2 %

Stage III

22

25.3 %

38

23.0 %

Stage I

6.9 %

22

13.3 %

Stage II

32

36.8 %

79

47.9 %

Stage III

49

56.3 %

64

38.8 %

pN0

27

31.0 %

61

37.0 %

pN1

16

18.4 %

48

29.1 %

pN2

31

35.6 %

36

21.8 %

pN3

13

14.9 %

20

12.1 %

ER status
Negative

22

25.3 %

71

43.0 %

Positive

29

33.3 %

85

51.5 %

Unknown

36

41.4 %

5.5 %

3554
[54
TNM

Time of recurrence

0.8540.993

0.033

* p values for global association

view of breast surgery, partial mastectomy was applied to

tumors of relatively small size and early stage breast cancer. The remnant breast after partial mastectomy generally
receives radiation therapy. Radiation therapy has been
reported to reduce breast cancer-related mortality [12].
Radiation therapy is suggested to affect the result in
addition to breast surgery.
Adjuvant chemotherapy has dramatically changed in the
past quarter of a century. Oral chemotherapy was the
mainstream in the 1980s at our institution. New drugs, such
as taxane, trastuzumab-combined chemotherapy, or
anthracycline-containing regimens, have become available
since 2000. Randomized trials have demonstrated that
these drugs improve survival. We did not identify any
specific drug as contributing to improved survival over
time, but we assume that advances in adjuvant chemotherapy in general have all contributed to this clear trend.
As an adjuvant endocrine therapy, selective estrogen
receptor modulators such as tamoxifen were approved in
1981, and third generation aromatase inhibitors have been
administered since 2001 in Japan. In this study, the number
of patients given aromatase inhibitors as adjuvant therapy
was quite small, such that their impact on improvements
over time appeared to be minimal. This supports the view
that adjuvant endocrine therapy did not affect chronological change.
The kinds of treatment provided after recurrence varies
with each individual. ER-negative patients usually receive

123

19801994
(n = 87)

p value

ER status
Negative

Table 5 Characteristics of recurrent patients according to recurrent

periods

pTNM

pN

PR status
Negative

21

24.1 %

63

38.2 %

Positive

20

23.0 %

44

26.7 %

Unknown

46

52.9 %

58

35.2 %

Negative

1.1 %

72

43.6 %

Positive

0.0 %

30

18.2 %

86

98.9 %

63

38.2 %

HER2 status

Unknown
Breast surgery
Breast-conserving surgery
Modified radical
mastectomy
Radical mastectomy

1.1 %

72

43.6 %

80

92.0 %

89

53.9 %

6.9 %

2.4 %

Recurrent organ
Soft tissue

37

42.5 %

65

39.4 %

Liver

5.7 %

4.8 %

Lung

10

11.5 %

34

20.6 %

Bone

15

17.2 %

34

20.6 %

More than 2 organs

20

23.0 %

24

14.5 %
25.5 %

Adjuvant chemotherapy
None

1.1 %

42

Oral monotherapy

48

55.2 %

72

43.6 %

Oral multitherapy

5.7 %

10

6.1 %

Intravenous

22

25.3 %

37

22.4 %

Unknown

11

12.6 %

2.4 %

Breast Cancer

(A) 1.0

Table 5 continued
19952008
(n = 165)

Adjuvant endocrine therapy

None

10.3 %

47

28.5 %

SERM

25

28.7 %

70

42.4 %

Aromatase inhibitor

0.0 %

18

10.9 %

LHRH analog ? SERM

1.1 %

16

9.7 %

Medroxyprogesterone
acetate

0.0 %

2.4 %

52

59.8 %

10

6.0 %

Unknown

0.8

al probabilitty
Surviva

19801994
(n = 87)

0.6
0.4
0.2
0.0
0

Trastuzumab use after recurrence

None

87

Used

10

12

14

16

18

14

16

18

Years after first recurrence

100.0 %
0.0 %

135

81.8 %

30

18.2 %

(B) 1.0

0.6
0.4
0.2
0.0
0

10

12

Years after first recurrence

(C) 1.0
0.8

Surviva
al probability

chemotherapy after recurrence. In recent years, many new

drugs for metastatic breast cancer have appeared, for
example, capecitabine, trastuzumab, vinorelbine, and
gemcitabine, in addition to taxane and anthracycline.
Trastuzumab has also improved the survival of HER2positive patients [13], although our study does not significantly support this. In addition to chemotherapy or
molecular-targeted therapy, wide varieties of endocrine
therapy are used after recurrence in many ER-positive
patients. Drugs that may benefit ER-positive patients
include aromatase inhibitors, tamoxifen, LHRH analogs,
high-dose toremifene, and medroxyprogesterone acetate.
We document the importance of the increasing number of
approved drugs allowing better survival after recurrence.
The survival curve showed improvement both in ERpositive and ER-negative subgroups. Especially in the ERnegative subgroup, survival after the first recurrence
increased significantly (p = 0.022). Our results differ from
reports by Andre et al. [14] and Shigematsu et al. [15].
They reported that the hormone receptor-positive subgroup
has better survival improvement. In our institution, adjuvant chemotherapy has changed from oral therapy to
intravenous multi-therapy. On univariate logistic analysis,
adjuvant chemotherapy is one of the factors contributing to
chronological survival changes. Our results indicate that
survival improvement has been led by the change in
adjuvant chemotherapy.
Pathological TNM was a factor contributing to changes
in survival over time. This prompted us to speculate that
more women are undergoing breast cancer screening,
including mammography, and therefore seeking treatment
in earlier stages of breast cancer, which has resulted in
better outcomes [16].
Disease-free intervals were also one of the factors significantly influencing chronological change. It is reasonable to assume that DFI was prolonged because of

al probabilitty
Surviva

0.8

0.6
0.4
0.2
0.0
0

10

Years after first recurrence

Fig. 2 Chronological changes in survival after first recurrence of

a all patients, b ER-positive subgroup, and c ER-negative subgroup

advances in adjuvant treatments and better patient awareness of breast cancer. Prolonged DFI also impacted the
observed improvement not only in survival over time, but
also in the survival after first recurrence. Certain factors
such as the attitudes of patients and their families, or the
method of follow-up, may affect the ongoing improvement
in survival. These results give us hope that survival time
after the first recurrence will be even longer in the future.

123

Breast Cancer
Table 6 Univariate logistic regression analysis for the factors contributing to chronological change, n = 252
Factors

Chronological changes
Odds radio

95 % CI

pTNM

p value
0.024*

Stage I

1.000

Stage II
Stage III

0.673
0.356

0.2501.815
0.1340.946

NS
0.038
\0.001*

Breast surgery
Breast-conserving surgery

1.000

Modified radical
mastectomy

0.015

0.0020.114 \0.001

Radical mastectomy

0.009

0.0010.097 \0.001

Adjuvant chemotherapy

0.014*

None

1.000

Oral monotherapy

0.036

00.228 \0.001

Oral multitherapy

0.051

00.525

Intravenous

0.041

00.278 \0.001

1.217

1.0851.364 \0.001

Disease-free interval (year)

0.007

* p values for global association

Conclusions
The series of patients examined in this study showed ERpositive status at the time of diagnosis to be associated with
increased OS, DFI, and survival time after the first recurrence of breast cancer. However, the site of recurrence, PR
status at the time of diagnosis, DFI, and recurrent year were
also factors predicting outcomes. Survival time after the
first recurrence of breast cancer has steadily increased in
recent decades, significantly in ER-negative subgroups. We
confirmed that advances in treatments and attitudes about
breast cancer have contributed to this improvement in
survival after first recurrence.
Conflict of interest
peting interests.

The authors declare that they have no com-

References
1. Kiba T, Inamoto T, Nishimura T, Ueno M, Yanagihara K, Teramukai S, Kato H, Toi M, Fukushima M. The reversal of
recurrence hazard rate between ER positive and negative breast
cancer patients with axillary lymph node dissection (pathological
stage IIII) 3 years after surgery. BMC Cancer. 2008;8:323.

123

2. Ueno M, Kiba T, Nishimura T, Kitano T, Yanagihara K, Yoshikawa K, Ishiguro H, Teramukai S, Fukushima M, Kato H, et al.
Changes in survival during the past two decades for breast cancer
at the Kyoto University Hospital. Eur J Surg Oncol.
2007;33(6):6969.
3. Giordano SH, Buzdar AU, Smith TL, Kau SW, Yang Y, Hortobagyi GN. Is breast cancer survival improving? Cancer.
2004;100(1):4452.
4. EBCTCG. Effects of chemotherapy and hormonal therapy for
early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005;365(9472):1687717.
5. Hortobagyi GN. Treatment of breast cancer. N Engl J Med.
1998;339(14):97484.
6. Clark GM, Sledge GW Jr, Osborne CK, McGuire WL. Survival
from first recurrence: relative importance of prognostic factors in
1,015 breast cancer patients. J Clin Oncol. 1987;5(1):5561.
7. Howell A, Barnes DM, Harland RN, Redford J, Bramwell VH,
Wilkinson MJ, Swindell R, Crowther D, Sellwood RA. Steroidhormone receptors and survival after first relapse in breast cancer.
Lancet. 1984;1(8377):58891.
8. Insa A, Lluch A, Prosper F, Marugan I, Martinez-Agullo A,
Garcia-Conde J. Prognostic factors predicting survival from first
recurrence in patients with metastatic breast cancer: analysis of
439 patients. Breast Cancer Res Treat. 1999;56(1):6778.
9. Koenders PG, Beex LV, Kloppenborg PW, Smals AG, Benraad TJ.
Human breast cancer: survival from first metastasis. Breast Cancer
Study Group. Breast Cancer Res Treat. 1992;21(3):17380.
10. Weichselbaum RR, Hellman S. Oligometastases revisited. Nat
Rev Clin Oncol. 2011;8(6):37882.
11. Thelen A, Benckert C, Jonas S, Lopez-Hanninen E, Sehouli J,
Neumann U, Rudolph B, Neuhaus P. Liver resection for metastases from breast cancer. J Surg Oncol. 2008;97(1):259.
12. Clarke M, Collins R, Darby S, Davies C, Elphinstone P, Evans E,
Godwin J, Gray R, Hicks C, James S, et al. Effects of radiotherapy and of differences in the extent of surgery for early breast
cancer on local recurrence and 15-year survival: an overview of
the randomised trials. Lancet. 2005;366(9503):2087106.
13. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire
WL. Human breast cancer: correlation of relapse and survival
with amplification of the HER-2/neu oncogene. Science.
1987;235(4785):17782.
14. Andre F, Slimane K, Bachelot T, Dunant A, Namer M, Barrelier
A, Kabbaj O, Spano JP, Marsiglia H, Rouzier R, et al. Breast
cancer with synchronous metastases: trends in survival during a
14-year period. J Clin Oncol. 2004;22(16):33028.
15. Shigematsu H, Kawaguchi H, Nakamura Y, Tanaka K, Shiotani
S, Koga C, Nishimura S, Taguchi K, Nishiyama K, Ohno S.
Significant survival improvement of patients with recurrent breast
cancer in the periods 20012008 vs. 19922000. BMC Cancer.
2011;11:118.
16. Berry DA, Cronin KA, Plevritis SK, Fryback DG, Clarke L,
Zelen M, Mandelblatt JS, Yakovlev AY, Habbema JD, Feuer EJ.
Effect of screening and adjuvant therapy on mortality from breast
cancer. N Engl J Med. 2005;353(17):178492.