Breast Cancer Research and Treatment

, Volume 111, Issue 2, pp 303–311

Distant disease-free interval, site of first relapse and post-relapse survival in BRCA1- and BRCA2-associated compared to sporadic breast cancer patients

Authors

    • Department of Medical Oncology, Rotterdam Family Cancer ClinicErasmus MC-Daniel den Hoed Cancer Center
  • Caroline Seynaeve
    • Department of Medical Oncology, Rotterdam Family Cancer ClinicErasmus MC-Daniel den Hoed Cancer Center
  • Hanne Meijers-Heijboer
    • Family Cancer Clinic, Department of Clinical GeneticsErasmus MC
    • Department of Clinical GeneticsVU Medical Center
  • J. Margriet Collee
    • Family Cancer Clinic, Department of Clinical GeneticsErasmus MC
  • Marian B. E. Menke-Pluymers
    • Family Cancer Clinic, Department of SurgeryErasmus MC-Daniel den Hoed Cancer Center
  • Carina C. M. Bartels
    • Family Cancer Clinic, Department of SurgeryErasmus MC-Daniel den Hoed Cancer Center
  • Madeleine M. A. Tilanus-Linthorst
    • Family Cancer Clinic, Department of SurgeryErasmus MC-Daniel den Hoed Cancer Center
  • Ans van den Ouweland
    • Family Cancer Clinic, Department of Clinical GeneticsErasmus MC
  • Bert van Geel
    • Family Cancer Clinic, Department of SurgeryErasmus MC-Daniel den Hoed Cancer Center
  • Cecile T. M. Brekelmans
    • Department of Medical Oncology, Rotterdam Family Cancer ClinicErasmus MC-Daniel den Hoed Cancer Center
  • Jan G. M. Klijn
    • Department of Medical Oncology, Rotterdam Family Cancer ClinicErasmus MC-Daniel den Hoed Cancer Center
Epidemiology

DOI: 10.1007/s10549-007-9781-7

Cite this article as:
Kriege, M., Seynaeve, C., Meijers-Heijboer, H. et al. Breast Cancer Res Treat (2008) 111: 303. doi:10.1007/s10549-007-9781-7

Abstract

Background Data on distant disease-free interval (DDFI) and the localization of the first distant metastasis (DM) in BRCA1- and BRCA2-associated breast cancer (BC) patients are as yet scarcely available. Patients and methods We identified 57 BRCA1-associated and 31 BRCA2-associated BC patients, diagnosed between 1980 and 2001, and developing DM disease before 2004, July 1. DDFI, the site(s) of first DM and post-relapse survival of these patients were compared with those of 192 sporadic BC patients. Results As compared to sporadic patients, BRCA1 patients developed less often bone DM (30% vs. 51%; P = 0.005), but tended to develop more often lung DM (26% vs. 16%; P = 0.07), and DM at multiple sites (44% vs. 32%; P = 0.11). In BRCA2-associated compared to sporadic patients, first DM more commonly occurred in lymph nodes (23% vs. 7%; P = 0.007) and at multiple sites (48% vs. 32%; P = 0.08). Adjuvant systemic therapy appeared to be most effective in BRCA2 mutation carriers. Post-relapse survival was worse for BRCA1- and better for BRCA2-associated patients as compared to sporadic patients, but differences disappeared after adjustment for ER-status, site of first DM and DDFI. Conclusion The site of first DM is different between BRCA1- and BRCA2-associated and sporadic BC patients. Differences in post-relapse survival could be explained by differences in site of first DM, in ER-status and in DDFI. Treatment efficacy may differ dependent on genetic status.

Keywords

BRCA1BRCA2Breast cancerDistant disease-free intervalPost-relapse survivalSite of first relapse

Abbreviations

BC

Breast cancer

DDFI

Distant disease-free interval

DM

Distant metastasis

ER

Estrogen receptor

PgR

Progesterone receptor

Introduction

Hereditary breast cancer due to a BRCA1 or BRCA2 mutation can be distinguished from sporadic BC by specific characteristics. BRCA1-associated breast cancer (BC) is more often estrogen receptor (ER), progesterone receptor (PgR) and HER2neu receptor negative (triple negative), is more often of the basal-like and medullary phenotype, and has more frequently a high mitotic count and histological grade [14]. Many of these tumor characteristics are associated with a less favorable prognosis. However, studies investigating overall survival in BRCA1-associated versus age-matched sporadic BC patients have reported contradictive results, with some studies describing a worse survival, and others an equal survival for BRCA1-associated BC patients [2, 510]. In a recent study from our institution comparing 223 BRCA1-associated with 446 sporadic BC patients (matched for age at and period of diagnosis), we found no significant differences in breast cancer specific survival between the two groups, while distant disease-free survival was less favorable in BRCA1-associated patients. However, when adjusted for tumor characteristics, administration of systemic treatment and bilateral (salpingo) oophorectomy (BSO), this difference disappeared [9].

Breast cancer due to a BRCA2 mutation is also of higher histological grade than sporadic breast cancer, although the difference is less pronounced as compared to BRCA1-related tumors [11]. In our experience, no difference in histological grade was found between BRCA2-associated (n = 103) and age-matched sporadic BC patients (n = 759) [12]. ER and PgR-status in BRCA2-associated BC is either not different or even more frequently positive as compared to sporadic BC [3, 4, 12]. With respect to HER2neu status, some studies reported a similar and others a lower percentage of HER2neu overexpression for BRCA2-associated as compared to sporadic tumors [3, 4, 13]. No significant differences in survival were observed between women with a BRCA2-related as compared to a sporadic breast cancer, after adjustment for age and stage at diagnosis [7, 8, 12, 14, 15].

While many studies have explored the tumor characteristics and survival of BRCA1-, and BRCA2-associated as compared to sporadic BC patients, less data are available on the time interval between BC diagnosis and the occurrence of distant metastasis, and the pattern of (first) metastasi(e)s. Previously, it has been reported that the duration of the distant disease-free interval (DDFI), and the localization of first metastasis are prognostic factors for post-relapse survival [1619]. Recent publications have suggested that basal-like tumors such as seen in BRCA1-mutation carriers may result in a different pattern of metastatic spread [20]. To our knowledge, only one study has reported on the spectrum of distant metastases occurring throughout the disease in BRCA1- (n = 15), BRCA2- (n = 12), and non-BRCA1/2-associated (n = 58) BC patients, respectively [21]. In this small study, lung and brain metastases were more, and bone metastases less common in BRCA1-associated compared to non-BRCA1/2 BC patients. In BRCA2 mutation patients, lung metastases were less frequently observed. To our knowledge, there are no data yet on the location of the first metastasis in mutation carriers as compared to sporadic BC patients. These data, however, might be important as this knowledge may have implications for (adjuvant) systemic therapy.

In order to obtain more data, we investigated the DDFI and site of first distant metastasi(e)s in BRCA1- and BRCA2-associated as compared to sporadic breast cancer patients. In addition, post-relapse survival was calculated for the three groups.

Patients and methods

From our institutional database we selected all female patients diagnosed with a primary invasive BC between 1980 and 2001, belonging to a proven BRCA1 or BRCA2 mutation family, with available data on histopathology and follow-up, and having a diagnosis of metastatic disease before July 2004. This selection included identified BRCA1/2 mutation carriers as well as BC patients from a BRCA1/2 mutation family with an unknown individual genetic status [9, 12]. Methods of DNA testing have previously been described [9, 22].

In this way, 223 breast cancer patients from 149 BRCA1 mutation families, and 103 patients from 73 BRCA2 mutation families were identified from the database with breast cancer diagnosed between 1980 and 2001. The 223 BRCA1-associated patients were individually matched for age at and year of breast cancer diagnosis (within 5-years) with two times more (n = 446) sporadic breast cancer patients, selected from the Erasmus MC-Daniel den Hoed Cancer registry. The medical files of the sporadic patients were checked for family history, and patients were excluded if they had at least two additional family members with breast cancer (irrespective of age at onset), or one additional relative with either breast cancer diagnosed below 55 years of age or ovarian cancer (any age) [9].

A further selection criteria for the controls also included that they developed DM before 1 July 2004. Of the 223 BRCA1-associated BC patients, 57 (26%) developed DM, and in the BRCA2 group, 31 of the 103 patients (30%) developed DM. The mean follow-up period after breast cancer diagnosis was 6.0 and 6.3 years for BRCA1- and BRCA2 patients, respectively. In total 446 sporadic patients were selected, of whom 192 (43%) developed DM, the mean follow-up time from BC onset being 5.4 years.

For each patient, the following items were collected from the database and/or the hospital records: birth date, date of breast cancer diagnosis, tumor size, axillary lymph node status, presence of metastases at diagnosis, histology, Bloom and Richardson differentiation grade, ER and PgR-status, type of surgery, radiotherapy, adjuvant systemic treatment, date of first metastasis, sites of first metastasis, vital status, and date of death, date of lost from follow-up or end of the study period (1-7-2004), whichever occurred first. Sites of first metastases were defined as all sites in which metastases were found within 3 months from the first diagnosis of metastatic disease. As all patients were diagnosed with breast cancer and most patients with DM before 2001, data on HER2neu-status were not yet available for most of the patients.

Statistics

The difference in mean DDFI between BRCA1-associated and sporadic patients was tested with a t-test. Further, differences in the spectrum of first distant metastases between BRCA1-associated and sporadic patients were tested by a χ2 test. All these analyses were stratified for ER-status and adjuvant systemic therapy (yes/no; hormonal and chemotherapy combined). The same analyses were performed for BRCA2-associated compared to sporadic patients. To calculate post-relapse survival, Kaplan–Meier survival curves were constructed for BRCA1-associated, BRCA2-associated and sporadic patients, respectively. Differences between curves were tested by a log-rank test. Further, a multivariate hazard ratio for the risk of death was calculated by using the Cox proportional hazard method. ER-status, site of first DM (visceral vs. bone plus soft tissue), and DDFI were included in the model as potential confounders. Longevity bias may occur if index BRCA1/2 tested patients are included, by the selection of long-living patients. To correct for longevity bias, the analyses were also performed for an “unselected” BRCA1- and BRCA2-associated group by excluding index patients undergoing DNA testing >2 years after breast cancer diagnosis [9, 12].

A two-sided P-value of less than 0.05 was considered to indicate statistical significance. All statistical analyses were performed with the use of STATA software (version 9.1).

Results

Patient and tumor characteristics

In Table 1, the patient and primary tumor characteristics for BRCA1-, BRCA2- and sporadic breast cancer patients, respectively, are presented. As there were no significant differences between characteristics of “all” BRCA1/2- and “unselected” BRCA1/2-associated patients, only the characteristics of the whole group of BRCA1- and BRCA2-associated patients and tumors are presented. Compared to sporadic cases, BRCA1-associated cancers were more often ER-negative (78% vs. 38%; P < 0.001) and PgR-negative (63% vs. 31%; P = 0.001), and more frequently lymph node negative (58% vs. 34%; P = 0.001). BRCA1-associated patients more often underwent breast-conserving therapy (53% vs. 37%; P = 0.02), although the number of T1 tumors was not significantly different in the two groups. Adjuvant chemotherapy was less often administered in BRCA1-associated patients (35% vs. 57%; P = 0.004), most likely reflecting the lower percentage of node-positive disease in this group. Further, also adjuvant hormonal therapy was less often administered in BRCA1-associated as compared to sporadic patients (4% vs. 12%; P = 0.08), most likely reflecting the lower frequency of ER-/PgR-positive tumors in addition to the higher incidence of node-negative disease in this patient group.
Table 1

Patient and primary tumor characteristics of patients who developed metastatic disease

 

BRCA1

BRCA2

Sporadic

N

%

P*

N

%

P*

N

%

Number

57

  

31

  

192

 

Mean age at diagnosis (years, range)

41

22–80

0.25

41

31–62

0.33

39

32–64

Period of diagnosis

  

0.60

  

0.26

  

    1980–1984

8

14

 

6

19

 

24

13

    1985–1989

12

21

 

11

35

 

45

23

    1990–1994

24

42

 

7

23

 

65

34

    >1994

13

23

 

7

23

 

58

30

T-status

  

0.52

  

0.97

  

    T1

24

44

 

9

35

 

59

37

    T2

23

43

 

14

54

 

82

52

    T3/T4

7

13

 

3

11

 

18

11

    Unknown

3

  

5

  

33

 

N-status

  

0.001

  

0.25

  

    N0

31

58

 

13

45

 

62

34

    N+

21

42

 

16

55

 

121

66

    Unknown

5

  

2

  

9

 

Histologic gradea

  

0.29

  

0.44

  

    I

0

0

 

0

0

 

2

2

    II

3

11

 

2

12

 

29

24

    III

23

89

 

15

88

 

90

74

    Unknown

31

  

14

  

71

 

ER-statusa

  

<0.001

  

0.01

  

    Positive

10

22

 

19

91

 

105

62

    Negative

35

78

 

2

9

 

64

38

    Unknown

12

  

10

  

23

 

PgR-statusa

  

0.001

  

0.08

  

    Positive

11

37

 

16

89

 

83

69

    Negative

19

63

 

2

11

 

37

31

    Unknown

27

  

13

  

72

 

Surgical therapy

  

0.02

  

0.56

  

    None

4

7

 

1

3

 

6

3

    BCT

30

53

 

14

47

 

70

37

    Mastectomy

23

40

 

15

50

 

116

60

    Unknown

  

1

  

 

Adjuvant chemotherapy

  

0.004

  

0.11

  

    No

35

65

 

18

58

 

81

43

    Yes

19

35

 

13

42

 

109

57

    Unknown

3

  

  

2

 

Adjuvant hormonal therapy

  

0.08

  

0.73

  

    No

50

96

 

27

90

 

165

88

    Yes

2

4

 

3

10

 

23

12

    Unknown

5

  

1

  

4

 

Adjuvant radiotherapy

  

0.84

  

0.81

  

    No

14

26

 

8

27

 

47

26

    Yes

40

74

 

22

73

 

144

74

    Unknown

3

  

1

  

4

 

*P-values were obtained by comparing to the sporadic group

aAs no pathological review was performed, data on tumor characteristics are incomplete, especially for histologic grade and ER- and PgR-status

BRCA2-associated as compared to sporadic tumors were significantly more often ER-positive (91% vs. 62%; P = 0.01), while there was a non-significant trend that BRCA2-associated tumors were more often PgR-positive (89% vs. 69%).

No additional significant differences in-patient and tumor characteristics between BRCA1- or BRCA2-related compared to sporadic breast cancer were observed. In general, adjuvant hormonal therapy was infrequently given, reflecting the insights in the Netherlands before 1994 at which time, if considered, pre-menopausal patients were only receiving adjuvant chemotherapy. In total 11 patients did not have primary surgery, because of metastatic disease, N3 or T4 disease at time of primary diagnosis.

The mean and median DDFI for all BRCA1- and BRCA2-associated patients, BRCA1- and BRCA2-unselected patients (without index-patients undergoing DNA testing >2 years after diagnosis of primary tumor), and sporadic patients, respectively, are presented in Table 2, stratified for ER-status and adjuvant systemic therapy.
Table 2

Distant disease free interval (DDFI)

 

Mean DDFI

Median DDFI

Number

Years (SD)

P*

Years

BRCA1-associated patients

    Total

57

2.8 (2.4)

0.47

2.1

    Unselected

50

2.5 (1.9)

0.15

2.0

    ER+

10

3.9 (2.8)

0.82

2.8

    ER−

28

2.5 (1.6)

0.99

2.3

    Adjuvant therapy

        No

28

3.1 (2.1)

0.52

2.9

        Yes

19

2.0 (1.2)

0.14

1.6

BRCA2-associated patients

    Total

31

4.3 (4.2)

0.04

2.7

    Unselected

24

4.1 (4.3)

0.13

2.5

    ER+

13

3.6 (2.8)

0.95

2.5

    ER−

2

3.6 (4.1)

0.50

3.6

    Adjuvant therapy

        No

12

3.4 (5.4)

0.90

1.9

        Yes

12

4.8 (3.1)

0.01

4.2

Sporadic patients

    Total

192

3.1 (2.8)

 

2.2

    ER+

105

3.6 (3.2)

 

2.6

    ER−

64

2.5 (2.1)

 

2.1

    Adjuvant therapy

        No

75

3.6 (3.3)

 

2.2

        yes

115

2.8 (2.4)

 

2.1

SD: standard deviation; ER: estrogen receptor

*P-values were obtained by comparing to the sporadic group

The mean DDFI was longer in the BRCA2-associated patients (4.3 vs. 3.1 years; P = 0.04), but not significantly different in the BRCA1-associated patients (2.8 years) compared to sporadic patients. When stratified for ER status, the DDFI was not different between the three groups. Strikingly, in patients treated with systemic adjuvant therapy, the DDFI was much longer in BRCA2 carriers (4.8 years) than in the sporadic breast cancer patients (2.8 years; P = 0.01 and BRCA1 carriers (2.0 years; P = 0.002). When using median values of DDFI a similar pattern in the differences between subgroups were found, but in general the median values are somewhat lower than the mean values.

Sites of relapse

In Table 3, the sites and pattern of first metastases for the three patient cohorts is presented for the total group, and for ER-positive and ER-negative patients. Separate data on the site of first metastases are given in Table 4 for patients who did and did not receive adjuvant systemic therapy. As several patients developed metastases at multiple sites, percentages can be over 100% in total.
Table 3

Site and pattern of first distant metastasi(e)s for all, ER-negative and ER-positive patients

 

All

ER+

ER−

BRCA1

BRCA2

Sporadic

BRCA1

BRCA2

Sporadic

BRCA1

BRCA2

Sporadic

N(%)

P*

N(%)

P*

N(%)

N(%)

P*

N(%)

P*

N(%)

N(%)

P*

N(%)

P*

N(%)

Number of women

57

 

31

 

192

10

 

19

 

105

35

 

2

 

64

First site

    Lymph nodes

6 (11)

0.43

7 (23)

0.007

14 (7)

1 (10)

0.48

4 (21)

0.01

5 (5)

4 (11)

0.94

1 (50)

0.10

7 (11)

    Skin

2 (4)

0.89

2 (7)

0.36

6 (3)

0 (0)

0.53

1 (5)

0.77

4 (4)

2 (6)

0.53

1 (50)

0.002

2 (3)

    Bone

17 (30)

0.005

17 (55)

0.70

98 (51)

6 (60)

0.95

10 (53)

0.60

62 (60)

8 (23)

0.14

1 (50)

0.72

24 (38)

    Liver

13 (23)

0.74

8 (26)

0.92

48 (25)

1 (10)

0.38

6 (32)

0.36

23 (22)

8 (23)

0.47

1 (50)

0.54

19 (30)

    Lung

15 (26)

0.07

6 (19)

0.60

30 (16)

1 (10)

0.83

4 (21)

0.31

13 (12)

11 (31)

0.15

0

0.50

12 (19)

    Pleura

7 (12)

0.69

1 (3)

0.20

20 (10)

0 (0)

0.89

1 (5)

0.42

12 (11)

4 (11)

0.75

0

0.65

6 (9)

    Brain

8 (14)

0.20

3 (10)

0.80

16 (8)

1 (10)

0.59

2 (11)

0.43

6 (6)

5 (14)

0.80

0

0.59

8 (13)

    Other

10 (18)

0.15

2 (7)

0.49

20 (10)

3 (30)

0.07

2 (11)

0.99

11 (11)

5 (14)

0.98

0

0.57

9 (14)

    Unknown

1 (2)

0.36

0

0.69

1 (1)

0

 

0

 

0

1 (3)

0.66

0

0.86

1 (2)

Multiple sites

    Total

25 (44)

0.11

15 (48)

0.08

62 (32)

3 (30)

0.98

11 (58)

0.02

32 (31)

16 (46)

0.34

2 (100)

0.07

23 (36)

    With bone

9 (16)

0.28

7 (23)

0.98

43 (22)

2 (20)

0.84

6 (32)

0.41

24 (23)

4 (11)

0.26

1 (50)

0.31

13 (20)

    Without bone

16 (28)

0.001

8 (26)

0.01

19 (10)

1 (10)

0.79

5 (26)

0.01

8 (8)

2 (34)

0.03

1 (50)

0.20

10 (16)

*P-values were obtained by comparing to the sporadic group

Table 4

Site and pattern of first distant metastasi(e)s for patients with and without adjuvant therapya

 

No adjuvant therapy

Adjuvant therapy

BRCA1

BRCA2

Sporadic

BRCA1

BRCA2

Sporadic

N(%)

P*

N(%)

P*

N(%)

N(%)

P*

N(%)

P*

N(%)

Number of women

34

 

16

 

75

20

 

15

 

115

First site

    Lymph nodes

1 (3)

0.10

2 (13)

0.93

10 (13)

5 (25)

<0.001

5 (33)

<0.001

4 (4)

    Skin

1 (3)

0.56

2 (13)

0.02

1 (1)

1 (5)

0.74

0

0.46

4 (4)

    Bone

9 (27)

0.03

7 (44)

0.76

36 (48)

7 (35)

0.12

10 (67)

0.35

62 (54)

    Liver

9 (27)

0.20

7 (44)

0.01

12 (16)

2 (10)

0.06

1 (7)

0.05

35 (30)

    Lung

11(32)

0.05

4 (25)

0.39

12 (16)

3 (15)

0.94

2 (13)

0.82

18 (16)

    Pleura

3 (9)

0.93

0

0.20

7 (9)

4 (20)

0.28

1 (7)

0.59

13 (11)

    Brain

6 (18)

0.22

2 (13)

0.70

7 (9)

2 (10)

0.74

1 (7)

0.87

9 (8)

    Other

6 (18)

0.69

1 (6)

0.37

11 (15)

4 (20)

0.09

1 (7)

0.87

9 (8)

    Unknown

0

 

0

 

0

1 (5)

0.16

0

0.72

1 (1)

Multiple sites

    Total

15 (44)

0.10

9 (56)

0.03

21 (28)

9 (45)

0.42

6 (40)

0.74

41 (36)

      With bone

5 (15)

0.73

3 (19)

0.89

13 (17)

3 (15)

0.29

4 (27)

0.96

30 (26)

      Without bone

10 (29)

0.02

6 (38)

0.007

8 (11)

6 (30)

0.01

2 (13)

0.65

11 (10)

*P-values were obtained by comparing to the sporadic group

aAdjuvant therapy: chemo-and hormonal therapy combined

The most frequent sites of first distant metastases in BRCA1-associated and sporadic BC patients were bone, lung, and liver, whereby the ranking order and percentages significantly differed between the two groups. Metastases in bone significantly less commonly occurred in BRCA1-associated than in sporadic patients (30% vs. 51%; P = 0.005). Further, non-significant trends were found for more frequent metastases in the lung (26% vs. 16%; P = 0.07) and at multiple sites (44% vs. 32%; P = 0.11) in BRCA1-associated patients. The lower frequency of bone metastases in BRCA1-associated patients disappeared after stratification for ER status with only a trend (23% vs. 38%; P = 0.14) in ER-negative patients. The lower incidence of bone metastases and higher incidence of lung metastases were especially observed in BRCA1-associated patients without adjuvant therapy (Table 4). No significant differences were observed for the other metastatic sites with the exception of a higher incidence of lymph node metastases (25% vs. 4%; ≤ 0.001) in adjuvantly treated BRCA1 mutation carriers.

As compared to sporadic patients, patients with a BRCA2 mutation more often developed lymph node metastases (23% vs. 7%; P = 0.007; Table 3), mainly in the adjuvantly treated patients (Table 4). Furthermore, there was a non-significant trend for more frequently metastases at multiple sites (48% vs. 32%; P = 0.08; Table 3). In the ER-positive group, the same differences between BRCA2 and sporadic patients were observed. Since only two BRCA2 mutation carriers had an ER-negative tumor, no conclusions for this subgroup could be drawn. The higher incidence of metastases at multiple sites in BRCA2 patients persisted in the group of patients not having been treated with adjuvant systemic therapy (56% vs. 28%; P = 0.03), but almost disappeared in the adjuvantly treated group (40% vs. 36%). Intriguingly, more liver metastases were observed in BRCA2-associated as compared to sporadic patients in the group not having been treated with adjuvant therapy (44% vs. 16%; P = 0.01), whereas less liver metastases were seen in BRCA2-associated patients in the patient cohort having received adjuvant therapy (7% vs. 30%; P = 0.05). Overall no difference in the incidence of liver metastases was observed between the total group BRCA2-associated and sporadic patients (Table 3).

Post-relapse survival

Kaplan–Meier post-relapse survival rates and hazard ratios (HR) for the risk of death for all BRCA1- and BRCA2-associated, BRCA1- and BRCA2-unselected, and sporadic patients, respectively, are shown in Table 5 and Fig. 1. BRCA1-associated patients had a worse post-relapse survival than sporadic patients, especially the BRCA1-unselected patients (P log-rank 0.04; HR univariate 1.43). The difference was especially shown in the first two years, but disappeared after 5 years of follow-up. However when we adjusted for ER-status, site of first metastasis and DDFI, this difference disappeared (HR multivariate 1.01; Table 5). BRCA2-associated patients had a borderline significantly better post-relapse survival (P = 0.05) than sporadic patients (Table 5; Fig 1). This difference also disappeared after adjustment for ER-status, site of first metastasis and DDFI.
Table 5

Post-relapse survival rates, hazard ratios (HR) and 95% confidence intervals (CI) for the different patient cohorts

 

BRCA1

BRCA1-unselected

BRCA2

BRCA2-unselected

Sporadic

PRS rate

95% CI

PRS rate

95% CI

PRS rate

95% CI

PRS rate

95% CI

PRS rate

95% CI

1 year

0.50

0.56–0.63

0.47

0.32–0.60

0.90

0.72–0.97

0.91

0.69–0.98

0.69

0.61–0.75

2 year

0.34

0.21–0.47

0.32

0.19–0.45

0.65

0.44–0.79

0.60

0.37–0.77

0.43

0.36–0.50

5 year

0.11

0.04–0.22

0.07

0.02–0.18

0.25

0.10–0.44

0.11

0.01–0.34

0.11

0.06–0.18

P-value*

0.22

 

0.04

 

0.05

 

0.31

   
 

HR

 

HR

 

HR

 

HR

   

Univariate

1.23

0.88–1.73

1.43

1.01–2.01

0.81

0.65–1.01

0.94

0.83–1.06

  

Multivariatea

0.86

0.56–1.31

1.01

0.66–1.54

0.92

0.70–1.20

1.00

0.86–1.17

  

PRS: post-relapse survival, CI: confidence interval, HR: hazard ratio

*P-values were obtained by comparing to the sporadic group

aAdjusted for ER-status, site first metastasis and distant disease free interval

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9781-7/MediaObjects/10549_2007_9781_Fig1_HTML.gif
Fig. 1

Kaplan–Meier curves for post-relapse survival of the sporadic, BRCA1 and BRCA2 group

Discussion

Our study confirms that BRCA1-associated tumors are more frequently hormone receptor negative and that in contrast BRCA2-associated tumors tend to be more frequently hormone receptor positive compared to sporadic breast cancer. However, currently there are scarcely any data on DDFI, sites of first distant metastases and post-relapse survival in BRCA1- and BRCA2-associated BC patients as compared with sporadic patients. At time at diagnosis of the primary tumor BRCA1 mutation carriers had significantly less lymph node metastases, while during follow-up lymph node metastases developed most frequently in BRCA2 mutation carriers. This may suggest that BRCA1-associated tumors tend to disseminate more frequently hematogenic, and BRCA2-associated tumors lymphogenic. In addition, BRCA1-associated tumors appeared to develop less frequently bone metastases than BRCA2-associated tumors. These differences in metastatic pattern can partly be explained by ER-status. It is well known that ER-positive patients develop more often metastasis to bone and soft tissue, while ER-negative tumors are associated with visceral disease [16, 17, 23, 24].

Our findings regarding the lower rate of bone metastases and the higher frequency of lung metastases in BRCA1-associated patients are in line with the results of Albiges et al. [21], although the latter study also reported a higher occurrence of brain metastases in BRCA1-associated patients, which we did not observe. Further, the less frequent occurrence of lung metastases in BRCA2-associated BC patients found by Albiges et al. was also at variance with results from our study. The observed differences might partly be explained by differences in study designs, and patient cohorts. We focused on the sites of the first distant metastases, while Albiges et al. considered all sites of metastatic involvement occurring during the course of metastatic disease. In our opinion, the first metastatic location is a different entity than the spectrum of metastatic locations occurring throughout the course of metastatic disease, as the latter is influenced by the result of the systemic therapy. Further, our sporadic control group consisted of BC patients without a family history of breast/ovarian cancer, matched for age at and year of breast cancer diagnosis to the BRCA1/2-associated cases, while Albiges et al. used an unmatched control group of BC patients with a family history of breast cancer, whereby no BRCA1 or BRCA2 mutation was identified (=non-BRCA1/2 cohort). As a consequence, the age at BC diagnosis in their control group was higher than in the BRCA1 group. Finally, the sample size in their study was much smaller than in our analysis.

The mean DDFI in BRCA1-associated patients was shorter than for sporadic patients, especially in the unselected BRCA1-associated cohort. In contrast, for the total group of BRCA2-asssociated patients, the DDFI was significantly longer than for sporadic patients. These findings can also be explained by differences in ER-status, as no differences could be observed within the subgroups stratified by ER status.

Similar observations were done for post-relapse survival i.e., BRCA2-associated patients showed a better post-relapse survival and patients with BRCA1-associated tumors a worse post-relapse survival in comparison with patients with sporadic tumors. These differences disappeared in the multivariate analysis including hormone receptor status, DDFI and site of relapse.

Regarding treatment effects, the most striking observation was the much higher mean DDFI in adjuvantly treated BRCA2 mutation carriers (4.8 years) versus those in adjuvantly treated BRCA1 mutation carriers (2.0 years; P = 0.002) and in patients with sporadic disease (2.8 years), while no major differences were found within the non-adjuvantly treated groups (3.4, 3.1, and 3.6 years respectively; Table 2). This observation suggest a greater sensitivity of BRCA2-associated tumors to systemic treatment, especially chemotherapy. Furthermore adjuvant systemic therapy diminished especially the occurrence of visceral disease (liver, lung, and brain) in comparison to the non-adjuvantly treated group. The efficacy of systemic treatment can directly best evaluated in patients measurable disease i.e., patients with measurable metastases or in the neo-adjuvant setting, The efficacy of first-line chemotherapy in BRCA1 and BRCA2 mutation carriers with metastatic disease is currently under evaluation in our Family Cancer Clinic. In vitro studies as well as small studies regarding neo-adjuvant chemotherapy suggest that BRCA1/2-associated BC may be more sensitive to chemotherapeutic agents causing double strand DNA breaks [2527]. Current therapeutic guidelines do not distinguish between BRCA1- and BRCA2-associated, and sporadic breast cancer. However, differences in disease entity and metastatic involvement between BRCA1- or BRCA2-associated patients and sporadic patients might ultimately have therapeutic consequences. For example, the lower frequency of bone metastases as was found in BRCA1-associated patients might ultimately have impact regarding treatment with biphosphonates in the adjuvant setting. It would be of interest to gather more data regarding this issue from previously performed and currently ongoing bisphosphonate studies [28]. Further, it has been suggested that because of the deficient DNA repair mechanism in BRCA1/2- associated tumors, different and tailored chemotherapeutic regimens will result in an improved outcome. In BRCA1 and BRCA2 carriers with metastatic disease, a randomized trial with carboplatin versus docetaxel is underway [29]. In the near future, it is expected that more precise delineation of the specific disease entity by means of more advanced immunochemical methods [30], molecular biological technologies, such as molecular profiling will provide more individualized treatment strategies aiming at improved outcome.

In conclusion, we found differences in metastatic pattern and post-relapse survival between BRCA1-, BRCA2-associated and sporadic tumors, which at least partly could be explained by differences in ER-status. In addition post-relapse survival was also influenced by DDFI and site of relapse. Most interestingly, adjuvant systemic treatment appears especially effective in BRCA2-associated tumors. For confirmation of these results additional studies are warranted.

Acknowledgements

The study was supported by a grant from the Dutch Cancer Society (DDHK 2004-3124) and Cancer Genomics Center. We thank Ellen Crepin, Jannet Blom and Elisabeth Huijskens for the collection of data.

Copyright information

© Springer Science+Business Media, LLC. 2007