Breast Cancer Res Treat (2013) 142:559–567 DOI 10.1007/s10549-013-2726-4
CLINICAL TRIAL
Adjuvant tamoxifen-induced mammographic breast density reduction as a predictor for recurrence in estrogen receptor-positive premenopausal breast cancer patients Kyung Lan Ko • In Suk Shin • Ji Young You So-Youn Jung • Jungsil Ro • Eun Sook Lee
•
Received: 30 May 2013 / Accepted: 3 October 2013 / Published online: 14 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Tamoxifen is known to reduce the risk of breast cancer in women at high risk and also reduces mammographic breast density (MD) in a preventive setting. We investigated the efficacy of MD reduction (MDR) for predicting recurrence in estrogen receptor (ER)-positive patients in an adjuvant setting. A total of 1,066 ER-positive breast cancer patients who were enrolled in this study underwent curative surgery and received adjuvant tamoxifen for at least 2 years at our institution between January 2003 and December 2008. Using a computerized system, a single radiologist reviewed all mammograms and classified MD patterns on the basis of the Breast Imaging Reporting and Data System. MDR was assessed using the baseline mammogram taken before surgery (preMD) and the followup mammogram taken after the start of adjuvant tamoxifen (postMD). MDR positivity was defined as downgrading of the postMD grade, with the preMD grade as a reference. Patients were divided into 2 groups, MDR-positive and MDR-negative, for statistical analysis. Patients who showed MDR after an average of 19 months of adjuvant tamoxifen treatment had a 65 % lower risk of recurrence than patients who did not show MDR. Furthermore, significant risk reduction according to MDR had a predictive power for any type of recurrence pattern including loco-regional recurrence (87 % reduction) and systemic recurrence (52 % reduction) in ER-positive breast cancer patients, especially in women B50 years. In our study, only 4 patients (0.4 %) Kyoung Lan Ko and In Suk Shin contributed equally to this paper. K. L. Ko I. S. Shin J. Y. You S.-Y. Jung J. Ro E. S. Lee (&) Center for Breast Cancer, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do 410-769, Republic of Korea e-mail: [email protected]
showed contralateral breast recurrence during the mean 61-month follow-up period and none of them experienced MDR. In conclusion, MDR during adjuvant tamoxifen therapy was independently associated with a lower risk of systemic and loco-regional recurrence in ER-positive breast cancer patients, especially in young women. For patients who do not experience MDR after approximately 1.5 years of tamoxifen therapy, more caution should be taken and other treatment strategies are warranted. Keywords Mammographic breast density Estrogen receptor positive Adjuvant tamoxifen therapy Breast cancer
Introduction As at least 70 % of breast cancers are estrogen receptor (ER)-positive, targeting estrogen action has been a mainstay of breast cancer treatment. The selective ER modulator tamoxifen, synthesized during the 1970s, was the most competitive inhibitor of estrogen–ER binding and became the first and most successful targeted cancer therapy [1]. The Early Breast Cancer Trialists’ Collaborative Group showed that adjuvant tamoxifen therapy almost halved the rate of annual recurrence and reduced mortality by 31 % [2]. Unfortunately, some patients with resistance to endocrine therapy especially tamoxifen have recurrent disease during or after this treatment. Many investigators are attempting to develop more specific biomarkers to predict endocrine resistance, because currently the ability to predict resistance is suboptimal. It has been hypothesized that breast density, which is known to correlate with breast cancer risk in population studies, may predict response to hormonal manipulations [3].
123
560
Breast Cancer Res Treat (2013) 142:559–567
Fig. 1 Schematic diagram showing the flow of patients through the study
Breast density is most commonly measured by mammography. Mammographic breast density (MD) varies among women because of variations in breast tissue composition and differences in the radiographic attenuation properties of stroma, epithelium, and fat [4]. All of these factors are influenced by endogenous estrogen concentration [5]. Western women in the highest MD category are thought to have a fourfold to sixfold increased risk of breast cancer [6]. For Asian women, breast cancer risk is thought to be increased by twofold to fivefold in women with a high breast density [7, 8]. Tamoxifen is known to reduce the risk of breast cancer in women at high risk and also reduces MD. A case–control study found that the change in MD over 12–18 months is a predictor of response to tamoxifen in a preventive setting [9]. The efficacy of MD reduction (MDR) in predicting response to tamoxifen in an adjuvant setting is not well known. A recent report showed that MDR, assessed by imaging analysis software, during adjuvant endocrine therapy is a significant predictor of recurrence of ERpositive breast cancer [10]. This is an important step toward establishing effective strategies for patients who will benefit less from endocrine therapy. We also investigated the efficacy of MDR in predicting breast cancer recurrence in the adjuvant setting using qualitative methods based on the widely used Breast Imaging Reporting and Data System (BI-RADS; American College of Radiology, Reston, VA, USA).
Methods Patients Our study is a retrospective one based on the existing prospective breast cancer database of the National Cancer Center, Goyang, Korea. A total of 2,402 ER-positive breast cancer patients who were enrolled in this study underwent curative surgery at our institution between January 2003 and December 2008. We investigated 1,526 patients who
123
received adjuvant tamoxifen for at least 2 years. Patients were excluded if their digital mammograms were not available or not appropriate for evaluation or if they had bilateral breast cancer or occult breast cancer. Finally, we collected the clinicopathologic information on 1,066 patients by reviewing the prospective database of our institution and the data of disease recurrence by additional medical record review (Fig. 1). Written informed consent was obtained prior to surgery from all patients, and this study was approved by the Institutional Review Board of the National Cancer Center, Goyang, Korea. Mammographic breast density assessment Using a computerized system, a single radiologist (K. Ko: 10 years of experience in interpreting mammograms) reviewed 2,132 preoperative and postoperative mammograms classified breast density patterns according to BIRADS. Four density grades were classified: (1) almost entirely fat (\25 % glandular), (2) scattered fibroglandular densities (25–50 % glandular), (3) heterogeneously dense (50–75 % glandular), and (4) extremely dense ([75 % glandular). MDR was determined according to a pair of digital mammograms: the baseline mammogram taken before surgery (preMD) and the follow-up mammogram taken 10–34 months (median, 19 months) after the start of adjuvant tamoxifen (postMD). MDR positivity was defined as the downgrading of the postMD grade, with the preMD grade as a reference. We relied on a single radiologist who is a specialist in breast imaging studies, thereby eliminating interobserver variability. We did not seek to measure reproducibility as the BI-RADS density classifications are standardized. Statistical analysis Patients were divided into 2 groups, MDR-positive and MDR-negative, to perform statistical analysis. To identify the association of factors with MDR, the t test and Chi square test were used. Body mass index (BMI) was
Breast Cancer Res Treat (2013) 142:559–567
561
Table 1 Clinicopathologic characteristics of patients
Table 1 continued
Variable number (%)
Total
MDR positive
MDR negative
Variable number (%)
Mean age, years ± SD
45.3 ± 7.6
44 ± 5.94
46 ± 8.11
Recurrence
Range
25–78
28–68
25–78
B50 [50
888 (83.3) 178 (16.7)
308 (91.1) 30 (8.9)
580 (79.7) 148 (20.3)
59 ± 17.6
59 ± 17.7
59 ± 17.5
Mean follow-up time, months ± SD Range
26–114
27–113
26–114
23.4 ± 3.2
22.89 ± 3.1
23.07 ± 3.2
15.6–50.2
17.5–35.5
15.6–50.2
1.7 ± 1.2
1.6 ± 1.2
1.7 ± 1.1
0.1–9.0
0.1–6
0.1–9
Negative
666 (61.5)
195 (57.7)
461 (63.3)
Positive
410 (38.5)
143 (42.3)
267 (36.7)
Mean BMI, kg/ m2 ± SD Range Mean tumor size, cm ± SD Range Lymph node status
Histologic grade 1/2
840 (78.8)
268 (79.3)
572 (78.6)
3
226 (21.2)
70 (20.7)
156 (21.4)
ER score (Allred score) Low (B6) 336 (31.6) High (C7)
729 (68.4)
117 (34.6)
220 (30.2)
221 (65.4)
508 (69.8)
PgR score (Allred score) Low (B5)
370 (34.7)
113 (33.4)
257 (35.3)
High (C6)
696 (65.3)
225 (66.6)
471 (64.7)
HER2 Negative
972 (91.2)
302 (89.3)
670 (92.0)
Positive
94 (8.83)
36 (10.7)
58 (8.0)
Ki-67 (%) \14
669 (62.8)
202 (59.8)
467 (64.1)
C14
397 (37.2)
136 (40.2)
261 (35.9)
Total
MDR positive
MDR negative
Total
67
10 (3.0)
57 (7.8)
Systemic
48
9 (2.7)
39 (5.4)
Loco-regional
17
1 (0.3)
15 (2.1)
Contralateral breast
4
0 (0)
4 (0.5)
BMI body mass index, ER estrogen receptor, PgR progesterone receptor, HER2 human epidermal growth factor receptor 2
calculated as weight/height2 (kg/m2) and both ER and progesterone receptor (PgR) expression levels were divided into binary covariates for an Allred score (low ER, B6; high ER, C7; low PgR, B5; high PgR, C7). Breasts were classified into 3 groups according to preMD and postMD grades: total fatty breasts (both ‘‘almost entirely fat’’ and ‘‘scattered fibroglandular tissue’’), heterogeneously dense breasts, and extremely dense breasts. If the preMD score was 1, it could not be 0 after treatment. We defined the total fatty breast group as including grades 1 and 2. Multivariate logistic regression analysis was performed to evaluate the factors related to MDR. Multivariate Cox regression analysis was used to compare recurrence-free survival according to MDR and to calculate survival curves. The association of MDR with disease-free survival according to patterns of recurrent disease (loco-regional, systemic, and contralateral recurrence) was analyzed. All statistical analyses were performed using Stata version 10.0 (Stata Corporation, College Station, Texas, USA) software, and factors with a p value of \0.05 were considered statistically significant.
Results
Lymphovascular invasion Negative
320 (30)
88 (26.0)
232 (31.9)
Positive
746 (70)
250 (74.0)
496 (68.1)
Patients and MD distribution The mean age of the 1,066 enrolled patients was 45.3 years (range 25–78 years; Table 1). One hundred and thirty-four patients (12.6 %) had ductal carcinoma in situ, and 175 patients (16.4 %) had received neoadjuvant chemotherapy. A total of 67 recurrences were recorded during the current period of available follow-up (mean 61.3 months). The distribution of patients according to BI-RADS MD category before and after tamoxifen therapy is shown in Table 2. Extremely dense breasts were noted in 422 patients (39.6 %) before treatment and in 273 patients (25.6 %) after an average of 19 months (10–36 months) of adjuvant tamoxifen. We analyzed multiple factors for
Operation Mastectomy
162 (15.2)
46 (13.6)
116 (15.9)
904 (84.8)
292 (86.4)
612 (84.1)
No
303 (28.4)
68 (20.1)
235 (32.3)
Yes (adjuvant/ neoadjuvant)
763 (71.6)(588(55.2)/ 175(16.4))
270 (79.9)
493 (67.7)
No
173
50 (14.8)
123 (16.9)
Yes
893
288 (85.2)
605 (83.1)
Breastconserving surgery Chemotherapy
Radiotherapy
123
562
Breast Cancer Res Treat (2013) 142:559–567
Table 2 Distribution of patients according to BI-RADS breast density category before and after treatment PostMDb \50 %
[75 %
50–75 %
Total
PreMDa \50 %
141
0
0
141
50–75 % [75 %
180 5
323 144
0 273
503 422
Total
326
467
273
1,066
BI-RADS: The American College of Radiology Breast Imaging Reporting and Data System, \50 %: grade 1 (almost entirely fat) and grade 2 (scattered fibroglandular tissue), 50–75 %: grade 3 (heterogeneously dense), [75 %: grade 4 (extremely dense) a
PreMD: initial mammographic density before treatment
b
PostMD: tamoxifen
follow-up
mammographic
density
after
adjuvant
association with preMD; we found that age and BMI were inversely associated with preMD (p \ 0.001; Fig. 2). MDR was observed in 338 patients (31.7 %); MD was downgraded by 2 grades in only 5 patients (0.6 %), whereas MDR was downgraded by 1 grade in the others (99.4 %). Factors associated with MD reduction Multiple variables were analyzed to identify associations with MDR. MDR-positive patients were significantly younger than MDR-negative patients (43.6 ± 5.9 vs. 46.2 ± 8.1 years; p \ 0.001), but BMI did not differ significantly between the 2 groups (23.4 ± 3.1 vs. 23.4 ± 3.3 kg/m2; p = 0.939). Time from the start of adjuvant tamoxifen to the follow-up mammogram was significantly longer in MDR-positive patients than in MDR-negative patients (19.7 ± 4.3 vs. 18.6 ± 4.2 months; p \ 0.001). A high preMD grade and chemotherapy administration were significantly associated with MDR positivity (p \ 0.001). To evaluate the independent effect of each factor associated with MDR positivity, we used multivariate logistic regression analysis (Table 3). We found that age as a continuous variable was inversely associated with MDR positivity [odds ratio (OR) = 0.97; p = 0.025). In addition, a long interval from the start of adjuvant tamoxifen to the follow-up mammogram (OR = 1.05; p = 0.001), a high preMD grade (OR = 6.30; p \ 0.001), and chemotherapy administration (OR = 1.67; p = 0.002) were significantly associated with MDR (Table 3). However, the ER score, PgR score, and Ki-67 level had no association with MDR. MD reduction and recurrence-free survival Sixty-seven of 1,066 patients (6.3 %) experienced tumor recurrence during follow-up (48 patients: systemic
123
Fig. 2 Association of preMD with age and BMI, a preMD and BMI, b preMD and age. BMI body mass index, preMD pretreatment mammographic breast density, preMD 1 almost entirely fatty breasts, preMD 2 breasts with scattered fibroglandular tissue, preMD 3 heterogeneously dense breasts, and preMD 4 extremely dense breasts. The dotted lines represent linear prediction lines. Linear regression analysis
recurrence; 16 patients: loco-regional recurrence; and 4 patients: contralateral breast recurrence; Table 1). The association of disease recurrence with MDR according to recurrence pattern is shown in Table 4. MDR was significantly associated with loco-regional recurrence, systemic recurrence, and total recurrence (p \ 0.05). MDR and total recurrence-free survival In multivariate Cox regression analyses adjusted for conventional prognostic factors [age, BMI, tumor size, lymph node status, ER score, PgR score, and human epidermal growth factor receptor 2 (HER2) status], MDR positivity remained an independent predictor of total recurrence-free survival risk [hazard ratio (HR) = 0.36; p = 0.003; Table 5, 6; Fig. 3a]. Large tumor size, lymph node
Breast Cancer Res Treat (2013) 142:559–567
563
Table 3 Factors related to mammographic density reduction Variable
Odds ratio (95 % CI)
p value
Table 5 Effect of mammographic density reduction (MDR) on recurrence-free survival Variable
Hazard ratio (95 % CI)
p value
Age, years (continuous)
0.97 (0.95, 0.99)
0.025
MDR
0.35 (0.17, 0.68)
0.002
Interval to 2nd mammography, months (continuous)a
1.05 (1.02, 1.09)
0.001
Age, years (continuous)
0.98 (0.95, 1.02)
0.500
BMI, kg/m2 (continuous)
1.06 (0.98, 1.14)
0.119
Tumor size, cm (continuous)
1.29 (1.07, 1.55)
0.006
ER score (Allred score) Low (B6)
1.00 (reference)
Lymph node positivity
2.48 (1.42, 4.32)
0.001
0.91 (0.68, 1.22)
0.522
High histologic grade
1.18 (0.67, 2.07)
0.561
Low (B5)
1.00 (reference)
0.442
HER2 positivity Ki-67 C 14 %
2.01 (1.06, 3.81) 1.02 (0.61, 1.72)
0.032 0.912
High (C6
1.03 (0.83, 1.50)
High (C7) PgR score (Allred score)
Ki-67 C 14 %
1.04 (0.78, 1.38)
0.815
BMI body mass index, HER2 human epidermal growth factor receptor 2
PreMDb Total fatty breasts
1.00
Total dense breasts
6.30 (3.09, 12.84) \ 0.001
Chemotherapy administration
MDR and loco-regional recurrence-free survival
1.67 (1.21, 2.31)
0.002
CI confidence interval, ER estrogen receptor, PgR progesterone receptor a
Interval between start of tamoxifen and 2nd mammography
b
PreMD: initial mammographic density before treatment, ‘‘Total fatty breasts’’ includes almost entirely fat breasts and breasts with scattered fibroglandular tissue, and ‘‘total dense breasts’’ includes heterogeneously dense breasts and extremely dense breasts
Table 4 Distribution of recurrence patterns and association with mammographic density reduction (MDR) Recurrence pattern (n)
MDR
p value
Negative (n = 728)
Positive (n = 338)
Total recurrence (67)
57
10
0.002
Systemic recurrence (48)
39
9
0.048
Loco-regional recurrence (16)
15
1
0.027
4
0
0.172
Contralateral breast recurrence (4)
metastasis, HER2 positivity, and a high PgR score were also significantly associated with total recurrence-free survival. MDR and systemic recurrence-free survival In multivariate Cox regression analyses adjusted for conventional prognostic factors (as above), MDR positivity remained an independent predictor of systemic recurrencefree survival risk (HR = 0.48; p = 0.048; Table 5, 6; Fig. 3b). Large tumor size, lymph node metastasis, and a high PgR score were also significantly associated with total recurrence-free survival.
In multivariate Cox regression analyses adjusted for conventional prognostic factors (as above), MDR positivity remained an independent predictor of loco-regional recurrence-free survival risk (HR = 0.13; p = 0.045; Table 5, 6 Fig. 3c). HER2 positivity was associated with loco-regional recurrence-free survival. MDR and recurrence-free survival according to age (£50 vs. [50 years) Subgroup analysis using multivariate Cox regression analyses adjusted for age (continuous), tumor size, lymph node status, and HER2 positivity was performed to identify the usefulness of MDR in predicting recurrence according to menopausal status. In the premenopausal group, MDR positivity was significantly associated with recurrence-free survival (HR = 0.37; p = 0.007), and age (continuous variable), large tumor size, lymph node metastasis, and HER2 positivity were also significant risk factors for recurrence-free survival (Table 7). However, in the postmenopausal group, MDR was not a significant predictor of recurrence-free survival (p = 0.376); a high PgR score was the sole significant predictor (p = 0.018). Subtype conversion after tamoxifen use in analysis of re-biopsy sample of patient with recurrence Sixty-seven of 1,066 patients (6.3 %) experienced tumor recurrence during follow-up (48 patients: systemic recurrence; 16 patients: loco-regional recurrence; and 4 patients: contralateral breast recurrence; Table 1). Of them, rebiopsies were performed in 43 patients. There were 17 local, 6 regional, and 20 systemic recurrences. We analyzed this data to find the subtype conversion. We observed a total of 14 subtype conversion cases out of the 43 rebiopsy patients (Table 8).
123
564
Breast Cancer Res Treat (2013) 142:559–567
Table 6 Effect of mammographic density reduction on recurrence-free survival Variable
Total recurrence
Systemic recurrence
HR (95 % CI)
p value
HR (95 % CI)
Age, years (continuous)
0.98 (0.96, 1.02)
0.402
0.99 (0.95, 1.03)
BMI, kg/m2 (continuous)
1.06 (0.99, 1.15)
0.092
1.06 (0.97, 1.16)
MDR positivity
0.36 (0.18, 0.70)
0.003
0.48 (0.23, 0.99)
Loco-regional recurrence p value
HR (95 % CI)
p value
0.576
0.98 (0.92, 1.04)
0.545
0.214
1.13 (0.97, 1.31)
0.116
0.048
0.13 (0.02, 0.96)
0.045 0.776
Size, cm (continuous)
1.30 (1.09, 1.55)
0.004
1.35 (1.11, 1.64)
0.002
1.06 (0.71, 1.60)
Lymph node positivity
2.46 (1.41, 4.27)
0.001
4.30 (2.07, 8.94)
\0.001
0.64 (0.22, 1.89)
0.418
HER2 positivity
1.94 (1.03, 3.64)
0.038
1.46 (0.67, 3.20)
0.344
4.70 (1.50, 14.78)
0.008
High ER score High PgR score
1.34 (0.77, 2.33) 0.50 (0.30, 0.83)
0.303 0.007
1.22 (0.64, 2.32) 0.51 (0.28, 0.93)
0.551 0.027
1.66 (0.53, 5.16) 0.35 (0.12, 1.02)
0.380 0.055
HR hazard ratio, CI confidence interval, BMI body mass index, MDR mammographic density reduction, HER2 human epidermal growth factor receptor 2, ER estrogen receptor, PgR progesterone receptor Fig. 3 Recurrence-free survival curves according to mammographic density reduction. a Total recurrence, b systemic recurrence, and c loco-regional recurrence
Discussion We have shown that MDR provides independent predictive information on any type of recurrence in ER-positive breast
123
cancer patients treated with adjuvant tamoxifen, especially in premenopausal women. Patients who showed a reduction in breast density according to BI-RADS MD classification after an average of 19 months of adjuvant tamoxifen
Breast Cancer Res Treat (2013) 142:559–567
565
Table 7 Subgroup analysis of mammographic density reduction on recurrence-free survival according to age Variable
Age B 50 years
Age [ 50 years
N = 888
N = 178
HR (95 % CI)
p value
HR (95 % CI)
p value
Age, years (continuous)
0.95 (0.90, 0.99)
0.043
1.05 (0.98, 1.13)
0.158
MDR positivity
0.37 (0.18, 0.76)
0.007
0.41 (0.52, 3.20)
0.395
Tumor size, cm (continuous) Lymph node positivity
1.30 (1.06, 1.59) 2.44 (1.31, 4.55)
0.011 0.005
1.23 (0.88, 1.72) 3.12 (0.998, 9.84)
0.270 0.053
HER2 positivity
2.04 (1.02, 4.08)
0.043
1.08 (0.23, 5.13)
0.920
High PgR score
0.74 (0.41, 1.31)
0.302
0.21 (0.06, 0.77)
0.018
HR hazard ratio, CI confidence interval, BMI body mass index, MDR mammographic density reduction, HER2 human epidermal growth factor receptor 2, PgR progesterone receptor Table 8 Recurrence pathologic subtype conversion rate and the pattern Conversion
n (%)
Total conversion
14 (32.6 %)
ER/PgR conversion (?) ? (-)
7 (16.3 %)
Recurrence pattern (n)
Local 5 Systemic 2
HER2 conversion (?) ? (-) (-) ? (?) Simultaneous conversion
1 (2.3 %)
Local 1
8 (18.6 %)
Local 6, regional 1, systemic 1
2 (4.6 %)
Local 2
treatment had a 65 % lower risk of recurrence than patients who did not show MDR. Furthermore, the significant risk reduction according to MDR has predictive power for any type of recurrence pattern, including loco-regional recurrence (87 % reduction) and systemic recurrence (52 % reduction). In our study, only 4 patients (0.4 %) showed contralateral breast recurrence during a mean 61-month follow-up period, and none of these patients showed MDR. We speculated that patients who showed MDR also had a lower risk of contralateral breast recurrence during adjuvant tamoxifen treatment, but as the recurrence incidence was relatively low, this study might not have sufficient power to detect a significant difference. ER-positive breast cancer is biologically heterogeneous, and this could be a possible explanation for variable clinical outcomes [11]. The Early Breast Cancer Trialists’ Collaborative Group showed that more than one-third of patients experienced disease relapse, almost half of whom relapsed during the tamoxifen treatment period [2]. Many investigators have attempted to identify biomarkers to predict this heterogeneity and endocrine resistance through examining gene expression in the context of global genomic, proteomic, or functional data and to increase the depth
of understanding of endocrine-resistant disease [12]. Nevertheless, these novel efforts have not been clearly proven to predict host endocrine resistance and need more investigation before use in the clinical setting. It is well known that MD represents a general marker of breast cancer risk [5, 6, 13]. A nested case–control study showed that the change in MD over 12–18 months is an excellent predictor of response to tamoxifen in the preventive setting [9]. Furthermore, a recent retrospective study showed that MD change during short-term use of adjuvant endocrine therapy was a significant predictor of long-term recurrence in women with ER-positive breast cancer. Patients who experienced\5 % absolute MDR and those with increased MD after approximately 1 year of endocrine therapy were at 1.92- and 2.26-fold greater risk of recurrence than patients with MDR of C10 % in a study using quantitative imaging analysis software to assess MD [10]. Our findings are in agreement with previous studies in many aspects; in particular, we showed that on average MDR over 19 months could be an independent risk factor for any type of recurrence pattern. Tamoxifen is the mostly widely used adjuvant endocrine treatment for ER-positive premenopausal breast cancer patients. Subgroup analysis according to age showed that patients B50 years who experienced MDR had a 63 % lower risk of recurrence. These results suggest that premenopausal women who did not experience MDR during adjuvant tamoxifen therapy have a high risk of disease recurrence, and other strategies are needed to reduce this risk such as switching to aromatase inhibitors. However, patients[50 years old who experienced MDR did not benefit from risk reduction according to our data. Our study was designed to evaluate the efficacy of MDR in patients treated solely with adjuvant tamoxifen, so the patients enrolled had received tamoxifen for more than 2 years regardless of menopausal status. Our institution’s guidelines recommend aromatase inhibitors as the first-
123
566
choice endocrine therapy for ER-positive postmenopausal breast cancer patients, and this could be a factor in this non-significant association. To evaluate the accurate effect of MDR in patients aged [50 years, another study design which includes any type of adjuvant endocrine therapy will be needed. We investigated factors that had an effect on MD, and found that age and BMI were inversely associated. These findings are consistent with a previous study that evaluated the association of hormonal factors with breast density among Korean women, and younger age, lower BMI, lower parity, and premenopausal status were significantly associated with dense breasts [14]. Other reports showed that molecular markers including ER, PgR, and Ki-67 were not associated with dense and non-dense tissue in the normal breast; our findings also showed that ER and PgR status, Ki-67 levels, and HER2 status had no association with breast density [15, 16]. These consistencies may support the reliability of our data. In addition, we found that young age, a long interval between the start of tamoxifen and the follow-up mammogram, a high preMD grade, and chemotherapy administration were significantly associated with MDR. Kim et al. [10] also showed that age \50 years, a high preMD grade, and a long interval to the second mammogram were significantly associated with MDR. Our results were consistent with previous studies of Kim et al. Furthermore, we found that chemotherapy administration was associated with MDR, but BMI and molecular markers including ER, PgR, HER2, and Ki-67 were not associated with MDR. It is well known that adjuvant chemotherapy can induce suppression of ovarian function in premenopausal breast cancer patients [17, 18]. This suppression may be the reason for our result that chemotherapy was significantly associated with MDR. These findings suggest that adjuvant tamoxifen-induced MDR is associated with host response to endocrine resistance, not with tumor biology including ER, PgR, HER2, and Ki-67 status. This study has several limitations. First, we measured MD using a qualitative method based on BI-RADS categories. Recently, quantitative assessment of mammograms using Cumulus or other similar methods of measurement have been used in research studies. These computer-assisted methods require the placement of a dichotomous threshold between dense and non-dense tissue that is likely to exist in reality and a trained observer; thus, the measurements are subjective [19]. Although the interobserver variances remain and this is not a precise method to use for risk-predicting models, the BI-RADS system is a quick, readily available, and widely used method for assessing breast density. Second, follow-up mammograms were not taken at consistent intervals; the range was 10–34 months. To date,
123
Breast Cancer Res Treat (2013) 142:559–567
there is no information regarding the specific time point after tamoxifen treatment at which maximal MD changes are documented. Therefore, a prospective study to evaluate the optimal timing of mammographic evaluation for tamoxifen-induced MDR is needed. In this retrospective study, most patients had their follow-up mammograms done around 19 months post treatment (median follow-up period). Regardless of these limitations, our data are consistent with previous studies, lending credence to our results. Otherwise, our report has many strengths compared to the previous study. Although the previous report included all general imaging techniques, this study used a single imaging modality, mammography, to facilitate specific evaluation and accurate comparison. The regular scoring by a single radiologist using the BI-RADS system eliminated interobserver variability, and the population number is large enough to buttress the previous results. Various medications that can cause potential bias were used in the previous study, but we evaluated tamoxifen use-only group and all patients were pre-menopausal women. That could reduce confounding effects above. To our knowledge, this is the first study analyzing ER contents systematically using Allred scoring system. In Allred scoring system of ER and PgR we applied a strict criteria which define that the high score is bigger than 7 and low is smaller than 6.
Conclusion In conclusion, MDR during adjuvant tamoxifen therapy was independently associated with a decreased risk of systemic and loco-regional recurrence in ER-positive breast cancer patients, especially in women less than 50 years. For patients who do not experience MDR after approximately 1.5 years of tamoxifen therapy, more caution should be taken and other treatment strategies would be considered. Acknowledgments This work was supported by grant from the National Cancer Center Korea (1211200-1). Conflict of interest interest.
The authors indicated no potential conflicts of
References 1. Jensen EV, Jordan VC (2003) The estrogen receptor: a model for molecular medicine. Clin Cancer Res 9(6):1980–1989 2. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (2005) Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet 365(9472):1687–1717. doi:10.1016/ S0140-6736(05)66544-0
Breast Cancer Res Treat (2013) 142:559–567 3. Prowell TM, Blackford AL, Byrne C, Khouri NF, Dowsett M, Folkerd E, Tarpinian KS, Powers PP, Wright LA, Donehower MG, Jeter SC, Armstrong DK, Emens LA, Fetting JH, Wolff AC, Garrett-Mayer E, Skaar TC, Davidson NE, Stearns V et al (2011) Changes in breast density and circulating estrogens in postmenopausal women receiving adjuvant anastrozole. Cancer Prev Res (Phila) 4(12):1993–2001. doi:10.1158/1940-6207.capr-11-0154 4. Johns PC, Yaffe MJ (1987) X-ray characterisation of normal and neoplastic breast tissues. Phys Med Biol 32(6):675–695 5. Boyd NF, Martin LJ, Yaffe MJ, Minkin S (2006) Mammographic density: a hormonally responsive risk factor for breast cancer. J Br Menopause Soc 12(4):186–193. doi:10.1258/136218006779 160436 6. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, Jong RA, Hislop G, Chiarelli A, Minkin S, Yaffe MJ (2007) Mammographic density and the risk and detection of breast cancer. New Engl J Med 356(3):227–236. doi:10.1056/NEJMoa062790 7. Ursin G, Ma H, Wu AH, Bernstein L, Salane M, Parisky YR, Astrahan M, Siozon CC, Pike MC (2003) Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomarkers Prev 12(4):332–338 8. Nagata C, Matsubara T, Fujita H, Nagao Y, Shibuya C, Kashiki Y, Shimizu H (2005) Mammographic density and the risk of breast cancer in Japanese women. Br J Cancer 92(12):2102–2106. doi:10.1038/sj.bjc.6602643 9. Cuzick J, Warwick J, Pinney E, Duffy SW, Cawthorn S, Howell A, Forbes JF, Warren RM (2011) Tamoxifen-induced reduction in mammographic density and breast cancer risk reduction: a nested case–control study. J Natl Cancer Inst 103(9):744–752. doi:10.1093/jnci/djr079 10. Kim J, Han W, Moon HG, Ahn SK, Shin HC, You JM, Han SW, Im SA, Kim TY, Koo HR, Chang JM, Cho N, Moon WK, Noh DY (2012) Breast density change as a predictive surrogate for response to adjuvant endocrine therapy in hormone receptor positive breast cancer. Breast Cancer Res 14(4):R102. doi:10. 1186/bcr3221 11. Bartlett JM, Bloom KJ, Piper T, Lawton TJ, van de Velde CJ, Ross DT, Ring BZ, Seitz RS, Beck RA, Hasenburg A, Kieback D, Putter H, Markopoulos C, Dirix L, Seynaeve C, Rea D (2012)
567
12.
13.
14.
15.
16.
17.
18.
19.
Mammostrat as an immunohistochemical multigene assay for prediction of early relapse risk in the tamoxifen versus exemestane adjuvant multicenter trial pathology study. J Clin Oncol. doi:10.1200/jco.2012.42.8896 Musgrove EA, Sutherland RL (2009) Biological determinants of endocrine resistance in breast cancer. Nat Rev Cancer 9(9):631–643. doi:10.1038/nrc2713 Vachon CM, Brandt KR, Ghosh K, Scott CG, Maloney SD, Carston MJ, Pankratz VS, Sellers TA (2007) Mammographic breast density as a general marker of breast cancer risk. Cancer Epidemiol Biomarkers Prev 16(1):43–49. doi:10.1158/10559965.epi-06-0738 Jeon JH, Kang JH, Kim Y, Lee HY, Choi KS, Jun JK, Oh DK, Lee CY, Ko K, Park EC (2011) Reproductive and hormonal factors associated with fatty or dense breast patterns among Korean women. Cancer Res Treat 43(1):42–48. doi:10.4143/crt. 2011.43.1.42 Verheus M, Maskarinec G, Erber E, Steude JS, Killeen J, Hernandez BY, Cline JM (2009) Mammographic density and epithelial histopathologic markers. BMC Cancer 9:182. doi:10.1186/ 1471-2407-9-182 Ghosh K, Brandt KR, Reynolds C, Scott CG, Pankratz VS, Riehle DL, Lingle WL, Odogwu T, Radisky DC, Visscher DW, Ingle JN, Hartmann LC, Vachon CM (2012) Tissue composition of mammographically dense and non-dense breast tissue. Breast Cancer Res Treat 131(1):267–275. doi:10.1007/s10549-011-1727-4 Bines J, Oleske DM, Cobleigh MA (1996) Ovarian function in premenopausal women treated with adjuvant chemotherapy for breast cancer. J Clin Oncol 14(5):1718–1729 Fornier MN, Modi S, Panageas KS, Norton L, Hudis C (2005) Incidence of chemotherapy-induced, long-term amenorrhea in patients with breast carcinoma age 40 years and younger after adjuvant anthracycline and taxane. Cancer 104(8):1575–1579. doi:10.1002/cncr.21385 Boyd NF, Martin LJ, Bronskill M, Yaffe MJ, Duric N, Minkin S (2010) Breast tissue composition and susceptibility to breast cancer. J Natl Cancer Inst 102(16):1224–1237. doi:10.1093/jnci/ djq239
123
CLINICAL TRIAL
Adjuvant tamoxifen-induced mammographic breast density reduction as a predictor for recurrence in estrogen receptor-positive premenopausal breast cancer patients Kyung Lan Ko • In Suk Shin • Ji Young You So-Youn Jung • Jungsil Ro • Eun Sook Lee
•
Received: 30 May 2013 / Accepted: 3 October 2013 / Published online: 14 November 2013 Ó Springer Science+Business Media New York 2013
Abstract Tamoxifen is known to reduce the risk of breast cancer in women at high risk and also reduces mammographic breast density (MD) in a preventive setting. We investigated the efficacy of MD reduction (MDR) for predicting recurrence in estrogen receptor (ER)-positive patients in an adjuvant setting. A total of 1,066 ER-positive breast cancer patients who were enrolled in this study underwent curative surgery and received adjuvant tamoxifen for at least 2 years at our institution between January 2003 and December 2008. Using a computerized system, a single radiologist reviewed all mammograms and classified MD patterns on the basis of the Breast Imaging Reporting and Data System. MDR was assessed using the baseline mammogram taken before surgery (preMD) and the followup mammogram taken after the start of adjuvant tamoxifen (postMD). MDR positivity was defined as downgrading of the postMD grade, with the preMD grade as a reference. Patients were divided into 2 groups, MDR-positive and MDR-negative, for statistical analysis. Patients who showed MDR after an average of 19 months of adjuvant tamoxifen treatment had a 65 % lower risk of recurrence than patients who did not show MDR. Furthermore, significant risk reduction according to MDR had a predictive power for any type of recurrence pattern including loco-regional recurrence (87 % reduction) and systemic recurrence (52 % reduction) in ER-positive breast cancer patients, especially in women B50 years. In our study, only 4 patients (0.4 %) Kyoung Lan Ko and In Suk Shin contributed equally to this paper. K. L. Ko I. S. Shin J. Y. You S.-Y. Jung J. Ro E. S. Lee (&) Center for Breast Cancer, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do 410-769, Republic of Korea e-mail: [email protected]
showed contralateral breast recurrence during the mean 61-month follow-up period and none of them experienced MDR. In conclusion, MDR during adjuvant tamoxifen therapy was independently associated with a lower risk of systemic and loco-regional recurrence in ER-positive breast cancer patients, especially in young women. For patients who do not experience MDR after approximately 1.5 years of tamoxifen therapy, more caution should be taken and other treatment strategies are warranted. Keywords Mammographic breast density Estrogen receptor positive Adjuvant tamoxifen therapy Breast cancer
Introduction As at least 70 % of breast cancers are estrogen receptor (ER)-positive, targeting estrogen action has been a mainstay of breast cancer treatment. The selective ER modulator tamoxifen, synthesized during the 1970s, was the most competitive inhibitor of estrogen–ER binding and became the first and most successful targeted cancer therapy [1]. The Early Breast Cancer Trialists’ Collaborative Group showed that adjuvant tamoxifen therapy almost halved the rate of annual recurrence and reduced mortality by 31 % [2]. Unfortunately, some patients with resistance to endocrine therapy especially tamoxifen have recurrent disease during or after this treatment. Many investigators are attempting to develop more specific biomarkers to predict endocrine resistance, because currently the ability to predict resistance is suboptimal. It has been hypothesized that breast density, which is known to correlate with breast cancer risk in population studies, may predict response to hormonal manipulations [3].
123
560
Breast Cancer Res Treat (2013) 142:559–567
Fig. 1 Schematic diagram showing the flow of patients through the study
Breast density is most commonly measured by mammography. Mammographic breast density (MD) varies among women because of variations in breast tissue composition and differences in the radiographic attenuation properties of stroma, epithelium, and fat [4]. All of these factors are influenced by endogenous estrogen concentration [5]. Western women in the highest MD category are thought to have a fourfold to sixfold increased risk of breast cancer [6]. For Asian women, breast cancer risk is thought to be increased by twofold to fivefold in women with a high breast density [7, 8]. Tamoxifen is known to reduce the risk of breast cancer in women at high risk and also reduces MD. A case–control study found that the change in MD over 12–18 months is a predictor of response to tamoxifen in a preventive setting [9]. The efficacy of MD reduction (MDR) in predicting response to tamoxifen in an adjuvant setting is not well known. A recent report showed that MDR, assessed by imaging analysis software, during adjuvant endocrine therapy is a significant predictor of recurrence of ERpositive breast cancer [10]. This is an important step toward establishing effective strategies for patients who will benefit less from endocrine therapy. We also investigated the efficacy of MDR in predicting breast cancer recurrence in the adjuvant setting using qualitative methods based on the widely used Breast Imaging Reporting and Data System (BI-RADS; American College of Radiology, Reston, VA, USA).
Methods Patients Our study is a retrospective one based on the existing prospective breast cancer database of the National Cancer Center, Goyang, Korea. A total of 2,402 ER-positive breast cancer patients who were enrolled in this study underwent curative surgery at our institution between January 2003 and December 2008. We investigated 1,526 patients who
123
received adjuvant tamoxifen for at least 2 years. Patients were excluded if their digital mammograms were not available or not appropriate for evaluation or if they had bilateral breast cancer or occult breast cancer. Finally, we collected the clinicopathologic information on 1,066 patients by reviewing the prospective database of our institution and the data of disease recurrence by additional medical record review (Fig. 1). Written informed consent was obtained prior to surgery from all patients, and this study was approved by the Institutional Review Board of the National Cancer Center, Goyang, Korea. Mammographic breast density assessment Using a computerized system, a single radiologist (K. Ko: 10 years of experience in interpreting mammograms) reviewed 2,132 preoperative and postoperative mammograms classified breast density patterns according to BIRADS. Four density grades were classified: (1) almost entirely fat (\25 % glandular), (2) scattered fibroglandular densities (25–50 % glandular), (3) heterogeneously dense (50–75 % glandular), and (4) extremely dense ([75 % glandular). MDR was determined according to a pair of digital mammograms: the baseline mammogram taken before surgery (preMD) and the follow-up mammogram taken 10–34 months (median, 19 months) after the start of adjuvant tamoxifen (postMD). MDR positivity was defined as the downgrading of the postMD grade, with the preMD grade as a reference. We relied on a single radiologist who is a specialist in breast imaging studies, thereby eliminating interobserver variability. We did not seek to measure reproducibility as the BI-RADS density classifications are standardized. Statistical analysis Patients were divided into 2 groups, MDR-positive and MDR-negative, to perform statistical analysis. To identify the association of factors with MDR, the t test and Chi square test were used. Body mass index (BMI) was
Breast Cancer Res Treat (2013) 142:559–567
561
Table 1 Clinicopathologic characteristics of patients
Table 1 continued
Variable number (%)
Total
MDR positive
MDR negative
Variable number (%)
Mean age, years ± SD
45.3 ± 7.6
44 ± 5.94
46 ± 8.11
Recurrence
Range
25–78
28–68
25–78
B50 [50
888 (83.3) 178 (16.7)
308 (91.1) 30 (8.9)
580 (79.7) 148 (20.3)
59 ± 17.6
59 ± 17.7
59 ± 17.5
Mean follow-up time, months ± SD Range
26–114
27–113
26–114
23.4 ± 3.2
22.89 ± 3.1
23.07 ± 3.2
15.6–50.2
17.5–35.5
15.6–50.2
1.7 ± 1.2
1.6 ± 1.2
1.7 ± 1.1
0.1–9.0
0.1–6
0.1–9
Negative
666 (61.5)
195 (57.7)
461 (63.3)
Positive
410 (38.5)
143 (42.3)
267 (36.7)
Mean BMI, kg/ m2 ± SD Range Mean tumor size, cm ± SD Range Lymph node status
Histologic grade 1/2
840 (78.8)
268 (79.3)
572 (78.6)
3
226 (21.2)
70 (20.7)
156 (21.4)
ER score (Allred score) Low (B6) 336 (31.6) High (C7)
729 (68.4)
117 (34.6)
220 (30.2)
221 (65.4)
508 (69.8)
PgR score (Allred score) Low (B5)
370 (34.7)
113 (33.4)
257 (35.3)
High (C6)
696 (65.3)
225 (66.6)
471 (64.7)
HER2 Negative
972 (91.2)
302 (89.3)
670 (92.0)
Positive
94 (8.83)
36 (10.7)
58 (8.0)
Ki-67 (%) \14
669 (62.8)
202 (59.8)
467 (64.1)
C14
397 (37.2)
136 (40.2)
261 (35.9)
Total
MDR positive
MDR negative
Total
67
10 (3.0)
57 (7.8)
Systemic
48
9 (2.7)
39 (5.4)
Loco-regional
17
1 (0.3)
15 (2.1)
Contralateral breast
4
0 (0)
4 (0.5)
BMI body mass index, ER estrogen receptor, PgR progesterone receptor, HER2 human epidermal growth factor receptor 2
calculated as weight/height2 (kg/m2) and both ER and progesterone receptor (PgR) expression levels were divided into binary covariates for an Allred score (low ER, B6; high ER, C7; low PgR, B5; high PgR, C7). Breasts were classified into 3 groups according to preMD and postMD grades: total fatty breasts (both ‘‘almost entirely fat’’ and ‘‘scattered fibroglandular tissue’’), heterogeneously dense breasts, and extremely dense breasts. If the preMD score was 1, it could not be 0 after treatment. We defined the total fatty breast group as including grades 1 and 2. Multivariate logistic regression analysis was performed to evaluate the factors related to MDR. Multivariate Cox regression analysis was used to compare recurrence-free survival according to MDR and to calculate survival curves. The association of MDR with disease-free survival according to patterns of recurrent disease (loco-regional, systemic, and contralateral recurrence) was analyzed. All statistical analyses were performed using Stata version 10.0 (Stata Corporation, College Station, Texas, USA) software, and factors with a p value of \0.05 were considered statistically significant.
Results
Lymphovascular invasion Negative
320 (30)
88 (26.0)
232 (31.9)
Positive
746 (70)
250 (74.0)
496 (68.1)
Patients and MD distribution The mean age of the 1,066 enrolled patients was 45.3 years (range 25–78 years; Table 1). One hundred and thirty-four patients (12.6 %) had ductal carcinoma in situ, and 175 patients (16.4 %) had received neoadjuvant chemotherapy. A total of 67 recurrences were recorded during the current period of available follow-up (mean 61.3 months). The distribution of patients according to BI-RADS MD category before and after tamoxifen therapy is shown in Table 2. Extremely dense breasts were noted in 422 patients (39.6 %) before treatment and in 273 patients (25.6 %) after an average of 19 months (10–36 months) of adjuvant tamoxifen. We analyzed multiple factors for
Operation Mastectomy
162 (15.2)
46 (13.6)
116 (15.9)
904 (84.8)
292 (86.4)
612 (84.1)
No
303 (28.4)
68 (20.1)
235 (32.3)
Yes (adjuvant/ neoadjuvant)
763 (71.6)(588(55.2)/ 175(16.4))
270 (79.9)
493 (67.7)
No
173
50 (14.8)
123 (16.9)
Yes
893
288 (85.2)
605 (83.1)
Breastconserving surgery Chemotherapy
Radiotherapy
123
562
Breast Cancer Res Treat (2013) 142:559–567
Table 2 Distribution of patients according to BI-RADS breast density category before and after treatment PostMDb \50 %
[75 %
50–75 %
Total
PreMDa \50 %
141
0
0
141
50–75 % [75 %
180 5
323 144
0 273
503 422
Total
326
467
273
1,066
BI-RADS: The American College of Radiology Breast Imaging Reporting and Data System, \50 %: grade 1 (almost entirely fat) and grade 2 (scattered fibroglandular tissue), 50–75 %: grade 3 (heterogeneously dense), [75 %: grade 4 (extremely dense) a
PreMD: initial mammographic density before treatment
b
PostMD: tamoxifen
follow-up
mammographic
density
after
adjuvant
association with preMD; we found that age and BMI were inversely associated with preMD (p \ 0.001; Fig. 2). MDR was observed in 338 patients (31.7 %); MD was downgraded by 2 grades in only 5 patients (0.6 %), whereas MDR was downgraded by 1 grade in the others (99.4 %). Factors associated with MD reduction Multiple variables were analyzed to identify associations with MDR. MDR-positive patients were significantly younger than MDR-negative patients (43.6 ± 5.9 vs. 46.2 ± 8.1 years; p \ 0.001), but BMI did not differ significantly between the 2 groups (23.4 ± 3.1 vs. 23.4 ± 3.3 kg/m2; p = 0.939). Time from the start of adjuvant tamoxifen to the follow-up mammogram was significantly longer in MDR-positive patients than in MDR-negative patients (19.7 ± 4.3 vs. 18.6 ± 4.2 months; p \ 0.001). A high preMD grade and chemotherapy administration were significantly associated with MDR positivity (p \ 0.001). To evaluate the independent effect of each factor associated with MDR positivity, we used multivariate logistic regression analysis (Table 3). We found that age as a continuous variable was inversely associated with MDR positivity [odds ratio (OR) = 0.97; p = 0.025). In addition, a long interval from the start of adjuvant tamoxifen to the follow-up mammogram (OR = 1.05; p = 0.001), a high preMD grade (OR = 6.30; p \ 0.001), and chemotherapy administration (OR = 1.67; p = 0.002) were significantly associated with MDR (Table 3). However, the ER score, PgR score, and Ki-67 level had no association with MDR. MD reduction and recurrence-free survival Sixty-seven of 1,066 patients (6.3 %) experienced tumor recurrence during follow-up (48 patients: systemic
123
Fig. 2 Association of preMD with age and BMI, a preMD and BMI, b preMD and age. BMI body mass index, preMD pretreatment mammographic breast density, preMD 1 almost entirely fatty breasts, preMD 2 breasts with scattered fibroglandular tissue, preMD 3 heterogeneously dense breasts, and preMD 4 extremely dense breasts. The dotted lines represent linear prediction lines. Linear regression analysis
recurrence; 16 patients: loco-regional recurrence; and 4 patients: contralateral breast recurrence; Table 1). The association of disease recurrence with MDR according to recurrence pattern is shown in Table 4. MDR was significantly associated with loco-regional recurrence, systemic recurrence, and total recurrence (p \ 0.05). MDR and total recurrence-free survival In multivariate Cox regression analyses adjusted for conventional prognostic factors [age, BMI, tumor size, lymph node status, ER score, PgR score, and human epidermal growth factor receptor 2 (HER2) status], MDR positivity remained an independent predictor of total recurrence-free survival risk [hazard ratio (HR) = 0.36; p = 0.003; Table 5, 6; Fig. 3a]. Large tumor size, lymph node
Breast Cancer Res Treat (2013) 142:559–567
563
Table 3 Factors related to mammographic density reduction Variable
Odds ratio (95 % CI)
p value
Table 5 Effect of mammographic density reduction (MDR) on recurrence-free survival Variable
Hazard ratio (95 % CI)
p value
Age, years (continuous)
0.97 (0.95, 0.99)
0.025
MDR
0.35 (0.17, 0.68)
0.002
Interval to 2nd mammography, months (continuous)a
1.05 (1.02, 1.09)
0.001
Age, years (continuous)
0.98 (0.95, 1.02)
0.500
BMI, kg/m2 (continuous)
1.06 (0.98, 1.14)
0.119
Tumor size, cm (continuous)
1.29 (1.07, 1.55)
0.006
ER score (Allred score) Low (B6)
1.00 (reference)
Lymph node positivity
2.48 (1.42, 4.32)
0.001
0.91 (0.68, 1.22)
0.522
High histologic grade
1.18 (0.67, 2.07)
0.561
Low (B5)
1.00 (reference)
0.442
HER2 positivity Ki-67 C 14 %
2.01 (1.06, 3.81) 1.02 (0.61, 1.72)
0.032 0.912
High (C6
1.03 (0.83, 1.50)
High (C7) PgR score (Allred score)
Ki-67 C 14 %
1.04 (0.78, 1.38)
0.815
BMI body mass index, HER2 human epidermal growth factor receptor 2
PreMDb Total fatty breasts
1.00
Total dense breasts
6.30 (3.09, 12.84) \ 0.001
Chemotherapy administration
MDR and loco-regional recurrence-free survival
1.67 (1.21, 2.31)
0.002
CI confidence interval, ER estrogen receptor, PgR progesterone receptor a
Interval between start of tamoxifen and 2nd mammography
b
PreMD: initial mammographic density before treatment, ‘‘Total fatty breasts’’ includes almost entirely fat breasts and breasts with scattered fibroglandular tissue, and ‘‘total dense breasts’’ includes heterogeneously dense breasts and extremely dense breasts
Table 4 Distribution of recurrence patterns and association with mammographic density reduction (MDR) Recurrence pattern (n)
MDR
p value
Negative (n = 728)
Positive (n = 338)
Total recurrence (67)
57
10
0.002
Systemic recurrence (48)
39
9
0.048
Loco-regional recurrence (16)
15
1
0.027
4
0
0.172
Contralateral breast recurrence (4)
metastasis, HER2 positivity, and a high PgR score were also significantly associated with total recurrence-free survival. MDR and systemic recurrence-free survival In multivariate Cox regression analyses adjusted for conventional prognostic factors (as above), MDR positivity remained an independent predictor of systemic recurrencefree survival risk (HR = 0.48; p = 0.048; Table 5, 6; Fig. 3b). Large tumor size, lymph node metastasis, and a high PgR score were also significantly associated with total recurrence-free survival.
In multivariate Cox regression analyses adjusted for conventional prognostic factors (as above), MDR positivity remained an independent predictor of loco-regional recurrence-free survival risk (HR = 0.13; p = 0.045; Table 5, 6 Fig. 3c). HER2 positivity was associated with loco-regional recurrence-free survival. MDR and recurrence-free survival according to age (£50 vs. [50 years) Subgroup analysis using multivariate Cox regression analyses adjusted for age (continuous), tumor size, lymph node status, and HER2 positivity was performed to identify the usefulness of MDR in predicting recurrence according to menopausal status. In the premenopausal group, MDR positivity was significantly associated with recurrence-free survival (HR = 0.37; p = 0.007), and age (continuous variable), large tumor size, lymph node metastasis, and HER2 positivity were also significant risk factors for recurrence-free survival (Table 7). However, in the postmenopausal group, MDR was not a significant predictor of recurrence-free survival (p = 0.376); a high PgR score was the sole significant predictor (p = 0.018). Subtype conversion after tamoxifen use in analysis of re-biopsy sample of patient with recurrence Sixty-seven of 1,066 patients (6.3 %) experienced tumor recurrence during follow-up (48 patients: systemic recurrence; 16 patients: loco-regional recurrence; and 4 patients: contralateral breast recurrence; Table 1). Of them, rebiopsies were performed in 43 patients. There were 17 local, 6 regional, and 20 systemic recurrences. We analyzed this data to find the subtype conversion. We observed a total of 14 subtype conversion cases out of the 43 rebiopsy patients (Table 8).
123
564
Breast Cancer Res Treat (2013) 142:559–567
Table 6 Effect of mammographic density reduction on recurrence-free survival Variable
Total recurrence
Systemic recurrence
HR (95 % CI)
p value
HR (95 % CI)
Age, years (continuous)
0.98 (0.96, 1.02)
0.402
0.99 (0.95, 1.03)
BMI, kg/m2 (continuous)
1.06 (0.99, 1.15)
0.092
1.06 (0.97, 1.16)
MDR positivity
0.36 (0.18, 0.70)
0.003
0.48 (0.23, 0.99)
Loco-regional recurrence p value
HR (95 % CI)
p value
0.576
0.98 (0.92, 1.04)
0.545
0.214
1.13 (0.97, 1.31)
0.116
0.048
0.13 (0.02, 0.96)
0.045 0.776
Size, cm (continuous)
1.30 (1.09, 1.55)
0.004
1.35 (1.11, 1.64)
0.002
1.06 (0.71, 1.60)
Lymph node positivity
2.46 (1.41, 4.27)
0.001
4.30 (2.07, 8.94)
\0.001
0.64 (0.22, 1.89)
0.418
HER2 positivity
1.94 (1.03, 3.64)
0.038
1.46 (0.67, 3.20)
0.344
4.70 (1.50, 14.78)
0.008
High ER score High PgR score
1.34 (0.77, 2.33) 0.50 (0.30, 0.83)
0.303 0.007
1.22 (0.64, 2.32) 0.51 (0.28, 0.93)
0.551 0.027
1.66 (0.53, 5.16) 0.35 (0.12, 1.02)
0.380 0.055
HR hazard ratio, CI confidence interval, BMI body mass index, MDR mammographic density reduction, HER2 human epidermal growth factor receptor 2, ER estrogen receptor, PgR progesterone receptor Fig. 3 Recurrence-free survival curves according to mammographic density reduction. a Total recurrence, b systemic recurrence, and c loco-regional recurrence
Discussion We have shown that MDR provides independent predictive information on any type of recurrence in ER-positive breast
123
cancer patients treated with adjuvant tamoxifen, especially in premenopausal women. Patients who showed a reduction in breast density according to BI-RADS MD classification after an average of 19 months of adjuvant tamoxifen
Breast Cancer Res Treat (2013) 142:559–567
565
Table 7 Subgroup analysis of mammographic density reduction on recurrence-free survival according to age Variable
Age B 50 years
Age [ 50 years
N = 888
N = 178
HR (95 % CI)
p value
HR (95 % CI)
p value
Age, years (continuous)
0.95 (0.90, 0.99)
0.043
1.05 (0.98, 1.13)
0.158
MDR positivity
0.37 (0.18, 0.76)
0.007
0.41 (0.52, 3.20)
0.395
Tumor size, cm (continuous) Lymph node positivity
1.30 (1.06, 1.59) 2.44 (1.31, 4.55)
0.011 0.005
1.23 (0.88, 1.72) 3.12 (0.998, 9.84)
0.270 0.053
HER2 positivity
2.04 (1.02, 4.08)
0.043
1.08 (0.23, 5.13)
0.920
High PgR score
0.74 (0.41, 1.31)
0.302
0.21 (0.06, 0.77)
0.018
HR hazard ratio, CI confidence interval, BMI body mass index, MDR mammographic density reduction, HER2 human epidermal growth factor receptor 2, PgR progesterone receptor Table 8 Recurrence pathologic subtype conversion rate and the pattern Conversion
n (%)
Total conversion
14 (32.6 %)
ER/PgR conversion (?) ? (-)
7 (16.3 %)
Recurrence pattern (n)
Local 5 Systemic 2
HER2 conversion (?) ? (-) (-) ? (?) Simultaneous conversion
1 (2.3 %)
Local 1
8 (18.6 %)
Local 6, regional 1, systemic 1
2 (4.6 %)
Local 2
treatment had a 65 % lower risk of recurrence than patients who did not show MDR. Furthermore, the significant risk reduction according to MDR has predictive power for any type of recurrence pattern, including loco-regional recurrence (87 % reduction) and systemic recurrence (52 % reduction). In our study, only 4 patients (0.4 %) showed contralateral breast recurrence during a mean 61-month follow-up period, and none of these patients showed MDR. We speculated that patients who showed MDR also had a lower risk of contralateral breast recurrence during adjuvant tamoxifen treatment, but as the recurrence incidence was relatively low, this study might not have sufficient power to detect a significant difference. ER-positive breast cancer is biologically heterogeneous, and this could be a possible explanation for variable clinical outcomes [11]. The Early Breast Cancer Trialists’ Collaborative Group showed that more than one-third of patients experienced disease relapse, almost half of whom relapsed during the tamoxifen treatment period [2]. Many investigators have attempted to identify biomarkers to predict this heterogeneity and endocrine resistance through examining gene expression in the context of global genomic, proteomic, or functional data and to increase the depth
of understanding of endocrine-resistant disease [12]. Nevertheless, these novel efforts have not been clearly proven to predict host endocrine resistance and need more investigation before use in the clinical setting. It is well known that MD represents a general marker of breast cancer risk [5, 6, 13]. A nested case–control study showed that the change in MD over 12–18 months is an excellent predictor of response to tamoxifen in the preventive setting [9]. Furthermore, a recent retrospective study showed that MD change during short-term use of adjuvant endocrine therapy was a significant predictor of long-term recurrence in women with ER-positive breast cancer. Patients who experienced\5 % absolute MDR and those with increased MD after approximately 1 year of endocrine therapy were at 1.92- and 2.26-fold greater risk of recurrence than patients with MDR of C10 % in a study using quantitative imaging analysis software to assess MD [10]. Our findings are in agreement with previous studies in many aspects; in particular, we showed that on average MDR over 19 months could be an independent risk factor for any type of recurrence pattern. Tamoxifen is the mostly widely used adjuvant endocrine treatment for ER-positive premenopausal breast cancer patients. Subgroup analysis according to age showed that patients B50 years who experienced MDR had a 63 % lower risk of recurrence. These results suggest that premenopausal women who did not experience MDR during adjuvant tamoxifen therapy have a high risk of disease recurrence, and other strategies are needed to reduce this risk such as switching to aromatase inhibitors. However, patients[50 years old who experienced MDR did not benefit from risk reduction according to our data. Our study was designed to evaluate the efficacy of MDR in patients treated solely with adjuvant tamoxifen, so the patients enrolled had received tamoxifen for more than 2 years regardless of menopausal status. Our institution’s guidelines recommend aromatase inhibitors as the first-
123
566
choice endocrine therapy for ER-positive postmenopausal breast cancer patients, and this could be a factor in this non-significant association. To evaluate the accurate effect of MDR in patients aged [50 years, another study design which includes any type of adjuvant endocrine therapy will be needed. We investigated factors that had an effect on MD, and found that age and BMI were inversely associated. These findings are consistent with a previous study that evaluated the association of hormonal factors with breast density among Korean women, and younger age, lower BMI, lower parity, and premenopausal status were significantly associated with dense breasts [14]. Other reports showed that molecular markers including ER, PgR, and Ki-67 were not associated with dense and non-dense tissue in the normal breast; our findings also showed that ER and PgR status, Ki-67 levels, and HER2 status had no association with breast density [15, 16]. These consistencies may support the reliability of our data. In addition, we found that young age, a long interval between the start of tamoxifen and the follow-up mammogram, a high preMD grade, and chemotherapy administration were significantly associated with MDR. Kim et al. [10] also showed that age \50 years, a high preMD grade, and a long interval to the second mammogram were significantly associated with MDR. Our results were consistent with previous studies of Kim et al. Furthermore, we found that chemotherapy administration was associated with MDR, but BMI and molecular markers including ER, PgR, HER2, and Ki-67 were not associated with MDR. It is well known that adjuvant chemotherapy can induce suppression of ovarian function in premenopausal breast cancer patients [17, 18]. This suppression may be the reason for our result that chemotherapy was significantly associated with MDR. These findings suggest that adjuvant tamoxifen-induced MDR is associated with host response to endocrine resistance, not with tumor biology including ER, PgR, HER2, and Ki-67 status. This study has several limitations. First, we measured MD using a qualitative method based on BI-RADS categories. Recently, quantitative assessment of mammograms using Cumulus or other similar methods of measurement have been used in research studies. These computer-assisted methods require the placement of a dichotomous threshold between dense and non-dense tissue that is likely to exist in reality and a trained observer; thus, the measurements are subjective [19]. Although the interobserver variances remain and this is not a precise method to use for risk-predicting models, the BI-RADS system is a quick, readily available, and widely used method for assessing breast density. Second, follow-up mammograms were not taken at consistent intervals; the range was 10–34 months. To date,
123
Breast Cancer Res Treat (2013) 142:559–567
there is no information regarding the specific time point after tamoxifen treatment at which maximal MD changes are documented. Therefore, a prospective study to evaluate the optimal timing of mammographic evaluation for tamoxifen-induced MDR is needed. In this retrospective study, most patients had their follow-up mammograms done around 19 months post treatment (median follow-up period). Regardless of these limitations, our data are consistent with previous studies, lending credence to our results. Otherwise, our report has many strengths compared to the previous study. Although the previous report included all general imaging techniques, this study used a single imaging modality, mammography, to facilitate specific evaluation and accurate comparison. The regular scoring by a single radiologist using the BI-RADS system eliminated interobserver variability, and the population number is large enough to buttress the previous results. Various medications that can cause potential bias were used in the previous study, but we evaluated tamoxifen use-only group and all patients were pre-menopausal women. That could reduce confounding effects above. To our knowledge, this is the first study analyzing ER contents systematically using Allred scoring system. In Allred scoring system of ER and PgR we applied a strict criteria which define that the high score is bigger than 7 and low is smaller than 6.
Conclusion In conclusion, MDR during adjuvant tamoxifen therapy was independently associated with a decreased risk of systemic and loco-regional recurrence in ER-positive breast cancer patients, especially in women less than 50 years. For patients who do not experience MDR after approximately 1.5 years of tamoxifen therapy, more caution should be taken and other treatment strategies would be considered. Acknowledgments This work was supported by grant from the National Cancer Center Korea (1211200-1). Conflict of interest interest.
The authors indicated no potential conflicts of
References 1. Jensen EV, Jordan VC (2003) The estrogen receptor: a model for molecular medicine. Clin Cancer Res 9(6):1980–1989 2. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (2005) Effects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet 365(9472):1687–1717. doi:10.1016/ S0140-6736(05)66544-0
Breast Cancer Res Treat (2013) 142:559–567 3. Prowell TM, Blackford AL, Byrne C, Khouri NF, Dowsett M, Folkerd E, Tarpinian KS, Powers PP, Wright LA, Donehower MG, Jeter SC, Armstrong DK, Emens LA, Fetting JH, Wolff AC, Garrett-Mayer E, Skaar TC, Davidson NE, Stearns V et al (2011) Changes in breast density and circulating estrogens in postmenopausal women receiving adjuvant anastrozole. Cancer Prev Res (Phila) 4(12):1993–2001. doi:10.1158/1940-6207.capr-11-0154 4. Johns PC, Yaffe MJ (1987) X-ray characterisation of normal and neoplastic breast tissues. Phys Med Biol 32(6):675–695 5. Boyd NF, Martin LJ, Yaffe MJ, Minkin S (2006) Mammographic density: a hormonally responsive risk factor for breast cancer. J Br Menopause Soc 12(4):186–193. doi:10.1258/136218006779 160436 6. Boyd NF, Guo H, Martin LJ, Sun L, Stone J, Fishell E, Jong RA, Hislop G, Chiarelli A, Minkin S, Yaffe MJ (2007) Mammographic density and the risk and detection of breast cancer. New Engl J Med 356(3):227–236. doi:10.1056/NEJMoa062790 7. Ursin G, Ma H, Wu AH, Bernstein L, Salane M, Parisky YR, Astrahan M, Siozon CC, Pike MC (2003) Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomarkers Prev 12(4):332–338 8. Nagata C, Matsubara T, Fujita H, Nagao Y, Shibuya C, Kashiki Y, Shimizu H (2005) Mammographic density and the risk of breast cancer in Japanese women. Br J Cancer 92(12):2102–2106. doi:10.1038/sj.bjc.6602643 9. Cuzick J, Warwick J, Pinney E, Duffy SW, Cawthorn S, Howell A, Forbes JF, Warren RM (2011) Tamoxifen-induced reduction in mammographic density and breast cancer risk reduction: a nested case–control study. J Natl Cancer Inst 103(9):744–752. doi:10.1093/jnci/djr079 10. Kim J, Han W, Moon HG, Ahn SK, Shin HC, You JM, Han SW, Im SA, Kim TY, Koo HR, Chang JM, Cho N, Moon WK, Noh DY (2012) Breast density change as a predictive surrogate for response to adjuvant endocrine therapy in hormone receptor positive breast cancer. Breast Cancer Res 14(4):R102. doi:10. 1186/bcr3221 11. Bartlett JM, Bloom KJ, Piper T, Lawton TJ, van de Velde CJ, Ross DT, Ring BZ, Seitz RS, Beck RA, Hasenburg A, Kieback D, Putter H, Markopoulos C, Dirix L, Seynaeve C, Rea D (2012)
567
12.
13.
14.
15.
16.
17.
18.
19.
Mammostrat as an immunohistochemical multigene assay for prediction of early relapse risk in the tamoxifen versus exemestane adjuvant multicenter trial pathology study. J Clin Oncol. doi:10.1200/jco.2012.42.8896 Musgrove EA, Sutherland RL (2009) Biological determinants of endocrine resistance in breast cancer. Nat Rev Cancer 9(9):631–643. doi:10.1038/nrc2713 Vachon CM, Brandt KR, Ghosh K, Scott CG, Maloney SD, Carston MJ, Pankratz VS, Sellers TA (2007) Mammographic breast density as a general marker of breast cancer risk. Cancer Epidemiol Biomarkers Prev 16(1):43–49. doi:10.1158/10559965.epi-06-0738 Jeon JH, Kang JH, Kim Y, Lee HY, Choi KS, Jun JK, Oh DK, Lee CY, Ko K, Park EC (2011) Reproductive and hormonal factors associated with fatty or dense breast patterns among Korean women. Cancer Res Treat 43(1):42–48. doi:10.4143/crt. 2011.43.1.42 Verheus M, Maskarinec G, Erber E, Steude JS, Killeen J, Hernandez BY, Cline JM (2009) Mammographic density and epithelial histopathologic markers. BMC Cancer 9:182. doi:10.1186/ 1471-2407-9-182 Ghosh K, Brandt KR, Reynolds C, Scott CG, Pankratz VS, Riehle DL, Lingle WL, Odogwu T, Radisky DC, Visscher DW, Ingle JN, Hartmann LC, Vachon CM (2012) Tissue composition of mammographically dense and non-dense breast tissue. Breast Cancer Res Treat 131(1):267–275. doi:10.1007/s10549-011-1727-4 Bines J, Oleske DM, Cobleigh MA (1996) Ovarian function in premenopausal women treated with adjuvant chemotherapy for breast cancer. J Clin Oncol 14(5):1718–1729 Fornier MN, Modi S, Panageas KS, Norton L, Hudis C (2005) Incidence of chemotherapy-induced, long-term amenorrhea in patients with breast carcinoma age 40 years and younger after adjuvant anthracycline and taxane. Cancer 104(8):1575–1579. doi:10.1002/cncr.21385 Boyd NF, Martin LJ, Bronskill M, Yaffe MJ, Duric N, Minkin S (2010) Breast tissue composition and susceptibility to breast cancer. J Natl Cancer Inst 102(16):1224–1237. doi:10.1093/jnci/ djq239
123
