Ann Surg Oncol DOI 10.1245/s10434-014-3832-1
ORIGINAL ARTICLE – BREAST ONCOLOGY
High Volumetric Breast Density Predicts Risk for Breast Cancer in Postmenopausal, but not Premenopausal, Korean Women In Hae Park, MD1, Kyungran Ko, MD1, Jungnam Joo, PhD2, Boram Park, MS2, So-Youn Jung, MD1, Seeyoun Lee, MD, PhD1, Youngmi Kwon, MD, PhD1, Han-Sung Kang, MD, PhD1, Eun Sook Lee, MD, PhD1, Keun Seok Lee, MD, PhD1, and Jungsil Ro, MD, PhD1 Center For Breast Cancer, National Cancer Center, Goyang-si, Gyeonggi-do, Korea; 2Biometric Research Branch, National Cancer Center, Goyang, Korea
1
ABSTRACT Purpose. We investigated the association between mammographic breast density and breast cancer risk in Korean women according to menopausal status and breast cancer subtypes. Methods. We enrolled 677 patients diagnosed with breast cancer and 1,307 healthy controls who participated in screening mammography at the National Cancer Center. Breast density was estimated using volumetric breast composition measurement. Results. Of the total population, 1,156 (58.3 %) women were postmenopausal. The risk of breast cancer increased progressively with the increment of volumetric density grade (VDG) in postmenopausal women (p \ 0.001). High breast density (VDG 4) was significantly associated with breast cancer compared with low breast density (VDG 1/2) regardless of body mass index. However, the association with parity and history of hormone replacement therapy (HRT) was only found in those with C2 children and those not receiving HRT. Breast density was positively associated with breast cancer risk regardless of histologic grade, tumor size, lymph node involvement, Ki67 index, and hormone receptor status. The association was more prominent in human epidermal growth factor receptor 2 (HER2)-positive tumors (VDG 1/2 vs. VDG 4 for HER2 normal, odds ratio [OR] 2.21, 95 % confidence interval [CI] 1.28–3.83, p \ 0.001; for HER2 positive, OR 8.63,
In Hae Park and Kyungran Ko contributed equally to this study. Ó Society of Surgical Oncology 2014 First Received: 14 April 2014 J. Ro, MD, PhD e-mail: [email protected]
95 % CI 3.26–22.83, p = 0.001; Pheterogeneity= 0.030). However, no significant association was found between breast density and breast cancer risk in premenopausal women except for those with large-sized tumors ([2 cm) and a Ki67 index [15 %. Conclusion. High volumetric breast density is significantly associated with the risk of breast cancer in postmenopausal women; however, these relationships were not found in premenopausal women.
Mammographic density is known as one of the strongest risk factors for breast cancer, which is the most common cancer in women worldwide.1,2 The radiographic appearance of the breast by mammography reflects diversities in breast tissue composition that vary among women. Previous results consistently report that high density confers increased relative risks of 4- to 6-fold for breast cancer compared with low density, independent of parity, family history, or use of postmenopausal hormone replacement therapy (HRT).1 In contrast to breast cancer associated with lower density, cancers arising from dense breasts are associated with poor prognostic factors, including largersized tumors, lymph node (LN) involvement, and negative hormone receptor (HR) status,2–4 although the underlying reasons for the association have not been clearly explained. Recently, mammographic density was considered as not only a breast cancer risk factor, but also a predictive marker for adjuvant tamoxifen therapy.5,6 However, studies regarding the relationship between mammographic density and breast cancer-related death provided rather mixed results.7–9 Across studies, methods of measuring breast density were diverse in operation, from manual to fully automated, targeting dense areas by mammography or volume of
I. H. Park et al.
density, or in a format of density data from categorical to continuous values. In general, the qualitative Wolfe’s classification with breast imaging reporting and data systems (BIRADS) has been used to estimate tissue density.10 These methods have a few limitations, such as low agreement rates among radiologists, and difficulty in accurately quantifying certain degrees of change.11,12 Thus, there have been increasing efforts to identify a quantitative measurement of mammographic density. In the initial stage, percent dense area was assessed manually or via computer-assisted methods which have provided similar predictive information of breast cancer risk as visual assessment.13 Methods that assess the volume of dense tissue were also developed to estimate the amount of fibroglandular tissue within the whole breast.14,15 Until now, there have been a paucity of studies with volumetric measurements of breast density in clinical settings, and it is necessary to validate the relationship of volumetric breast density (VBD) with various clinical characteristics. There have been few studies that addressed the relationship between mammographic density and breast cancer risk in Asian women. In this study, we assess the association between VBD and breast cancer risk with well-known risk factors by comparing results from healthy women with those from Korean breast cancer patients. Furthermore, we investigate the relationship between breast density and the characteristics of breast cancer in women with breast cancer. PATIENTS AND METHODS Patients and Controls Patients had invasive breast cancer that was pathologically diagnosed from September 2012 to August 2013 at the National Cancer Center (NCC), Korea. Bilateral mammograms were taken before breast surgery or systemic treatment. We collected clinicopathological data, including tumor size; nodal status; estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 (HER2) status; and Ki-67 index (%) in tumor cells. The control group was composed of consecutive subjects who had screening mammograms at the NCC during the same period. Demographic information was obtained on a selfadministered questionnaire at the time of the screening mammography, which included questions about menopausal status, age, history of postmenopausal HRT, parity, height, weight, and breast health history. We excluded women who had a previous diagnosis of breast cancer or breast surgery for any reason. The study was approved by the NCC Institutional Review Board (IRB) before
proceeding (NCCNCS12657). The NCC IRB granted a waiver of written documentation of informed consent from all participants. Volumetric Breast Density Analysis Volpara software (version 1.5.1, Matakina Technology) was used to measure volumetric breast density. It is a computerized algorithm that calculates the x-ray attenuation at each pixel and converts the attenuation to an estimate of the tissue composition to create a density map. By adding up total values in the density map, the software can calculate the volume of fibroglandular tissue in the breast, and breast density is the percentage of fibroglandular tissue volume. Absolute dense volume was divided into four categories for analysis: 0–4.7 % as volumetric density grade (VDG) 1; 4.8–7.9 % as VDG 2; 8.0–15.0 % as VDG 3; and 15.1 or greater as VDG 4. Statistical Analysis Baseline characteristics of cases and controls were compared by Wilcoxon rank-sum test for continuous variables, and Chi-squared test for categorical variables. Overall, the percentage of missing data is very low in cases and less than 5 % in controls (Table 1). Subjects with missing data were excluded in the subsequent, corresponding analysis. Unconditional logistic regression was used to estimate the association between breast density and breast cancer risk while adjusting for potential confounders. Crude odds ratios (ORs) and corresponding adjusted ORs (adjusted for various clinical factors) were calculated. A Wald test was used to compare the association between breast density and breast cancer risk. We conducted tests for trends based on logistic regression with categorical explanatory variables. To test for trends, variables were coded with an increasing score then included in the regression model. The multinomial logistic regression test was performed to test for differences in trends across three volumetric breast-density levels by breast cancer tumor characteristics compared with controls (i.e. histologic grade, tumor size, LN involvement, HR positivity, HER2 status, and Ki67 index) while adjusting for breast cancer risk factors, including HRT, body mass index (BMI), and parity. To test whether the effect of breast density on the risk of breast cancer differs according to tumor characteristics, a heterogeneity test was performed using a likelihood ratio test.2 In this test, the likelihood of a model with a common coefficient of breast density for case groups was compared with that of a model with separate coefficients for each case group. All analyses were performed separately for postmenopausal and premenopausal women
Volumetric Density and Breast Cancer Risk TABLE 1 Characteristics of the study population Variables
Total (n = 1,984) Cases (%)
No. of subjects
677
Ageb [years; median (range)]
49 (23–88)
BMI [kg/m2; n (%)] \18.5
Postmenopausal (n = 1,156)
Controls (%)
p value
1,307 55 (39–89) \0.001
Cases (%)
Premenopausal (n = 828)
Controls (%) p value
303
853
56 (40–88)
60 (41–89)
\0.001
\0.001
a
Cases (%)
Controls (%) p valuea
374
454
44 (23–56)
45 (39–57)
12 (1.8)
25 (1.9)
1 (0.3)
16 (1.9)
11 (2.9)
9 (2.0)
18.5–24.9
442 (65.3)
905 (69.2)
175 (57.8)
564 (66.1)
267 (71.4)
341 (75.1)
C25
223 (32.9)
316 (24.2)
127 (41.9)
231 (27.1)
96 (25.7)
85 (18.7)
0 (0.0)
61 (4.7)
0 (0.0)
42 (4.9)
0 (0.0)
19 (4.2)
Missing
0.002
a
\0.001
0.070
Postmenopausal HRT [n (%)] 245 (80.9)
575 (67.4)
\2 years
None
NA
NA
NA
29 (9.6)
118 (13.8)
2–4.9 years
16 (5.3)
57 (6.7)
C5 years
12 (4.0)
63 (7.4)
Missing
1 (0.3)
40 (4.7)
0.005
NA
NA
NA
Parity [n (%)] 0–1
195 (28.8)
159 (12.2)
C2
481 (71.1)
1148 (87.8)
Missing
1 (0.2)
\0.001
0 (0.0)
\0.001
69 (22.8)
82 (9.6)
126 (33.7)
77 (17.0)
233 (76.9)
771 (90.4)
248 (66.3)
377 (83.0)
1 (0.3)
0 (0.0)
0 (0.0)
0 (0.0)
\0.001
Volumetric breast density VDG \0.001
3 (1)
64 (7.5)
\0.001
1
3 (0.4)
66 (5.1)
0 (0.0)
2 (0.4)
2
78 (11.5)
298 (22.8)
64 (21.1)
270 (31.7)
14 (3.7)
28 (6.2)
3
294 (43.4)
538 (41.2)
177 (58.4)
394 (46.2)
117 (31.3)
144 (31.7)
4
302 (44.6)
405 (31.0)
59 (19.5)
125 (14.7)
243 (65.0)
280 (61.7)
0.232
BMI body mass index, VDG volumetric density grade, NA not available, HRT hormone replacement therapy a
p-Value is based on the Wilcoxon rank sum test for age, and Chi-squared test for other categorical variables
b
Age at mammography
because of their different characteristics. All statistical tests were two-sided and p values \0.05 were considered statistically significant. All statistical analyses were carried out using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA) and STATA version 12.0 (StataCorp LP, College Station, TX, USA). RESULTS Characteristics of the Study Population During the study period, volumetric density measures were obtained from 1,984 women, including 677 breast cancer patients and 1,307 control subjects. The median age of the control group was 55 years (range 39–89), and that of case group was 49 years (range 23–88). The patient
population was younger than the control subjects because the peak incidence of breast cancer in Korea is observed in women in their late 40s.16 In postmenopausal women, cases were more likely to have high BMI and low parity (Table 1). Meanwhile, there was no statistically significant difference in BMI in premenopausal women between cases and control subjects. However, a low fertility rate was still observed in the premenopausal case group (p \ 0.001). The volumetric density category with greatest prevalence among postmenopausal women was VDG 3 (cases 58.4 %; controls 46.2 %), whereas VDG 4 was the most common in premenopausal women (cases 65.0 %; controls 61.7 %). In postmenopausal women, there was a difference in breast density between cases and controls; however, no difference was found in premenopausal women (Table 1).
I. H. Park et al.
Breast Cancer Risk Associated with Breast Density In all postmenopausal patients, high breast density was significantly associated with breast cancer (Table 2). Furthermore, the association between breast cancer and breast density was strongest in magnitude among women with a high BMI (C25 kg/m2), with a more than fourfold increase for VDG 4 compared with VDG 1/2 (OR 4.53; p \ 0.001). While there was a positive association between VDG and breast cancer risk in patients with a low BMI, the impact of breast density was much higher in patients with a high BMI (Pheterogeneity = 0.002). There was a significant trend for increasing breast cancer risk in patients with high breast density for women with C2 children (VDG 1/2 vs. VDG 4; OR 3.40; p \ 0.001), but the association was not significant for women with\2 children (VDG 1/2 vs. VDG 4; OR 1.79; p = 0.371). A positive association between breast density and cancer risk was also found with respect to the use of postmenopausal HRT, which was more prominent in patients with no history of postmenopausal HRT (VDG 1/2 vs. VDG 4 for postmenopausal HRT users, OR 2.28; for postmenopausal HRT non-users, OR 3.46; Pheterogeneity = 0.415). Although overall there was no association between breast density and breast cancer risk in premenopausal women (p = 0.125), there was a trend of increasing breast cancer risk as incremental breast density (VDG 1/2 vs. VDG 4; OR 3.08) in the high BMI group (Table 2). Association between Breast Density and Tumor Characteristics Among postmenopausal patients, 129 (42.9 %) had tumors [2 cm, and 119 patients (39.4 %) had nodal involvement. Overall, 117 tumors (39.4 %) were histological grade 3, and 163 tumors (54.9 %) showed a [15 % Ki67 index. HR status was positive in 213 tumors (70.5 %) and HER2 positivity was confirmed in 79 tumors (27.3 %). Overall, a positive association between breast density and breast cancer was found in postmenopausal women when adjusted for age, BMI, parity, and postmenopausal HRT (Table 3). Breast density was positively associated with histologic grade (VDG 1/2 vs. VDG 4: for grade 1/2, OR 2.69, p \ 0.001; for grade 3, OR 3.23, p \ 0.001; Pheterogeneity = 0.790), tumor size (VDG 1/2 vs. VDG 4: for B2 cm, OR 2.17, p = 0.002; for [2 cm, OR 4.86, p \ 0.001; Pheterogeneity = 0.062), LN involvement (VDG 1/2 vs. VDG 4: for negative LN, OR 2.40, p = 0.001; for positive LN, OR 4.92, p \ 0.001; Pheterogeneity = 0.128), and Ki67 index (VDG 1/2 vs. VDG 4: for B15 %, OR 2.46, p = 0.001; for [15 %, OR 3.67, p \ 0.001; Pheterogeneity = 0.485). Furthermore, the strong associations between breast density and the characteristics of breast
cancer were found in both HR-positive and HRnegative tumors (VDG 1/2 vs. VDG 4 for HR positive, OR 2.76, p \ 0.001; for HR negative, OR 3.93, p \ 0.001; Pheterogeneity = 0.489). The association between breast density and breast cancer risk was more prominent in HER2-positive tumors (VDG 1/2 vs. VDG 4: for HER2 normal, OR 2.21, p \ 0.001; for HER2 positive, OR 8.63, p = 0.001; Pheterogeneity = 0.030). Contrariwise, the associations between breast density and breast cancer risk according to tumor characteristics were generally not found in premenopausal patients when adjusted for age, BMI, and parity (Table 4). Even when the difference in subtype-specific association of breast density by Ki67 index and tumor size were statistically significant (for tumor size, Pheterogeneity = 0.017; for Ki67, Pheterogeneity = 0.007), there was a lack of significance in trends within the subgroup, mainly because there were few premenopausal cases with lower breast density (VDG 1/2).
DISCUSSION Extensive mammographic density is associated with an increased risk for developing abnormal breast lesions, including invasive breast cancer.17 The interpretation of density as a risk factor for breast cancer is complicated by the fact that the density is confounded by other risk factors, such as age, parity, history of postmenopausal HRT, or BMI. Breast density tends to fall along with increasing BMI, parity, and age.13 Therefore, the proportion of the disease attributable to breast density should be assessed with several interaction factors. Breast density could be quantitatively measured using fully automated volumetric approaches.18,19 Recently, Gweon et al. reported good comparability for breast-density categories between the fully automated volumetric method and the radiologist-assigned BIRADS system (q = 0.765; p \ 0.001).20 Its use could exclude the user variability and time-consuming density estimation, and could evaluate the breast as a 3-dimensional organ. In our study, we investigated breast densities with volumetric measurements of consecutive breast cancer patients and compared the results with those of healthy controls who had screening mammograms at the NCC. In Korea, a National Health Screening program is provided for women aged over 40 years; therefore, there was a difference in the median age between our two study groups.16 Compared with the control group, breast cancer in postmenopausal women was associated with higher BMI, lower parity, and a higher proportion of VDG 3 or 4. Previous studies reported that even though breast density was higher in Asian women, these differences
Volumetric Density and Breast Cancer Risk TABLE 2 Breast cancer risk associated with volumetric breast density by menstrual status, parity, and BMI
Factors
N
Adjusted OR (95 % CI)a
p value
\0.001
Case
Control
302
774
67
302
1.00 (reference)
VDG 3
176
356
2.64 (1.85–3.78)
VDG 4
59
116
3.07 (1.89–4.99)
57
224
Pbheterogeneity
Postmenopausal All VDG 1/2
Postmenopausal HRT Yes
0.415
VDG 1/2
14
87
1.00 (reference)
VDG 3
31
97
2.57 (1.17–5.66)
12
40
2.28 (0.83–6.29)
245
550
VDG 4 No
53
215
1.00 (reference)
VDG 3
VDG 1/2
145
259
2.67 (1.78–3.99)
VDG 4
47
76
3.46 (1.98–6.06)
175
555
0.078
\0.001
BMI
0.002
\25 kg/m2 VDG 1/2
25
154
1.00 (reference)
VDG 3
98
289
1.75 (1.06–2.88)
VDG 4
52
112
2.15 (1.21–3.83)
C25 kg/m2
127
219
VDG 1/2
42
148
1.00 (reference)
VDG 3
78
67
3.74 (2.29–6.12)
VDG 4
7
4
4.53 (1.20–17.12)
69
69
0.011
\0.001
Parity \2 children
0.238
VDG 1/2
12
20
1.00 (reference)
VDG 3
39
33
1.81 (0.67–4.91) 1.79 (0.54–5.91)
VDG 4
18
16
C2 children
233
705
VDG 1/2
55
282
1.00 (reference)
VDG 3
137
323
2.77 (1.88–4.07)
VDG 4
41
100
3.40 (1.98–5.83)
374
435
0.371
\0.001
Premenopausal All VDG 1/2
14
29
1.00 (reference)
VDG 3
117
138
1.87 (0.91–3.86)
VDG 4
243
268
2.05 (0.99–4.23)
278
350
0.125
BMI
0.099
\25 VDG 1/2
5
11
1.00 (reference)
VDG 3
64
92
1.36 (0.42–4.38)
VDG 4
209
247
1.35 (0.43–4.26)
96
85
C25 VDG 1/2
BMI body mass index, OR odds ratio, CI confidence interval, VDG volumetric density grade, HRT hormone replacement therapy a
Multivariable logistic regression model adjusted for age (continuous), BMI (categorical), parity (categorical), and postmenopausal HT use (postmenopausal women only)
9
18
1.00 (reference)
VDG 3
53
46
2.08 (0.83–5.19)
VDG 4
34
21
3.08 (1.15–8.30)
126
74
0.029
Parity \2 children VDG 1/2 VDG 3
0.576 5
3
1.00 (reference)
37
23
0.94 (0.18–4.85) 1.04 (0.20–5.40)
VDG 4
84
48
C2 children
248
361
VDG 1/2
9
26
1.00 (reference)
VDG 3
80
115
2.23 (0.97–5.16)
VDG 4
159
220
2.45 (1.06–5.67)
b
Pheterogeneity values were calculated using a likelihood ratio test
0.800
0.839
0.102
I. H. Park et al. TABLE 3 Associations between breast density and breast cancer characteristics in postmenopausal women
Factors
Adjusted OR (95 % CI)a
N Case
Control
180
774
p value
Histologic grade 1/2 VDG 1/2 VDG 3 VDG 4 3
0.790 40
302
1.00 (reference)
112
356
2.99 (1.94–4.59) 2.69 (1.46–4.93)
28
116
117
774
VDG 1/2
26
302
1.00 (reference)
VDG 3
61
356
2.05 (1.21–3.47)
VDG 4
30
116
3.23 (1.65–6.35)
172
774
\0.001
\0.001
Tumor size B2 cm VDG 1/2 VDG 3 VDG 4 [2 cm
0.062 43
302
1.00 (reference)
100
356
2.24 (1.47–3.43)
29
116
2.17 (1.20–3.93)
129
774
VDG 1/2
24
302
1.00 (reference)
VDG 3
76
356
3.42 (2.02–5.81)
VDG 4
29
116
4.86 (2.43–9.70)
183
774
0.002
\0.001
Lymph node involvement Negative VDG 1/2 VDG 3 VDG 4 Positive
0.128 49
302
1.00 (reference)
101
356
2.09 (1.38–3.15)
33
116
2.40 (1.35–4.25)
119
774
VDG 1/2
18
302
1.00 (reference)
VDG 3
75
356
4.19 (2.35–7.45)
VDG 4
26
116
4.92 (2.36–10.26)
213
774
0.001
\0.001
HR status Positive VDG 1/2
0.489 50
302
1.00 (reference)
VDG 3
125
356
2.58 (1.73–3.85)
VDG 4
38
116
2.76 (1.59–4.78)
89
774
Negative VDG 1/2
17
302
1.00 (reference)
VDG 3
51
356
2.82 (1.53–5.22)
VDG 4
21
116
3.93 (1.79–8.66)
79
774
\0.001
\0.001
HER2 status Positive VDG 1/2 VDG 3 VDG 4 Negative VDG 1/2
OR odds ratio, CI confidence interval, VDG volumetric density grade, HR hormone receptor, HER2 human epidermal growth factor receptor 2, HRT hormone replacement therapy, BMI body mass index a
Adjusted for age (continuous), postmenopausal HRT, BMI (categorical), and parity (categorical)
b
Pheterogeneity values were calculated using a likelihood ratio test
Pbheterogeneity
0.030 7
302
1.00 (reference)
50
356
6.22 (2.69–14.42) 8.63 (3.26–22.83)
22
116
210
774
59
302
1.00 (reference)
VDG 3
116
356
2.04 (1.38–3.00)
VDG 4
35
116
2.21 (1.28–3.83)
134
774
VDG 1/2
36
302
1.00 (reference)
VDG 3
78
356
2.42 (1.52–3.85) 2.46 (1.25–4.85)
0.001
\0.001
Ki67
0.485
B15 %
VDG 4 [15 %
20
116
163
774
VDG 1/2
30
302
1.00 (reference)
VDG 3
95
356
2.91 (1.81–4.69)
VDG 4
38
116
3.67 (1.99–6.79)
0.001
\0.001
Volumetric Density and Breast Cancer Risk TABLE 4 Association between breast density and breast cancer characteristics in premenopausal women
Factors
Adjusted OR (95 % CI)a
N Case
Control
226
435
p value
Histologic grade 1/2
0.653
VDG 1/2
11
29
1.00 (reference)
VDG 3
70
138
1.46 (0.67–3.19)
145
268
1.65 (0.75–3.62)
142
435
VDG 4 3 VDG 1/2
3
29
VDG 3
45
138
3.37 (0.95–11.97)
1.00 (reference)
VDG 4
94
268
3.59 (1.01–12.79)
200
435
0.236
0.147
Tumor size B2 cm VDG 1/2 VDG 3 VDG 4 [2 cm VDG 1/2
0.017 8
29
1.00 (reference)
66
138
1.59 (0.67–3.79)
126
268
1.38 (0.58–3.29)
173
435
6
29
1.00 (reference)
VDG 3
51
138
2.20 (0.83–5.87)
VDG 4
116
268
3.28 (1.22–8.78)
225
435
0.982
0.008
Lymph node involvement Negative VDG 1/2
0.610 7
29
1.00 (reference)
VDG 3
66
138
1.98 (0.80–4.89)
VDG 4
152
268
2.28 (0.92–5.62)
149
435
Positive VDG 1/2
7
29
1.00 (reference)
VDG 3
51
138
1.77 (0.70–4.51)
VDG 4
91
268
1.80 (0.70–4.63)
297
435
0.112
0.405
HR status Positive
0.751
VDG 1/2
12
29
1.00 (reference)
VDG 3
91
138
1.72 (0.81–3.66)
VDG 4
194
268
1.96 (0.92–4.19)
77
435
Negative VDG 1/2
2
29
VDG 3
26
138
2.87 (0.62–13.21)
1.00 (reference)
VDG 4
49
268
2.64 (0.57–12.28)
61
435
0.538
0.120
HER2 status Positive VDG 1/2 VDG 3 VDG 4 Negative
0.448 0
29
19
138
1.00 (reference) NA
42
268
NA
303
435
VDG 1/2
13
29
1.00 (reference)
VDG 3
92
138
1.53 (0.73–3.22)
VDG 4
198
268
1.75 (0.83–3.69)
170
435
NA
0.140
Ki67
0.007
B15 % OR odds ratio, CI confidence interval, VDG volumetric density grade, HR hormone receptor, HER2 human epidermal growth factor receptor 2, NA not applicable, BMI body mass index a
Adjusted for age (continuous), BMI (categorical), and parity (categorical)
b
Pheterogeneity values were calculated using a likelihood ratio test
Pbheterogeneity
VDG 1/2 VDG 3 VDG 4 [15 % VDG 1/2
5
29
1.00 (reference)
72
138
3.13 (1.13–8.67)
93
268
2.11 (0.75–5.92)
197
435
8
29
1.00 (reference)
VDG 3
45
138
1.26 (0.52–3.09)
VDG 4
144
268
2.19 (0.90–5.30)
0.708
0.006
I. H. Park et al.
disappeared when controlling for BMI and age.21,22 Most premenopausal women in our study population fell into the VDG 4 group and there was no significant difference in breast density between the case and control groups, in contrast to previous studies by Thomas et al. who reported that dense breast was an important risk factor for breast cancer regardless of age and menopausal status.23 In addition, breast density was an important risk factor for breast cancer in postmenopausal women when adjusted for BMI, parity, and a history of postmenopausal HRT, which was not detected in the premenopausal group. The cause of these conflicting results could possibly reside in the large proportion of young women who had higher breast density in our study population. We also addressed the association between breast density and breast cancer risk according to the characteristics of breast cancer. In this study, there was a strong association between dense breasts and large tumor size, axillary LN involvement, high histologic grade, and high Ki67 index in postmenopausal patients. In addition, there was a remarkable similarity in the magnitude of the associations between HR-positive and HR-negative breast cancer (Pheterogeneity = 0.489). However, there was a strong positive association with HER2-positive breast cancer with breast density compared with HER2-normal breast cancer (Pheterogeneity = 0.030). Overall, there was no objective association between mammographic density and subtype-specific breast cancer in premenopausal patients. Findings by prior studies that have examined associations of breast density with characteristics of breast cancer have been inconsistent.2,24,25 These conflicting results came about as a result of a different study population, study design, and variable methods of density measurement. Mammographic dense areas of the breast were known to be made up of greater proportions of epithelium and stromal components which could affect tumor microenvironment promoting cancer initiation and metastasis.26,27 Recent pooled analysis by Bertrand et al. showed that mammographic density itself was a strong risk factor for all subtypes of breast cancer, and the higher density was associated with tumor aggressiveness especially in women aged \55 years.24 In spite of that, there has been a lack of evidence supporting the relationship between mammographic density and breast cancer mortality or survival.3,9 Such discordances between mechanisms and clinical results may originate from the heterogeneities of the study population and study design. Therefore, further studies are needed to find specific groups in which mammographic density could be used as a prognostic marker.
CONCLUSIONS High VBD is associated with a risk of breast cancer with specific characteristics in postmenopausal women; however, these relationships were not found in premenopausal Korean women. ACKNOWLEDGMENT This study was funded by NCC Grant 1310340. The funding source had no role in the study design, data analysis, data collection, and interpretation. We are thankful to all participants in this study. DISCLOSURE There are no disclosures of any commercial interest or financial support.
REFERENCES 1. McCormack V, dosSantosSilva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomark Prev. 2006;15:1159–69. 2. Yaghjyan L, Colditz G, Collins L, et al. Mammographic breast density and subsequent risk of breast cancer in postmenopausal women according to tumor characteristics. J Natl Cancer Inst. 2011;103:1179–89. 3. Conroy S, Pagano I, Kolonel L, et al. Mammographic density and hormone receptor expression in breast cancer: the Multiethnic Cohort Study. Cancer Epidemiol. 2011;35:448–52. 4. Sala E, Solomon L, Warren R, et al. Size, node status and grade of breast tumours: association with mammographic parenchymal patterns. Eur Radiol. 2000;10:157–61. 5. Li J, Humphreys K, Eriksson L, et al. Mammographic density reduction is a prognostic marker of response to adjuvant tamoxifen therapy in postmenopausal patients with breast cancer. J Clin Oncol. 2013;31:2249–56. 6. Kim J, Han W, Moon H-G, et al. Breast density change as a predictive surrogate for response to adjuvant endocrine therapy in hormone receptor positive breast cancer. Breast Cancer Res. 2012;14:R102. 7. Porter G, Evans A, Cornford E, et al. Influence of mammographic parenchymal pattern in screening-detected and interval invasive breast cancers on pathologic features, mammographic features, and patient survival. AJR Am J Roentgenol. 2007;188:676–83. 8. Olsen A, Bihrmann K, Jensen MB, et al. Breast density and outcome of mammography screening: a cohort study. Br J Cancer. 2009;100:1205–8. 9. Gierach G, Ichikawa L, Kerlikowske K, et al. Relationship between mammographic density and breast cancer death in the Breast Cancer Surveillance Consortium. J Natl Cancer Inst. 2012;104:1218–27. 10. Vanel D. The American College of Radiology (ACR) Breast Imaging and Reporting Data System (BI-RADS): a step towards a universal radiological language? Eur J Radiol. 2007;61:183. 11. Kerlikowske K, Grady D, Barclay J, et al. Variability and accuracy in mammographic interpretation using the American College of Radiology Breast Imaging Reporting and Data System. J Natl Cancer Inst. 1998;90:1801–09. 12. Martin K, Helvie M, Zhou C, et al. Mammographic density measured with quantitative computer-aided method: comparison with radiologists’ estimates and BI-RADS categories. Radiology. 2006;240:656–65.
Volumetric Density and Breast Cancer Risk 13. Assi V, Warwick J, Cuzick J, Duffy S. Clinical and epidemiological issues in mammographic density. Nat Rev Clin Oncol. 2012;9:33–40. 14. Highnam R, Brady M, Yaffe MJ, et al. Robust breast composition measurement – VolparaTM. Girona: Springer; 2010. p. 342–349. 15. Highnam R, Sauber N, Destounis S, et al. Breast density into clinical practice. Berlin: Springer; 2012. p. 466–473. 16. Park JH, Yang HK, Nam EJ, et al. Cancer facts & figures 2013. President of National Cancer Center, Minister for Health and Welfare; 2013. 17. Boyd N, Martin L, Yaffe M, Minkin S. Mammographic density and breast cancer risk: current understanding and future prospects. Breast Cancer Res. 2011;13:223. 18. Aitken Z, McCormack V, Highnam R, et al. Screen-film mammographic density and breast cancer risk: a comparison of the volumetric standard mammogram form and the interactive threshold measurement methods. Cancer Epidemiol Biomarkers Prev. 2010;19:418–28. 19. Jeffreys M, Warren R, Highnam R, Davey Smith G. Breast cancer risk factors and a novel measure of volumetric breast density: cross-sectional study. Br J Cancer. 2008;98:210–16. 20. Gweon H, Youk J, Kim J-A, Son E. Radiologist assessment of breast density by BI-RADS categories versus fully automated volumetric assessment. AJR Am J Roentgenol. 2013;201:692–7.
21. del Carmen M, Halpern E, Kopans D, et al. Mammographic breast density and race. AJR Am J Roentgenol. 2007; 188:1147–50. 22. Ursin G, Ma H, Wu A, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomark Prev. 2003;12:332–38. 23. Thomas D, Carter R, Bush W, et al. Risk of subsequent breast cancer in relation to characteristics of screening mammograms from women less than 50 years of age. Cancer Epidemiol Biomarkers Prev. 2002;11:565–71. 24. Bertrand K, Tamimi R, Scott C, et al. Mammographic density and risk of breast cancer by age and tumor characteristics. Breast Cancer Res. 2013;15:104. 25. Antoni S, Sasco A, dosSantosSilva I, McCormack V (2013) Is mammographic density differentially associated with breast cancer according to receptor status? A meta-analysis. Breast Cancer Res Treat. 137:337–47. 26. Ghosh K, Brandt KR, Reynolds C, et al. Tissue composition of mammographically dense and non-dense breast tissue. Breast Cancer Res Treat. 2012;131:267–75. 27. Mao Y, Keller ET, Garfield DH, et al. Stromal cells in tumor microenvironment and breast cancer. Cancer Metastasis Rev. 2013;32:303–12.
ORIGINAL ARTICLE – BREAST ONCOLOGY
High Volumetric Breast Density Predicts Risk for Breast Cancer in Postmenopausal, but not Premenopausal, Korean Women In Hae Park, MD1, Kyungran Ko, MD1, Jungnam Joo, PhD2, Boram Park, MS2, So-Youn Jung, MD1, Seeyoun Lee, MD, PhD1, Youngmi Kwon, MD, PhD1, Han-Sung Kang, MD, PhD1, Eun Sook Lee, MD, PhD1, Keun Seok Lee, MD, PhD1, and Jungsil Ro, MD, PhD1 Center For Breast Cancer, National Cancer Center, Goyang-si, Gyeonggi-do, Korea; 2Biometric Research Branch, National Cancer Center, Goyang, Korea
1
ABSTRACT Purpose. We investigated the association between mammographic breast density and breast cancer risk in Korean women according to menopausal status and breast cancer subtypes. Methods. We enrolled 677 patients diagnosed with breast cancer and 1,307 healthy controls who participated in screening mammography at the National Cancer Center. Breast density was estimated using volumetric breast composition measurement. Results. Of the total population, 1,156 (58.3 %) women were postmenopausal. The risk of breast cancer increased progressively with the increment of volumetric density grade (VDG) in postmenopausal women (p \ 0.001). High breast density (VDG 4) was significantly associated with breast cancer compared with low breast density (VDG 1/2) regardless of body mass index. However, the association with parity and history of hormone replacement therapy (HRT) was only found in those with C2 children and those not receiving HRT. Breast density was positively associated with breast cancer risk regardless of histologic grade, tumor size, lymph node involvement, Ki67 index, and hormone receptor status. The association was more prominent in human epidermal growth factor receptor 2 (HER2)-positive tumors (VDG 1/2 vs. VDG 4 for HER2 normal, odds ratio [OR] 2.21, 95 % confidence interval [CI] 1.28–3.83, p \ 0.001; for HER2 positive, OR 8.63,
In Hae Park and Kyungran Ko contributed equally to this study. Ó Society of Surgical Oncology 2014 First Received: 14 April 2014 J. Ro, MD, PhD e-mail: [email protected]
95 % CI 3.26–22.83, p = 0.001; Pheterogeneity= 0.030). However, no significant association was found between breast density and breast cancer risk in premenopausal women except for those with large-sized tumors ([2 cm) and a Ki67 index [15 %. Conclusion. High volumetric breast density is significantly associated with the risk of breast cancer in postmenopausal women; however, these relationships were not found in premenopausal women.
Mammographic density is known as one of the strongest risk factors for breast cancer, which is the most common cancer in women worldwide.1,2 The radiographic appearance of the breast by mammography reflects diversities in breast tissue composition that vary among women. Previous results consistently report that high density confers increased relative risks of 4- to 6-fold for breast cancer compared with low density, independent of parity, family history, or use of postmenopausal hormone replacement therapy (HRT).1 In contrast to breast cancer associated with lower density, cancers arising from dense breasts are associated with poor prognostic factors, including largersized tumors, lymph node (LN) involvement, and negative hormone receptor (HR) status,2–4 although the underlying reasons for the association have not been clearly explained. Recently, mammographic density was considered as not only a breast cancer risk factor, but also a predictive marker for adjuvant tamoxifen therapy.5,6 However, studies regarding the relationship between mammographic density and breast cancer-related death provided rather mixed results.7–9 Across studies, methods of measuring breast density were diverse in operation, from manual to fully automated, targeting dense areas by mammography or volume of
I. H. Park et al.
density, or in a format of density data from categorical to continuous values. In general, the qualitative Wolfe’s classification with breast imaging reporting and data systems (BIRADS) has been used to estimate tissue density.10 These methods have a few limitations, such as low agreement rates among radiologists, and difficulty in accurately quantifying certain degrees of change.11,12 Thus, there have been increasing efforts to identify a quantitative measurement of mammographic density. In the initial stage, percent dense area was assessed manually or via computer-assisted methods which have provided similar predictive information of breast cancer risk as visual assessment.13 Methods that assess the volume of dense tissue were also developed to estimate the amount of fibroglandular tissue within the whole breast.14,15 Until now, there have been a paucity of studies with volumetric measurements of breast density in clinical settings, and it is necessary to validate the relationship of volumetric breast density (VBD) with various clinical characteristics. There have been few studies that addressed the relationship between mammographic density and breast cancer risk in Asian women. In this study, we assess the association between VBD and breast cancer risk with well-known risk factors by comparing results from healthy women with those from Korean breast cancer patients. Furthermore, we investigate the relationship between breast density and the characteristics of breast cancer in women with breast cancer. PATIENTS AND METHODS Patients and Controls Patients had invasive breast cancer that was pathologically diagnosed from September 2012 to August 2013 at the National Cancer Center (NCC), Korea. Bilateral mammograms were taken before breast surgery or systemic treatment. We collected clinicopathological data, including tumor size; nodal status; estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 (HER2) status; and Ki-67 index (%) in tumor cells. The control group was composed of consecutive subjects who had screening mammograms at the NCC during the same period. Demographic information was obtained on a selfadministered questionnaire at the time of the screening mammography, which included questions about menopausal status, age, history of postmenopausal HRT, parity, height, weight, and breast health history. We excluded women who had a previous diagnosis of breast cancer or breast surgery for any reason. The study was approved by the NCC Institutional Review Board (IRB) before
proceeding (NCCNCS12657). The NCC IRB granted a waiver of written documentation of informed consent from all participants. Volumetric Breast Density Analysis Volpara software (version 1.5.1, Matakina Technology) was used to measure volumetric breast density. It is a computerized algorithm that calculates the x-ray attenuation at each pixel and converts the attenuation to an estimate of the tissue composition to create a density map. By adding up total values in the density map, the software can calculate the volume of fibroglandular tissue in the breast, and breast density is the percentage of fibroglandular tissue volume. Absolute dense volume was divided into four categories for analysis: 0–4.7 % as volumetric density grade (VDG) 1; 4.8–7.9 % as VDG 2; 8.0–15.0 % as VDG 3; and 15.1 or greater as VDG 4. Statistical Analysis Baseline characteristics of cases and controls were compared by Wilcoxon rank-sum test for continuous variables, and Chi-squared test for categorical variables. Overall, the percentage of missing data is very low in cases and less than 5 % in controls (Table 1). Subjects with missing data were excluded in the subsequent, corresponding analysis. Unconditional logistic regression was used to estimate the association between breast density and breast cancer risk while adjusting for potential confounders. Crude odds ratios (ORs) and corresponding adjusted ORs (adjusted for various clinical factors) were calculated. A Wald test was used to compare the association between breast density and breast cancer risk. We conducted tests for trends based on logistic regression with categorical explanatory variables. To test for trends, variables were coded with an increasing score then included in the regression model. The multinomial logistic regression test was performed to test for differences in trends across three volumetric breast-density levels by breast cancer tumor characteristics compared with controls (i.e. histologic grade, tumor size, LN involvement, HR positivity, HER2 status, and Ki67 index) while adjusting for breast cancer risk factors, including HRT, body mass index (BMI), and parity. To test whether the effect of breast density on the risk of breast cancer differs according to tumor characteristics, a heterogeneity test was performed using a likelihood ratio test.2 In this test, the likelihood of a model with a common coefficient of breast density for case groups was compared with that of a model with separate coefficients for each case group. All analyses were performed separately for postmenopausal and premenopausal women
Volumetric Density and Breast Cancer Risk TABLE 1 Characteristics of the study population Variables
Total (n = 1,984) Cases (%)
No. of subjects
677
Ageb [years; median (range)]
49 (23–88)
BMI [kg/m2; n (%)] \18.5
Postmenopausal (n = 1,156)
Controls (%)
p value
1,307 55 (39–89) \0.001
Cases (%)
Premenopausal (n = 828)
Controls (%) p value
303
853
56 (40–88)
60 (41–89)
\0.001
\0.001
a
Cases (%)
Controls (%) p valuea
374
454
44 (23–56)
45 (39–57)
12 (1.8)
25 (1.9)
1 (0.3)
16 (1.9)
11 (2.9)
9 (2.0)
18.5–24.9
442 (65.3)
905 (69.2)
175 (57.8)
564 (66.1)
267 (71.4)
341 (75.1)
C25
223 (32.9)
316 (24.2)
127 (41.9)
231 (27.1)
96 (25.7)
85 (18.7)
0 (0.0)
61 (4.7)
0 (0.0)
42 (4.9)
0 (0.0)
19 (4.2)
Missing
0.002
a
\0.001
0.070
Postmenopausal HRT [n (%)] 245 (80.9)
575 (67.4)
\2 years
None
NA
NA
NA
29 (9.6)
118 (13.8)
2–4.9 years
16 (5.3)
57 (6.7)
C5 years
12 (4.0)
63 (7.4)
Missing
1 (0.3)
40 (4.7)
0.005
NA
NA
NA
Parity [n (%)] 0–1
195 (28.8)
159 (12.2)
C2
481 (71.1)
1148 (87.8)
Missing
1 (0.2)
\0.001
0 (0.0)
\0.001
69 (22.8)
82 (9.6)
126 (33.7)
77 (17.0)
233 (76.9)
771 (90.4)
248 (66.3)
377 (83.0)
1 (0.3)
0 (0.0)
0 (0.0)
0 (0.0)
\0.001
Volumetric breast density VDG \0.001
3 (1)
64 (7.5)
\0.001
1
3 (0.4)
66 (5.1)
0 (0.0)
2 (0.4)
2
78 (11.5)
298 (22.8)
64 (21.1)
270 (31.7)
14 (3.7)
28 (6.2)
3
294 (43.4)
538 (41.2)
177 (58.4)
394 (46.2)
117 (31.3)
144 (31.7)
4
302 (44.6)
405 (31.0)
59 (19.5)
125 (14.7)
243 (65.0)
280 (61.7)
0.232
BMI body mass index, VDG volumetric density grade, NA not available, HRT hormone replacement therapy a
p-Value is based on the Wilcoxon rank sum test for age, and Chi-squared test for other categorical variables
b
Age at mammography
because of their different characteristics. All statistical tests were two-sided and p values \0.05 were considered statistically significant. All statistical analyses were carried out using SAS version 9.2 (SAS Institute Inc., Cary, NC, USA) and STATA version 12.0 (StataCorp LP, College Station, TX, USA). RESULTS Characteristics of the Study Population During the study period, volumetric density measures were obtained from 1,984 women, including 677 breast cancer patients and 1,307 control subjects. The median age of the control group was 55 years (range 39–89), and that of case group was 49 years (range 23–88). The patient
population was younger than the control subjects because the peak incidence of breast cancer in Korea is observed in women in their late 40s.16 In postmenopausal women, cases were more likely to have high BMI and low parity (Table 1). Meanwhile, there was no statistically significant difference in BMI in premenopausal women between cases and control subjects. However, a low fertility rate was still observed in the premenopausal case group (p \ 0.001). The volumetric density category with greatest prevalence among postmenopausal women was VDG 3 (cases 58.4 %; controls 46.2 %), whereas VDG 4 was the most common in premenopausal women (cases 65.0 %; controls 61.7 %). In postmenopausal women, there was a difference in breast density between cases and controls; however, no difference was found in premenopausal women (Table 1).
I. H. Park et al.
Breast Cancer Risk Associated with Breast Density In all postmenopausal patients, high breast density was significantly associated with breast cancer (Table 2). Furthermore, the association between breast cancer and breast density was strongest in magnitude among women with a high BMI (C25 kg/m2), with a more than fourfold increase for VDG 4 compared with VDG 1/2 (OR 4.53; p \ 0.001). While there was a positive association between VDG and breast cancer risk in patients with a low BMI, the impact of breast density was much higher in patients with a high BMI (Pheterogeneity = 0.002). There was a significant trend for increasing breast cancer risk in patients with high breast density for women with C2 children (VDG 1/2 vs. VDG 4; OR 3.40; p \ 0.001), but the association was not significant for women with\2 children (VDG 1/2 vs. VDG 4; OR 1.79; p = 0.371). A positive association between breast density and cancer risk was also found with respect to the use of postmenopausal HRT, which was more prominent in patients with no history of postmenopausal HRT (VDG 1/2 vs. VDG 4 for postmenopausal HRT users, OR 2.28; for postmenopausal HRT non-users, OR 3.46; Pheterogeneity = 0.415). Although overall there was no association between breast density and breast cancer risk in premenopausal women (p = 0.125), there was a trend of increasing breast cancer risk as incremental breast density (VDG 1/2 vs. VDG 4; OR 3.08) in the high BMI group (Table 2). Association between Breast Density and Tumor Characteristics Among postmenopausal patients, 129 (42.9 %) had tumors [2 cm, and 119 patients (39.4 %) had nodal involvement. Overall, 117 tumors (39.4 %) were histological grade 3, and 163 tumors (54.9 %) showed a [15 % Ki67 index. HR status was positive in 213 tumors (70.5 %) and HER2 positivity was confirmed in 79 tumors (27.3 %). Overall, a positive association between breast density and breast cancer was found in postmenopausal women when adjusted for age, BMI, parity, and postmenopausal HRT (Table 3). Breast density was positively associated with histologic grade (VDG 1/2 vs. VDG 4: for grade 1/2, OR 2.69, p \ 0.001; for grade 3, OR 3.23, p \ 0.001; Pheterogeneity = 0.790), tumor size (VDG 1/2 vs. VDG 4: for B2 cm, OR 2.17, p = 0.002; for [2 cm, OR 4.86, p \ 0.001; Pheterogeneity = 0.062), LN involvement (VDG 1/2 vs. VDG 4: for negative LN, OR 2.40, p = 0.001; for positive LN, OR 4.92, p \ 0.001; Pheterogeneity = 0.128), and Ki67 index (VDG 1/2 vs. VDG 4: for B15 %, OR 2.46, p = 0.001; for [15 %, OR 3.67, p \ 0.001; Pheterogeneity = 0.485). Furthermore, the strong associations between breast density and the characteristics of breast
cancer were found in both HR-positive and HRnegative tumors (VDG 1/2 vs. VDG 4 for HR positive, OR 2.76, p \ 0.001; for HR negative, OR 3.93, p \ 0.001; Pheterogeneity = 0.489). The association between breast density and breast cancer risk was more prominent in HER2-positive tumors (VDG 1/2 vs. VDG 4: for HER2 normal, OR 2.21, p \ 0.001; for HER2 positive, OR 8.63, p = 0.001; Pheterogeneity = 0.030). Contrariwise, the associations between breast density and breast cancer risk according to tumor characteristics were generally not found in premenopausal patients when adjusted for age, BMI, and parity (Table 4). Even when the difference in subtype-specific association of breast density by Ki67 index and tumor size were statistically significant (for tumor size, Pheterogeneity = 0.017; for Ki67, Pheterogeneity = 0.007), there was a lack of significance in trends within the subgroup, mainly because there were few premenopausal cases with lower breast density (VDG 1/2).
DISCUSSION Extensive mammographic density is associated with an increased risk for developing abnormal breast lesions, including invasive breast cancer.17 The interpretation of density as a risk factor for breast cancer is complicated by the fact that the density is confounded by other risk factors, such as age, parity, history of postmenopausal HRT, or BMI. Breast density tends to fall along with increasing BMI, parity, and age.13 Therefore, the proportion of the disease attributable to breast density should be assessed with several interaction factors. Breast density could be quantitatively measured using fully automated volumetric approaches.18,19 Recently, Gweon et al. reported good comparability for breast-density categories between the fully automated volumetric method and the radiologist-assigned BIRADS system (q = 0.765; p \ 0.001).20 Its use could exclude the user variability and time-consuming density estimation, and could evaluate the breast as a 3-dimensional organ. In our study, we investigated breast densities with volumetric measurements of consecutive breast cancer patients and compared the results with those of healthy controls who had screening mammograms at the NCC. In Korea, a National Health Screening program is provided for women aged over 40 years; therefore, there was a difference in the median age between our two study groups.16 Compared with the control group, breast cancer in postmenopausal women was associated with higher BMI, lower parity, and a higher proportion of VDG 3 or 4. Previous studies reported that even though breast density was higher in Asian women, these differences
Volumetric Density and Breast Cancer Risk TABLE 2 Breast cancer risk associated with volumetric breast density by menstrual status, parity, and BMI
Factors
N
Adjusted OR (95 % CI)a
p value
\0.001
Case
Control
302
774
67
302
1.00 (reference)
VDG 3
176
356
2.64 (1.85–3.78)
VDG 4
59
116
3.07 (1.89–4.99)
57
224
Pbheterogeneity
Postmenopausal All VDG 1/2
Postmenopausal HRT Yes
0.415
VDG 1/2
14
87
1.00 (reference)
VDG 3
31
97
2.57 (1.17–5.66)
12
40
2.28 (0.83–6.29)
245
550
VDG 4 No
53
215
1.00 (reference)
VDG 3
VDG 1/2
145
259
2.67 (1.78–3.99)
VDG 4
47
76
3.46 (1.98–6.06)
175
555
0.078
\0.001
BMI
0.002
\25 kg/m2 VDG 1/2
25
154
1.00 (reference)
VDG 3
98
289
1.75 (1.06–2.88)
VDG 4
52
112
2.15 (1.21–3.83)
C25 kg/m2
127
219
VDG 1/2
42
148
1.00 (reference)
VDG 3
78
67
3.74 (2.29–6.12)
VDG 4
7
4
4.53 (1.20–17.12)
69
69
0.011
\0.001
Parity \2 children
0.238
VDG 1/2
12
20
1.00 (reference)
VDG 3
39
33
1.81 (0.67–4.91) 1.79 (0.54–5.91)
VDG 4
18
16
C2 children
233
705
VDG 1/2
55
282
1.00 (reference)
VDG 3
137
323
2.77 (1.88–4.07)
VDG 4
41
100
3.40 (1.98–5.83)
374
435
0.371
\0.001
Premenopausal All VDG 1/2
14
29
1.00 (reference)
VDG 3
117
138
1.87 (0.91–3.86)
VDG 4
243
268
2.05 (0.99–4.23)
278
350
0.125
BMI
0.099
\25 VDG 1/2
5
11
1.00 (reference)
VDG 3
64
92
1.36 (0.42–4.38)
VDG 4
209
247
1.35 (0.43–4.26)
96
85
C25 VDG 1/2
BMI body mass index, OR odds ratio, CI confidence interval, VDG volumetric density grade, HRT hormone replacement therapy a
Multivariable logistic regression model adjusted for age (continuous), BMI (categorical), parity (categorical), and postmenopausal HT use (postmenopausal women only)
9
18
1.00 (reference)
VDG 3
53
46
2.08 (0.83–5.19)
VDG 4
34
21
3.08 (1.15–8.30)
126
74
0.029
Parity \2 children VDG 1/2 VDG 3
0.576 5
3
1.00 (reference)
37
23
0.94 (0.18–4.85) 1.04 (0.20–5.40)
VDG 4
84
48
C2 children
248
361
VDG 1/2
9
26
1.00 (reference)
VDG 3
80
115
2.23 (0.97–5.16)
VDG 4
159
220
2.45 (1.06–5.67)
b
Pheterogeneity values were calculated using a likelihood ratio test
0.800
0.839
0.102
I. H. Park et al. TABLE 3 Associations between breast density and breast cancer characteristics in postmenopausal women
Factors
Adjusted OR (95 % CI)a
N Case
Control
180
774
p value
Histologic grade 1/2 VDG 1/2 VDG 3 VDG 4 3
0.790 40
302
1.00 (reference)
112
356
2.99 (1.94–4.59) 2.69 (1.46–4.93)
28
116
117
774
VDG 1/2
26
302
1.00 (reference)
VDG 3
61
356
2.05 (1.21–3.47)
VDG 4
30
116
3.23 (1.65–6.35)
172
774
\0.001
\0.001
Tumor size B2 cm VDG 1/2 VDG 3 VDG 4 [2 cm
0.062 43
302
1.00 (reference)
100
356
2.24 (1.47–3.43)
29
116
2.17 (1.20–3.93)
129
774
VDG 1/2
24
302
1.00 (reference)
VDG 3
76
356
3.42 (2.02–5.81)
VDG 4
29
116
4.86 (2.43–9.70)
183
774
0.002
\0.001
Lymph node involvement Negative VDG 1/2 VDG 3 VDG 4 Positive
0.128 49
302
1.00 (reference)
101
356
2.09 (1.38–3.15)
33
116
2.40 (1.35–4.25)
119
774
VDG 1/2
18
302
1.00 (reference)
VDG 3
75
356
4.19 (2.35–7.45)
VDG 4
26
116
4.92 (2.36–10.26)
213
774
0.001
\0.001
HR status Positive VDG 1/2
0.489 50
302
1.00 (reference)
VDG 3
125
356
2.58 (1.73–3.85)
VDG 4
38
116
2.76 (1.59–4.78)
89
774
Negative VDG 1/2
17
302
1.00 (reference)
VDG 3
51
356
2.82 (1.53–5.22)
VDG 4
21
116
3.93 (1.79–8.66)
79
774
\0.001
\0.001
HER2 status Positive VDG 1/2 VDG 3 VDG 4 Negative VDG 1/2
OR odds ratio, CI confidence interval, VDG volumetric density grade, HR hormone receptor, HER2 human epidermal growth factor receptor 2, HRT hormone replacement therapy, BMI body mass index a
Adjusted for age (continuous), postmenopausal HRT, BMI (categorical), and parity (categorical)
b
Pheterogeneity values were calculated using a likelihood ratio test
Pbheterogeneity
0.030 7
302
1.00 (reference)
50
356
6.22 (2.69–14.42) 8.63 (3.26–22.83)
22
116
210
774
59
302
1.00 (reference)
VDG 3
116
356
2.04 (1.38–3.00)
VDG 4
35
116
2.21 (1.28–3.83)
134
774
VDG 1/2
36
302
1.00 (reference)
VDG 3
78
356
2.42 (1.52–3.85) 2.46 (1.25–4.85)
0.001
\0.001
Ki67
0.485
B15 %
VDG 4 [15 %
20
116
163
774
VDG 1/2
30
302
1.00 (reference)
VDG 3
95
356
2.91 (1.81–4.69)
VDG 4
38
116
3.67 (1.99–6.79)
0.001
\0.001
Volumetric Density and Breast Cancer Risk TABLE 4 Association between breast density and breast cancer characteristics in premenopausal women
Factors
Adjusted OR (95 % CI)a
N Case
Control
226
435
p value
Histologic grade 1/2
0.653
VDG 1/2
11
29
1.00 (reference)
VDG 3
70
138
1.46 (0.67–3.19)
145
268
1.65 (0.75–3.62)
142
435
VDG 4 3 VDG 1/2
3
29
VDG 3
45
138
3.37 (0.95–11.97)
1.00 (reference)
VDG 4
94
268
3.59 (1.01–12.79)
200
435
0.236
0.147
Tumor size B2 cm VDG 1/2 VDG 3 VDG 4 [2 cm VDG 1/2
0.017 8
29
1.00 (reference)
66
138
1.59 (0.67–3.79)
126
268
1.38 (0.58–3.29)
173
435
6
29
1.00 (reference)
VDG 3
51
138
2.20 (0.83–5.87)
VDG 4
116
268
3.28 (1.22–8.78)
225
435
0.982
0.008
Lymph node involvement Negative VDG 1/2
0.610 7
29
1.00 (reference)
VDG 3
66
138
1.98 (0.80–4.89)
VDG 4
152
268
2.28 (0.92–5.62)
149
435
Positive VDG 1/2
7
29
1.00 (reference)
VDG 3
51
138
1.77 (0.70–4.51)
VDG 4
91
268
1.80 (0.70–4.63)
297
435
0.112
0.405
HR status Positive
0.751
VDG 1/2
12
29
1.00 (reference)
VDG 3
91
138
1.72 (0.81–3.66)
VDG 4
194
268
1.96 (0.92–4.19)
77
435
Negative VDG 1/2
2
29
VDG 3
26
138
2.87 (0.62–13.21)
1.00 (reference)
VDG 4
49
268
2.64 (0.57–12.28)
61
435
0.538
0.120
HER2 status Positive VDG 1/2 VDG 3 VDG 4 Negative
0.448 0
29
19
138
1.00 (reference) NA
42
268
NA
303
435
VDG 1/2
13
29
1.00 (reference)
VDG 3
92
138
1.53 (0.73–3.22)
VDG 4
198
268
1.75 (0.83–3.69)
170
435
NA
0.140
Ki67
0.007
B15 % OR odds ratio, CI confidence interval, VDG volumetric density grade, HR hormone receptor, HER2 human epidermal growth factor receptor 2, NA not applicable, BMI body mass index a
Adjusted for age (continuous), BMI (categorical), and parity (categorical)
b
Pheterogeneity values were calculated using a likelihood ratio test
Pbheterogeneity
VDG 1/2 VDG 3 VDG 4 [15 % VDG 1/2
5
29
1.00 (reference)
72
138
3.13 (1.13–8.67)
93
268
2.11 (0.75–5.92)
197
435
8
29
1.00 (reference)
VDG 3
45
138
1.26 (0.52–3.09)
VDG 4
144
268
2.19 (0.90–5.30)
0.708
0.006
I. H. Park et al.
disappeared when controlling for BMI and age.21,22 Most premenopausal women in our study population fell into the VDG 4 group and there was no significant difference in breast density between the case and control groups, in contrast to previous studies by Thomas et al. who reported that dense breast was an important risk factor for breast cancer regardless of age and menopausal status.23 In addition, breast density was an important risk factor for breast cancer in postmenopausal women when adjusted for BMI, parity, and a history of postmenopausal HRT, which was not detected in the premenopausal group. The cause of these conflicting results could possibly reside in the large proportion of young women who had higher breast density in our study population. We also addressed the association between breast density and breast cancer risk according to the characteristics of breast cancer. In this study, there was a strong association between dense breasts and large tumor size, axillary LN involvement, high histologic grade, and high Ki67 index in postmenopausal patients. In addition, there was a remarkable similarity in the magnitude of the associations between HR-positive and HR-negative breast cancer (Pheterogeneity = 0.489). However, there was a strong positive association with HER2-positive breast cancer with breast density compared with HER2-normal breast cancer (Pheterogeneity = 0.030). Overall, there was no objective association between mammographic density and subtype-specific breast cancer in premenopausal patients. Findings by prior studies that have examined associations of breast density with characteristics of breast cancer have been inconsistent.2,24,25 These conflicting results came about as a result of a different study population, study design, and variable methods of density measurement. Mammographic dense areas of the breast were known to be made up of greater proportions of epithelium and stromal components which could affect tumor microenvironment promoting cancer initiation and metastasis.26,27 Recent pooled analysis by Bertrand et al. showed that mammographic density itself was a strong risk factor for all subtypes of breast cancer, and the higher density was associated with tumor aggressiveness especially in women aged \55 years.24 In spite of that, there has been a lack of evidence supporting the relationship between mammographic density and breast cancer mortality or survival.3,9 Such discordances between mechanisms and clinical results may originate from the heterogeneities of the study population and study design. Therefore, further studies are needed to find specific groups in which mammographic density could be used as a prognostic marker.
CONCLUSIONS High VBD is associated with a risk of breast cancer with specific characteristics in postmenopausal women; however, these relationships were not found in premenopausal Korean women. ACKNOWLEDGMENT This study was funded by NCC Grant 1310340. The funding source had no role in the study design, data analysis, data collection, and interpretation. We are thankful to all participants in this study. DISCLOSURE There are no disclosures of any commercial interest or financial support.
REFERENCES 1. McCormack V, dosSantosSilva I. Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis. Cancer Epidemiol Biomark Prev. 2006;15:1159–69. 2. Yaghjyan L, Colditz G, Collins L, et al. Mammographic breast density and subsequent risk of breast cancer in postmenopausal women according to tumor characteristics. J Natl Cancer Inst. 2011;103:1179–89. 3. Conroy S, Pagano I, Kolonel L, et al. Mammographic density and hormone receptor expression in breast cancer: the Multiethnic Cohort Study. Cancer Epidemiol. 2011;35:448–52. 4. Sala E, Solomon L, Warren R, et al. Size, node status and grade of breast tumours: association with mammographic parenchymal patterns. Eur Radiol. 2000;10:157–61. 5. Li J, Humphreys K, Eriksson L, et al. Mammographic density reduction is a prognostic marker of response to adjuvant tamoxifen therapy in postmenopausal patients with breast cancer. J Clin Oncol. 2013;31:2249–56. 6. Kim J, Han W, Moon H-G, et al. Breast density change as a predictive surrogate for response to adjuvant endocrine therapy in hormone receptor positive breast cancer. Breast Cancer Res. 2012;14:R102. 7. Porter G, Evans A, Cornford E, et al. Influence of mammographic parenchymal pattern in screening-detected and interval invasive breast cancers on pathologic features, mammographic features, and patient survival. AJR Am J Roentgenol. 2007;188:676–83. 8. Olsen A, Bihrmann K, Jensen MB, et al. Breast density and outcome of mammography screening: a cohort study. Br J Cancer. 2009;100:1205–8. 9. Gierach G, Ichikawa L, Kerlikowske K, et al. Relationship between mammographic density and breast cancer death in the Breast Cancer Surveillance Consortium. J Natl Cancer Inst. 2012;104:1218–27. 10. Vanel D. The American College of Radiology (ACR) Breast Imaging and Reporting Data System (BI-RADS): a step towards a universal radiological language? Eur J Radiol. 2007;61:183. 11. Kerlikowske K, Grady D, Barclay J, et al. Variability and accuracy in mammographic interpretation using the American College of Radiology Breast Imaging Reporting and Data System. J Natl Cancer Inst. 1998;90:1801–09. 12. Martin K, Helvie M, Zhou C, et al. Mammographic density measured with quantitative computer-aided method: comparison with radiologists’ estimates and BI-RADS categories. Radiology. 2006;240:656–65.
Volumetric Density and Breast Cancer Risk 13. Assi V, Warwick J, Cuzick J, Duffy S. Clinical and epidemiological issues in mammographic density. Nat Rev Clin Oncol. 2012;9:33–40. 14. Highnam R, Brady M, Yaffe MJ, et al. Robust breast composition measurement – VolparaTM. Girona: Springer; 2010. p. 342–349. 15. Highnam R, Sauber N, Destounis S, et al. Breast density into clinical practice. Berlin: Springer; 2012. p. 466–473. 16. Park JH, Yang HK, Nam EJ, et al. Cancer facts & figures 2013. President of National Cancer Center, Minister for Health and Welfare; 2013. 17. Boyd N, Martin L, Yaffe M, Minkin S. Mammographic density and breast cancer risk: current understanding and future prospects. Breast Cancer Res. 2011;13:223. 18. Aitken Z, McCormack V, Highnam R, et al. Screen-film mammographic density and breast cancer risk: a comparison of the volumetric standard mammogram form and the interactive threshold measurement methods. Cancer Epidemiol Biomarkers Prev. 2010;19:418–28. 19. Jeffreys M, Warren R, Highnam R, Davey Smith G. Breast cancer risk factors and a novel measure of volumetric breast density: cross-sectional study. Br J Cancer. 2008;98:210–16. 20. Gweon H, Youk J, Kim J-A, Son E. Radiologist assessment of breast density by BI-RADS categories versus fully automated volumetric assessment. AJR Am J Roentgenol. 2013;201:692–7.
21. del Carmen M, Halpern E, Kopans D, et al. Mammographic breast density and race. AJR Am J Roentgenol. 2007; 188:1147–50. 22. Ursin G, Ma H, Wu A, et al. Mammographic density and breast cancer in three ethnic groups. Cancer Epidemiol Biomark Prev. 2003;12:332–38. 23. Thomas D, Carter R, Bush W, et al. Risk of subsequent breast cancer in relation to characteristics of screening mammograms from women less than 50 years of age. Cancer Epidemiol Biomarkers Prev. 2002;11:565–71. 24. Bertrand K, Tamimi R, Scott C, et al. Mammographic density and risk of breast cancer by age and tumor characteristics. Breast Cancer Res. 2013;15:104. 25. Antoni S, Sasco A, dosSantosSilva I, McCormack V (2013) Is mammographic density differentially associated with breast cancer according to receptor status? A meta-analysis. Breast Cancer Res Treat. 137:337–47. 26. Ghosh K, Brandt KR, Reynolds C, et al. Tissue composition of mammographically dense and non-dense breast tissue. Breast Cancer Res Treat. 2012;131:267–75. 27. Mao Y, Keller ET, Garfield DH, et al. Stromal cells in tumor microenvironment and breast cancer. Cancer Metastasis Rev. 2013;32:303–12.
