Sage Journals: Discover world-class research

Abstract

This study aimed to analyze the correlation between mammographic density obtained by density analysis software (DAS)/radiologists visual (RV) classification with molecular subtype, and the expression levels of estrogen receptor (ER), progesterone receptor (PR), Ki67 antigen (Ki-67), p53 gene (p53), and human epidermal growth factor receptor-2 (HER2). A total of 688 breast cancer patients with digital mammography and complete molecular pathological results in Tianjin Medical University Cancer Institute and Hospital between February 2015 and February 2016 were collected. The DAS-density grade (DASD) and the radiologists visually classified density grade (RVD) were evaluated by 3 radiologists. The correlation between density grade and the expression levels of ER, PR, Ki-67, p53, HER2 and breast cancer molecular subtype (PMS) were analyzed. The agreement between DASD and RVD was explored. ER, PR and HER-2 positive rate were significantly different among patients with different RVD grades ( $P<$ 0.05). HER2 positive rates showed an increasing trend following RVD upgrading ( $P_{\textit{trend}}<$ 0.05). HER-2 positive rate in RVD D1 $+$ D2 was 7.69%, which was higher than that in D3 $+$ D4 ( $P<$ 0.05). The ER and Ki-67 expressions in patients were markedly different among DASD ( $P=$ 0.009 and 0.002) and RVD ( $P=$ 0.012 and 0.036) with different grades. The kappa value of each DASD to RVD was 0.31 ( $P<$ 0.01). The RVD 3 proportion was 14.58% (63/432) in HER2 Over-expressing subtype, which was apparently higher than RVD1 (2.43%, 1/41) ( $P<$ 0.05). Breast density may be partial correlated with molecular pathology in breast cancer.

Keywords

Breast cancer mammographic density molecular pathology correlation

1. Introduction

Mammographic breast density is one of the strongest independent risk factors for breast cancer, which is even significant than age or family history [1]. Dense breast tissue not only increases the risk of breast cancer, but also makes cancer difficult to detect. Over the past 20 years, a number of reports focused on breast density measurements [2, 3, 4] because of its characteristics in breast cancer [5]. In most studies, epithelial and adipose tissue content was determined by X-ray attenuation, which in turn was used in the breast density detection given that breast density is the percentage of fibrous and glandular tissue in the breast [6]. Breast density is most often assessed by radiologist using bilateral mammogram images, whereas its accuracy is affected by radiologist’s experience [7]. The breast thickness and exposure factors are not considered when breast density is assessed by visual classification. On the contrary, density analysis software (DAS), fully automated measurement software approved by the FDA, takes filtration, target material, and imaging parameters (such as kV, mAs, and breast thickness) into consideration to calculate breast density [8].

Elsamany et al. showed that breast density is closely related to progression-free survival in patients diagnosed with malignant tumor [9]. Low breast density is associated with better survival outcomes in metastatic breast cancer patients over 40 years old [9]. In addition, numerous factors may focus on molecular pathology and prognosis of breast cancer. Indeed, HER2, p53, and Ki-67 have been shown to be important biomarkers for assessing the prognosis of invasive ductal carcinoma. HER2 and p53 overexpression was found to be associated with a worse prognosis in breast cancer [10]. Although breast density and biomarkers have been identified to be closely related to prognosis, no studies explore the linear relationship between breast density and biomarkers. In this study, we investigated the potential correlation between DASD and radiologists visually classified density grade (RVD) with pathological molecular characteristics in breast cancer.

2. Materials and methods

2.1 Subjects

A total of 688 invasive breast cancer patients with mean age at 46.4 (35–76) years old at our hospital between February 2015 and February 2016 were enrolled. All the subjects received digital mammography views [craniocaudal (CC) and mediolateral oblique (MLO) projections]. This research was performed in Tianjin Medical University Cancer Institute and Hospital and approved by the Research Ethics Committee of Tianjin Medical University Cancer Institute and Hospital. All the participants had signed informed consent.

2.2 Pathological classification

All patients with breast cancer were confirmed pathologically as invasive breast cancer by biopsy. They were divided into five subtypes according to genetic analysis and biological characteristics of the tumor (The 2011 St. Gallen International Expert Consensus). Specifically, they were type 1: Lumina A, ER and/or PR positive, HER2 negative, Ki-67 $<$ 14%; Type 2: Lumina B1, ER and/or PR positive, HER2 negative, Ki-67 $>$ 14%; Type 3: Lumina B2, ER and/or PR positive, HER2 overexpression (positive); Type 4: HER2 overexpression but ER and PR negative; Type 5: triple negative breast cancer, ER, PR, and HER2 negative. The breast density was measured by Density analysis software (DAS, Quantra ${}^{\text{TM}}$ , Hologic). The mammograms were collected using Hologic Selenia equipment and assessed by three radiologists (5, 8, and 30 years of reading experience for radiologist 1, 2, and 3, respectively) for RVD (Table 1). Based on ACR BI-RADS, the subjects were divided into 4 density grades (DASD). DASD 1: The breasts are almost entirely fatty (Fig. 1a); DASD 2: There are scattered areas in fibroglandular density (Fig. 1b); DASD 3: The breasts were heterogeneously dense, which may be obscure small masses (Fig. 1c); and DASD 4: The breasts were extremely dense, which lowers the sensitivity of mammography (Fig. 1d). The agreement of DASD and RVD was analyzed, too.

Table 1
Density grades evaluated by three radiologists

Radiologists	RVD 1	RVD 2	RVD 3	RVD 4
1	45	178	433	32
2	41	182	425	40
3	39	180	433	36

$P=$ 0.71.

Figure 1.

Breast RVD 1-4 based on visual classification, respectively.

In total 688 cases, there were 50, 241, 270, and 127 cases in DASD 1-4, respectively. In addition, the case number was 41, 180, 432, and 35 according to RVD criteria, respectively (Table 1).

2.3 Analysis of parenchymal density by DAS

The DAS algorithm was developed to calculate the density of breast [11]. This algorithm takes the parameter of each image, such as kVp, mAs, and target, into the overall consideration of breast density. The two breast densities were calculated separately and the mean density of two breasts was used in the study. The consistent of density classification of glandular tissue between DASD and RVD was examined.

2.4 ER and PR expression detected by immunohistochemical staining

ER and PR were localized in the nuclei of the tumor cells. Monoclonal antibody 1D5 (M7047, DakoCytomation, Carpinteria, CA) was adopted to detect ER, while monoclonal anti-PR antibody 636 (M3569, DakoCytomation) was used for PR detection. The slides were incubated with primary antibody and labeled polymer (Envision ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}\textregistered}$ $+$ dual link) for 30 minutes, respectively, and further stained by 3, 3-diaminoben- zidine (DAB), hematoxylin, and hematoxylin for 5 minutes, respectively. The positive criterion was defined as the yellow or brown particles appeared in cell nucleus. Positive nuclear reaction for ER and PR, irrespective of percentage of reactive cells, was recorded as positive [12]. Totally, one hundred tumor cells were observed per high magnification field to calculate the proportion of positive cells. ER and PR positivity was defined as $>$ 1% of tumor cell nuclei staining.

2.5 HER2 expression detected by the immunohistochemical method

HercepTest was used for the antigen retrieval of HER2 [12]. The criterion of positively-stained tumor cells was that brown-yellow particles located on the cytomembrane after staining. One hundred tumor cells were observed per high magnification field to calculate the proportion of positive cells. The tissues were graded at four levels according to color development: 1) negative grade (-), no membrane or background staining observed; 2) weakly positive grade (+), 25%–50% of membrane staining in tumor cells; 3) positive (++), 51%–75% of membrane staining in tumor cells; and 4) strongly positive (+++), $>$ 75% of membrane staining in tumor cells. As opposed to low (1+) or none (-) for HER2 in more than 10% of the tumor cells was used as the cut-off for tumor’s HER2 positivity

2.6 Ki-67 and p53 expression

The slides were incubated overnight with a mouse monoclonal antibody against Ki-67 antigen (MIB-1, 1:100; Dako, Glostrup, Denmark) and a mouse monoclonal antibody against the p53 antigen clone JC70A (pre-diluted, Dako). The Envision Dual link system-HRP (ready to use, Dako) was selected as the secondary antibody. Then the slides were incubated in 3, 3-diaminobenzidine tetrahydrochloride for 10 min as a substrate chromogen solution to produce a brown color. Finally, the slides were counterstained with Mayer’s hematoxylin [13].

Slides were evaluated under a light microscope. The criterion of positively-stained tumor cells was that the nucleus showed brown-yellow by anti-Ki-67 antibody. The proliferation index (PI) of tumor cells was calculated as follows: PI $=$ (number of positive cells/100) $\times$ 100%. In non-tumoral tissues, the Ki-67 labeling was weaker when it was limited to the basal layer, whereas it was higher in the upper layers. Cases were considered Ki67-positive if the sum of the percentage of Ki67-positive nuclei across cores $>$ 14%. p53 was localized in tumor cell nuclei. The positive criterion was the cell nucleus particles showing color in brown-yellow. The p53 staining reactions were analyzed in 100 cancer cells of each sample (original magnification, $\times$ 400).

3. Statistical method

The kappa statistic was used to measure the agreement between DASD and RVD: 0.00–0.20 indicated slight agreement, 0.21–0.40 referred to fair agreement, 0.41–0.60 implied moderate agreement, 0.61–0.80 showed substantial agreement, and 0.81–1.00 exhibited almost perfect agreement. One-way ANOVA was adopted to analyze the difference of the biomarkers of each density grade among groups. Student t-test was performed to calculate the significant changes of the biomarkers between grades.

Chi square test was selected to analyze the four DASD and RVD proportions in different molecular subtype. A $P<$ 0.05 was considered to be statistically significant. All analyses except the test of Cochran-Armitage were performed using SAS software (version 9.2, SAS Institute, Cary, NC, USA).

4. Results

4.1 Biomarker profiles according to DASD and RVD

Biomarker qualitative analysis results showed that no statistically significant difference in positive rate of biomarker was found among DASD density ( $P>$ 0.05).

As the breast density increased, pathological biomarkers did not increase in the same trend ( $P_{\textit{trend}}>$ 0.05). ER, PR and HER-2 positive rate was statistical different among RVD groups ( $P<$ 0.05). HER-2 positive rate exhibited an increasing trend as the RVD upgrade ( $P_{\textit{trend}}<$ 0.05) (Table 2).

Table 2
Association of breast density with the tumor marker of breast cancer patients

Biomarker	DASD	Cases	Positive rate	OR (95% CI)	RVD	Biomarker	Positive rate	OR (95% CI)
	Total	532			Total	532
ER	1	38	76.00%	1.00 (referent)	1	36	87.80%	1.00 (referent)
	2	192	79.67%	0.81 (0.39–1.66)	2	144	80.00%	1.80 (0.66–4.91)
	3	206	76.30%	0.98 (0.49–1.99)	3	321	74.31%	2.49 (0.95–6.51)
	4	96	75.59%	1.02 (0.48–2.20)	4	31	88.575	0.92 ${}^{\Delta}$ (0.23–3.77)
	$P_{\textit{trend}}$	0.487	$P\chi^{2}$	0.757	$P_{\textit{trend}}$	0.195	$P\chi^{2}$	0.026
PR	Total	471			Total	471
	1	38	76.00%	1.00 (referent)	1	32	78.05%	1.00 (referent)
	2	164	68.05%	1.47 (0.74–3.00)	2	130	72.22%	1.37 (0.61–3.07)
	3	178	65.93%	1.64 (0.81–3.28)	3	280	64.81%	1.93 (0.90–4.15)
	4	91	71.65%	1.25 (0.59–2.67)	4	29	82.86% ${}^{\Delta}$	0.74 (0.23–2.32)
	$P_{\textit{trend}}$	0.813	$P\chi^{2}$	0.433	$P_{\textit{trend}}$	0.236	$P\chi^{2}$	0.030
HER-2	Total	276			Total	276
	1	22	44.00%	1.00 (referent)	1	13	31.71%	1.00 (referent)
	2	88	36.51%	1.37 (0.74–2.53)	2	61	33.89%	1.10 (5.22–2.32)
	3	107	39.63%	1.20 (0.65–2.20)	3	188	43.52% ${}^{*\#}$	0.73 (0.36–1.49)
	4	59	46.46%	0.91 (0.47–1.75)	4	14	40.00%	0.85 (0.33–2.21)
	$P_{\textit{trend}}$	0.242	$P\chi^{2}$	0.288	$P_{\textit{trend}}$	0.006	$P\chi^{2}$	0.024
Ki-67	Total	597			Total	597
	1	45	90.00%	1.00 (referent)	1	34	82.93%	1.00 (referent)
	2	203	84.23%	1.69 (0.63–4.52)	2	154	85.56%	0.82 (0.33–2.04)
	3	239	88.52%	1.17 (0.43–3.16)	3	378	87.50%	0.69 (0.29–1.64)
	4	110	86.61%	1.39 (0.48–3.99)	4	31	88.57%	0.62 (0.17–2.35)
	$P_{\textit{trend}}$	0.7224	$P\chi^{2}$	0.470	$P_{\textit{trend}}$	0.674	$P\chi^{2}$	0.787

Table 3

Differences comparison of biomarker expression in different breast density analyzed by density analysis software (mean $\pm$ SD)

DAS	n	ER	PR	HER2	Ki-67	p53
D1	50	64.84 $\pm$ 37.48	41.36 $\pm$ 37.23	1.44 $\pm$ 0.86	27.56 $\pm$ 18.59	17.58 $\pm$ 28.33
D2	241	64.59 $\pm$ 36.05	39.97 $\pm$ 36.62	1.22 $\pm$ 0.89	32.66 $\pm$ 20.53	19.00 $\pm$ 29.27
D3	270	54.59 $\pm$ 37.25 ${}^{*}$	36.94 $\pm$ 35.97	1.32 $\pm$ 0.83	37.94 $\pm$ 23.33 ${}^{*}$	21.54 $\pm$ 31.01
D4	127	55.69 $\pm$ 37.25 ${}^{*}$	45.63 $\pm$ 39.28 ${}^{\Delta}$	1.43 $\pm$ 0.88 ${}^{\#}$	41.28 $\pm$ 26.52 ${}^{*}$	26.01 $\pm$ 34.24 ${}^{*}$
$F$ value		63.34	1.63	2.02	4.934	1.695
$P$ value		0.009	0.182	0.109	0.002	0.167

${}^{*}$ Compared to D1, $P<$ 0.05; ${}^{\#}$ Compared to D2, $P<$ 0.05; ${}^{\Delta}$ Compared to D3, $P<$ 0.05.

Table 4

The differences comparison of biomarker expression in different breast density analyzed visually (mean $\pm$ SD)

RVD	n	ER	PR	HER2	Ki-67	p53
D1	41	71.58 $\pm$ 28.92	43.29 $\pm$ 35.81	1.22 $\pm$ 0.75	38.50 $\pm$ 22.48	20.34 $\pm$ 31.41
D2	180	63.77 $\pm$ 36.19	40.17 $\pm$ 35.54	1.21 $\pm$ 0.85	32.68 $\pm$ 21.12	17.04 $\pm$ 26.72
D3	432	55.79 $\pm$ 38.55 ${}^{*{\#}}$	38.51 $\pm$ 37.72	1.38 $\pm$ 0.89 ${}^{\#}$	37.70 $\pm$ 23.55	22.64 $\pm$ 32.27
D4	35	60.40 $\pm$ 31.02	52.23 $\pm$ 34.99 ${}^{\Delta}$	1.28 $\pm$ 0.83	38.30 $\pm$ 23.57 ${}^{\#}$	25.71 $\pm$ 32.30
$F$ value		3.655	1.62	1.76	2.87	1.66
$P$ value		0.012	0.183	0.15	0.036	0.173

${}^{*}$ : Compared to D1, there was significant difference, $P<$ 0.05; ${}^{\#}$ : Compared to D2, there was significant difference, $P<$ 0.05; ${}^{\Delta}$ : Compared to D3, there was significant difference, $P<$ 0.05.

Table 5

Agreement analysis of breast density grade between DASD and RVD

	RVD
	1	2	3	4	Total
DASD
1	15	27	8	0	50
2	22	115	104	0	241
3	4	34	224	8	270
4	0	4	96	27	127
Total	41	180	432	35	688

Table 6

Molecular subtypes proportion of different DASDs

DASD	Molecular subtype
	Lumina A	Lumina B	Her-2 over	Triple
			expression	negative
1 ( $n=$ 50)	3 (6.00)	33 (66.00)	4 (8.00)	10 (20.00)
2 ( $n=$ 241)	33 (13.69)	156 (64.73)	29 (12.03)	23 (9.54)
3 ( $n=$ 270)	25 (9.26)	184 (68.15)	30 (11.11)	31 (11.48) ${}^{*}$
4 ( $n=$ 127)	10 (7.87)	87 (68.50)	20 (15.75)	10 (7.87) ${}^{*}$
$\chi^{2}$ value	5.12	0.867	2.63	6.07
$P$ value	0.16	0.833	0.45	0.11

${}^{*}$ Compared to D1, $P<$ 0.05.

Figure 2.

RVD 2 but DASD 1(13%) (Left); RVD 3 but DASD 4 (76%) (Right).

The ER and Ki-67 expressions in patients were significantly different among four DASD ( $P=$ 0.009 and 0.002) and RVD ( $P=$ 0.012 and 0.036). ER expression in D2 (64.59 $\pm$ 36.05%) was significant higher than that in DASD 3 (54.59 $\pm$ 37.25%) and DASD 4 (55.69 $\pm$ 37.25%) ( $P<$ 0.05). HER2 and p53 expressions in DASD 2 were 1.22 $\pm$ 0.89% and 19.00 $\pm$ 29.27%, which were obviously lower than those in DASD 4 as 1.43 $\pm$ 0.88% and 26.01 $\pm$ 34.24% ( $P<$ 0.05) (Table 3). ER expression in RVD 3 was markedly higher than RVD 1 and RVD 2 ( $P<$ 0.05). HER2 and Ki-67 expressions in RVD 2 were apparently lower than that in RVD2 and RVD4 ( $P<$ 0.05) (Table 4).

4.2 The agreement of DASD and RVD

The density agreement between DASD and RVD was showed in Table 5. The kappa value in each density grade between them was 0.31, indicating a fair agreement and the density was quite distinctive between DASD and RVD. For example (Fig. 2), the density of case 1 was classified into RVD 2, while it was defined as DASD 1 (13% of density). In addition, the density of case 2 was classified into RVD 3 but was defined as DASD 4 (76% of density).

4.3 DASD and RVD proportion in different molecular subtypes

DASD 1 proportion in Triple negative was 20.00% (10/50), which was significant higher than DASD3 (11.48%, 31/270) and DSAD4 (7.87%, 10/127) ( $P<$ 0.05) (Table 6). RVD 3 proportion in HER2 Over-expressing subtype was 14.58% (63/432), which was obviously higher than RVD1 (2.43%, 1/41) ( $P<$ 0.05) (Table 7). There was no statistical difference among DASD groups and RVD groups. Moreover, there was no statistically significant difference between Lumina and non-lumina ratios between DASD and RVD ( $P>$ 0.05). The HER-2 positive rate in D1 $+$ D2 of DASD and RVD was markedly higher than D3 $+$ D4 ( $P<$ 0.05).

Table 7
Molecular subtypes proportion of different RVDs

RVD	Molecular subtype
	Lumina A	Lumina B	Her-2 over	Triple
			expression	negative
1 ( $n=$ 41)	5 (12.20)	30 (73.17)	1 (2.43)	5 (12.20)
2 ( $n=$ 180)	23 (12.78)	120 (66.67)	16 (8.90)	21 (11.67)
3 ( $n=$ 432)	39 (9.03)	283 (65.51)	63 (14.58) ${}^{*}$	47 (10.88)
4 ( $n=$ 35)	4 (11.43)	27 (77.14)	3 (8.57)	1 (8.57%)
$\chi^{2}$ value	2.15	2.765	6.98	2.52
$P$ value	0.54	0.42	0.07	0.47

${}^{*}$ Compared to D1, $P<$ 0.05.

5. Discussion

Numerous studies showed that breast density decreases following age, BMI, and perimenopause, but increases according to menopause, smoking, hormone therapy, and excessive alcohol consumption [14]. Breast density detected by mammography reflected the percentage of breast volume that is occupied by dense breast tissue. Various researches indicated that breast density is a high-risk factor for breast cancer. High breast density may not only suggest an early breast cancer risk, but also an independent risk factor in addition to age and BRCA [15]. High breast density was most significant among all risk factors due to its prevalence [3, 16, 17, 18, 19, 20]. A substantial percentage of women with high-density breasts develop breast cancer, which is independent of age, menopausal status, and hormone replacement therapy [19]. Thus, it is necessary to amend the screening process (higher frequency of screening combined ultrasonography and magnetic resonance) for women with high breast density. Objective evaluation of breast density contributes to the assessment of breast cancer risk and severity, which facilitates the development of a more rational screening and treatment regimen. It is therefore critical to report breast density to guide women with high breast density for further examinations, such as ultrasonography or MRI. In Connecticut, the government has legislation requiring radiologists to report breast density and communicate the findings with patients [21]. Although the applications are still immature, the legislation will also effectively promote the development of personalized diagnostic imaging and management of patients with high-density breasts.

To objectively assess breast density, ACR published breast imaging reporting and data systems in 2013. However, a large number of studies found low agreement in the mammogram reading between different radiologists [22, 23]. Recently, computer-aided detection was widely used in the mammogram examination of the breasts, which separated the measurement of density and fat regions from visual observation [24, 25]. Among 63 biopsies and aspirations, three lesions were malignant (all BI-RADS category 4, diagnosed with biopsy). All three cancers were smaller than 1 cm, were found in postmenopausal patients, and were solid masses. However, there are some limitations with the method of binary computation. Its calculation classified the mammogram image into fat or substance, which did not take the exposure parameters, the details of half-value layers, and the thickness of solid breasts into account [26]. DAS can overcome these limitations and considered the kVp, mAs, target, filtered material, energy deposition, and other factors together. These factors were reflected on the pixels in the detection, which was used to calculate the entire breast volume, fibrous and glandular tissue to obtain the breast density from mammography. The first comprehensive independent study on DAS was accomplished by Ciatto et al. [23], which analyzed 481 cases of mammography reported by 11 radiologists and the performance of breast density assessment by DAS using two classifications (density 1–2, and density 3–4). The new version of DAS (version 2.1.1) can classify density into four classes with good reproducibility on the analysis of breast volume and density and can be treated as an important aid to BI-RADs rating [27, 28, 29, 30]. Classification of breast density by DAS and radiologists by visual observation showed a fair agreement with a kappa value of 0.233, which was consistent with the previous reports [8, 31].

A variety of receptors played important roles in regulating the growth and differentiation of breast cells. Among the receptors, ER and PR were present in large amounts in the nuclei of mammary epithelial cells. Some cells still retain these receptors during tumorigenesis. There is still controversy regarding the effects of neoadjuvant chemotherapy on the ER and PR expression. It was revealed that compared with the ER and PR positive patients, the negative ones are more sensitive to neoadjuvant chemotherapy [32].

HER2 was a gene that can play a role in the breast cancer development. HER2 receptors normally regulated breast cell growth, division, and self-repairing. Breast cancer in patients with HER2-positive tended to grow faster and was more likely to metastasis and recurrence compared with HER2-negative breast cancers [33]. Ki-67, a nuclear-associated antigen, was associated with cell proliferation and mitosis. A number of studies indicated it as a breast cancer evaluation index for prognosis and clinical outcome of neoadjuvant therapy [34, 35]. Expression of Ki-67 and the proliferation index (PI) were significantly reduced after neoadjuvant chemotherapy. The clinical outcome of neoadjuvant chemotherapy in breast cancer patients with high Ki-67 expression was better than that of low expression [36]. P53, a tumor suppressor gene, can be divided into wild-type and mutant p53. When the p53 gene is mutated, it loses regulation of cell growth, apoptosis, and cell repair [36]. The histopathologic correlation of mammographic density were thought to represent dense connective tissues in addition to epithelial cells, but the amount of connective tissue was far larger than glandular tissue and contributed more to the variability in percentage of dense area [37]. It was considered that the extracellular matrix (ECM) contributes to neoplastic progression and disruptions in the ECM may precede epithelial changes. Therefore, it may enhance tumor formation through epithelial-stromal interactions [38].

Biomarker data of this study indicated that there was no difference in the positive rate of all biomarkers among different DASD, and the positive rate did not elevate with the increase of density. The positive rate of HER-2 was increasing with the increase of RVD density. Research showed that the positive rate of biomarker (ER, PR, HER-2, Ki-67, etc.) in patients with breast cancer augmented with the increase of density [39], which was not consistent with the results of our study. ER expression was reduced in trend, whereas PR, HER2, and p53 expressions were enhanced with DAS densities. Ki-67 exhibited no statistical tendency. Therefore, DASD may reflect biomarker expression levels to predict patients’ outcome. Nishimura and others [24, 40] reported the results of 372 patients with breast cancer and showed that the mean survival rate of TNBC, HER2 overexpression, Lumina A and Lumina B types were 63.3, 64.4, 73.4, and 77.6 months, respectively. The prognosis of TNBC and HER2 overexpression were the worst among these sub-types. The results from this study demonstrated that there was no significant difference in the four different molecular subtypes among DASD and RVD groups ( $P>$ 0.05). However, DASD 1 proportion in Triple negative was significant higher than DASD3 and DSAD4, while RVD 3 proportion in HER2 Over-expressing was obviously higher than RVD 1.

Breast cancer is a heterogeneous disease composed of at least four major subtypes, namely luminal-A and luminal-B breast cancer, basal-like, and HER2-like breast cancer [41]. The subgroups were differed by expressions of ER and PR, HER2 expression/amplification status, and the proliferative activity of the tumor [42]. Basal-like breast cancer was often approximated with the triple negative breast cancer subtype [43]. Triple negative breast cancer was characterized by an adverse prognosis, particularly in case of limited sensitivity against neoadjuvant chemotherapy [44]. It was well described that molecular breast cancer subtypes were associated with significant differences in prognosis [45]. Similarly, a significant association between age at diagnosis and prognosis was well known [46]. Younger age at diagnosis was generally considered to be associated with an adverse disease prognosis [47].

Though number of studies indicated that high breast density is a high-risk factor for breast cancer, this study showed high breast density was not associated with worse molecular subtypes. In triple negative patients, lowest density accounted for the highest proportion. ER and Ki-67 expressions in patients were significantly different among DASD groups and among RVD groups, while it was hard to find the trend as the density increased. These changes in trend were not clear in RVD. DAS, taking kVp, mAs, target and filter materials, and energy deposited into account in the image detector in each pixel (as a surrogate for the thickness of breast tissue traversed) to calculate the volume of the entire breast, the fibroglandular tissue, and the resultant mammographic density, overcame the limitation of RVD. This study indicated DASD and RVD agreement was not ideal enough.

6. Conclusion

DAS is an easy-to-be used a computer program to calculate breast density as it takes kVp, mAs, target and filter materials, and energy deposited into account. Unideal DASD and RVD agreement was reasonable. Though two density assessments were inconsistent, the ER and Ki-67 expression of both insisted significant difference. This study found that there was no strong correlation between density and molecular subtypes. However, lowest DASD account for the highest proportion in triple negative patients. Since triple negative was the worst prognosis molecular subtype, DASD 1 may indicate worse prognosis.

Footnotes

Acknowledgments

This work was supported by the The Science and Technology Project of Anticancer, Tianjin, China [grant numbers 12ZCDZSY16000]; The Chinese National Key Scientific and Technoloical Project [grant numbers 2013BAI09B08].

Conflict of interest

All authors declare that they have no conflict of interest.

References

McLean

K.E.

and Stone

, Role of breast density measurement in screening for breast cancer, Climacteric (2018), 1–7.

Assi

Warwick

Cuzick

and Duffy

S.W.

, Clinical and epidemiological issues in mammographic density, Nat Rev Clin Oncol 9 (2011), 33–40.

McCormack

V.A.

and dos Santos Silva

, Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis, Cancer Epidemiol Biomarkers Prev 15 (2006), 1159–1169.

Boyd

N.F.

Byng

J.W.

Jong

R.A.

Fishell

E.K.

Little

L.E.

Miller

A.B.

Lockwood

G.A.

Tritchler

D.L.

and Yaffe

M.J.

, Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study, J Natl Cancer Inst 87 (1995), 670–675.

Boyd

N.F.

Guo

Martin

L.J.

Sun

Stone

Fishell

Jong

R.A.

Hislop

Chiarelli

Minkin

and Yaffe

M.J.

, Mammographic density and the risk and detection of breast cancer, N Engl J Med 356 (2007), 227–236.

Johns

P.C.

and Yaffe

M.J.

, X-ray characterisation of normal and neoplastic breast tissues, Phys Med Biol 32 (1987), 675–695.

Regini

Mariscotti

Durando

Ghione

Luparia

Campanino

P.P.

Bianchi

C.C.

Bergamasco

Fonio

and Gandini

, Radiological assessment of breast density by visual classification (BI-RADS) compared to automated volumetric digital software (Quantra): implications for clinical practice, Radiol Med 119 (2014), 741–749.

Pahwa

Hari

Thulkar

and Angraal

, Evaluation of breast parenchymal density with QUANTRA software, Indian J Radiol Imaging 25 (2015), 391–396.

Elsamany

Alzahrani

Elkhalik

S.A.

Elemam

Rawah

Farooq

M.U.

H.A.

and F

K.O.

, Prognostic value of mammographic breast density in patients with metastatic breast cancer, Med Oncol 31 (2014), 96.

10.

Yamashita

Nishio

Toyama

Sugiura

Zhang

Kobayashi

and Iwase

, Coexistence of HER2 over-expression and p53 protein accumulation is a strong prognostic molecular marker in breast cancer, Breast Cancer Res 6 (2004), R24–30.

11.

Yuan

Giger

M.L.

Bhooshan

and Sennett

C.A.

, Multimodality computer-aided breast cancer diagnosis with FFDM and DCE-MRI, Acad Radiol 17 (2010), 1158–1167.

12.

Kinsella

M.D.

Nassar

Siddiqui

M.T.

and Cohen

, Estrogen receptor (ER), progesterone receptor (PR), and HER2 expression pre- and post-neoadjuvant chemotherapy in primary breast carcinoma: a single institutional experience, Int J Clin Exp Pathol 5 (2012), 530–536.

13.

Kobayashi

Iwaya

Moriya

Yamasaki

Tsuda

Yamamoto

and Matsubara

, A simple immunohistochemical panel comprising 2 conventional markers, Ki67 and p53, is a powerful tool for predicting patient outcome in luminal-type breast cancer, BMC Clin Pathol 13 (2013), 5.

14.

Haars

van Noord

P.A.

van Gils

C.H.

Grobbee

D.E.

and Peeters

P.H.

, Measurements of breast density: no ratio for a ratio, Cancer Epidemiol Biomarkers Prev 14 (2005), 2634–2640.

15.

Passaperuma

Warner

Hill

K.A.

Gunasekara

and Yaffe

M.J.

, Is mammographic breast density a breast cancer risk factor in women with BRCA mutations? J Clin Oncol 28 (2010), 3779–3783.

16.

Ciatto

Bernardi

Calabrese

Durando

Gentilini

M.A.

Mariscotti

Monetti

Moriconi

Pesce

Roselli

Stevanin

Tapparelli

and Houssami

, A first evaluation of breast radiological density assessment by QUANTRA software as compared to visual classification, Breast 21 (2012), 503–506.

17.

Choi

Woo

O.H.

Shin

H.S.

Jang

Cho

K.R.

and Seo

B.K.

, Comparison of two-dimensional synthesized mammogram (2DSM) and conventional full-field digital mammogram (FFDM) for evaluation of breast cancer, Clin Imaging 43 (2017), 170–174.

18.

Veronesi

Boyle

Goldhirsch

Orecchia

and Viale

, Breast cancer, Lancet 365 (2005), 1727–1741.

19.

Boyd

N.F.

Rommens

J.M.

Vogt

Lee

Hopper

J.L.

Yaffe

M.J.

and Paterson

A.D.

, Mammographic breast density as an intermediate phenotype for breast cancer, Lancet Oncol 6 (2005), 798–808.

20.

Mandelson

M.T.

Oestreicher

Porter

P.L.

White

Finder

C.A.

Taplin

S.H.

and White

, Breast density as a predictor of mammographic detection: comparison of interval- and screen-detected cancers, J Natl Cancer Inst 92 (2000), 1081–1087.

21.

Hooley

R.J.

Greenberg

K.L.

Stackhouse

R.M.

Geisel

J.L.

Butler

R.S.

and Philpotts

L.E.

, Screening US in patients with mammographically dense breasts: initial experience with Connecticut Public Act 09-41, Radiology 265 (2012), 59–69.

22.

Spayne

M.C.

Gard

C.C.

Skelly

Miglioretti

D.L.

Vacek

P.M.

and Geller

B.M.

, Reproducibility of BI-RADS breast density measures among community radiologists: a prospective cohort study, Breast J 18 (2012), 326–333.

23.

Ciatto

Houssami

Apruzzese

Bassetti

Brancato

Carozzi

Catarzi

Lamberini

M.P.

Marcelli

Pellizzoni

Pesce

Risso

Russo

and Scorsolini

, Categorizing breast mammographic density: intra- and interobserver reproducibility of BI-RADS density categories, Breast 14 (2005), 269–275.

24.

Freedman

San Martin

O’Gorman

Eckert

Lippman

M.E.

S.C.

Walls

E.L.

and Zeng

, Digitized mammography: a clinical trial of postmenopausal women randomly assigned to receive raloxifene, estrogen, or placebo, J Natl Cancer Inst 93 (2001), 51–56.

25.

Martin

K.E.

Helvie

M.A.

Zhou

Roubidoux

M.A.

Bailey

J.E.

Paramagul

Blane

C.E.

Klein

K.A.

Sonnad

S.S.

and Chan

H.P.

, Mammographic density measured with quantitative computer-aided method: comparison with radiologists’ estimates and BI-RADS categories, Radiology 240 (2006), 656–665.

26.

Kopans

D.B.

, Basic physics and doubts about relationship between mammographically determined tissue density and breast cancer risk, Radiology 246 (2008), 348–353.

27.

Morrish

O.W.

Tucker

Black

Willsher

Duffy

S.W.

and Gilbert

F.J.

, Mammographic breast density: comparison of methods for quantitative evaluation, Radiology 275 (2015), 356–365.

28.

Singh

J.M.

Fallenberg

E.M.

Diekmann

Renz

D.M.

Witlandt

Bick

and Engelken

, Volumetric breast density assessment: reproducibility in serial examinations and comparison with visual assessment, Rofo 185 (2013), 844–848.

29.

Engelken

Singh

J.M.

Fallenberg

E.M.

Bick

Bottcher

and Renz

D.M.

, Volumetric breast composition analysis: reproducibility of breast percent density and fibroglandular tissue volume measurements in serial mammograms, Acta Radiol 55 (2014), 32–38.

30.

Schmachtenberg

Hammann-Kloss

Bick

and Engelken

, Intraindividual comparison of two methods of volumetric breast composition assessment, Acad Radiol 22 (2015), 447–452.

31.

Zhou

X.L.

Fan

Yang

and Yu

M.X.

, The clinical significance of PR, ER, NF-kappa B, and TNF-alpha in breast cancer, Dis Markers 2014 (2014), 494581.

32.

Petit

Wilt

Velten

Millon

Rodier

J.F.

Borel

Mors

Haegele

Eber

and Ghnassia

J.P.

, Comparative value of tumour grade, hormonal receptors, Ki-67, HER-2 and topoisomerase II alpha status as predictive markers in breast cancer patients treated with neoadjuvant anthracycline-based chemotherapy, Eur J Cancer 40 (2004), 205–211.

33.

Harbeck

, Advances in targeting HER2-positive breast cancer, Curr Opin Obstet Gynecol 30 (2018), 55–59.

34.

Mannell

, The role of Ki-67 in breast cancer, S Afr J Surg 54 (2016), 10–13.

35.

Han

Wang

Tian

Yao

Yuan

Qian

Zou

Zhou

and Yang

, Prognostic values of Ki-67 in neoadjuvant setting for breast cancer: a systematic review and meta-analysis, Future Oncol 13 (2017), 1021–1034.

36.

Wang

X.Z.

Liu

Sun

J.J.

Zuo

W.S.

D.W.

S.G.

D.B.

and Yu

Z.Y.

, Correlation between p53 and epidermal growth factor receptor expression in breast cancer classification, Genet Mol Res 14 (2015), 4282–4290.

37.

Boyd

N.F.

, Mammographic density and risk of breast cancer, Am Soc Clin Oncol Educ Book (2013).

38.

Maskarinec

Pagano

I.S.

Little

M.A.

Conroy

S.M.

Park

S.Y.

and Kolonel

L.N.

, Mammographic density as a predictor of breast cancer survival: the Multiethnic Cohort, Breast Cancer Res 15 (2013), R7.

39.

Yaghjyan

Pettersson

Colditz

G.A.

Collins

L.C.

Schnitt

S.J.

Beck

A.H.

Rosner

Vachon

and Tamimi

R.M.

, Postmenopausal mammographic breast density and subsequent breast cancer risk according to selected tissue markers, Br J Cancer 113 (2015), 1104–1113.

40.

Nishimura

Osako

Okumura

Tashima

Toyozumi

and Arima

, Changes in the ER, PgR, HER2, p53 and Ki-67 biological markers between primary and recurrent breast cancer: discordance rates and prognosis, World J Surg Oncol 9 (2011), 131.

41.

Perou

C.M.

Sorlie

Eisen

M.B.

van de Rijn

Jeffrey

S.S.

Rees

C.A.

Pollack

J.R.

Ross

D.T.

Johnsen

Akslen

L.A.

Fluge

Pergamenschikov

Williams

Zhu

S.X.

Lonning

P.E.

Borresen-Dale

A.L.

Brown

P.O.

and Botstein

, Molecular portraits of human breast tumours, Nature 406 (2000), 747–752.

42.

Reis-Filho

J.S.

and Pusztai

, Gene expression profiling in breast cancer: classification, prognostication, and prediction, Lancet 378 (2011), 1812–1823.

43.

Carey

Winer

Viale

Cameron

and Gianni

, Triple-negative breast cancer: disease entity or title of convenience? Nat Rev Clin Oncol 7 (2010), 683–692.

44.

Sorlie

Perou

C.M.

Tibshirani

Aas

Geisler

Johnsen

Hastie

Eisen

M.B.

van de Rijn

Jeffrey

S.S.

Thorsen

Quist

Matese

J.C.

Brown

P.O.

Botstein

Lonning

P.E.

and Borresen-Dale

A.L.

, Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications, Proc Natl Acad Sci U S A 98 (2001), 10869–10874.

45.

Prat

Ellis

M.J.

and Perou

C.M.

, Practical implications of gene-expression-based assays for breast oncologists, Nat Rev Clin Oncol 9 (2011), 48–57.

46.

Kast

Link

Friedrich

Petzold

Niedostatek

Schoffer

Werner

Klug

S.J.

Werner

Gatzweiler

Richter

Baretton

and Wimberger

, Impact of breast cancer subtypes and patterns of metastasis on outcome, Breast Cancer Res Treat 150 (2015), 621–629.

47.

Colak

Nofal

Albakheet

Nirmal

Jeprel

Eldali

Al-Tweigeri

Tulbah

Ajarim

Malik

O.A.

Inan

M.S.

Kaya

Park

B.H.

and Bin Amer

S.M.

, Age-specific gene expression signatures for breast tumors and cross-species conserved potential cancer progression markers in young women, PLoS One 8 (2013), e63204.

The correlation between mammographic densities and molecular pathology in breast cancer

Abstract

Keywords

1. Introduction

2. Materials and methods

2.1 Subjects

2.2 Pathological classification

Table 1 Density grades evaluated by three radiologists

2.4 ER and PR expression detected by immunohistochemical staining

2.5 HER2 expression detected by the immunohistochemical method

2.6 Ki-67 and p53 expression

3. Statistical method

4. Results

4.1 Biomarker profiles according to DASD and RVD

Table 2 Association of breast density with the tumor marker of breast cancer patients

4.3 DASD and RVD proportion in different molecular subtypes

Table 7 Molecular subtypes proportion of different RVDs

6. Conclusion

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Density grades evaluated by three radiologists

Table 2
Association of breast density with the tumor marker of breast cancer patients

Table 7
Molecular subtypes proportion of different RVDs