The Association of Measured Breast Tissue Characteristics with Mammographic Density and Other Risk Factors for Breast Cancer (original) (raw)

Skip Nav Destination

Research Articles| February 25 2005

Tong Li;

1Epidemiology and Statistics and Divisions of

Search for other works by this author on:

Limei Sun;

1Epidemiology and Statistics and Divisions of

Search for other works by this author on:

Naomi Miller;

3Department of Pathology, Princess Margaret Hospital, Toronto, Ontario, Canada

Search for other works by this author on:

Trudey Nicklee;

3Department of Pathology, Princess Margaret Hospital, Toronto, Ontario, Canada

Search for other works by this author on:

Jennifer Woo;

3Department of Pathology, Princess Margaret Hospital, Toronto, Ontario, Canada

Search for other works by this author on:

Lee Hulse-Smith;

3Department of Pathology, Princess Margaret Hospital, Toronto, Ontario, Canada

Search for other works by this author on:

Ming-Sound Tsao;

3Department of Pathology, Princess Margaret Hospital, Toronto, Ontario, Canada

Search for other works by this author on:

Rama Khokha;

2Experimental Therapeutics, Ontario Cancer Institute and

Search for other works by this author on:

Lisa Martin;

1Epidemiology and Statistics and Divisions of

Search for other works by this author on:

Norman Boyd

1Epidemiology and Statistics and Divisions of

Search for other works by this author on:

Requests for reprints: Norman F. Boyd, Division of Epidemiology and Statistics, Ontario Cancer Institute, 610 University Avenue, Toronto, Ontario, Canada M5G 2M9. Phone: 416-946-2945; Fax: 416-946-2024. E-mail: boyd@oci.utoronto.ca

Received: June 30 2004

Revision Received: September 09 2004

Accepted: September 20 2004

Online ISSN: 1538-7755

Print ISSN: 1055-9965

American Association for Cancer Research

2005

Cancer Epidemiol Biomarkers Prev (2005) 14 (2): 343–349.

Article history

Revision Received:

September 09 2004

Accepted:

September 20 2004

• Split-Screen
• Views Icon Views
  • Article contents
  • Figures & tables
  • Video
  • Audio
  • Supplementary Data
  • Peer Review
• Open the PDF for in another window
• Tools Icon Tools
• Search Site
• Article Versions Icon Versions
  • Version of Record February 25 2005

Abstract

Background: We have examined the relationships between the measured properties of breast tissue and mammographic density and other risk factors for breast cancer, using breast tissue obtained at forensic autopsy and not selected for the presence of abnormalities.

Methods: We used randomly selected tissue blocks taken from breast tissue slices obtained by s.c. mastectomy at the time of forensic autopsy to measure histologic features using quantitative microscopy. The proportions of the biopsy occupied by cells (estimated by nuclear area), glandular structures, and collagen were determined. These measurements were examined in relation to the percent density in the faxitron image of the tissue slice from which the biopsy was taken and other risk factors for breast cancer.

Results: The percent mammographic density was associated with the proportion of the area of the biopsy occupied by nuclei, both epithelial and nonepithelial, and by collagen and the area of glandular structures. Several other risk factors for breast cancer, notably body weight, parity, and number of births, and menopausal status, that are associated with variations in mammographic density, were also associated with differences in one or more of these tissue features.

Conclusion: All risk factors for breast cancer must ultimately exert their influence by an effect on the breast, and these findings suggest that, for some risk factors, this influence includes an effect on the number of cells and the quantity of collagen.

Introduction

The radiographic appearance, or mammographic pattern, of the female breast varies between individuals because of differences in the relative amounts of fat, connective and epithelial tissue, and the different X-ray attenuation characteristics of these tissues (1). Fat is radiologically lucent and seems dark on a mammogram. Connective and epithelial tissues are radiologically dense and seem light, an appearance that we refer to as “mammographic densities.” These variations in radiological appearance are associated with 4- to 6-fold differences in risk of breast cancer, and extensive mammographic densities may be responsible for a large proportion of the disease (2).

Mammographic density is inversely related to age, and among women of the same age, is less extensive in those who are parous, who have a higher body mass index, or who are postmenopausal and is more extensive in those taking hormone replacement therapy (3-5). These factors also influence risk of breast cancer. All risk factors for the disease must ultimately exert their influence through an effect on the breast The association of these risk factors with mammographic density suggests a potential mechanism for their effect on risk.

However, the biological basis for the association of mammographic density with risk of breast cancer is not understood. The published literature suggests that mammographic density is associated with epithelial and stromal proliferation, either of which might account for the association of density with risk of breast cancer (2). However, this literature is based entirely on examination of breast tissue obtained by surgical biopsy or mastectomy. The results of these studies may be biased, however, because the histologic features of breast tissue from subjects known or suspected of having breast disease may not be representative of all women. In this paper, we examine the relationships between the measured properties of breast tissue and mammographic density and other risk factors for breast cancer, using breast tissue obtained at forensic autopsy and not selected for the presence of abnormalities. This tissue was collected by Bartow et al., who have previously described the association of breast density with breast lesions of various types, using the qualitative nomenclature of diagnostic pathology (6-8). We describe here the quantitative association of measured breast tissue features with radiological percent density.

Methods

Methods of Collection of Original Study Material

This study used tissue blocks from a forensic autopsy series of 519 women collected at the New Mexico Office of the Medical Investigator between December 1978 and December 1983 by Bartow et al. (9, 10). The methods used to collect this material have been described in detail elsewhere by Bartow et al. and will be given only briefly here (9, 10).

The women were between 15 and 90 years of age and included 229 non-Hispanic Whites, 156 Hispanics, and 134 American Indians. Bilateral s.c. mastectomy was done on all cases at the time of autopsy. Whole-breast anteroposterior mammograms using compression were obtained on a Phillips low-dose film screen system. The specimens were then sectioned at 1-cm intervals and the slices examined with high-dose nonscreen (Faxitron, Wheeling, IL) radiography. Whole-breast mammograms were made for 466 cases and section radiography by faxitron for 486 (excluding pregnant subjects). No difference in radiographic architecture was found between fresh and formalin-fixed specimens. Ethnicity, weight, and height were known for all subjects. Age at menarche, age at first birth, menopause, and prior hormone use were known for some subjects. Menstrual status at the time of death was determined by reference to the histology of the endometrium and was available for most subjects.

Samples of breast tissue were taken from the nipple and, using two methods of sampling, from all four quadrants of each breast. One sample was guided by the pathologists' impression of areas that were representative of the overall breast tissue, and the other sample was a random selection from a randomly selected slice. For each subject, eight tissue blocks were made from randomly selected sites in the breast, four from each quadrant of each breast (9, 11).

Method of Sampling for the Present Study

We selected tissue samples for inclusion in the present study by stratifying all of the available material according to radiological classification. This was done using Wolfe's classification of the entire breast from which the samples came. The number of subjects in each category was as follows: Nl, 153; P1, 90; P2, 113; DY, 130. Subjects were sampled randomly from each Wolfe category to give ∼60 subjects per group. Histologic sections of the tissue block randomly selected from the upper outer quadrant of each breast were examined for autolysis, and those with sufficient preservation of structure used for the histologic measurements described below. If the section examined did show autolysis, other random samples from the same breast were examined, and if none were adequate the subject was excluded and another subject selected. Biopsies from a total of 78 subjects were excluded because of autolysis. Of these, 53 were from subjects with the Nl or Pl pattern and 25 from the P2 or DY categories (χ2, P = 0.02). The composition of the final sample included was as follows: NI, 66; PI, 59; P2, 60; DY, 51. The faxitron mammogram for the formalin-fixed tissue slice from which the biopsy had been taken was used in analysis.

Measurement of Histologic Features

Serial sections, 3-μm-thick were cut from formalin fixed tissue, embedded in paraffin blocks, mounted on silinated slides, and dried in a 60°C oven. The first section was stained with routine H&E, the second with Masson's trichrome, where nuclei were stained blue-black; muscle, RBC, fibrin, and cytoplasmic granules stained red; and collagen stained green.

Measurement of Total Area of Histologic Sections. Each H&E slide was scanned using Polaroid SprintScan 35 Plus with PathScan Enabler (Meyer Instruments, Houston, TX). These images were then analyzed with Microcomputer Imaging Device version M2 software (Imaging Research Inc., St. Catharines, Ontario). The total area for each section was estimated by outlining the tissue section, including all necrotic areas, and fat cells.

Selection of Areas for Measurement. To reduce the time required to analyze each section, 80 fields of view were selected for each tissue section. This number of fields had been used in previous work and can be accommodated by most biopsies and provides a manageable work load (12). Using the tissue section area obtained from the SprintScan image, a systematic randomization pattern was then used to select fields to be measured. Each field was the area covered by the view of a 10× 0.25 N.A. Achrostigmat objective on a Zeiss Axioskop light microscope, equipped with a Sony DXC 970MD 3-chip CCD color video camera, a Biopoint motorized stage, and all controlled by Microcomputer Imaging Device software. The 10× yielded a camera field view of 0.7 mm2 or 994.3 × 751.7 μm, and a total of 80 fields of view were measured in each section. This approach is illustrated in Fig. 1. If a section had fewer than 80 fields of view, the entire section was measured.

Figure 1.

Examples of Faxitron images of tissue slices (top left); example of random selection of fields to be measured in histological section (top right); examples of histological sections without (bottom left), and with (bottom right), thresholds set for measuring nuclear area.

Figure 1.

Examples of Faxitron images of tissue slices (top left); example of random selection of fields to be measured in histological section (top right); examples of histological sections without (bottom left), and with (bottom right), thresholds set for measuring nuclear area.

Close modal

Total Nuclei. Total nuclear area was measured in the H&E-stained sections. Nuclei were counted by segmenting the hematoxylin-stained nuclei through a threshold setting that created a mask over the nuclei. Microcomputer Imaging Device then calculated the area under the nuclear mask in square microns. An example of nuclear threshold segmentation, with nuclear mask is shown in Fig. 1.

Epithelial and Nonepithelial Nuclear Area. Epithelial and nonepithelial nuclear areas were measured in the H&E-stained sections. Epithelial nuclei were first isolated by manually drawing around the epithelial tissue or glandular tissue. Epithelial nuclei were counted by segmenting the hematoxylin stained nuclei from the remaining cellular material by setting the same threshold used for total nuclei above. The resulting epithelial nuclear mask gave epithelial nuclear area in square micrometers. Nonepithelial nuclear area was calculated by subtracting epithelial nuclear area from total nuclear area.

Total Collagen and Glandular Area. Using Polaroid SprintScan 35 Plus with PathScan Enabler, images of Masson's trichrome–stained tissue sections were acquired. The entire area of the section was then analyzed for collagen and glandular area using the Microcomputer Imaging Device software program. The collagen area was obtained by setting a threshold to segment out the green-stained collagen. The mask created by the threshold was then translated into square microns of collagen. A second threshold was set to segment the purple stained (red cytoplasm and blue black nuclei) glandular area. This second mask yielded the output of glandular area in square micrometers.

Reliability of Measurements. The reliability of the measurement of histologic features was assessed by remeasuring a randomly selected 10% of the material. Test-retest reliability was >0.96 for all of the measurements except nonepithelial nuclear area which was 0.86.

Statistical Methods

Linear regression was used to examine the association among breast tissue measures made from histologic sections, percent mammographic density, and selected breast cancer risk factors. For categorical risk factors, analysis of covariance was used. The assumptions of normality for response variables were inspected and transformations were made when necessary. Percent mammographic density is known to be strongly influenced by age, and to avoid confounding by this important variable, all univariate models were adjusted for age (3, 13). Furthermore, to examine potential modification of the results by age, all analyses were carried out after division of the study population into those ages <50 or >50 years. Selected results are shown separately for these age groups.

The transformations applied were as follows: for univariate models adjusted for age, percent mammographic density, percent epithelial area and percent collagen area were square root transformed, and percent nuclear area, percent nonepithelial area, and percent glandular area were log transformed when they served as response variables in each model. Because the associations between independent variables and response variable in each model were not changed by these transformations, the results from the untransformed model are shown for ease of interpretation.

There were no missing values in the variables of age, height, weight, and ethnicity. The value of menopausal status was missing for four subjects, and value of breast size was missing for one subject. For all other variables, at least one third of the values were missing, All analyses were done on the original data without any replacement or imputation of missing values.

Results

Distribution of Breast Cancer Risk Factors and Association with Mammographic Density

Biopsies from a total of 236 subjects were examined, and Tables 1 and 2 show the distribution of selected demographic, anthropometric, menstrual, and reproductive risk factors for breast cancer as well as the association of these variables with mammographic density in these subjects.

Table 1.

Distribution of participants' demographic characteristics and their association with mammographic density for continuous variables

*

Univariate linear regression analysis of mammographic density against the corresponding demographic variables listed in the left column, adjusted for age.

†

Variables with missing data.

Table 2.

Distribution of participants' demographic characteristics and their association with mammographic density for categorical variables

*

Univariate linear regression analysis of mammographic density against the corresponding demographic variables listed in the left column, adjusted for age.

†

_P_s of overall effect for each variable.

‡

Variables with missing data.

For continuous variables (Table 1), the mean age of the subjecrs was 43 years, the age range was 75 years (15-90 years), and the interquartile range 24 years (29-53 years). Mean height was 1.6 m and mean weight 61 kg. Age at menarche was known for 72 subjects and the average was 12.7 years. Age at first birth was known for 82 subjects and the average was 22 years. Number of births was known for 139 subjects who had on average two children. Statistically significant associations, all of them negative, were found on univariate analysis, adjusted for age, between percent mammographic density and the continuous variables of age, weight, body mass index, and number of births.

For categorical variables (Table 2), 34% were classified as Hispanic, 47% as Anglo, and 19% as Native American. Native Americans had less mammographic density (mean = 27%) than Anglos or Hispanics (means = 36% and 34%, respectively), differences that were statistically significant after age adjustment. Thirty-seven percent were postmenopausal. Parity was known for 164 of the subjects, and of these 71% were parous. Forty percent had ever used oral contraceptives, and 21% had ever used hormone replacement therapy. Statistically significant, negative associations with percent mammographic density were found on univariate analysis with menopausal status and parity. Use of oral contraceptives was associated with a greater percent density that was of borderline significance after adjustment for age. Hormone replacement therapy and a family history of breast cancer were not associated with mammographic density. Smaller breast size was significantly associated with a greater percent density. Similar results were seen in those ages <50 years and those ages ≥50 years with the exceptions of menopausal status, number of births, and parity, all of which were more strongly associated with percent density in the younger age group (data not shown).

Association of Breast Tissue Measurements with Mammographic Density

Table 3 and Fig. 2 show the association between each of the breast tissue measurements made from histologic sections and expressed as a percentage of the total area of the section and percent mammographic density in the tissue slice from which the section was taken. Greater percent mammographic density was associated with a significantly greater total nuclear area, a greater nuclear area of both epithelial and nonepithelial cells, a greater proportion of collagen, and a greater area of glandular structures.

Table 3.

Linear regression analysis of mammograpbic density against each of the breast tissue measurement

*

Univariate linear regression analysis of mammographic density against the corresponding biomarkers, adjusted for age.

†

_R_2 for the linear regression model.

Figure 2.

Boxplots of percent density verse age and each histological measure. The first, second, third, and fourth quantiles were chosen as the category cut points for age and each tissue measure. Median (line), interquartile range (column), 1.5 times the interquartile range (whiskers) and outliers.

Figure 2.

Boxplots of percent density verse age and each histological measure. The first, second, third, and fourth quantiles were chosen as the category cut points for age and each tissue measure. Median (line), interquartile range (column), 1.5 times the interquartile range (whiskers) and outliers.

Close modal

Similar associations were found in those ages <50 years and those ages ≥50 years with the exception of epithelial nuclear area and glandular area. Both measures were significantly associated with percent density only in those aged less than 50.

Association of Risk Factors with Breast Tissue Measurements

Table 4 shows the association of risk factors and the breast tissue measurements. The risk factors shown are those that were associated with mammographic density in Table 1. The variables found on univariate analysis to be associated with percent density were in general also associated, in the same direction, with one or more of the breast tissue measures. Age, body mass index, and postmenopausal status were significantly and inversely associated with mammographic density and all of the breast tissue measurements. Weight was inversely and significantly associated with mammographic density, and with all of the measurements except glandular area. Parity and number of births were associated inversely only with percent collagen. Compared with Native Americans, Hispanics and Anglos had a larger area of collagen. Ever use of oral contraceptive or HRT was not significantly associated with any of the tissue measurements. Smaller breast size was associated with greater total nuclear area, epithelial and nonepithelial nuclear areas, and greater areas of collagen and glands in the biopsy. Of the variables not associated with percent mammographic density in Tables 1 and 2, none was significantly associated with any of the breast tissue measurements.

Table 4.

Univariate linear regression analysis of molecular biomarkers against risk factors, adjusted for age

*

Univariate linear regression analysis of the corresponding molecular biomarkers listed in the top row against the corresponding variables in the left column and adjusted for age (i.e., biomarker = the corresponding variable listed in the left column + age).

†

Variables with missing data.

‡

Native American serves as the reference group.

§

Postmenopausal versus premenopausal.

∥

Parous versus nulliparous.

**

With versus without family history.

††

Large size serves as the reference group.

Similar results were seen in those ages <50 and ≥50 years, with the exceptions of menopausal status, number of births and parity, all of which were more strongly associated with percent collagen in the younger age group (data not shown).

Discussion

These results, that are based on random samples from breast tissue obtained at forensic autopsy, show that mammographic density is associated with the proportion of the area of the biopsy occupied by nuclei, both epithelial and nonepithelial, and by collagen and the area of glandular structures. Several other risk factors for breast cancer, that are associated with variations in mammographic density, were also associated with one or more of these tissue features. All risk factors for breast cancer must ultimately exert their influence by an effect on the breast, and these findings suggest that, for some risk factors, notably parity and number of births and menopausal status, this influence includes an effect on the number of cells and the quantity of collagen that is in the same direction as the effect of these variables on risk. Similar findings have been reported by Gertig et al. (14), using methods of measurement similar to those used here, applied to surgical biopsies. They found that age, menopausal status, and time since last birth were associated with variations in the proportion of the biopsy occupied by epithelium and/or stroma.

Body weight has repeatedly been found to be inversely associated with breast density (2) and is here also found to be inversely associated with the histologic features that are related to density. This is consistent with other data that show leanness to be associated with an increased risk of breast cancer in premenopausal women, but inconsistent with the association of obesity with an increase in risk in the postmenopausal women (15). Thus, the effect of obesity on risk of breast cancer after the menopause does not seem to be mediated through density.

The relationship between histologic and radiological features of the breast has been examined in nine other studies (16-25, 26). All have found an association between mammographic density and proliferation of either stroma or epithelium, or both, the two types of tissue in the breast with X-ray attenuation characteristics responsible for radiologically dense breast tissue. Six of the nine found epithelial proliferation, with or without atypia, to be associated with radiological densities (9, 18-21, 26), and a further study found cytologic atypia in nipple aspirate fluid associated with densities (25). All of the six studies that reported specifically on the stroma described an association between stromal fibrosis and mammographic densities (16, 18, 19, 23, 26).

The study of Bartow et al. (9), that used the same forensic autopsy material as the present study, is unique in that it is the only study in which there has been no selection of subjects or histologic material for the presence of breast disease. Using Wolfe's classification of the radiological appearance of the breast and qualitative methods of classifying histology, the more dense P2 and DY patterns were found to be associated with increasing fibrosis, that was periductal, interductal, and intralobular in distribution, with microcalcification, and also with epithelial hyperplasia (9, 11). The present results show quantitative relationships between the stromal and epithelial elements classified in this earlier work and radiological density.

These results are also similar to those that we obtained in a pilot study, based on surgical biopsy material from 92 subjects, in which we used immunohistochemistry and quantitative microscopy to examine the relationship of growth factors and stromal regulatory proteins to mammographic densities (27). Tissue sections from formalin-fixed paraffin blocks of breast tissue surrounding benign lesions were examined, half from breasts with little or no radiological density, and half from breasts with extensive densities. Subjects with extensive dense breast tissue were matched individually according to age (within 2 years) to subjects with little or no dense tissue, their tissue sections stained with antibodies that were markers for cell nuclei, collagen and selected growth factors and stromal regulatory proteins, and measured using the same methods of quantitative microscopy as were used in the present study. We found that, compared with subjects with low breast density, breast tissue from subjects with high density had a greater nuclear area, an area of stained collagen about twice as great, an area of staining for insulin-like growth factor-I about 30% greater, and an area of tissue inhibitor of metalloproteinase-3 staining about twice as great. Differences in nuclear area and area of stained collagen were statistically significant, whereas differences in stained areas for insulin-like growth factor-1 and tissue inhibitor of metalloproteinase-3 were of borderline significance.

The results of these studies show that a radiological feature of the breast that is associated with an increased risk of breast cancer is associated with a greater cellularity of breast tissue, and with greater stained areas of collagen, and of some growth factors and regulatory proteins that are known to play a role in breast carcinogenesis. These results are consistent with our hypothesis that the stromal and epithelial proliferation that contributes to mammographically dense breast tissue indicates the activity of growth factors on the breast. Further support for this hypothesis comes from finding that mammographic density is also associated with higher blood levels of the breast mitogens insulin-like growth factor-I (in premenopausal women) and prolactin (in postmenopausal women; refs. 28-30). Blood levels of these mitogens are also associated with risk of breast cancer (31, 32).

In addition to endocrine influences, stromal cells, epithelium, and matrix components communicate and are influenced by means of several paracrine growth factors (33-35). As well, extracellular signals by matrix proteins such as collagen, laminin, and fibronectin can directly affect breast epithelial proliferation and differentiation (36, 37). The extent of stromal matrix degradation, and hence net matrix deposition, is largely determined by the opposing actions of matrix degrading enzymes called matrix metalloproteinases 1 to 24 and their natural tissue inhibitors (tissue inhibitor of metalloproteinases 1-4; ref. 38). Furthermore, these secreted proteins can also influence growth factor bioavailability (39, 40). Therefore, breast-specific expression of MMP/TlMP genes may have a dual role as they are able to directly affect both the stromal matrix as well as growth factors.

Traditionally, matrix has been viewed as an ultrastructure of molecules capable of providing support for cells and tissue. However, it is now realized that in addition to providing supportive architecture, the matrix contributes to apoptosis, gene expression, cell adhesion and migration, growth factor bioavailability, and angiogenesis (36, 37, 41, 42). In vitro and in vivo studies have shown that breast epithelial cell-matrix contacts are necessary for proper differentiation, proliferation, polarity, and maintenance (43-46). These influences can be mediated through cell surface receptors specific for matrix components, such as integrins and/or the sequestering and release of growth factors (47-49). Increasing evidence also indicates that matrix actively participates in the control of most of the successive stages of breast tumors from appearance to progression to metastasis (50), and alteration in the structure and composition of matrix proteins accompanies changes in breast tissue associated with development, estrous cycle, pregnancy, lactation, and involution (51, 52). The age-related decline in the prevalence of mammographic density is likely to reflect the reduction in epithelium and stroma in the breast, and the increase in fat, that is well described and referred to as involution (53). Pregnancy and the menopause are also associated with a reduction in mammographic density, and as shown here, with a reduction in cellularity and in the amount of collagen.

Age, parity, and the menopause, however, explain only about 20% of the variation in mammographic density (54). Mammographic density has been shown to be highly heritable, and that after adjustment for age, parity, and menopausal status, an additive genetic model explains about 60% of the residual variance (5). The amounts of stroma, epithelium, and fat present in the breast at a given age, that give rise to radiological density, are likely therefore also to be largely heritable. The genes responsible for mammographic density have not yet been identified but those responsible for the proliferative activity, maintenance, and regulation of breast epithelium, stroma, and extracellular matrix and fat seem likely candidates.

Grant support: Susan G. Komen Foundation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Acknowledgments

We thank the generous assistance of Dr. Sue Bartow and the New Mexico Cancer Registry in providing access to the material on which this research was based.

References

1

Ingleby H, Gerson-Cohen J. Comparative anatomy, pathology and roentgenology of the breast. Philadelphia: University of Philadelphia Press; 1960.

2

Boyd NF, Lockwood GA, Byng J, Tritchler DL, Yaffe M. Mammographic densities and breast cancer risk.

Cancer Epidemiol Biomarkers Prev

1998

;

7

:

1133

–44.

3

Grove JS, Goodman MJ, Gilbert F. Factors associated with mammographic pattern.

Br J Radiol

1985

;

58

:

21

–5.

4

Greendale GA, Reboussin BA, Sie A, et al. Effects of estrogen and estrogen-progestin on mammographic parenchymal density.

Ann Intern Med

1999

;

130

:

262

–9.

5

Boyd NF, Dite GS, Stone J, et al. Heritability of mammographic density, a risk factor for breast cancer.

N Engl J Med

2002

;

347

:

886

–94.

6

Bartow SA, Pathak DR, Mettler FA. Radiographic microcalcification and parenchymal pattern as indicators of histologic “high-risk” benign breast disease.

Cancer

1990

;

66

:

1721

–5.

7

Bartow SA, Pathak DR, Black WC, Key CR, Teaf SR. Prevalence of benign, atypical and malignant breast lesions in populations at different risk for breast cancer. A forensic autopsy study.

Cancer

1987

;

60

:

2751

–60.

8

Pathak DR, Pike MC, Key CR, Teaf SR, Bartow SA. Parity factors and prevalence of fibrocystic breast change in a forensic autopsy series.

Br J Cancer

1991

;

63

:

1005

–9.

9

Bartow SA, Mettler FA, Black WC. Correlations between radiographic patterns and morphology of the female breast.

Rad Patterns Morph

1997

;

13

:

263

–75.

10

Hart BL, Steinbock RT, Mettler FA, Pathak DR, Bartow SA. Age and race related changes in mammographic parenchymal patterns.

Cancer

1989

;

63

:

2537

–9.

11

Bartow SA, Pathak DR, Black WC, Key CR, Teaf SR. Prevalence of benign, atypical and malignant breast lesions in populations at different risk for breast cancer. A forensic autopsy study.

Cancer

1987

;

60

:

2751

–60.

12

Guo YP, Martin LJ, Hanna W, et al. Growth factors and stromal matrix proteins associated with mammographic densities.

Cancer Epidemiol Biomarkers Prev

2001

;

70

:

243

–8.

13

Boyd NF, Dite GS, Stone J, et al. Heritability of mammographic density, a risk factor for breast cancer.

N Engl J Med

2002

;

347

:

886

–94.

14

Gertig DM, Stillman IE, Byrne C, et al. Association of age and reporductive factors with benign breast tissue composition.

Cancer Epidemiol Biomarkers Prev

1999

;

8

:

873

–9.

15

Hunter DJ, Willett WC. Diet, body size, and breast cancer.

Epidemiol Rev

1993

;

15

:

110

–32.

16

Fisher E, Palekar A, Kim W, Redmond C. The histopathology of mammographic patterns.

Am J Clin Pathol

1978

;

69

:

421

–6.

17

Bartow SA, Mettler FA, Black WC. Correlations between radiographic patterns and morphology of the female breast.

Rad Patterns Morph

1997

;

13

:

263

–75.

18

Wellings SR, Wolfe J. Correlative studies of the histological and radiographic appearance of the breast parenchyma.

Radiology

1978

;

729

:

299

–306.

19

Bright R, Morrison A, Brisson J, et al. Relationship between mammographic and histologic features of breast tissue in women with benign biopsies.

Cancer

1988

;

61

:

266

–71.

20

Urbanski S, Jensen HM, Cooke G, et al. The association of histological and radiological indicators of breast cancer risk.

Br J Cancer

1988

;

58

:

474

–9.

21

Bland KI, Kuhns JG, Buchanan JB, et al. A clinicopathologic correlation of mammographic parenchymal patterns and associated risk factors for human mammary carcinoma.

Ann Surg

1982

;

195

:

582

–94.

22

Moskowitz M, Gartside P, McLaughlin C. Mammographic patterns as markers for high-risk benign breast disease and incidence cancers.

Radiology

1980

;

134

:

293

–5.

23

Arthur JE, Ellis IO, Flowers C, Roebuck E, Elston CW, Blarney RW. The relationship of high risk mammographic patterns to histological risk factors for development of cancer in the human breast.

Br J Radiol

1990

;

63

:

845

–9.

24

Boyd NF, Jensen H, Cooke G, Lee Han HW. Relationship between mammographic and histological risk factors for breast cancer.

J Natl Cancer Inst

1992

;

84

:

1170

–9.

25

Lee MM, Petrakis NL, Wrensch MR, King EB, Miike R, Sickles E. Association of abnormal nipple aspirate cytology and mammographic pattern and density.

Cancer Epidemiol Biomarkers Prev

1994

;

3

:

33

–6.

26

Boyd NF, Jensen H, Cooke G, Lee Han HW. Relationship between mammographic and histological risk factors for breast cancer.

J Natl Cancer Inst

1992

;

84

:

1170

–9.

27

Guo YP, Martin LJ, Hanna W, et al. Growth factors and stromal matrix proteins associated with mammographic densities.

Cancer Epidemiol Biomarkers Prev

2001

;

10

:

243

–8.

28

Byrne C, Colditz GA, Willet WC, Speizer FE, Pollak M, Hankinson SE. Plasma insulin-like growth factor (IGF) I, IGF-binding protein 3, and mammographic density.

Cancer Res

2000

;

60

:

3744

–8.

29

Boyd NF, Stone J, Martin LJ, et al. The association of breast mitogens with mammographic densities.

Br J Cancer

2002

;

87

:

876

–82.

30

Byrne C, Hankinson SE, Colditz GA, Willett W, Speizer FE. Plasma prolactin and mammographic density in postmenopausal women.

Proc AACR

200l

;

42

:

153

.

31

Hankinson SE, Willett WC, Michaud DS, et al. Plasma prolactin levels and subsequent risk of breast cancer in postmenopausal women.

J Natl Cancer Inst

1999

;

97

:

629

–34.

32

Hankinson SE, Willett WC, Colditz GA, et al. Circulating concentrations of insulin-like growth factor-l and risk of breast cancer.

Lancet

1998

;

351

:

1393

–6.

33

Cullen KJ, Lippman ME. Stromal-epithelial interactions in breast cancer. In: Dickson RB, Lippman ME, editors. Boston: Kluwer Associates; 1991. p. 413–31.

34

Sakakura T. New aspects of stroma-parenchyma relations in mammary gland differentiation.

Int Rev Cytol

1991

;

125

:

165

–202.

35

Dickson RB, Lippman ME. Growth factors in breast cancer.

Endocr Rev

1995

;

76

:

559

–89.

36

Hay ED. Extracellular matrix alters epithelial differentiation.

Curr Opin Cell Biol

1993

;

5

:

1029

–35.

37

Lin CQ, Bissell MJ. Multi-faceted regulation of cell differentiation by extracellular matrix.

FASEB J

1993

;

7

:

737

–43.

38

Woessner JF. Matrix metalloproteinases and their inhibitors in connective tissue remodeling [review].

FASEB J

1991

;

5

:

2145

–54.

39

Martin DC, Sanchez-Sweatman OH, Ho AT, Inderdeo DS, Khokha R. Transgenic TIMP-I inhibits simian virus 40 T antigen-induced hepatocarcinogenesis by impairment of hepatocellular proliferation and tumor angiogenesis.

Lab Invest

1999

;

79

:

225

–34.

40

Fowlkes JL, Thrailkill KM, Serra DM, Suzuki K, Nagase H. Matrix metalloproteinases as insulin-like growth factor binding protein-degrading proteinases [review].

Progress in Growth Factor Research

1995

;

6

:

255

–63.

41

Hojilla CV, Mohammed FF, Khokha R. Matrix metalloproteinases and their tissue inhibitors direct cell fate during cancer development.

Br J Cancer

2003

;

8

:

1817

–21.

42

Overall CM, Lopez-Otin C. Strategies for MMP inhibition in cancer: innovations for the post-trial era.

Nat Rev

2002

;

2

:

657

–72.

43

Petersen OW, Ronnov-Jessen L, Howlett AR, Bissell MJ. interaction with basement membrane serves to rapidly distinguish growth and differentiation pattern of normal and malignant human breast epithelial cells.

Proc Natl Acad Sci U S A

1992

;

89

:

9064

–8.

44

Berdichevsky F, Gilbert C, Shearer M, Taylor-Papadimitriou J. Collagen-induced rapid morphogenesis of human mammary epithelial cells: the role of the α2 β1 integrin.

J Cell Sci

1992

;

102

:

437

–46.

45

Darcy KM, Black JD, Hahm HA, Ip MM. Mammary organoids from immature virgin rats undergo ductal and alveolar morphogenesis when grown within a reconstituted basement membrane.

Exp Cell Res

1991

;

196

:

49

–65.

46

Li ML, Aggeler J, Farson DA, Hatier C, Hassell J, Bissell MJ. Influence of a reconstituted basement membrane and its components on casein gene expression and secretion in mouse mammary epithelial cells.

Proc Natl Acad Sci U S A

1987

;

84

:

136

–40.

47

Damsky CH, Werb Z. Signal transduction by integrin receptors for extracellular matrix: cooperative processing of extracellular information [review].

Curr Opin Cell Biol

1992

;

4

:

772

–81.

48

Juliano RL, Haskill S. Signal transduction from the extracellular matrix [review].

J Cell Biol

1993

;

120

:

577

–85.

49

Taipale J, Keski-Oja J. Growth factors in the extracellular matrix [review].

FASEB J

1997

;

11

:

51

–9.

50

Lochter A, Bissell MJ. Involvement of extracellular matrix constituents in breast cancer.

Semin Cancer Biol

1995

;

6

:

165

–73.

51

Silberstein GB, Daniel CW. Glycosaminoglycans in the basal lamina and extracellular matrix of the developing mouse mammary duct.

Dev Biol

1982

;

90

:

215

–22.

52

Bissell MJ, Hall HG. The mammary gland: development, regulation, and function. New York: Plenum Press; 1987.

53

Henson DE, Tarone RE. Involution and the etiology of breast cancer.

Cancer

1994

;

74

:

424

–9.

54

Vachon CM, Kuni CC, Anderson K. Association of mammographically defined percent breast density with epidemiologic risk factors for breast cancer (United States).

Cancer Causes Control

2001

;

17

:

653

–62.

American Association for Cancer Research

2005

1,180 Views

266 Web of Science

Citing articles via

Email alerts