Background:

Endogenous hormones are associated with breast cancer risk, but little is known about their role on breast tissue composition, a strong risk predictor. This study aims to investigate the relationship between growth and sex hormone levels and breast tissue composition in young nulliparous women.

Methods:

A cross-sectional study of 415 young (age ∼21.5 years) nulliparous women from an English prebirth cohort underwent a MRI examination of their breasts to estimate percent-water (a proxy for mammographic percent density) and provided a blood sample to measure plasma levels of growth factors (insulin-like growth factor-I, insulin-like growth factor-II, insulin growth factor-binding protein-3, growth hormone) and, if not on hormonal contraception (n = 117) sex hormones (dehydroepiandrosterone, androstenedione, testosterone, estrone, estadiol, sex hormone–binding globulin, prolactin). Testosterone (n = 330) and sex hormone–binding globulin (n = 318) were also measured at age 15.5 years. Regression models were used to estimate the relative difference (RD) in percent-water associated with one SD increment in hormone levels.

Results:

Estradiol at age 21.5 and sex hormone–binding globulin at age 21.5 were positively associated with body mass index (BMI)-adjusted percent-water [RD (95% confidence interval (CI)): 3% (0%–7%) and 3% (1%–5%), respectively]. There was a positive nonlinear association between androstenedione at age 21.5 and percent-water. Insulin-like growth factor-I and growth hormone at age 21.5 were also positively associated with BMI-adjusted percent-water [RD (95% CI): 2% (0%–4%) and 4% (1%–7%), respectively].

Conclusions:

The findings suggest that endogenous hormones affect breast tissue composition in young nulliparous women.

Impact:

The well-established associations of childhood growth and development with breast cancer risk may be partly mediated by the role of endogenous hormones on breast tissue composition.

The risk of developing breast cancer has been associated with several hormone-related exposures. Early age of menarche, late age of menopause, and use of hormonal contraceptives and menopausal hormone therapy (HT), factors that act as proxy measures for an increased lifetime exposure to endogenous and exogenous sex hormone levels, are well-established risk factors for breast cancer (1–3). Such associations are in line with findings from prospective studies, which have consistently shown that higher circulating levels of estrogens and/or androgens in both pre- (4, 5) and postmenopausal women (6) are associated with increased breast cancer risk. Associations of risk with other hormones have been less consistent. Previous studies have reported positive associations between prolactin levels and breast cancer risk, particularly among postmenopausal women (7) and postmenopausal HT users (8). Circulating levels of growth factors, such as insulin-like growth factor-I (IGF-I), have been associated with increased breast cancer risk in both pre- and postmenopausal women (9, 10).

Mammographic density, a measure of the relative amounts of radio-dense fibroglandular tissue to fat tissue in the breast as seen on a mammogram (i.e., the relative amounts of white radio-dense areas to black nonradio-dense areas), for a woman's age and body mass index (BMI) is one of the strongest predictors of breast cancer risk (11). Mammographic percent-density increases with combined estrogen–progesterone HT use (12, 13) and decreases with tamoxifen use (14) and the menopause (15). However, current evidence on the relationship between breast tissue composition and circulating levels of endogenous sex hormones and growth factors is mixed. BMI strongly influences the relationship between endogenous sex hormones and mammographic density as obese women have, on average, higher estradiol levels and lower sex hormone–binding globulin (SHBG; ref. 6) as well as lower percent mammographic density (15). Studies that have accounted for BMI have found sex hormone levels to be positively associated with mammographic density in premenopausal women (16–18). Higher circulating levels of IGF-I, insulin growth factor–binding protein-3 (IGFBP-3), IGF-I/IGFBP-3 molar ratio, and growth hormone (GH) have been associated with increased mammographic density in premenopausal women in some (19), but not all (20), studies.

To our knowledge, no study has yet investigated the role of circulating sex hormones and growth factors on breast tissue composition in women whose breast tissue has not been influenced by reproductive-related events, partly because mammography involves exposure to ionizing radiation and hence cannot be performed in healthy young women. Herein, we investigate the relationship between sex and growth hormone measures collected in adolescence and early adulthood and breast tissue composition measured by radiation-free MRI (16, 21, 22), in young nulliparous women within a British prebirth cohort.

Study population

The Avon Longitudinal Study of Parents and Children (ALSPAC) is a prospective prebirth cohort of 14,775 children born in Avon, England, between April 1, 1991, and December 31, 1992 (23, 24). The cohort represents 72% of the eligible population (23). For the current primarily cross-sectional study, nulliparous women born from singleton pregnancies who participated in at least one follow-up survey were invited to attend an MRI examination of their breasts at the University of Bristol Clinical Research and Imaging Centre between June 2011 and November 2014. Women who had MRI contraindications (e.g., pregnancy, metal implants), or who had been diagnosed with cancer or a hormone-related disease, were excluded. In all, 500 of 2,530 (20%) eligible women invited to participate attended an MRI examination (Fig. 1). The highly demanding nature of the study (e.g., time and travel to the MRI examination center) and relocation away from the study area (i.e., to attend university) contributed to the low response rate. Nevertheless, participants and nonparticipants were similar in relation to body size and available hormone measurements. For example, mean BMI and median testosterone blood levels at approximately 15.5 years were 21.3 kg/m2 (SD = 3.1 kg/m2) and 0.82 nmol/L [interquartile range (IQR) = 0.42 nmol/L], respectively, among women who underwent the MRI examination, and 21.8 kg/m2 (SD = 3.8 kg/m2) and 0.80 nmol/L (IQR = 0.45 nmol/L), respectively, among those who did not.

The study received approval from the South West Frenchay NRES Committee (11/SW/0051), the ALSPAC Law and Ethics Committee, and the London School of Hygiene and Tropical Medicine ethics committee. Participants provided written informed consent.

Blood sample collection at age 21.5 years

A blood sample was taken from the participants on the same day they had their MRI breast examination, age, on average, 21.5 [mean = 21.5 (SD = 0.92); Table 1] years. The samples were processed within 24 hours, and plasma aliquots stored at −80°C. The blood sample for 2% of participants was collected after the MRI examination at their general practitioner clinics and sent by post to the ALSPAC laboratory. Figure 1 provides details on the number of samples that contributed to each hormone measurement, and reasons for exclusion.

Hormone measurements at age 15.5 years

Previously conducted serum measurements of total testosterone and SHBG levels at age 15.5 [mean = 15.4 (SD = 0.25)] years were available for a subset of participants (Fig. 1). Total testosterone was measured using Agilent triple quadrupole 6410 liquid chromatography/mass spectrometry equipment with an electrospray ionization source operating in positive ion mode (Agilent Technologies). SHBG was measured using a Cobas Auto Analyzer (Roche Diagnostics), and SHBG reagent using the manufacturer's calibrators and quality control material.

Hormone measurements at age 21.5 years

Measurements of plasma sex steroid hormones and SHBG at age 21.5 were restricted to women who had not used hormone contraception during the 3 months prior to blood sample collection (n = 117, Fig. 1), and were performed at the Royal Marsden Hospital (laboratory of M. Dowsett, Institute of Cancer Research, London, United Kingdom). Radioimmunoassay was used to determine plasma concentrations of testosterone (Indirect RIA, IM1087, Beckman Coulter), SHBG (IRMA-RIA-4184, DRG International, Inc), dehydroepiandrosterone (DHEA, RIA DSL 8900, Beckman Coulter), androstenedione (DSL 3800, Beckman Coulter), and prolactin (MG12161, IBL International). Assay sensitivities were 0.10 nmol/L, 0.41 nmol/L, 0.21 nmol/L, 0.31 nmol/L, and 10 mIU/L, respectively. Between and within assay variation were, respectively, 5.1% and 13% for testosterone at 1.1 nmol/L, 3.6% and 1.0% for SHBG at 75 nmol/L, 2.4% and 0.9% for DHEA at 21 nmol/L, 8.3% and 4.9% for androstenedione at 3.1 nmol/L, and 5.7% and 2.5% for prolactin at 503 mIU/L. Estrone and estradiol were measured by indirect radioimmunoassay methods (25, 26) and the sensitivity of the assays was 15 pmol/L and 3.0 pmol/L, respectively; their between and within assay variation were 16.0% and 6.6% at 118 pmol/L estrone, and 12.0% and 5.9% at 266 pmol/L estradiol.

Plasma levels of IGF-I, insulin-like growth factor-II (IGF-II), IGFBP-3, and GH were measured in samples from 406 participants (Fig. 1) at the University of Bristol (laboratory of Prof. J.M.P. Holly, University of Bristol, Bristol, United Kingdom) by in-house radioimmunoassay method (IGF-I, IGF-II, IGFBP-3), and an ELISA (GH, Quantikine kit DGH00, R&D Systems), respectively. Measurement errors for IGFs are known to be low/moderate, with within and between coefficients of variation of 3.4% and 13.7%, 2.8% and 7.4%, 3.9% and 11.7%, and 3.1% and 7.8% for IGF-I, IGF-II, IGFBP-3, and GH, respectively (27). Molar ratios of IGF-I and IGF-II to IGFBP-3 were calculated as [0.13 × (IGF-I or IGF-II)/(0.025 × IGFBP-3).

If the sample volume was too small for all laboratory measurements, priority was given to the IGFs, GH, and prolactin assays. Laboratory staff were blind to the characteristics of the participants.

Participants in the MRI study for whom adolescent sex hormone measurements were, and were not available, were similar with respect to anthropometric and breast measures. For example, mean age at menarche, BMI, and percent-water at age 21.5 years were 12.7 (SD = 1.00) years, 24.1 (SD = 4.3) kg/m2, and 42.0% (SD = 10.6%), respectively, among those with testosterone measurement at age 15.5 (n = 330; Table 1), compared with 12.8 (SD = 1.1) years, 23.6 (SD = 4.4) kg/m2, and 41.4% (SD = 9.8%) among those without such measurements. Similarly, among women who contributed to the hormone analyses, those with any sex hormone measurements (n = 117; Table 1) were comparable with those for whom these were not possible [mean BMI = 23.3 (SD = 3.8) kg/m2 and percent-water = 42.8% (SD = 10.4%) at age 21.5 compared with BMI = 23.8 (SD = 4.1) kg/m2) and percent-water of 42.2% (SD = 9.8%), respectively].

Breast tissue composition assessment

Women underwent an examination of both breasts using a 3T Siemens Skyra MRI system. For each participant, a set of T1-weighted VIBE 3-D images (≈176 images/woman), with a voxel size of 0.76 × 0.76 × 0.90 mm3, and T2-weighted trans-axial images (≈40 images/woman), with in-plane resolution 0.85 × 0.85 mm2 and slice thickness of 4 mm were obtained. Fully automated algorithms were developed to estimate breast volume using T1-weighted and T2-weighted images, while fat/water segmentations were completed on T2-weighted images. Left–right average estimates of volumes (cm3) of breast, water, and fat (the latter two correspond to mammographic-dense and nondense tissues, respectively), as well as percent-water, were generated. In the same women, percent-water and mammographic percent-density are highly correlated (range r = 0.76–0.85; refs. 16, 21, 22). Of the 500 participants who had an MRI examination, valid breast parameters were produced for 491. A full description of the methodology, and of its validity, is given in Doran and colleagues (28).

Covariates

Information on anthropometric measures were collected from annual clinical assessments participants attended between ages 7 to 13 years and at ages 15 and 17 years. Annual questionnaires sent to participants between the ages 8 and 17 years collected pubertal and menstruation data. At the time of the MRI examination and blood sample collection for this study, participants completed a short questionnaire on menstrual-related variables, and anthropometric measurements were taken using a standard protocol. The study website contains details of all the data that are available through a fully searchable data dictionary (29).

Statistical analysis

MRI percent-water was the primary outcome of interest in the analysis given its well-established associations with breast cancer risk but, whenever appropriate, hormone associations with water and fat volumes were also investigated. Continuous hormone measurements were standardized (z-scores), and MRI breast tissue measures were log-transformed to achieve near-normal distributions. The relationship between participants’ hormone measurements and breast tissue measures were examined using linear regression models. Exponentiated regression coefficients are shown; these coefficients represent the relative difference (RD) in breast-tissue measures associated with a unit increase in the exposure of interest (i.e., one SD for hormone levels). Nonlinear relationships between hormone levels and breast tissue measures were investigated and, if appropriate, are reported. Two types of models were fitted to the data. The first adjusted for assay batch number and storage time (number of months between blood sample being taken and analyzed), and age, phase of menstrual cycle, and hormone contraceptive use at MRI examination and blood sample collection (Table 1). Phase of menstrual cycle (follicular, luteal, irregular) was estimated for women not using hormone contraception by calculating the number of days since last menstrual period (date of MRI – start of last menstrual period), and luteal (day 14–17 to 28–31) and follicular (day 0 to 14–17) phase, or an “irregular period” (32+ days) was defined using average length of menstrual cycle. BMI at age 15.5 years, when the adolescent blood sample was taken, age of menarche, and history of contact with a doctor regarding their periods, were also investigated as possible confounding factors but inclusion of these additional variables in the models had little effect on the magnitude of the estimates—hence, these models are not presented. The second type of models were further adjusted for BMI at age 21.5 years as it is unclear whether this variable should be treated as a confounder or a mediator for the hormone–breast tissue associations. We also examined whether BMI at 21.5 years modified the hormone–breast tissue associations.

Data analyses were conducted in STATA, version 14 (StataCorp LP). All tests of significance are two-sided.

Study subjects

Table 1 presents the characteristics of the participants and their distributions of hormone measurements and MRI breast tissue composition estimates.

Age of menarche was weakly correlated with circulating estradiol (r = 0.19, P = 0.044), while BMI at age 21.5 years was inversely correlated with SHBG at ages 15.5 (r = −0.16, P = 0.005) and 21.5 years (r = −0.29, P = 0.001), and with IGF-II at 21.5 years (r = −0.10, P = 0.040; Supplementary Table S1). BMI at age 21.5 years was associated with MRI breast measures, with a unit increase in BMI being associated with a 4% [RD = −0.04; 95% confidence interval (CI): −0.05 to −0.04] lower percent-water but a 14% (0.14; 0.13, 0.15) and a 7% (0.07; 0.06, 0.08) higher fat and water volumes, respectively (unadjusted estimates).

Plasma sex hormones and MRI breast measures

Estradiol levels at age 21.5 were positively associated with percent-water, reflecting a more marked inverse association of this hormone with fat than water volume (Fig. 1, model 1). These associations strengthened upon further adjustment for BMI at 21.5 years (Fig. 1, model 2), such that a one SD (=283.6 pmol/L) increase in circulating estradiol was associated with a 3% [RD = 1.03; 95% CI 1.00–1.07] higher percent-water reflecting a 12% (0.88; 0.81–0.98) lower fat volume and a 7% (0.93; 0.86–1.01) lower water volume. Overall, there was no association between estrone levels and MRI breast measures; however, analyses stratified according to BMI at MRI showed evidence that this variable modified the estrone–percent-water association, with estrone being positively associated with percent-water (1.11; 1.03–1.20) in overweight/obese (BMI ≥ 25 kg/m2) participants, but not in underweight/normal weight women (Pinteraction = 0.018; Supplementary Table S2).

SHBG levels at ages 15.5 and 21.5 were only weakly correlated (r = 0.36, P < 0.001). Nevertheless, SHBG levels at both these ages were positively associated with percent-water, driven mainly by inverse associations with fat volume (Fig. 2, model 1). A one SD increase in SHBG levels at ages 15.5 (SD = 32.0 nmol/L) and at 21.5 (SD = 32.9 nmol/L) was accompanied, respectively, by a 6% (1.06; 1.03–1.09) and a 7% (1.07; 1.02–1.12) higher percent-water. These SHBG–percent-water associations were, however, attenuated upon further adjustment for BMI at MRI, particularly so for adult SHBG (Fig. 2, model 2).

There was no clear evidence of an association between DHEA and percent-water before (0.96; 0.92–1.01; Fig. 1, model 1) or after further adjustment for BMI at MRI. There was, however, an inverse association between DHEA and water volume that was strengthened upon further adjustment for BMI at MRI (0.91; 0.84–0.99). There was also evidence of a nonlinear association between androstenedione and percent-water; relative to women in the lowest quartile those in the second lowest had a 4% (0.96; 0.87–1.06) lower percent-water while those in the second and first highest quartiles had, respectively, a 2% (1.02; 0.92–1.13) and 1% (1.01; 0.92–1.12) higher percent-water (Supplementary Table S3). Plasma testosterone levels at age 15.5 were weakly correlated with levels at age 21.5 (r = 0.45; P < 0.001) despite the different measurement methodologies used (30). Neither adolescent nor young adult testosterone levels were found to be associated with MRI breast measures.

Circulating levels of prolactin were not associated with MRI breast measures overall, but there was evidence that BMI at MRI modified the association between prolactin and fat volume, with high levels of this hormone being associated with lower fat volume in overweight/obese participants only (Pinteraction = 0.04; Supplementary Table S2).

There was no evidence that phase of menstrual cycle modified the sex hormone–percent-water associations but the power of the study to detect interactions was low.

Plasma growth factors and MRI breast measures

Plasma levels of IGF-I and IGF-II were weakly correlated with each other (r = 0.37, P = <0.001) and with IGFBP-3 (r = 0.27, P = <0.001 and 0.39, P = <0.001, respectively). GH was not correlated with IGF-I (r = 0.01, P = 0.56), IGF-II (r = −0.01, P = 0.91), or IGFBP-3(r = 0.04, P = 0.38). The relationship between circulating levels of growth factors and MRI breast measures are presented in Fig. 3. IGF-I and its IGF-I/IGFBP-3 molar ratio (molar ratio I) were both positively associated with percent-water, with these associations persisting upon further adjustment for BMI at MRI (1.02; 1.00, 1.04 and 1.02; 1.00, 1.04, respectively). The positive relationship between IGF-I and percent-water was driven by a more markedly lower fat volume (0.94; 0.90–0.99) than water volume (0.97; 0.92–1.02). Neither IGF-II, nor its IGF-II/IGFBP-3 molar ratio (molar ratio II), were found to be associated with MRI breast measures (Fig. 3). GH levels were also associated, albeit in a nonlinear fashion, with percent-water (Fig. 4). In BMI-adjusted models, the top two GH quartiles were positively associated with percent-water, compared with the lowest GH quartile, and appeared to have a plateau effect (Supplementary Table S3). Collapsing quartiles into a binary category, GH levels ≥1.65 ng/mL were associated with a 4% (1.04; 1.01–1.08) higher percent-water relative to levels <1.65 ng/mL.

There was weak evidence that IGFBP-3 was inversely associated with percent-water, with the magnitude of this association being little affected by further adjustment for BMI at MRI. This association was driven by a nonlinear inverse association between IGFBP-3 and water volume, with young women in the second and third highest quartiles of the IGFBP-3 distribution having, respectively, a 5% (0.95; 0.90–1.00) and a 4% (0.94; 0.90–0.99) lower water volume compared with those in the lowest quartile in BMI-adjusted models (Supplementary Table S3).

Variability in MRI percent-water

Sex hormones explained little of the variability in percent-water being lowest for testosterone at age 15.5 (r2 adjusted for age, assay batch, storage time, and the other hormone levels: 0.06%) and highest for SHBG at age 21.5 (adjusted r2 = 7.7%). Growth hormone levels accounted for even less (adjusted r2 < 1.6% for all). In contrast, BMI at MRI explained 54.4% of the variability in percent-water.

Main findings

In this unique study with endogenous sex and growth hormones measurements taken in adolescence and young adulthood, we found evidence of positive associations between estradiol at age 21.5 years, and SHBG at ages 15.5 and 21.5 years, with MRI percent-water at age 21.5 years. The magnitude of the young adult SHBG–percent-water association was slightly attenuated upon adjustment for BMI at MRI; in contrast, the magnitude of the adolescent SHBG and young adult estradiol associations with percent-water were little affected upon such adjustment. Testosterone levels at ages 15.5 or 21.5 were not associated with any MRI breast measures, but androstenedione level at 21.5 years was associated, albeit in a nonlinear fashion, with percent-water. Young adult circulating levels of IGF-I and of GH were also associated with increases in percent-water, with these associations persisting after adjustment for concurrent BMI.

Strengths and weaknesses

To our knowledge, this is the first study to examine breast tissue composition in young women prior to it being affected by reproductive-related events (e.g., pregnancies, breastfeeding). Other strengths of the study include the use of objective, and previously validated, methods of assessing volumetric breast tissue composition (28). The unique prebirth cohort design meant that prospective standardized data on a wide range of childhood and adolescence variables, including certain adolescent hormone measurements, could be used alongside hormone measurements performed when participants were aged 21.5. But the study had some weaknesses. First, the participation rate was low (≈20%), albeit comparable with that of a similar MRI breast tissue composition study of young women by Boyd and colleagues (16), reflecting the demanding nature of the study for participants (time and travel to the MRI examination center). Second, blood samples could not be collected, or the amount collected was insufficient, for 16% of the participants. In addition, a high proportion of the participants had been on hormonal contraception in the 3 months prior to blood collection and hence were excluded from the sex hormone analyses. However, there was no evidence that participants for whom hormone measurements were available were a biased sample as they did not differ from the rest of the active ALSPAC cohort with regard to anthropometric or adolescent hormone measures, although by age 15.5 years, the cohort was affected by socially driven attrition (31). Furthermore, although response biases might have affected the sample's representativeness, it is unlikely that it would have distorted the associations between hormone levels and MRI breast measures. Third, hormone and MRI breast measurements were taken at a single point in time and may not characterize long-term average levels. Plasma sex hormone levels change throughout the menstrual cycle, some markedly and some modestly. Small variations in breast density throughout the menstrual cycle have also been reported (32). It was logistically impossible to time the blood collection and the MRI examination to the menstrual cycle, but the participants’ menstrual phase was accounted for in the analysis. A previous investigation (33) has shown that a single blood sample is sufficient to reliably characterize average levels of androgens, estrone sulfate, prolactin, and IGFs in premenopausal women over a 2- to 3-year period; however, reproducibility of a single measurement of estradiol and estrone was somewhat lower. GH secretion by the pituitary gland occurs in pulses throughout the day but again it was logistically impossible to time the blood collection to a particular time of the day (e.g., fasting morning samples). However, any errors in the measurement of plasma hormone/growth factors levels is likely to be nondifferential as they were performed independently of the MRI ascertainment and hence the observed magnitude of the hormone–breast measure associations is likely to be an underestimation of the true magnitude of these associations. Fourth, hormone levels in young adulthood were measured at the time of the MRI breast examination making it impossible to establish temporality. The only exceptions were for testosterone and SHBG as plasma levels of these hormones at ages 15.5 years were also available, with adolescent and young adult levels yielding associations with percent-water of similar direction and magnitude. Finally, the high number of statistical tests performed may have led to spurious associations.

Interpretation of the findings

The study reveals positive associations between MRI percent-water in young women and concurrent circulating levels of IGF-I, and its molar ratio, consistent with mammography- and MRI-based findings reported by others (18, 19, 34), albeit not all (16, 20, 35), studies in premenopausal women. These associations are also in line with the reported positive association between IGF-I and breast cancer risk, albeit possibly restricted to estrogen receptor–positive tumors (9, 10), suggesting that the role of these hormones in the etiology of this cancer may be mediated through their effect on breast tissue composition at young ages. This study also revealed a positive nonlinear association between GH and percent-water. GH is the main mediator of postnatal somatic growth, exerting its effect directly by binding to receptors located on the membrane of target cells as well as indirectly by stimulating the production of IGF factors by the liver and other target organs, including the breast. A positive GH–percent-water association was reported by another study on the determinants of MRI breast tissue composition in young women (16) although, in contrast to our study, it found no evidence of an association with IGF-I. Our findings are also consistent with previously reported positive associations of childhood growth and adult height with breast cancer risk (36) and breast density, as ascertained by percent mammographic density (37, 38) or MRI percent-water (16), including with our previous findings from the ALSPAC cohort showing strong associations between childhood growth and MRI percent-water (39).

Our finding of a positive estradiol–percent-water association is consistent with results from two other studies which examined sex hormone levels throughout the menstrual cycle in relation to percent mammographic density (17, 34). However, the only other study to our knowledge to have examined the role of circulating sex hormones on MRI breast tissue composition in young women found that percent-water was not associated with follicular levels of this hormone, but was inversely associated with luteal levels of free estradiol and free testosterone (16). Studies of the associations between sex hormone levels and mammographic density in postmenopausal women have produced mixed results (40, 41).

This study identified positive associations between SHBG at ages 15.5 and 21.5, and MRI percent-water at age 21.5, consistent with Boyd and colleagues (16). The direction of the observed association is, however, the opposite of the reported SHBG–breast cancer risk association (4, 5). The biological basis for the opposing effects on breast tissue composition and cancer risk is unclear. Plasma SHBG not only binds to circulating steroids, thus regulating their bioavailability and access to target cells, but it also mediates the uptake of steroid molecules into cells through binding to cell membrane receptors. The latter might stimulate intracellular messengers to cell proliferation and thus contribute to mitogenesis in the breast.

Conclusion

The study findings suggest that endogenous sex hormones and growth factors affect breast tissue composition in young women. In particular, the associations of MRI percent breast water with GH and IGF-I, which are key mediators of postnatal somatic growth, suggests a potential biological mechanism for the well-established associations of childhood growth and adult height with breast tissue composition and, ultimately, breast cancer risk.

No potential conflicts of interest were disclosed.

Conception and design: M.O. Leach, D.J. Hawkes, I. dos-Santos-Silva

Development of methodology: R. Denholm, B.L. De Stavola, J.H. Hipwell, S.J. Doran, M.O. Leach, D.J. Hawkes, E. Folkerd, M. Dowsett, I. dos-Santos-Silva

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.M.P. Holly, I. dos-Santos-Silva

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Denholm, B.L. De Stavola, J.H. Hipwell, S.J. Doran, J.M.P. Holly, E. Folkerd, M. Dowsett, I. dos-Santos-Silva

Writing, review, and/or revision of the manuscript: R. Denholm, B.L. De Stavola, J.H. Hipwell, J.M.P. Holly, M.O. Leach, D.J. Hawkes, E. Folkerd, M. Dowsett, I. dos-Santos-Silva

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Denholm, J.H. Hipwell

Study supervision: M.O. Leach, I. dos-Santos-Silva

This work was funded by a Cancer Research UK project grant C405/A12730 (to I. dos-Santos-Silva). The Medical Research Council, the Wellcome Trust grant 102215/2/13/2, and the University of Bristol provided core support for the Avon Longitudinal Study of Parents and Children (ALSPAC). The Cancer Imaging Centre at UCL and King's College London is supported by Cancer Research UK and The Engineering and Physical Science Research Council (to D. Hawkes). J.H. Hipwell was funded by a European Union Seventh Framework Programme grant FP7-ICT-2011-9, 601040 and the Engineering and Physical Science Research Council grant EP/K020439/1. The UCL segmentation code is part of the UCL NifTK Translational Medical Imaging Platform. The Cancer Imaging Centre at the ICR and the RMH (to M.O. Leach and S.J. Doran) is supported by Cancer Research UK and the Engineering and Physical Science Research Council in association with the Medical Research Council and UK Department of Health grants C1060/A10334 and C1060/A16464; and the National Health Service funding to the National Institute for Health Research Biomedical Research Centre at the RMH. M.O. Leach and M. Dowsett are National Institute for Health Research Emeritus Senior Investigators.

We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team; the University of Bristol Clinical Research and Imaging Centre; and our study nurses, Elizabeth Folkes and Sally Pearce, for recruiting and conducting the MRI breast examinations. We would also like to acknowledge Dr. Maria Schmidt, who assisted us in establishing the MRI protocol, and Drs. Amanda Eng and Marta C. Busana, who assisted in the setup and coordination of the study.

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.

1.
Collaborative Group on Hormonal Factors in Breast Cancer
. 
Menarche, menopause, and breast cancer risk: individual participant meta-analysis, including 118 964 women with breast cancer from 117 epidemiological studies
.
Lancet Oncol
2012
;
13
:
1141
51
.
2.
Collaborative Group on Hormonal Factors in Breast Cancer
. 
Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer
.
Lancet
1997
;
350
:
1047
59
.
3.
Collaborative Group on Hormonal Factors in Breast Cancer
. 
Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53 297 women with breast cancer and 100 239 women without breast cancer from 54 epidemiological studies
.
Lancet
1996
;
347
:
1713
27
.
4.
Endogenous
H
,
Breast Cancer Collaborative Group
,
Key
TJ
,
Appleby
PN
,
Reeves
GK
,
Travis
RC
, et al
Sex hormones and risk of breast cancer in premenopausal women: a collaborative reanalysis of individual participant data from seven prospective studies
.
Lancet Oncol
2013
;
14
:
1009
19
.
5.
Kaaks
R
,
Tikk
K
,
Sookthai
D
,
Schock
H
,
Johnson
T
,
Tjonneland
A
, et al
Premenopausal serum sex hormone levels in relation to breast cancer risk, overall and by hormone receptor status—results from the EPIC cohort
.
Int J Cancer
2014
;
134
:
1947
57
.
6.
Key
TJ
,
Appleby
PN
,
Reeves
GK
,
Roddam
A
,
Dorgan
JF
,
Longcope
C
, et al
Body mass index, serum sex hormones, and breast cancer risk in postmenopausal women
.
J Natl Cancer Inst
2003
;
95
:
1218
26
.
7.
Tworoger
SS
,
Eliassen
AH
,
Zhang
X
,
Qian
J
,
Sluss
PM
,
Rosner
BA
, et al
A 20-year prospective study of plasma prolactin as a risk marker of breast cancer development
.
Cancer Res
2013
;
73
:
4810
9
.
8.
Tikk
K
,
Sookthai
D
,
Johnson
T
,
Rinaldi
S
,
Romieu
I
,
Tjonneland
A
, et al
Circulating prolactin and breast cancer risk among pre- and postmenopausal women in the EPIC cohort
.
Ann Oncol
2014
;
25
:
1422
8
.
9.
Endogenous
H
,
Breast Cancer Collaborative Group
,
Key
TJ
,
Appleby
PN
,
Reeves
GK
,
Roddam
AW
. 
Insulin-like growth factor 1 (IGF1), IGF binding protein 3 (IGFBP3), and breast cancer risk: pooled individual data analysis of 17 prospective studies
.
Lancet Oncol
2010
;
11
:
530
42
.
10.
Kaaks
R
,
Johnson
T
,
Tikk
K
,
Sookthai
D
,
Tjonneland
A
,
Roswall
N
, et al
Insulin-like growth factor I and risk of breast cancer by age and hormone receptor status–A prospective study within the EPIC cohort
.
Int J Cancer
2014
;
134
:
2683
90
.
11.
McCormack
VA
,
dos Santos Silva
I
. 
Breast density and parenchymal patterns as markers of breast cancer risk: a meta-analysis
.
Cancer Epidemiol Biomarkers Prev
2006
;
15
:
1159
69
.
12.
Crandall
C
,
Palla
S
,
Reboussin
BA
,
Ursin
G
,
Greendale
GA
. 
Positive association between mammographic breast density and bone mineral density in the Postmenopausal Estrogen/Progestin Interventions Study
.
Breast Cancer Res
2005
;
7
:
R922
8
.
13.
Byrne
C
,
Ursin
G
,
Martin
CF
,
Peck
JD
,
Cole
EB
,
Zeng
D
, et al
Mammographic density change with estrogen and progestin therapy and breast cancer risk
.
J Natl Cancer Inst
2017
;
109
:
1
7
.
14.
Martin
LJ
,
Minkin
S
,
Boyd
NF
. 
Hormone therapy, mammographic density, and breast cancer risk
.
Maturitas
2009
;
64
:
20
6
.
15.
Burton
A
,
Maskarinec
G
,
Perez-Gomez
B
,
Vachon
C
,
Miao
H
,
Lajous
M
, et al
Mammographic density and ageing: a collaborative pooled analysis of cross-sectional data from 22 countries worldwide
.
PLoS Med
2017
;
14
:
e1002335
.
16.
Boyd
N
,
Martin
L
,
Chavez
S
,
Gunasekara
A
,
Salleh
A
,
Melnichouk
O
, et al
Breast-tissue composition and other risk factors for breast cancer in young women: a cross-sectional study
.
Lancet Oncol
2009
;
10
:
569
80
.
17.
Iversen
A
,
Frydenberg
H
,
Furberg
AS
,
Flote
VG
,
Finstad
SE
,
McTiernan
A
, et al
Cyclic endogenous estrogen and progesterone vary by mammographic density phenotypes in premenopausal women
.
Eur J Cancer Prev
2016
;
25
:
9
18
.
18.
Boyd
NF
,
Stone
J
,
Martin
LJ
,
Jong
R
,
Fishell
E
,
Yaffe
M
, et al
The association of breast mitogens with mammographic densities
.
Br J Cancer
2002
;
87
:
876
82
.
19.
dos Santos Silva
I
,
Johnson
N
,
De Stavola
B
,
Torres-Mejia
G
,
Fletcher
O
,
Allen
DS
, et al
The insulin-like growth factor system and mammographic features in premenopausal and postmenopausal women
.
Cancer Epidemiol Biomarkers Prev
2006
;
15
:
449
55
.
20.
Rice
MS
,
Tworoger
SS
,
Rosner
BA
,
Pollak
MN
,
Hankinson
SE
,
Tamimi
RM
. 
Insulin-like growth factor-1, insulin-like growth factor-binding protein-3, growth hormone, and mammographic density in the Nurses' Health Studies
.
Breast Cancer Res Treat
2012
;
136
:
805
12
.
21.
Khazen
M
,
Warren
RM
,
Boggis
CR
,
Bryant
EC
,
Reed
S
,
Warsi
I
, et al
A pilot study of compositional analysis of the breast and estimation of breast mammographic density using three-dimensional T1-weighted magnetic resonance imaging
.
Cancer Epidemiol Biomarkers Prev
2008
;
17
:
2268
74
.
22.
Thompson
DJ
,
Leach
MO
,
Kwan-Lim
G
,
Gayther
SA
,
Ramus
SJ
,
Warsi
I
, et al
Assessing the usefulness of a novel MRI-based breast density estimation algorithm in a cohort of women at high genetic risk of breast cancer: the UK MARIBS study
.
Breast Cancer Res
2009
;
11
:
R80
.
23.
Boyd
A
,
Golding
J
,
Macleod
J
,
Lawlor
DA
,
Fraser
A
,
Henderson
J
, et al
Cohort Profile: the ‘children of the 90s’–the index offspring of the Avon Longitudinal Study of Parents and Children
.
Int J Epidemiol
2013
;
42
:
111
27
.
24.
Fraser
A
,
Macdonald-Wallis
C
,
Tilling
K
,
Boyd
A
,
Golding
J
,
Davey Smith
G
, et al
Cohort profile: the Avon Longitudinal Study of Parents and Children: ALSPAC mothers cohort
.
Int J Epidemiol
2013
;
42
:
97
110
.
25.
Brand
JS
,
Chan
MF
,
Dowsett
M
,
Folkerd
E
,
Wareham
NJ
,
Luben
RN
, et al
Cigarette smoking and endogenous sex hormones in postmenopausal women
.
J Clin Endocrinol Metab
2011
;
96
:
3184
92
.
26.
Dowsett
M
,
Goss
PE
,
Powles
TJ
,
Hutchinson
G
,
Brodie
AM
,
Jeffcoate
SL
, et al
Use of the aromatase inhibitor 4-hydroxyandrostenedione in postmenopausal breast cancer: optimization of therapeutic dose and route
.
Cancer Res
1987
;
47
:
1957
61
.
27.
Allen
NE
,
Roddam
AW
,
Allen
DS
,
Fentiman
IS
,
Dos Santos Silva
I
,
Peto
J
, et al
A prospective study of serum insulin-like growth factor-I (IGF-I), IGF-II, IGF-binding protein-3 and breast cancer risk
.
Br J Cancer
2005
;
92
:
1283
7
.
28.
Doran
SJ
,
Hipwell
JH
,
Denholm
R
,
Eiben
B
,
Busana
M
,
Hawkes
DJ
, et al
Breast MRI segmentation for density estimation: do different methods give the same results and how much do differences matter?
Med Phys
2017
;
44
:
4573
92
.
29.
University of Bristol. Avon Longitudinal Study of Parents and Children (ALSPAC).
Available from
: http://www.bris.ac.uk/alspac/researchers/data-access/data-dictionary/.
30.
Tran
TS
,
Center
JR
,
Seibel
MJ
,
Eisman
JA
,
Kushnir
MM
,
Rockwood
AL
, et al
Relationship between serum testosterone and fracture risk in men: a comparison of RIA and LC-MS/MS
.
Clin Chem
2015
;
61
:
1182
90
.
31.
Howe
LD
,
Tilling
K
,
Galobardes
B
,
Lawlor
DA
. 
Loss to follow-up in cohort studies: bias in estimates of socioeconomic inequalities
.
Epidemiology
2013
;
24
:
1
9
.
32.
White
E
,
Velentgas
P
,
Mandelson
MT
,
Lehman
CD
,
Elmore
JG
,
Porter
P
, et al
Variation in mammographic breast density by time in menstrual cycle among women aged 40–49 years
.
J Natl Cancer Inst
1998
;
90
:
906
10
.
33.
Missmer
SA
,
Spiegelman
D
,
Bertone-Johnson
ER
,
Barbieri
RL
,
Pollak
MN
,
Hankinson
SE
. 
Reproducibility of plasma steroid hormones, prolactin, and insulin-like growth factor levels among premenopausal women over a 2- to 3-year period
.
Cancer Epidemiol Biomarkers Prev
2006
;
15
:
972
8
.
34.
Walker
K
,
Fletcher
O
,
Johnson
N
,
Coupland
B
,
McCormack
VA
,
Folkerd
E
, et al
Premenopausal mammographic density in relation to cyclic variations in endogenous sex hormone levels, prolactin, and insulin-like growth factors
.
Cancer Res
2009
;
69
:
6490
9
.
35.
Rinaldi
S
,
Biessy
C
,
Hernandez
M
,
Lesueur
F
,
dos-Santos-Silva
I
,
Rice
MS
, et al
Circulating concentrations of insulin-like growth factor-I, insulin-like growth factor-binding protein-3, genetic polymorphisms and mammographic density in premenopausal Mexican women: results from the ESMaestras cohort
.
Int J Cancer
2014
;
134
:
1436
44
.
36.
Ahlgren
M
,
Melbye
M
,
Wohlfahrt
J
,
Sorensen
TI
. 
Growth patterns and the risk of breast cancer in women
.
N Engl J Med
2004
;
351
:
1619
26
.
37.
Lope
V
,
Perez-Gomez
B
,
Moreno
MP
,
Vidal
C
,
Salas-Trejo
D
,
Ascunce
N
, et al
Childhood factors associated with mammographic density in adult women
.
Breast Cancer Res Treat
2011
;
130
:
965
74
.
38.
Sellers
TA
,
Vachon
CM
,
Pankratz
VS
,
Janney
CA
,
Fredericksen
Z
,
Brandt
KR
, et al
Association of childhood and adolescent anthropometric factors, physical activity, and diet with adult mammographic breast density
.
Am J Epidemiol
2007
;
166
:
456
64
.
39.
Denholm
R
,
De Stavola
B
,
Hipwell
JH
,
Doran
SJ
,
Busana
MC
,
Leach
MO
, et al
Growth trajectories, breast size, and breast tissue composition in a British prebirth cohort of young women
.
Am J Epidemiol
2018
;
187
:
1259
68
.
40.
Warren
R
,
Skinner
J
,
Sala
E
,
Denton
E
,
Dowsett
M
,
Folkerd
E
, et al
Associations among mammographic density, circulating sex hormones, and polymorphisms in sex hormone metabolism genes in postmenopausal women
.
Cancer Epidemiol Biomarkers Prev
2006
;
15
:
1502
8
.
41.
Tamimi
RM
,
Byrne
C
,
Colditz
GA
,
Hankinson
SE
. 
Endogenous hormone levels, mammographic density, and subsequent risk of breast cancer in postmenopausal women
.
J Natl Cancer Inst
2007
;
99
:
1178
87
.