Hindawi Journal of Oncology Volume 2019, Article ID 1781762, 8 pages https://doi.org/10.1155/2019/1781762 Research Article Exploring the Role of Breast Density on Cancer Prognosis among Women Attending Population-Based Screening Programmes Laia Domingo,1,2 Maria Sala ,1,2 Javier Louro,1,2 Marisa Baré,2,3,4 Teresa Barata,5 Joana Ferrer,6 Maria Carmen Carmona-Garcia,7,8,9 Mercè Comas,1,2 Xavier Castells,1,2,3 and CAMISS Study Group10 1 Department of Epidemiology and Evaluation, IMIM (Hospital del Mar Medical Research Institute), Passeig Marı́tim, 25-29, 08003 Barcelona, Spain 2 Research Network on Health Services in Chronic Diseases (REDISSEC), Barcelona, Spain 3 Department of Paediatrics, Obstetrics and Gynaecology, Preventive Medicine and Public Health, Universitat Autònoma de Barcelona (UAB), 08193 Bellaterra, Barcelona, Spain 4 Cancer Screening and Clinical Epidemiology, Corporació Sanitària Parc Taulı́, 08208 Sabadell, Spain 5 General Directorate of Health Care Programmes, Canary Islands Health Service, C/Juan XXIII, 13 35005, Las Palmas de Gran Canaria, Spain 6 Department of Radiology, Hospital de Santa Caterina, C/Dr. Castany, s/n, 17190 Salt, Girona, Spain 7 Epidemiology Unit and Girona Cancer Registry, Oncology Coordination Plan, Department of Health, Catalan Institute of Oncology, C/Sol, 15, Barcelona 17004, Girona, Spain 8 Girona Biomedical Research Institute (IDIBGI), C/Dr Castany s/n, 17190 Salt, Girona, Spain 9 Department of Medical Oncology, Catalan Institute of Oncology, University Hospital Dr. Josep Trueta, Av França s/n, 17007, Barcelona, Girona, Spain 10 IMIM (Hospital del Mar Medical Research Institute), Passeig Marı́tim, 25-29, 08003 Barcelona, Spain Correspondence should be addressed to Maria Sala; 92601@parcdesalutmar.cat Received 26 April 2019; Accepted 5 August 2019; Published 27 November 2019 Guest Editor: Chia-Jung Li Copyright © 2019 Laia Domingo et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background. Our aim was to assess the role of breast density on breast cancer mortality and recurrences, considering patient and tumour characteristics and the treatments received among women attending population-based screening programmes. Methods. We conducted a retrospective cohort study among women aged 50–69 years attending population-based screening programmes, diagnosed with invasive breast cancer between 2000 and 2009, and followed up to 2014. Breast density was categorised as low density (≤25% dense tissue), intermediate density (25–50%), and high density (≥50%). Cox proportional hazards regression models were fitted to estimate the adjusted hazard ratios (aHR) and 95% confidence intervals (95% CI) for death and recurrences, adjusting by patient characteristics, mode of detection (screen-detected vs. interval cancer), and tumour features. Results. The percentage of deaths and recurrences was higher among women with intermediate- and high-density breasts than among women with low-density breasts (p � 0.011 for death; p � 0.037 for recurrences). Adjusted Cox proportional hazards regression models revealed that women with intermediate- and high-density breasts had a higher risk of death than women with low-density breasts, being statistically significant for intermediate densities (aHR � 2.19 [95% CI: 1.16–4.13], aHR � 1.44 [95% CI: 0.67–3.1], re- spectively). No association was found between breast density and recurrences. Conclusions. Breast density was associated with a higher risk of death, but not of recurrences, among women participating in breast cancer screening. These findings reinforce the need to improve screening sensitivity among women with dense breasts and to routinely assess breast density, not only for its role as a risk factor for breast cancer but also for its potential influence on cancer prognosis. 2 Journal of Oncology 1. Introduction Canary Islands; CAMISS retrospective cohort). All of them were diagnosed with breast cancer between 2000 and 2009 Mammographic breast density is defined as the relative and were followed up until June 2014. The study included amount of radiolucent elements (fatty tissue) and radi- asymptomatic women with cancers detected in routine opaque elements of the breast (fibroglandular tissue). It has screening mammograms and symptomatic women with become a key element in breast cancer screening because of cancers detected between two screening mammograms its dual effect on breast cancer risk: high breast density (interval cancers). impairs the detection of abnormalities in the breast, de- Mammography screening in Spain follows the recom- creasing the sensitivity of mammography [1], and is also an mendations of the European Guidelines for quality assur- independent risk factor for breast cancer, as most cancers ance in breast cancer screening and diagnosis [17], offering develop in the glandular parenchyma [2]. More recently, all women aged 50 to 69 years free biennial screening. Two breast density has been postulated as a robust candidate for mammographic projections (mediolateral oblique and cra- tailoring screening intervals, suggesting that annual niocaudal views) are made, using the BI-RADS (Breast screening may be more effective than biennial screening for Imaging Reporting and Data System) classification for women at high risk due to dense breasts in combination with mammogram reading [18]. other risk factors [3]. However, such an approach has not As breast density is not routinely recorded by all par- been implemented in any screening programme, since it ticipating screening programmes, we determined breast requires more individual-level data, among many other density for a subsample of cases. Sample size was calculated unresolved issues and challenges [4]. to estimate a hazard ratio of 2.5 [15], with a mortality rate of Variations in breast density during a woman’s lifetime 14.5% (from the whole CAMISS cohort). With 5% signifi- may be influenced by several internal and external factors cance level and 80% power, 55 subjects were needed in the related to the hormonal environment. Breast density is high-density group. The subsample included all interval inversely associated with age, with premenopausal women cancers with available screening and diagnostic mammo- younger than 50 years being more likely to have dense grams and a random sample of screen-detected cancers, breasts [5, 6]. The use of hormone replacement therapy slows matched by screening programme and year of cancer di- the age-related trend to fatty tissue, especially for those agnosis. After the breast density assessment, this resulted in women taking a combination of oestrogen and progestin 375 invasive breast cancers, 79 of them assigned to the high- components [7]. In addition, some studies have reported density group, thus assuring enough sample size for the that tumours developing in dense breasts are more likely to analysis. express hormone receptors such as oestrogen receptor (ER) The study protocol was approved by the Ethics Com- and progesterone receptor (PR) [8, 9], suggesting a positive mittee of Parc de Salut Mar, Barcelona (CEIC Parc de Salut association with stromal composition and the oestrogenic Mar). Specific patient consent was not required. microenvironment. However, whereas increased breast density is a well- recognised risk factor for breast cancer, the relationship 2.2. Breast Density Assessment. For the purpose of this study, between breast density and breast cancer prognosis is still breast density was retrospectively evaluated by three expe- controversial. Some studies have reported an increased risk rienced radiologists who followed a consensus-based pro- of death for women with dense breasts [10, 11], while others tocol, as detailed elsewhere [19]. In brief, each radiologist have found an inverse association [12] or no relationship determined the breast density of the cancer-free breast at the [13]. In addition, only few studies have been restricted to the moment of diagnosis using Boyd’s scale, a semiquantitative context of mammography screening [14–16], also with score of six categories using percentages of density: A: 0%; B: contradictory results. Because this population has particular 1–10%; C: 10–25%; D: 25–50%; E: 50–75%; F: 75–100% [20]. characteristics (e.g., average-risk women, women over 45/50 For statistical purposes, breast density was collapsed into low years, mostly postmenopausal), performing studies focused (≤25% density), intermediate (25–50% density), and high on this population may provide useful information to better density (≥50% density). understand the relationship between breast density and cancer prognosis and to eventually provide individually 2.3. Study Variables. Patient information, including age at tailored screening strategies. diagnosis, menopausal status, hormone replacement therapy Our aim was to assess the role of breast density on (ever/never), and first-degree family history of breast cancer, mortality and recurrences, taking into account patient and was obtained from the databases of the screening pro- tumour characteristics and the treatments received among grammes. To obtain information on the burden of disease at women attending population-based screening programmes. diagnosis, we manually reviewed clinical records to identify the presence of comorbidities and construct the Charlson 2. Materials and Methods comorbidity index (CCI) [21]. The CCI was stratified into three categories: CCI � 0, CCI � 1, and CCI ≥ 2. 2.1. Setting and Study Population. This study was carried out Information on mode of detection was obtained from the among a retrospective cohort of 1,086 women with breast screening programme databases and by merging data with cancer, aged between 50 and 69 years, who underwent breast population-based cancer registries, the hospital minimum cancer screening in two Spanish regions (Catalonia and the basic dataset, and hospital-based cancer registries. We Journal of Oncology 3 differentiated between breast cancers detected by routine Survival curves for overall mortality and for recurrences screening mammograms (i.e., screen-detected cancers) and were generated by using the Kaplan–Meier method and were cancers detected between 2 screening mammograms, or compared by the log-rank test. Recurrence-free survival and within 24 months for women who reached the upper age overall survival were plotted by breast density categories. 5- limit (i.e., interval cancers). Further details on the identi- year and 10-year survival rates and their 95% confidence fication of interval cancers are explained elsewhere [19]. intervals (95% CI) were computed. Tumour-related information, including tumour size, lymph We fitted two multivariate Cox proportional hazards node involvement, focality, histological type, histological regression models to estimate the hazard ratios (HR) and grade, and biomarker expression, was retrieved from the their 95% CI for death and recurrences using a stepwise cancer registries, hospital-based registries, and clinical backward variable selection approach. The initial model records. Biomarker expression included information on ER, included all predictors. In the final models, we forced to PR, human epidermal growth factor receptor 2 (Her2), p53, include age, screening programme, and CCI as adjusting and Ki67 status. The positivity criteria for biomarker ex- variables, although they were not statistically significant. The pression followed international recommendations and their proportional hazards assumption was ascertained by as- updates throughout the study period [22, 23]. Tumours were sessment of log-log survival plots. To test the statistical classified into the following four phenotypes based on the significance of breast density variable as a whole, we per- expression of ER, PR, and Her2: (1) luminal A: ER+/Her2− formed a Wald test in both models. or PR+/Her2− ; (2) luminal B: ER+/Her2+ or PR+/Her2+; All statistical tests were two-sided. p values <0.05 were (3) Her2: ER− /PR− /Her2+; and (4) triple-negative: ER− , considered statistically significant. Analyses were performed PR− , Her2− [24]. using the statistical software IBM SPSS Statistics version 23.0 From the review of the clinical records, we obtained (Armonk, NY, USA) and R statistical software version 3.3.2 information on the treatments received. We considered two (http://www.r-project.org). types of surgery: radical (including all the mastectomies performed, whether radical or simple) and conservative. 3. Results Information on breast surgery and axillary lymph node dissection (ALND) treatments was collapsed into a single A total of 375 invasive breast cancers were included in this explicative variable. Information on adjuvant treatment was study, most of them detected among women with low- categorised as follows: chemotherapy, radiotherapy, and density breasts (51.2%, 27.7%, and 21.1% of tumours de- hormonal therapy; radiotherapy and hormonal therapy; and tected in women with low-, intermediate-, and high-density other treatments. breasts). Patient characteristics by breast density categories are summarized in Table 1. Percentages of women with low 2.4. Follow-Up Information. Information on recurrences breast density were highest among older and post- (including locoregional and distant recurrences), second menopausal women. No differences were observed between breast neoplasms, and vital status at the end of follow-up a family history of breast cancer, the use of hormone re- (alive or dead) was obtained from the cancer registries and placement therapy or comorbidities, and breast density clinical records. Locoregional recurrence was defined as categories. disease recurrence within the ipsilateral breast or chest wall, Tumour characteristics according to breast density in the ipsilateral axillary nodes, internal mammary nodes, or categories are shown in Table 2. Screen-detected cancers supraclavicular nodes. Distant recurrence was defined as were more common among women with low-density disease recurrence in sites other than the breast or regional breasts, whereas interval cancers were more frequent in lymph nodes (bone, skin, or visceral metastasis). A second intermediate- and high-density breasts. Tumours detected in neoplasm was considered as a second primary carcinoma low-density breasts showed a trend to be smaller, node- developing in the ipsilateral or contralateral breast. negative, unifocal, and triple-negative. No differences were Overall survival was computed from the date of breast observed among the treatments received, although the cancer diagnosis to death from any cause. Patients were percentage of radical surgery tended to be higher among censored at the date of their last hospital visit. Recurrence- women with dense breasts. free survival was computed from the date of breast cancer Kaplan–Meier curves revealed poorer overall survival diagnosis to the first locoregional or distant recurrence, (p � 0.010) and poorer relapse-free survival (p � 0.032) whichever occurred first. Women lost to follow-up or those among women with high-density breasts (Figure 1). 5-year who died were censored either at the last visit or at death. overall survival rate for women with low breast density was The median follow-up period was 8.7 years (interquartile 0.97 (95% CI: 0.94–1.00), whereas figures for women with range (IQR): 7.2–10.6). high breast density were 0.83 (95% CI: 0.72–0.96). The same pattern was observed at 10 years of follow-up. Recurrence- free survival rate at 5 years was 0.97 (0.94–1.00) for women 2.5. Statistical Analyses. Descriptive analyses of patient and with low breast density and 0.81 (0.69–0.95) for women with tumour characteristics and the treatments received high breast density (Table 3). according to breast density categories were explored using Adjusted Cox proportional hazards regression models contingency tables. revealed that breast density was statistically significant for 4 Journal of Oncology Table 1: Patient-related characteristics by breast density categories. Total n � 375 Low breast density (<25%) Intermediate breast density (25–50%) High breast density (>50%) (%) n � 192 (%) n � 104 (%) n � 79 (%) Age groups (years) 50–54 106 (28.3) 34 (17.7) 33 (31.7)a 39 (49.4) 55–59 102 (27.2) 50 (26) 33 (31.7) 19 (24.1) 60–64 103 (27.5) 64 (33.3) 24 (23.1) 15 (19) 65–70 64 (17.1) 44 (22.9) 14 (13.5) 6 (7.6) Menopausal status Premenopause 31 (13.3) 5 (4.7) 9 (12.3) 17 (32.1) Menopause 202 (86.7) 102 (95.3) 64 (87.7) 36 (67.9) Hormone replacement therapy No 192 (85) 86 (84.3) 60 (84.5) 46 (86.8) Yes 34 (15) 16 (15.7) 11 (15.5) 7 (13.2) Family history of breast cancer No 200 (86.6) 90 (85.7) 63 (86.3) 47 (88.7) Yes 31 (13.4) 15 (14.3) 10 (13.7) 6 (11.3) Charlson comorbidity index 0 285 (76) 143 (74.5) 79 (76) 63 (79.7) 1 53 (14.1) 31 (16.1) 13 (12.5) 9 (11.4) 2 37 (9.9) 18 (9.4) 12 (11.5) 7 (8.9) predicting mortality (Wald test p value � 0.050) but not for the screening interval was 3 years, and the survival analyses predicting recurrences (Wald test p value � 0.499). Women were not adjusted. A recent study carried out among the with intermediate- and high-density breasts had a higher Dutch population also reported no relationship between risk of death than women with low-density breasts, reaching breast density and survival [16]. In this study, as pointed out statistical significance for intermediate densities (Table 4) by the authors, the lack of tumour-related information may (aHR � 2.19 [95% CI: 1.16–4.13], aHR � 1.44 [95% CI: confound the results shown. In addition, the definition for 0.67–3.10], for intermediate and high densities, re- high density includes ≥25% of dense tissue, differing from spectively). Tumours arising as interval cancers (aHR � 1.96 most of the published studies. Other works conducted in [95% CI: 1.09–3.52] and node-positive tumours were also nonscreening populations have also found contradictory associated with a higher risk of death (aHR � 2.73 [95% CI: results, some of them reporting positive associations between 1.55–4.81]) in the adjusted model (data not shown). breast density and mortality [10, 11] and others finding no Breast density showed no association with the risk of association [13] or even a negative association [12]. recurrences (aHR � 1.43 [95% CI: 0.71–2.89]; aHR � 1.47 Some authors have hypothesized that the association [95% CI: 0.71–3.08], for intermediate and high densities, between higher density and worse survival would be respectively) (Table 5). Node-positive tumours showed an explained by the diagnosis delay due to the masking effect. In increased risk of recurrences in the adjusted analysis that sense, we do observe a higher percentage of larger, (aHR � 3.96 [95% CI: 2.12–7.39]) (data not shown). node-positive, and interval cancers among women with high-density breasts in the descriptive data. In the adjusted 4. Discussion analyses, lymph node involvement and detection as an in- terval cancer were also associated with mortality, along with The results of the current study suggest that higher breast intermediate breast densities so that the current results density is associated with a greater risk of death in women would support this hypothesis, since breast density as well as participating in breast cancer screening, while breast density other factors related to diagnostic delay remained associated showed no association with the risk of recurrences. with the risk of death. These findings reinforce the need to The positive association between dense tissue and risk of improve screening sensitivity among women with dense death is consistent with some [14, 15], but not all [16, 25], breasts, which is currently been proposed by means of prior studies conducted among screened women. Based on shifting the conventional one-size-fits-all screening ap- the data of Swedish women, Chiu et al. found that dense proach towards more personalized screening strategies tissue increased mortality from breast cancer in addition to based on the individual risk of breast cancer. increasing breast cancer risk and the likelihood of more Other authors have postulated that the relationship be- aggressive tumours [14]. Based on Danish data, Olsen et al. tween breast density and survival may be explained by the also found a positive association between dense tissue and tumour characteristics of cancers arising in epithelial tissue. It death, although they reported lower case fatality among has been suggested an increased proliferation and growth tumours developing in dense breasts [15]. By contrast, a factors in dense tissue [26, 27] that may be involved in study carried out in the UK [25] reported no relationship pathways that lead to more aggressive tumours. Nevertheless, between breast density and survival. In that study, however, the evidence supporting this hypothesis is not conclusive and Journal of Oncology 5 Table 2: Tumour characteristics by breast density categories. Total Low breast density Intermediate breast density High breast density n � 375 (%) (<25%) n � 192 (%) (25–50%) n � 104 (%) (>50%) n � 79 (%) Mode of detection Screen-detected cancers 195 (52) 113 (58.9) 45 (43.3) 37 (46.8) Interval cancer 180 (48) 79 (41.1) 59 (56.7) 42 (53.2) Tumour size <20 mm 199 (61.8) 108 (56.3) 53 (51) 38 (48.1) ≥20 mm 123 (38.2) 58 (30.2) 33 (31.7) 32 (40.5) Lymph node involvement Negative 218 (66.1) 117 (60.9) 60 (57.7) 41 (51.9) Positive 112 (33.9) 51 (26.6) 30 (28.8) 31 (39.2) Focality Unifocal 301 (84.6) 155 (87.1) 80 (80) 66 (84.6) Multifocal and/or multicentric 55 (15.4) 23 (12.9) 20 (20) 12 (15.4) Histological type Ductal 303 (80.8) 152 (79.2) 86 (82.7) 65 (82.3) Lobular 39 (10.4) 22 (11.5) 8 (7.7) 9 (11.4) Others 33 (8.8) 18 (9.4) 10 (9.6) 5 (6.3) Histological grade I 83 (23.9) 44 (22.9) 18 (17.3) 21 (26.6) II 134 (38.5) 64 (33.3) 40 (38.5) 30 (38) III 116 (33.3) 60 (31.3) 35 (33.7) 21 (26.6) NA 15 (4.3) 9 (4.7) 3 (2.9) 3 (3.8) Oestrogen receptor Negative 88 (23.5) 46 (24) 28 (26.9) 14 (17.7) Positive 287 (76.5) 146 (76) 76 (73.1) 65 (82.3) Progesterone receptor Negative 150 (40.1) 70 (36.5) 45 (43.3) 35 (44.3) Positive 224 (59.9) 121 (63) 59 (56.7) 44 (55.7) HER2 Negative 261 (79.1) 133 (69.3) 70 (67.3) 58 (73.4) Positive 69 (20.9) 35 (18.2) 21 (20.2) 13 (16.5) Ki67 Negative 114 (60.3) 62 (32.3) 29 (27.9) 23 (29.1) Positive 75 (39.7) 49 (25.5) 15 (14.4) 11 (13.9) Tumour phenotype Luminal A 155 (48) 79 (41.1) 38 (36.5) 38 (48.1) Luminal B 93 (28.8) 46 (24) 26 (25) 21 (26.6) HER2 31 (9.6) 11 (5.7) 12 (11.5) 8 (10.1) Triple-negative 44 (13.6) 29 (15.1) 11 (10.6) 4 (5.1) Treatment Conservative surgery only or 97 (26.4) 50 (26.4) 28 (28) 19 (24.1) with sentinel lymph node biopsy Conservative surgery with 182 (49.5) 100 (52.9) 44 (44) 38 (48.1) axillary lymph node dissection Radical surgery with or without 79 (21.5) 35 (18.5) 24 (24) 20 (25.3) lymphadenectomy No surgery and/or adjuvant 10 (2.7) 4 (2.1) 4 (4) 2 (2.5) treatment Adjuvant treatment after surgery Chemotherapy, radiotherapy, 122 (32.5) 54 (28.1) 36 (34.6) 32 (40.5) and hormonal therapy Radiotherapy and hormonal 123 (32.8) 68 (35.4) 35 (33.7) 20 (25.3) therapy Other treatments 130 (34.7) 70 (36.5) 33 (31.7) 27 (34.2) seems contradictory to the overrepresentation (although characteristics, and clinical outcomes, are warranted to elu- nonsignificant) of triple-negative cancers in low-density cidate the mechanisms through which breast density and breasts, observed in the current descriptive data and in prognosis are associated. previous works [19, 28]. Further studies conducted in larger Contrasting with prior series, we did not find association cohorts, with information on breast density, tumour between breast density and the risk of recurrences [29–31]. 6 Journal of Oncology Table 3: 5-year and 10-year survival rates for overall survival and recurrence-free survival. 5-year survival rate (95% CI) 10-year survival rate (95% CI) Overall survival Low breast density (<25%) 0.97 (0.94–1.00) 0.95 (0.91–0.99) Intermediate breast density (25–50%) 0.89 (0.80–0.99) 0.82 (0.71–0.96) High breast density (>50%) 0.83 (0.72–0.96) 0.78 (0.65–0.93) Recurrence-free survival Low breast density (<25%) 0.97 (0.94–1.00) 0.92 (0.85–0.98) Intermediate breast density (25–50%) 0.93 (0.86–1.00) 0.88 (0.79–0.99) High breast density (>50%) 0.81 (0.69–0.95) 0.78 (0.65–0.93) 100 100 80 80 Relapses (%) Survival (%) 60 60 40 40 20 20 Log-rank test = 9.2 (p value = 0.010) Log-rank test = 6.9 (p value = 0.032) 0 0 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Follow-up (years) Follow-up (years) Breast density <25% Breast density <25% Breast density 25%–50% Breast density 25%–50% Breast density ≥50% Breast density ≥50% (a) (b) Figure 1: Survival and recurrence-free survival by breast density. (a) Overall survival; log-rank test � 0.010. (b) Recurrence-free survival; log- rank test � 0.032. Table 4: Unadjusted and adjusted hazard ratios for death. Those studies reported an increased risk of locoregional Number of deaths Unadjusted HR Adjusted HR ∗ recurrences, but not for distant metastasis or death. Un- fortunately, our study sample was not large enough to Breast density <25% 20 Ref. Ref. replicate the analysis for different types of recurrence. Be- 25–50% 24 2.48 (1.37–4.49) 2.19 (1.16–4.13) sides, the populations considered in these works differed >50% 15 1.89 (0.97–3.7) 1.44 (0.67–3.10) from ours, since they included study periods prior to ours, ∗ The final model included breast density, mode of detection, lymph node which could involve different treatment schemes. In addi- involvement, age, Charlson comorbidity index, and screening programme. tion, the study by Park et al. [30] only included patients HR: hazard ratio; aHR: adjusted hazard ratio. undergoing breast-conserving surgery and radiotherapy, whereas we included both patients receiving and not re- ceiving radiotherapy, which is strongly related to the risk of Table 5: Unadjusted and adjusted hazard ratios for recurrences. recurrences [29]. Our adjusted analyses revealed that the only factor associated with recurrences was the lymph node Number of Unadjusted Adjusted HR∗ involvement at diagnosis. recurrences HR Our study is limited by the relatively small number of Breast density events in some categories, which prevented us from in- ≤25% 19 Ref. Ref. 1.72 1.43 cluding all breast density categories of Boyd’s scale in the 25–50% 16 adjusted model. Nevertheless, most studies assessing the (0.89–3.35) (0.71–2.89) 1.47 effect of breast density on mortality outcomes collapse breast ≥50% 17 2.34 (1.21–4.5) density data into two or three categories, making our data (0.71–3.08) ∗ Adjusted by breast density, progesterone receptor, lymph node in- more comparable with those of previous works. Second, we volvement, age, Charlson comorbidity index, and screening programme. were not able to explore breast-specific cancer mortality. HR: hazard ratio; aHR: adjusted hazard ratio. Previous studies exploring both breast-specific cancer Journal of Oncology 7 mortality and mortality from other causes found that the Conflicts of Interest latter was not associated with breast density [11]. Therefore, the impact of analysing all causes of death together may lead The authors declare that they have no conflicts of interest. to underestimation of the effect of breast density on mor- tality. Finally, we used a qualitative classification for breast Acknowledgments density assessment, which is known to have moderate in- terobserver concordance [32]. However, to minimise mis- The authors acknowledge the contribution of Francesc classification, breast density assessment was centralised and Macià, Imma Collet, and the Hospital del Mar Tumour performed by a panel of experienced radiologists, specially Registry (Barcelona) in providing tumour-related data. trained for the study [19]. The authors also would like to acknowledge the dedica- The current study is strengthened by the homogeneity of tion and support of the entire CAMISS Study Group (in the study population included. Restricting the study to alphabetical order): IMIM (Hospital del Mar Medical screening participants allowed us to explore the effect of Research Institute), Barcelona: Xavier Castells, Mercè breast density on a relatively homogeneous group of patients Comas, Laia Domingo, Francesc Macià, Marta Roman, in terms of age range and tumour stage. Thus, the con- Anabel Romero, Maria Sala; Canary Islands Health clusions drawn from the current work are robust and in- Service: Teresa Barata, Isabel Diez de la Lastra, Mariola de formative within the framework of population-based la Vega; Corporació Sanitaria Parc Taulı́, Sabadell: Marisa screening and are of interest for tailored screening strategies. Baré, Núria Torà; Hospital Santa Caterina, Girona: Joana In addition, the availability of data on comorbidities, patient Ferrer; Epidemiology Unit and Girona Cancer Registry: and tumour characteristics, and the treatments received Maria Carmen Carmona-Garcia; Hospital Galdakao- allowed us to control for important prognostic factors and to Usansolo, Vizcaya: Susana Garcı́a, Maximina Martı́n, explore—for the first time among women participating in Miren Orive, Maria Amparo Valverde; Canary Islands breast cancer screening—the effect of breast density on Foundation for Health Research: Jeanette Pérez, Amado mortality considering both patient and tumour character- Rivero, Cristina Valcárcel; Hospital Costa del Sol, Uni- istics and treatments received. versity of Málaga: Marı́a del Carmen Padilla, Maximino Redondo, Teresa Téllez, Irene Zarcos; Hospital Uni- 5. Conclusion versitario Donostia/Biodonostia: Cristina Churruca, Amaia Perales, Javier Recio, Irune Ruiz, Cristina Sar- In conclusion, our findings reveal that increased breast asqueta, Jose Marı́a Urraca; Instituto Oncológico de density was associated with worse survival outcomes among Guipúzcoa-Onkologikoa: Maria Jesús Michelena; Hos- women participating in breast cancer screening. This as- pital Universitario Basurto: Julio Moreno; Hospital sociation seems to be mainly explained as a result of the Universitario Cruces: Gaizka Mallabiabarrena, Patricia masking effect of dense tissue, although an underlying bi- Cobos, Borja Otero; Hospital Universitario Txagorritxu: ological mechanism in the stroma composition may also Javier Gorostiaga, Itsaso Troya. This study was supported play a role. These findings reinforce the need to improve by grants from Instituto de Salud Carlos III FEDER (grant screening sensitivity among women with dense breasts by numbers: PS09/01153, PI12/00387, PI11/01296, PI15/ means of more personalized screening approaches as well as 00098, and PI16/00244) and by the Research Network on the importance to routinely assess and record information Health Services in Chronic Diseases (Instituto de Salud on breast density during the screening process, both because Carlos III) (REDISSEC: RD12/0001/0015, RD12/0001/ of its utility as a predictive factor for breast cancer and 0007, and RD16/0001/0013). because of its role in breast cancer prognosis. References Abbreviations [1] D. van der Waal, T. M. Ripping, A. L. M. Verbeek, and ALND: Axillary lymph node dissection M. J. M. Broeders, “Breast cancer screening effect across breast BI-RADS: Breast Imaging Reporting and Data System density strata: a case-control study,” International Journal of CCI: Charlson comorbidity index Cancer, vol. 140, no. 1, pp. 41–49, 2017. ER: Oestrogen receptor [2] N. F. Boyd, H. Guo, L. J. Martin et al., “Mammographic Her2: Human epidermal growth factor receptor 2 density and the risk and detection of breast cancer,” New HR: Hazard ratios England Journal of Medicine, vol. 356, no. 3, pp. 227–236, 2007. IQR: Interquartile range [3] A. Trentham-Dietz, K. Kerlikowske, N. K. Stout et al., “Tai- PR: Progesterone receptor loring breast cancer screening intervals by breast density and TNM: Tumour-node-metastasis. risk for women aged 50 years or older: collaborative modeling of screening outcomes,” Annals of Internal Medicine, vol. 165, Data Availability no. 10, pp. 700–712, 2016. [4] L. Rainey, D. van der Waal, A. Jervaeus et al., “Are we ready The retrospective observational data used to support the for the challenge of implementing risk-based breast cancer findings of this study are available from the corresponding screening and primary prevention?,” The Breast, vol. 39, author on reasonable request. pp. 24–32, 2018. 8 Journal of Oncology [5] C. M. Checka, J. E. Chun, F. R. Schnabel, J. Lee, and H. Toth, results of population-based mammography screening in “The relationship of mammographic density and age: im- Spain,” Breast Cancer Research, vol. 16, no. 1, p. R3, 2014. plications for breast cancer screening,” American Journal of [20] N. F. Boyd, J. W. Byng, R. A. Jong et al., “Quantitative Roentgenology, vol. 198, no. 3, pp. W292–W295, 2012. classification of mammographic densities and breast cancer [6] M. E. Work, L. L. Reimers, A. S. Quante, K. D. Crew, risk: results from the Canadian National Breast Screening A. Whiffen, and M. B. Terry, “Changes in mammographic Study,” JNCI Journal of the National Cancer Institute, vol. 87, density over time in breast cancer cases and women at high no. 9, pp. 670–675, 1995. risk for breast cancer,” International Journal of Cancer, [21] M. E. Charlson, P. Pompei, K. L. Ales, and C. R. MacKenzie, vol. 135, no. 7, pp. 1740–1744, 2014. “A new method of classifying prognostic comorbidity in [7] F. J. B. van Duijnhoven, P. H. M. Peeters, R. M. L. Warren longitudinal studies: development and validation,” Journal of et al., “Postmenopausal hormone therapy and changes in Chronic Diseases, vol. 40, no. 5, pp. 373–383, 1987. mammographic density,” Journal of Clinical Oncology, vol. 25, [22] A. C. Wolff, M. E. H. Hammond, J. N. Schwartz et al., no. 11, pp. 1323–1328, 2007. “American Society of Clinical Oncology/College of American [8] L. Yaghjyan, G. A. Colditz, L. C. Collins et al., “Mammo- Pathologists guideline recommendations for human epider- graphic breast density and subsequent risk of breast cancer in mal growth factor receptor 2 testing in breast cancer,” Journal postmenopausal women according to tumor characteristics,” of Clinical Oncology, vol. 25, no. 1, pp. 118–145, 2006. [23] M. E. H. Hammond, D. F. Hayes, M. Dowsett et al., JNCI Journal of the National Cancer Institute, vol. 103, no. 15, “American Society of Clinical Oncology/College of American pp. 1179–1189, 2011. Pathologists guideline recommendations for immunohisto- [9] J. Ding, R. Warren, A. Girling, D. Thompson, and D. Easton, chemical testing of estrogen and progesterone receptors in “Mammographic density, estrogen receptor status and other breast cancer (unabridged version),” Archives of Pathology & breast cancer tumor characteristics,” The Breast Journal, Laboratory Medicine, vol. 134, pp. e48–e72, 2010. vol. 16, no. 3, pp. 279–289, 2010. [24] A. Goldhirsch, W. C. Wood, A. S. Coates et al., “Strategies for [10] Å. Olsson, H. Sartor, S. Borgquist, S. Zackrisson, and subtypes-dealing with the diversity of breast cancer: highlights J. Manjer, “Breast density and mode of detection in relation to of the St Gallen international expert consensus on the primary breast cancer specific survival: a cohort study,” BMC Cancer, therapy of early breast cancer 2011,” Annals of Oncology, vol. 14, p. 229, 2014. vol. 22, no. 8, pp. 1736–1747, 2011. [11] G. Maskarinec, I. S. Pagano, M. A. Little, S. M. Conroy, [25] G. J. R. Porter, A. J. Evans, E. J. Cornford et al., “Influence of S.-Y. Park, and L. N. Kolonel, “Mammographic density as a mammographic parenchymal pattern in screening-detected predictor of breast cancer survival: the multiethnic cohort,” and interval invasive breast cancers on pathologic features, Breast Cancer Research, vol. 15, no. 1, p. R7, 2013. mammographic features, and patient survival,” American [12] A. Masarwah, P. Auvinen, M. Sudah et al., “Very low Journal of Roentgenology, vol. 188, no. 3, pp. 676–683, 2007. mammographic breast density predicts poorer outcome in [26] N. F. Boyd, L. J. Martin, M. J. Yaffe, and S. Minkin, patients with invasive breast cancer,” European Radiology, “Mammographic density and breast cancer risk: current vol. 25, no. 7, pp. 1875–1882, 2015. understanding and future prospects,” Breast Cancer Research, [13] G. L. Gierach, L. Ichikawa, K. Kerlikowske et al., “Relationship vol. 13, p. 223, 2011. between mammographic density and breast cancer death in [27] M. Ellingjord-Dale, E. Lee, E. Couto et al., “Polymorphisms in the Breast Cancer Surveillance Consortium,” JNCI: Journal of hormone metabolism and growth factor genes and mam- the National Cancer Institute, vol. 104, no. 16, pp. 1218–1227, mographic density in Norwegian postmenopausal hormone 2012. therapy users and non-users,” Breast Cancer Research, vol. 14, [14] S. Y. H. Chiu, S. Duffy, A. M. F. Yen, L. Tabar, R. A. Smith, and no. 5, p. R135, 2012. H. H. Chen, “Effect of baseline breast density on breast cancer [28] V. A. Kirsh, A. M. Chiarelli, S. A. Edwards et al., “Tumor incidence, stage, mortality, and screening parameters: 25-year characteristics associated with mammographic detection of follow-up of a Swedish mammographic screening,” Cancer breast cancer in the Ontario breast screening program,” JNCI Epidemiology Biomarkers & Prevention, vol. 19, no. 5, Journal of the National Cancer Institute, vol. 103, no. 12, pp. 1219–1228, 2010. pp. 942–950, 2011. [15] A. H. Olsen, K. Bihrmann, M.-B. Jensen, I. Vejborg, and [29] T. Cil, E. Fishell, W. Hanna et al., “Mammographic density and E. Lynge, “Breast density and outcome of mammography the risk of breast cancer recurrence after breast-conserving screening: a cohort study,” British Journal of Cancer, vol. 100, surgery,” Cancer, vol. 115, no. 24, pp. 5780–5787, 2009. no. 7, pp. 1205–1208, 2009. [30] C. C. Park, J. Rembert, K. Chew, D. Moore, and K. Kerlikowske, [16] D. van der Waal, A. L. M. Verbeek, and M. J. M. Broeders, “High mammographic breast density is independent predictor “Breast density and breast cancer-specific survival by de- of local but not distant recurrence after lumpectomy and ra- tection mode,” BMC Cancer, vol. 18, 2018. diotherapy for invasive breast cancer,” International Journal of [17] European Commission and Directorate-General for Health Radiation Oncology · biology · physics, vol. 73, no. 1, pp. 75–79, and Consumer Protection, European Guidelines for Quality 2009. Assurance in Breast Cancer Screening and Diagnosis, N. Perry [31] L. Eriksson, K. Czene, L. U. Rosenberg, S. Törnberg, and E. Puthaar, Eds., Office for Official Publications of the K. Humphreys, and P. Hall, “Possible influence of mam- mographic density on local and locoregional recurrence of European Communities, Brussels, Belgium, 2006. breast cancer,” Breast Cancer Research, vol. 15, no. 4, 2013. [18] E. A. Sickles and C. J. B. L. D’Orsi, “ACR BI-RADS ® [32] R. R. Winkel, M. von Euler-Chelpin, M. Nielsen et al., “Inter- ® mammography,” in ACR BI-RADS Atlas, Breast Imaging Reporting and Data System, American College of Radiology, observer agreement according to three methods of evaluating mammographic density and parenchymal pattern in a case Reston, VA, USA, 2013. control study: impact on relative risk of breast cancer,” BMC [19] L. Domingo, D. Salas, R. Zubizarreta et al., “Tumor phenotype Cancer, vol. 15, p. 274, 2015. and breast density in distinct categories of interval cancer: MEDIATORS of INFLAMMATION The Scientific Gastroenterology Journal of World Journal Hindawi Publishing Corporation Research and Practice Hindawi Hindawi Diabetes Research Hindawi Disease Markers Hindawi www.hindawi.com Volume 2018 http://www.hindawi.com www.hindawi.com Volume 2018 2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of International Journal of Immunology Research Hindawi Endocrinology Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com BioMed PPAR Research Hindawi Research International Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of Obesity Evidence-Based Journal of Stem Cells Complementary and Journal of Ophthalmology Hindawi International Hindawi Alternative Medicine Hindawi Hindawi Oncology Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2013 Parkinson’s Disease Computational and Mathematical Methods in Medicine Behavioural Neurology AIDS Research and Treatment Oxidative Medicine and Cellular Longevity Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018