Adherence to the WCRF/AICR cancer prevention recommendations and mammographic density Authors: Julia Fernández-Morata, Marina Pollán, Nerea Fernández de Larrea Baz, Vanessa Pachón-Olmos, Emma Ruíz-Moreno, Javier García-Pérez, Adela Castelló-Pastor, María Ángeles Sierra, Pilar Lucas, Rafael Llobet, Agostina Stradella, Blanca Cantos, Teresa Ramón y Cajal, Marta Santisteban, Miguel Ángel Seguí, Ana Santaballa Bertrán, Mónica Granja, Julia Camps-Herrero, Sabela Recalde, Miriam Méndez, Nuria Calvo Verges, Beatriz Pérez-Gómez, Roberto Pastor-Barriuso, Virginia Lope Published: 2026 ABSTRACT Background Mammographic density (MD) is a known predictor of breast cancer risk and may be influenced by modifiable lifestyle factors. This study aimed to assess the association between adherence to the 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) cancer prevention recommendations prior to diagnosis and MD in women diagnosed with breast cancer.Methods We analyzed data from 759 breast cancer patients recruited from eight Spanish hospitals enrolled in the case-case Breast Cancer & Density Association Study. Participants completed an epidemiological and a food frequency questionnaire. A standardized score was constructed to assess adherence to six WCRF/AICR recommendations. MD percentage was measured using the validated DM-Scan software on the digital mammogram of the contralateral, tumor-free, breast obtained prior to any treatment. Standardized prevalences and standardized prevalence ratios (SPRs) for four MD categories and for MD ≥ 50% across categories of adherence to the WCRF/AICR recommendations were estimated based on multinomial and binary multivariable logistic regression models.Results Overall, adherence to the WCRF/AICR recommendations was not consistently associated with MD. However, the standardized prevalence of MD ≥ 50% was higher among women with high adherence compared to those with low adherence (SPR high vs. low = 1.73; 95% CI = 1.03–2.91). This positive association was particularly evident with the physical activity recommendation (SPR fully met vs. no/partially met = 1.51; 95% CI = 1.02–2.25).Conclusions Breast cancer participants with greater adherence to the WCRF/AICR recommendations prior to diagnosis, particularly to physical activity guidelines, showed higher percent MD at diagnosis. This association likely reflects lower absolute breast adiposity rather than a greater amount of fibroglandular tissue, and would therefore not imply an increased carcinogenic risk. These findings highlight the complex relationship between lifestyle behaviors and breast tissue composition.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13058-026-02304-w. Background Breast cancer is the most frequent cancer among women worldwide, with approximately 2.3 million new diagnoses in 2022, and remains a leading cause of cancer-related mortality [1]. In Spain, it is the most frequently diagnosed tumor and the second with the highest mortality rate in women, accounting for 29% of new cases [2] and 14% of all cancer deaths [3] in 2024. Although survival rates have improved [4], the disease continues to represent a significant burden on public health, underscoring the urgent need for effective primary prevention and early detection strategies. Mammographic density (MD), which refers to the amount of dense fibroglandular tissue in the breast, is recognized as an independent risk factor for breast cancer. A meta-analysis combining eight studies observed a 13% increased risk of breast cancer for a 10-percentage point increase in MD [5]. MD is a dynamic trait that can be influenced by both non-modifiable factors (such as age, race, and genetic factors) and modifiable factors (such as weight gain, physical activity, alcohol consumption, and certain dietary habits) [6–8]. The World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) updated in 2018 their evidence-based lifestyle recommendations for cancer prevention, focusing on maintaining a healthy body weight, engaging in regular physical activity, and eating a healthy diet [9]. Previous meta-analyses have shown reductions in breast cancer risk of around 10% for each point increase in the adherence score to these ten recommendations [10, 11]. However, to the best of our knowledge, only one previous study has explored the relationship between adherence to the 2007 WCRF/AICR recommendations and MD [12]. Understanding whether and how current lifestyle recommendations to prevent cancer influence MD may provide insights into the mechanisms linking modifiable lifestyles and breast cancer. This research examines the association between adherence to the 2018 WCRF/AICR cancer prevention recommendations prior to diagnosis and percent MD at diagnosis in the tumor-free contralateral breast in a sample of women with breast cancer in Spain. Materials and methods Study population This research is part of the Breast Cancer & Density Association Study (BCDAS), a multicenter case-case study. A total of 1021 women diagnosed with breast cancer were recruited from the oncology departments in eight hospitals across four regions of Spain: Madrid, Catalonia, Navarre and the Valencian Region. Participants were recruited ambispectively, including women diagnosed between January 2014 and March 2019 who were subsequently interviewed between October 2016 and December 2019. The study inclusion criteria were: women with a histologically confirmed diagnosis of breast cancer (invasive or ductal in situ), aged 18 years or older, and able to complete a telephone-based epidemiological and dietary questionnaire administered by trained interviewers. Women were excluded if they had undergone surgery on their tumor-free breast prior to their cancer diagnosis, including procedures such as tissue removal, breast reconstruction, or augmentation. Women with a synchronous bilateral tumor were also excluded. Recruitment was performed by medical oncologists, who obtained written informed consent from all participants. The study protocol complied with the principles of the Declaration of Helsinki and received approval from the Research Ethics and Animal Welfare Committee of the Instituto de Salud Carlos III (CEI PI 32_2015-v2, June 2015). Among 1021 women diagnosed with breast cancer and initially recruited, 66 did not meet the eligibility criteria, 19 declined to participate, 156 were excluded due to unavailable MD data, 14 due to implausible caloric intake, 3 due to missing body mass index (BMI) information at diagnosis, and 4 due to incomplete data on one or more WCRF/AICR recommendations. The final sample included 759 participants. Study variables All breast cancer participants answered an epidemiological questionnaire that gathered detailed information on anthropometric measurements, sociodemographic data, personal and family medical history, lifestyle factors, such as tobacco and alcohol consumption, hormonal and reproductive health, and physical activity. Participants reported the types of leisure-time physical activities they had engaged in during the year prior to diagnosis for at least three months, specifying the frequency (hours per week) and duration (months) of each activity. Total weekly physical activity was then calculated by summing the minutes spent on each activity. Dietary habits during the five years prior to diagnosis were assessed using a food frequency questionnaire, tailored to the Spanish population and used in previous research [13, 14]. The ALEVINT 1.0 software [15] was used to calculate total energy intake in kilocalories per day and to convert each food item response into average daily or weekly consumption in grams. Following recruitment, hospital researchers completed a separate questionnaire, which included clinical and pathological data. Tumor subtypes were classified into three categories based on histopathological findings: Hormone receptor positive tumors (HR + : estrogen receptor positive and/or progesterone receptor positive with human epidermal growth factor receptor 2 (HER2) negative); HER2 positive tumors (HER2 + : irrespective of estrogen receptor or progesterone receptor status); and triple negative tumors (estrogen receptor negative, progesterone receptor negative and HER2 negative). 2018 WCRF/AICR score construction The 2018 standardised WCRF/AICR score was operationalized based on the previously published guidelines by Shams-White et al. [16, 17] and adapted to the BCDAS population. The Shams-White score comprises eight components: (1) be a healthy weight, (2) be physically active, (3) eat a diet rich in wholegrains, vegetables, fruit, and beans, (4) limit consumption of fast food and other processed foods high in fat, starches or sugar, (5) limit consumption of red and processed meat, (6) limit consumption of sugar-sweetened drinks, (7) limit alcohol consumption, and (8) for mothers, breastfeed their babies (optional component). Given the strong inverse relationship between BMI and MD percentage (women with a higher BMI have a greater amount of fatty tissue and, consequently, a lower percentage of dense tissue visible on mammography) [6, 7], for this study BMI was considered a confounding factor and was not included in the construction of the score. The breastfeeding component was also not included due to a lack of data. Consequently, the adherence score used in this study comprised six components. Scoring was applied following the established scoring guidelines [16, 17] as follows: 1 point was assigned when a recommendation was fully met, 0.5 points when it was partially met, and 0 points when it was not met. For the recommendation that included two sub-components (eating a diet rich in wholegrains, vegetables, fruit, and beans), the total score of one point was divided equally between both components (fruits and vegetables, and total fiber). An adapted variable for ultra-processed food consumption was defined according to the NOVA classification system [16], which categorizes foods according to the extent and purpose of food processing, rather than in terms of nutrient composition [18]. The percentage of total calories from ultra-processed foods was calculated for each participant and was categorized into tertiles within the study population, as recommended by the expert panel [16]. For the physical activity component, we operationalized total moderate-to-vigorous physical activity using leisure-time physical activity. Finally, physical activity cut-off points, measured in minutes per week, were derived from the Spanish national guidelines [19] in line with Shams-White et al.’s proposal [16, 17]. The total adherence score was calculated by summing the individual scores of each component. As a result, the overall score ranged from 0 (no recommendations met) to 6 (all recommendations fully met). Details on the operationalization, scoring, and adherence levels to the 2018 WCRF/AICR recommendations in the BCDAS study are presented in Supplementary Table 1. Based on the distribution of the total score, adherence was classified into three categories using rounded tertiles: low adherence (< 2.50), moderate adherence (2.50–3.25), and high adherence (≥ 3.25). Mammographic density assessment MD was assessed on the digital mammogram of the contralateral, tumor-free breast, obtained closest to the time of cancer diagnosis and prior to the initiation of any oncological treatment. A trained radiologist, blinded to all clinical and pathological information, evaluated the percent MD on the craniocaudal view using DM-Scan, a semi-automated software tool that estimates MD on two-dimensional digital mammograms and DICOM images. Previous studies have demonstrated that DM-Scan provides highly reproducible and valid measurements [20, 21]. Statistical analyses The 2018 WCRF/AICR score was described by means and standard deviations (SD) and compared across categories of participants characteristics using the ANOVA F-test. The prevalence of MD ≥ 50% was described using frequencies and percentages and compared across participants characteristics using logistic regression models adjusted for age and BMI, with statistical significance assessed using the Wald test. MD was categorized according to the Boyd scale, but combining the two extreme categories (A + B and E + F), due to small numbers: < 10% (reference), 10%– < 25%, 25%– < 50% and ≥ 50%. The prevalence of each of the four categories by tertiles of adherence to the WCRF/AICR recommendations was standardized to the distribution of selected confounding factors in the overall sample of participants, including recruitment region (Catalonia, Madrid, the Valencian Region, and Navarre), age at diagnosis (modeled as a continuous variable using splines), BMI at diagnosis (Kg/m2, modeled as a continuous variable using splines), educational level (primary education or less, high school/vocational training, university graduate), smoking status at diagnosis (non-smoker, ex-smoker, smoker), energy intake (< 1500, 1500–2000, > 2000 kcal/day), age at menarche (< 12, ≥ 12 years), age at first birth (nulliparous, < 25, 25–29, ≥ 30 years), contraceptives use (never, past use), menopausal status at diagnosis (pre/perimenopausal, postmenopausal), hormone replacement therapy use (never, past use), family history of breast cancer (none, second degree only, first degree), and prior breast biopsies with negative results (no, yes). We adjusted for total energy intake to better reflect differences in diet composition rather than differences in total food consumption. This standardization was accomplished by first fitting a multinomial logistic regression model relating the four MD categories with the tertiles of adherence to the 2018 WCRF/AICR recommendations, and adjusting for the above confounding factors, and then calculating the average predicted probabilities of having a MD < 10%, 10–25%, 25–50%, and ≥ 50%, as if every participant were in each adherence category with their confounding factors unchanged [22, 23]. We also estimated the standardized prevalence ratios (SPRs) with their 95% confidence intervals (95% CIs) for each of the four MD categories, comparing moderate and high adherence to the WCRF/AICR recommendations with low adherence. To assess possible non-linear associations, SPRs and their 95% CIs for each MD category as a smooth function of the 2018 WCRF/AICR score were obtained from a multinomial logistic regression model based on restricted cubic splines for adherence score, with knots at scores of 1.50, 2.75, and 4.50 (5th, 50th, and 95th percentiles). Standardized prevalences and SPRs of having a MD ≥ 50% stratifying by menopausal status at diagnosis, BMI at diagnosis, and breast cancer subtype were also calculated, fitting a binary logistic model and adjusting for the previously mentioned risk factors. Finally, for each specific recommendation we estimated the standardized prevalences of having a MD ≥ 50% in participants who did not meet or partially met that recommendation and in those who fully met it, fitting a binary logistic model adjusted for the above-mentioned factors and for the adherence score obtained from the remaining recommendations. SPRs with their 95% CIs for having a MD ≥ 50% were also assessed, comparing women who fully met the recommendation with those who did not meet or partially met it. Statistical analyses were performed using STATA/MP 18.0 (StataCorp LLC, College Station, TX, USA) and the figure was generated using R, version 4 (R Foundation for Statistical Computing). Results As shown in Supplementary Table 1, the recommendations most frequently followed by participants were limiting sugar-sweetened drinks consumption (55% did not consume any) and eating a diet rich in fruits and vegetables (51% consumed more than 400 g/day). In contrast, participants were less likely to meet the recommendations regarding a high-fiber diet (only 2% consumed at least 30 g/day) and limiting red and processed meat consumption (only 10% fully met it). Table 1 shows the sociodemographic, lifestyle and clinical characteristics of the study population at diagnosis, along with the mean WCRF/AICR score and the percentage of women with MD ≥ 50% according to these characteristics. About half of the participants (49%) were interviewed during the first year after diagnosis. The mean MD was 26.2% (SD = 17.4). The average adherence score was 2.9 out of 6 points (SD = 0.9), and 10% of women did not fully adhere to any of the recommendations. Participants had a mean age of 55 years at diagnosis (SD = 11.3), being 62% of them postmenopausal. A university degree had been completed by 33% of the women, 23% reported being smokers at the time of diagnosis, and 17% were nulliparous. Higher adherence to the WCRF/AICR recommendations was observed in older participants, women with lower levels of education, non-smokers, those with lower energy intake, having had their first child before the age of 25, women who did not use oral contraceptives or had used hormone replacement therapy, postmenopausal women, and those with first-degree relatives with breast cancer. Regarding MD, younger women, those with lower BMI, and women with menarche at age 12 or older showed a greater prevalence of high MD (≥ 50%).Table 1Mean 2018 WCRF/AICR score and percentage of women with mammographic density ≥ 50% by characteristics at diagnosis of breast cancer patientsTotalWCRF/AICR adherence scoreMammographic density ≥ 50%n(%)mean(SD)p-valuean(%)bp-valuecTotal759(100.0)2.92(0.95) 79(10.4) Recruiting regionCatalonia380(50.1)2.89(0.95)0.28625(6.6)0.021Madrid222(29.2)2.91(0.91)30(13.5)Valencian Region60(7.9)3.14(0.92)2(3.3)Navarre97(12.8)2.96(1.01)22(22.7)Time between diagnosis and interviewd, years < 1372(49.0)2.92(0.91)0.41834(9.1)0.2491–2278(36.6)2.89(0.93)30(10.8)3–5109(14.4)3.03(1.09)15(13.8)Agee, years < 49234(30.8)2.64(0.91) < 0.00155(23.5) < 0.00149–60271(35.7)2.91(0.96)23(8.5) > 60254(33.5)3.20(0.89)1(0.4)Educational levelPrimary education or less269(35.4)3.04(0.92)0.04610(3.7)0.120High school / vocational training237(31.2)2.84(1.00)22(9.3)University graduate253(33.3)2.88(0.91)47(18.6)Body mass index, kg/m2 < 25376(49.5)2.90(0.95)0.76771(18.9) < 0.00125–29.9247(32.5)2.93(0.94)7(2.8) ≥ 30136(17.9)2.97(0.95)1(0.7)Energy intake, kcal/day < 1500243(32.0)3.23(0.85) < 0.00127(11.1)0.4561500–1999307(40.4)2.97(0.95)28(9.1) ≥ 2000209(27.5)2.50(0.91)24(11.5)Smoking statusNon-smoker374(49.3)3.04(0.95)0.00128(7.5)0.216Ex-smoker212(28.0)2.89(0.96)21(9.9)Smoker172(22.7)2.71(0.88)30(17.4)Age at menarche, years < 12205(27.3)2.90(0.94)0.6879(4.4)0.004 ≥ 12546(72.7)2.93(0.95)69(12.6)Age at first birth, yearsNulliparous131(17.3)2.88(1.03)0.07017(13)0.308 < 25208(27.4)3.04(0.95)8(3.8)25–29214(28.2)2.94(0.95)15(7) ≥ 30205(27.0)2.80(0.88)39(19)Contraceptives useNever284(37.4)3.11(0.96) < 0.00123(8.1)0.667Past use475(62.6)2.81(0.92)56(11.8)Menopausal statusPre/perimenopausal290(38.2)2.70(0.93) < 0.00161(21)0.119Postmenopausal469(61.8)3.06(0.93)18(3.8)Hormone replacement therapy useNever682(90.1)2.90(0.94)0.03776(11.1)0.658Past use75(9.9)3.14(1.01)3(4)Family history of breast cancerNone445(58.6)2.91(0.93)0.03754(12.1)0.202Second degree only169(22.3)2.82(0.98)13(7.7)First degree145(19.1)3.09(0.95)12(8.3)Previous breast biopsiesNo600(79.1)2.95(0.93)0.07360(10)0.342Yes159(20.9)2.80(1.01)19(11.9)Tumor subtypeHR + 473(68.8)2.95(0.93)0.57045(9.5)0.225HER2 + 132(19.2)2.85(0.99)18(13.6)Triple negative83(12.1)2.95(0.88)9(10.8)Abbreviations: HR + = hormone receptor positive; HER2 + = human epidermal growth factor receptor 2 positive; SD = standard deviationap-values derived from one-way ANOVA (F-test) comparing mean scores across categoriesbPercentages in the MD columns correspond to row percentages, representing the prevalence of high MD within each category of patient characteristics. The complementary percentages within each category correspond to the prevalence of MD < 50%cp-values derived from logistic regression models adjusted for age and BMI at diagnosis; p-values correspond to Wald tests for overall heterogeneity across categoriesdInterview refers to the epidemiological and dietary assessmenteIn tertiles Table 1Mean 2018 WCRF/AICR score and percentage of women with mammographic density ≥ 50% by characteristics at diagnosis of breast cancer patientsTotalWCRF/AICR adherence scoreMammographic density ≥ 50%n(%)mean(SD)p-valuean(%)bp-valuecTotal759(100.0)2.92(0.95) 79(10.4) Recruiting regionCatalonia380(50.1)2.89(0.95)0.28625(6.6)0.021Madrid222(29.2)2.91(0.91)30(13.5)Valencian Region60(7.9)3.14(0.92)2(3.3)Navarre97(12.8)2.96(1.01)22(22.7)Time between diagnosis and interviewd, years < 1372(49.0)2.92(0.91)0.41834(9.1)0.2491–2278(36.6)2.89(0.93)30(10.8)3–5109(14.4)3.03(1.09)15(13.8)Agee, years < 49234(30.8)2.64(0.91) < 0.00155(23.5) < 0.00149–60271(35.7)2.91(0.96)23(8.5) > 60254(33.5)3.20(0.89)1(0.4)Educational levelPrimary education or less269(35.4)3.04(0.92)0.04610(3.7)0.120High school / vocational training237(31.2)2.84(1.00)22(9.3)University graduate253(33.3)2.88(0.91)47(18.6)Body mass index, kg/m2 < 25376(49.5)2.90(0.95)0.76771(18.9) < 0.00125–29.9247(32.5)2.93(0.94)7(2.8) ≥ 30136(17.9)2.97(0.95)1(0.7)Energy intake, kcal/day < 1500243(32.0)3.23(0.85) < 0.00127(11.1)0.4561500–1999307(40.4)2.97(0.95)28(9.1) ≥ 2000209(27.5)2.50(0.91)24(11.5)Smoking statusNon-smoker374(49.3)3.04(0.95)0.00128(7.5)0.216Ex-smoker212(28.0)2.89(0.96)21(9.9)Smoker172(22.7)2.71(0.88)30(17.4)Age at menarche, years < 12205(27.3)2.90(0.94)0.6879(4.4)0.004 ≥ 12546(72.7)2.93(0.95)69(12.6)Age at first birth, yearsNulliparous131(17.3)2.88(1.03)0.07017(13)0.308 < 25208(27.4)3.04(0.95)8(3.8)25–29214(28.2)2.94(0.95)15(7) ≥ 30205(27.0)2.80(0.88)39(19)Contraceptives useNever284(37.4)3.11(0.96) < 0.00123(8.1)0.667Past use475(62.6)2.81(0.92)56(11.8)Menopausal statusPre/perimenopausal290(38.2)2.70(0.93) < 0.00161(21)0.119Postmenopausal469(61.8)3.06(0.93)18(3.8)Hormone replacement therapy useNever682(90.1)2.90(0.94)0.03776(11.1)0.658Past use75(9.9)3.14(1.01)3(4)Family history of breast cancerNone445(58.6)2.91(0.93)0.03754(12.1)0.202Second degree only169(22.3)2.82(0.98)13(7.7)First degree145(19.1)3.09(0.95)12(8.3)Previous breast biopsiesNo600(79.1)2.95(0.93)0.07360(10)0.342Yes159(20.9)2.80(1.01)19(11.9)Tumor subtypeHR + 473(68.8)2.95(0.93)0.57045(9.5)0.225HER2 + 132(19.2)2.85(0.99)18(13.6)Triple negative83(12.1)2.95(0.88)9(10.8)Abbreviations: HR + = hormone receptor positive; HER2 + = human epidermal growth factor receptor 2 positive; SD = standard deviationap-values derived from one-way ANOVA (F-test) comparing mean scores across categoriesbPercentages in the MD columns correspond to row percentages, representing the prevalence of high MD within each category of patient characteristics. The complementary percentages within each category correspond to the prevalence of MD < 50%cp-values derived from logistic regression models adjusted for age and BMI at diagnosis; p-values correspond to Wald tests for overall heterogeneity across categoriesdInterview refers to the epidemiological and dietary assessmenteIn tertiles Mean 2018 WCRF/AICR score and percentage of women with mammographic density ≥ 50% by characteristics at diagnosis of breast cancer patients Abbreviations: HR + = hormone receptor positive; HER2 + = human epidermal growth factor receptor 2 positive; SD = standard deviation ap-values derived from one-way ANOVA (F-test) comparing mean scores across categories bPercentages in the MD columns correspond to row percentages, representing the prevalence of high MD within each category of patient characteristics. The complementary percentages within each category correspond to the prevalence of MD < 50% cp-values derived from logistic regression models adjusted for age and BMI at diagnosis; p-values correspond to Wald tests for overall heterogeneity across categories dInterview refers to the epidemiological and dietary assessment eIn tertiles Standardized prevalences of MD categories (< 10%; 10–25%; 25–50% and ≥ 50%) according to tertiles of the 2018 WCRF/AICR score are shown in Table 2. The standardized prevalence of having a high MD (≥ 50%) was 73% higher among women with higher adherence to the recommendations compared to those with lower adherence (SPR high vs low = 1.73; 95%CI = 1.03–2.91). Figure 1 depicts the SPRs and their 95%CIs for the four MD categories as a smoothed function of the 2018 WCRF/AICR score. The SPR for MD < 10% decreased with increasing adherence up to a score of three points, after which it increased. For MD between 10 and 25%, the SPR decreased progressively. For MD between 25 and 50%, the SPR showed an increasing trend up to a score of approximately 3 points, followed by a decrease, and for MD ≥ 50% the SPR showed a linear increasing trend.Table 2Standardized prevalences and prevalence ratios of mammographic density categories by tertiles of the 2018 WCRF/AICR score prior to diagnosisMammographic density (%) < 10%10–25%25–50% ≥ 50%Adherence to WCRF/AICR RecommendationsNnStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)TOTAL74713727026278Low (< 2.50)2323619.6 (14.9–25.4)1.00 (ref.)7936.9 (30.9–43.5)1.00 (ref.)9436.0 (30.6–41.7)1.00 (ref.)237.5 (5.1–10.8)1.00 (ref.)Moderate (2.50–3.25)2874315.8 (12.3–20.2)0.81 (0.56–1.16)10436.7 (31.7–42.0)0.99 (0.80–1.24)10735.5 (30.5–40.7)0.99 (0.79–1.22)3312.0 (9.0–15.7)1.60 (0.99–2.61)High (> 3.25)2285819.9 (15.9–24.7)1.01 (0.71–1.46)8734.7 (29.1–40.9)0.94 (0.73–1.21)6132.4 (26.6–38.9)0.90 (0.70–1.17)2212.9 (9.2–17.9)1.73 (1.03–2.91)Abbreviations: N = Number of women; n = number of women with the corresponding percentage of mammographic densityaStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer and previous breast biopsies in the entire sample of women with breast cancerFig. 1Standardized prevalence ratios for mammographic density < 10% (A), 10–25% (B), 25–50% (C), and ≥ 50% (D) as a smooth function of pre-diagnosis adherence score to the 2018 WCRF/AICR recommendations among women with breast cancer. Curves represent standardized prevalence ratios (solid lines) and their 95% confidence intervals (dashed lines) obtained from a multinomial logistic regression model based on restricted cubic splines for adherence to the WCRF/AICR recommendations with knots at scores of 1.50, 2.75, and 4.50 (5th, 50th, and 95th percentiles). The reference value (prevalence ratio = 1) was set at a score of 2. Prevalence ratios were standardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer and previous breast biopsies in the entire sample of women with breast cancer. Histogram represents the distribution of adherence score to the WCRF/AICR recommendations in the overall sample Table 2Standardized prevalences and prevalence ratios of mammographic density categories by tertiles of the 2018 WCRF/AICR score prior to diagnosisMammographic density (%) < 10%10–25%25–50% ≥ 50%Adherence to WCRF/AICR RecommendationsNnStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)nStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)TOTAL74713727026278Low (< 2.50)2323619.6 (14.9–25.4)1.00 (ref.)7936.9 (30.9–43.5)1.00 (ref.)9436.0 (30.6–41.7)1.00 (ref.)237.5 (5.1–10.8)1.00 (ref.)Moderate (2.50–3.25)2874315.8 (12.3–20.2)0.81 (0.56–1.16)10436.7 (31.7–42.0)0.99 (0.80–1.24)10735.5 (30.5–40.7)0.99 (0.79–1.22)3312.0 (9.0–15.7)1.60 (0.99–2.61)High (> 3.25)2285819.9 (15.9–24.7)1.01 (0.71–1.46)8734.7 (29.1–40.9)0.94 (0.73–1.21)6132.4 (26.6–38.9)0.90 (0.70–1.17)2212.9 (9.2–17.9)1.73 (1.03–2.91)Abbreviations: N = Number of women; n = number of women with the corresponding percentage of mammographic densityaStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer and previous breast biopsies in the entire sample of women with breast cancer Standardized prevalences and prevalence ratios of mammographic density categories by tertiles of the 2018 WCRF/AICR score prior to diagnosis Abbreviations: N = Number of women; n = number of women with the corresponding percentage of mammographic density aStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer and previous breast biopsies in the entire sample of women with breast cancer Standardized prevalence ratios for mammographic density < 10% (A), 10–25% (B), 25–50% (C), and ≥ 50% (D) as a smooth function of pre-diagnosis adherence score to the 2018 WCRF/AICR recommendations among women with breast cancer. Curves represent standardized prevalence ratios (solid lines) and their 95% confidence intervals (dashed lines) obtained from a multinomial logistic regression model based on restricted cubic splines for adherence to the WCRF/AICR recommendations with knots at scores of 1.50, 2.75, and 4.50 (5th, 50th, and 95th percentiles). The reference value (prevalence ratio = 1) was set at a score of 2. Prevalence ratios were standardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer and previous breast biopsies in the entire sample of women with breast cancer. Histogram represents the distribution of adherence score to the WCRF/AICR recommendations in the overall sample When stratifying by menopausal status, BMI, and breast cancer subtype (Supplementary Table 2), estimates suggested possible differences in the association between adherence to the recommendations and high percent MD across menopausal groups, with a borderline evidence of effect modification (p for homogeneity = 0.05). No clear evidence of effect modification was observed by BMI or tumor subtype. Given the limited statistical power for subgroup analyses, these findings should be interpreted cautiously and not as subgroup-specific effects. Finally, Table 3 shows standardized prevalences and SPRs of having MD ≥ 50% by adherence to specific recommendations. Among these, adherence to the physical activity recommendation was the most clearly associated with a higher prevalence of MD ≥ 50% (SPR fully met vs. no/partially met = 1.51; 95%CI = 1.02–2.25).Table 3Standardized prevalences and prevalence ratios of mammographic density ≥ 50% by adherence to each of the WCRF/AICR Recommendations prior to diagnosisAdherence to WCRF/AICR recommendationsMammographic density ≥ 50%NStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)Be physically activeNot met / partially met579.5 (7.6–11.7)1.00 (ref.)Fully met (≥ 300 min/week)2114.3 (10.2–19.7)1.51 (1.02–2.25)Eat a diet rich in fruits and vegetablesNot met / partially met439.9 (7.7–12.8)1.00 (ref.)Fully met (≥ 400 g/day)3511.1 (8.3–14.7)1.12 (0.74–1.68)Eat a diet rich in fiberNot met / partially met7510.3 (8.6–12.4)1.00 (ref.)Fully met (≥ 30 g/day)316.0 (5.4–38.7)1.55 (0.55–4.33)Limit consumption of “fast foods”Not met / partially met5710.2 (8.2–12.7)1.00 (ref.)Fully met (< 17.1% of energy intake)2111.1 (7.6–15.8)1.08 (0.69–1.69)Limit consumption of red and processed meatNot met / partially met6910.1 (8.3–12.2)1.00 (ref.)Fully met (< 500 and < 21 g/week)914.0 (8.2–22.7)1.38 (0.80–2.40)Limit consumption of sugar-sweetened drinksNot met / partially met399.5 (7.2–12.4)1.00 (ref.)Fully met (no consumption)3911.6 (8.9–14.9)1.22 (0.82–1.81)Limit alcohol consumptionNot met / partially met5410.1 (8.1–12.6)1.00 (ref.)Fully met (no consumption)2411.2 (8.0–15.5)1.11 (0.73–1.68)Abbreviations: N = Number of women with mammographic density ≥ 50%aStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer, previous breast biopsies and the score of adherence to the other recommendations in the entire sample of women with breast cancer Table 3Standardized prevalences and prevalence ratios of mammographic density ≥ 50% by adherence to each of the WCRF/AICR Recommendations prior to diagnosisAdherence to WCRF/AICR recommendationsMammographic density ≥ 50%NStandardized prevalencea % (95% CI)Standardized prevalence ratioa (95% CI)Be physically activeNot met / partially met579.5 (7.6–11.7)1.00 (ref.)Fully met (≥ 300 min/week)2114.3 (10.2–19.7)1.51 (1.02–2.25)Eat a diet rich in fruits and vegetablesNot met / partially met439.9 (7.7–12.8)1.00 (ref.)Fully met (≥ 400 g/day)3511.1 (8.3–14.7)1.12 (0.74–1.68)Eat a diet rich in fiberNot met / partially met7510.3 (8.6–12.4)1.00 (ref.)Fully met (≥ 30 g/day)316.0 (5.4–38.7)1.55 (0.55–4.33)Limit consumption of “fast foods”Not met / partially met5710.2 (8.2–12.7)1.00 (ref.)Fully met (< 17.1% of energy intake)2111.1 (7.6–15.8)1.08 (0.69–1.69)Limit consumption of red and processed meatNot met / partially met6910.1 (8.3–12.2)1.00 (ref.)Fully met (< 500 and < 21 g/week)914.0 (8.2–22.7)1.38 (0.80–2.40)Limit consumption of sugar-sweetened drinksNot met / partially met399.5 (7.2–12.4)1.00 (ref.)Fully met (no consumption)3911.6 (8.9–14.9)1.22 (0.82–1.81)Limit alcohol consumptionNot met / partially met5410.1 (8.1–12.6)1.00 (ref.)Fully met (no consumption)2411.2 (8.0–15.5)1.11 (0.73–1.68)Abbreviations: N = Number of women with mammographic density ≥ 50%aStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer, previous breast biopsies and the score of adherence to the other recommendations in the entire sample of women with breast cancer Standardized prevalences and prevalence ratios of mammographic density ≥ 50% by adherence to each of the WCRF/AICR Recommendations prior to diagnosis Abbreviations: N = Number of women with mammographic density ≥ 50% aStandardized to the overall distribution of age, recruiting region, body mass index, energy intake, educational level, smoking status, age at menarche, age at first birth, contraceptives use, menopausal status, hormone replacement therapy use, family history of breast cancer, previous breast biopsies and the score of adherence to the other recommendations in the entire sample of women with breast cancer Discussion The results of the present study suggest that adherence to the 2018 WCRF/AICR cancer prevention recommendations prior to breast cancer diagnosis was not generally associated with MD at diagnosis. However, a higher prevalence of high MD (≥ 50%) was observed among patients with greater overall adherence, particularly those who met the recommendation of being physically active. Numerous epidemiological studies have evaluated the relationship between adherence to the WCRF/AICR recommendations and breast cancer risk or survival [10, 11]. However, to our knowledge, only one previous study has specifically investigated the relationship between such adherence and MD [12]. Contrary to our results, Castelló et al. found that greater adherence to the 2007 WCRF/AICR guidelines was significantly associated with lower MD, especially among postmenopausal women [12]. This discrepancy may be partly explained by methodological differences, including population characteristics (their participants were healthy women recruited through population-based screening programs), MD assessment method (MD was visually read by a single radiologist), and operationalization of the score. Our observation that breast cancer women with higher adherence to the WCRF/AICR recommendations showed greater MD in the tumor-free breast could be explained by lower body and breast fat mass in these women (even at equal BMI), given that “be physically active” was the more clearly associated recommendation, and exercise interventions in breast cancer populations reduce percent body fat and improves lean mass [24]. Since MD is a relative measure, representing the proportion of fibroglandular (dense) tissue to total breast tissue (fibroglandular + adipose), a reduction in breast adiposity can lead to a higher percentage of dense fibroglandular tissue on mammography, even if the absolute amount of fibroglandular tissue remains unchanged. Nevertheless, neither direct measures of body composition nor absolute dense/non-dense area in the mammogram are available in our study, precluding any direct verification of this explanation. Although the relationship between physical activity and MD remains inconclusive, with majority of studies having found no association [25], Trinh et al. reported that, in a population-based cross-sectional study of 38,913 Swedish women, higher levels of physical activity were associated with lower absolute dense and non-dense volume, but with higher percent dense volume [26], which would be in line with our hypothesis. Anyway, these findings should not be interpreted like evidence of a possible harmful effect of the WCRF/AICR recommendations, but rather as a methodological consequence of how MD has been measured. Notably, given the substantial evidence supporting that higher levels of physical activity are associated with a reduced risk of breast cancer [27] and an improved health-related quality of life among breast cancer patients [28], these results must not lead to changes in lifestyle counselling or clinical management from a clinical and public health perspective. Clinicians should continue to encourage healthy lifestyle habits, including regular physical activity, among both women with breast cancer and the general population. This study has several limitations that should be considered when interpreting the findings. First, the cross-sectional design limits our ability to establish temporal relationships between adherence to the WCRF/AICR cancer prevention recommendations and MD, thereby precluding causal interpretations. Another important limitation is the inability to assess the absolute MD of the participants, since different mammogram devices were used in the different study settings and DICOM pixel-size metadata were unavailable. In addition, because this is a case–case study restricted to breast cancer patients, the results cannot be directly extrapolated to primary prevention settings. Survivorship bias may also have influenced our results, as women with more aggressive disease or poorer health may have been less likely to survive to be recruited and/or less willing to participate, particularly given the ambispective study design. This could have resulted in a selective sample in which both lifestyle behaviors and MD may differ from those of the overall population of incident breast cancer cases. The use of self-reported data may entail recall bias, particularly among participants interviewed more than one year after diagnosis. Although questionnaires explicitly referred to pre-diagnostic time windows (five years and one year prior to diagnosis for diet and physical activity, respectively), responses may have been influenced by post-diagnosis behavioral changes, potentially leading to misclassification. Moreover, heterogeneity in interval between diagnosis and interview (with 49% of participants interviewed within the first year, 37% 1–2 years after diagnosis, and 14% after 3–5 years) may have introduced additional variability. However, findings from the sensitivity analysis restricted to women interviewed within one year of diagnosis (Supplementary Table 3) were consistent with those from the main analysis. Furthermore, the relatively small number of participants with high MD limited the ability to accurately estimate its prevalence according to each recommendation and to perform stratified analyses by menopausal status, BMI, and histological subtype. Consequently, these results should be interpreted with caution, as the robustness of the conclusions could be compromised. Lastly, although we controlled for a wide range of potential confounding factors, residual confounding cannot be entirely ruled out. Despite these limitations, the study has several strengths that should be considered. To our knowledge, this is the first study to date evaluating the association between adherence to the 2018 WCRF/AICR cancer prevention recommendations and MD. Furthermore, the use of direct standardization to estimate prevalence and prevalence ratios allowed us to control for a broad range of confounders. In addition, MD was measured on a continuous scale using a validated computer-assisted method and evaluated by a single experienced reader, favoring internal consistency. Finally, the standardized scoring system by Shams-White et al. provides a consistent and reproducible method to quantify adherence to the WCRF/AICR recommendations. It enables comparisons across studies, reduces subjectivity, and supports rigorous analysis of lifestyle and cancer-related outcomes. Conclusions Our findings suggest that adherence to the 2018 WCRF/AICR cancer prevention recommendations included in our score (which excluded the BMI component) is not consistently associated with MD in women with breast cancer. However, higher overall adherence, particularly to the physical activity recommendation, was associated with a higher prevalence of high MD, possibly due to reduced breast adipose tissue. Because percent MD is a relative measure, this increase should not be interpreted as indicating a higher absolute amount of fibroglandular tissue or greater carcinogenic potential, nor as suggesting that healthier behaviors increase breast cancer risk. Importantly, physical activity is consistently associated with a reduced breast cancer risk, despite its association with a higher percent MD in our study. Furthermore, these findings should not be extrapolated to cancer-free women, the general population, or primary prevention contexts. Overall, our results reflect the complex interplay between lifestyle behaviors and breast tissue composition, and underscore the need for additional studies with larger sample sizes that consider both relative MD and absolute measures of fibroglandular and adipose tissue. Supplementary Information Supplementary Material 1. Supplementary Material 1.