Mammographic Density Measured with Quantitative Computer-aided Method: Comparison with Radiologists' Estimates and BI-RADS Categories

Objectives: This study aimed to evaluate the impact of digital breast tomosynthesis (DBT) on BI-RADS categorization compared to conventional mammography (MMG), across different mammographic breast density types. Methods: In this retrospective study, 520 female patients aged 35–85 years who underwent both MMG and DBT between 2023 and 2025 were included. Standard craniocaudal (CC) views were acquired for MMG and mediolateral oblique (MLO) views for DBT. Synthetic MLO (sMLO) images were automatically generated by the device from DBT data. Patients were categorized according to the breast density (Type A–D), and all images were reviewed in consensus by two experienced radiologists. Breast Imaging Reporting and Data System (BI-RADS) categories, lesion types, were recorded. BI-RADS categorizations from MMG and DBT were compared overall and within density subgroups. Results: The mean age was 53.0 ± 9.9 years. Breast density distribution was: Type A (3.8%, n=20), B (28.5%, n= 148), C (43.5%, n=226), and D (24.2%, n=126). BI-RADS classifications differed significantly between MMG and DBT, especially in dense breasts. Of the BI-RADS 2 cases on MMG, 21.6% (n=64) were reclassified as BI-RADS 3 or 4 with DBT, while 6.3% (n=7) of BI-RADS 3 cases were downgraded. BI-RADS 5 categorization showed complete agreement between modalities, though 33% (n=5) of BI-RADS 4 cases were downgraded. At least one pathological finding was observed in 91.3% (n=475) of patients, most commonly nodular opacities (57.1%, n=297), vascular wall calcifications (13.3%, n=69), and coarse calcifications (5.4%, n=28). McNemar analysis revealed significant reclassification between BI-RADS 2 and 3 in Types C and D (P<0.001). Conclusions: DBT significantly alters BI-RADS categorization in dense breasts. These findings support the potential role of DBT as a complementary tool in screening protocols, particularly for patients with heterogeneously dense or extremely dense breast tissue.

To develop a scoring system allowing quantification of the probability of malignancy of suspicious microcalcifications based on Breast Imaging Reporting and Data System (BI-RADS). A total of 163 microcalcification lesions surgically excised in 147 women (aged 28-65 years) were included. Two radiologists independently reviewed each lesion with BI-RADS (4th edition). The interobserver agreement and positive predictive value (PPV) for each descriptor were determined and multivariate analysis was used to develop a scoring system. The scores were compared between benign and malignant lesions and among BI-RADS categories by using the two-sample t-test or the analysis of variance. To assess the discriminative power of the scoring system, the area under the receiver-operating characteristic curve (AUC) and cutoff values for categorization were determined. For the test of the scoring system, the validation data set from a different facility was applied. Interobserver agreement was fair to moderate for distribution, morphology, and category (κ = 0.45, 0.40, and 0.37). PPVs were significantly different among BI-RADS descriptors and categories (P < 0.0001). Of the scoring system developed, the AUC was 0.75. The scores between benign and malignant lesions were significantly different. By means of cutoff values, PPV in category 4a, 4b, 4c, and 5 was 7.0%, 15.0%, 44.8%, and 83.3%, respectively. For the validation data set, the AUC was 0.79 and the PPV in category 4a, 4b, 4c, and 5 was 9.4%, 24.1%, 62.5%, and 80.0%, respectively. In suspicious microcalcifications, our scoring system based on BI-RADS (4th edition) could help to derive a specific final category with good stratification of the probability of malignancy.

This study aimed to evaluate the performance of large language models (LLMs) and multimodal LLMs in interpreting the Breast Imaging Reporting and Data System (BI-RADS) categories and providing clinical management recommendations for breast radiology in text-based and visual questions. This cross-sectional observational study involved two steps. In the first step, we compared ten LLMs (namely ChatGPT 4o, ChatGPT 4, ChatGPT 3.5, Google Gemini 1.5 Pro, Google Gemini 1.0, Microsoft Copilot, Perplexity, Claude 3.5 Sonnet, Claude 3 Opus, and Claude 3 Opus 200K), general radiologists, and a breast radiologist using 100 text-based multiple-choice questions (MCQs) related to the BI-RADS Atlas 5th edition. In the second step, we assessed the performance of five multimodal LLMs (ChatGPT 4o, ChatGPT 4V, Claude 3.5 Sonnet, Claude 3 Opus, and Google Gemini 1.5 Pro) in assigning BI-RADS categories and providing clinical management recommendations on 100 breast ultrasound images. The comparison of correct answers and accuracy by question types was analyzed using McNemar's and chi-squared tests. Management scores were analyzed using the Kruskal- Wallis and Wilcoxon tests. Claude 3.5 Sonnet achieved the highest accuracy in text-based MCQs (90%), followed by ChatGPT 4o (89%), outperforming all other LLMs and general radiologists (78% and 76%) (P < 0.05), except for the Claude 3 Opus models and the breast radiologist (82%) (P > 0.05). Lower-performing LLMs included Google Gemini 1.0 (61%) and ChatGPT 3.5 (60%). Performance across different categories of showed no significant variation among LLMs or radiologists (P > 0.05). For breast ultrasound images, Claude 3.5 Sonnet achieved 59% accuracy, significantly higher than other multimodal LLMs (P < 0.05). Management recommendations were evaluated using a 3-point Likert scale, with Claude 3.5 Sonnet scoring the highest (mean: 2.12 ± 0.97) (P < 0.05). Accuracy varied significantly across BI-RADS categories, except Claude 3 Opus (P < 0.05). Gemini 1.5 Pro failed to answer any BI-RADS 5 questions correctly. Similarly, ChatGPT 4V failed to answer any BI-RADS 1 questions correctly, making them the least accurate in these categories (P < 0.05). Although LLMs such as Claude 3.5 Sonnet and ChatGPT 4o show promise in text-based BI-RADS assessments, their limitations in visual diagnostics suggest they should be used cautiously and under radiologists' supervision to avoid misdiagnoses. This study demonstrates that while LLMs exhibit strong capabilities in text-based BI-RADS assessments, their visual diagnostic abilities are currently limited, necessitating further development and cautious application in clinical practice.

Objectives: To assess the performance of elastography (ES) and ultrasound (US) in predicting the malignancy of breast lesions and to compare their combined diagnostic value with that of magnetic resonance imaging (MRI).Materials and Methods: The study prospectively enrolled 242 female patients with dense breasts treated in 35 heath care facilities in China between November 2018 and October 2019. Based on conventional US and elastography, radiologists classified the degree of suspicion of breast lesions according to the US Breast Imaging Reporting and Data System (BI-RADS) criteria. The diagnostic value was compared between US BI-RADS and MRI BI-RADS, with pathological results used as the reference standard.Results: The results demonstrated that irregular tumor shape, a nonparallel growth orientation, indistinct margins, angular contours, microcalcifications, color Doppler flow and ES score on US imaging were significantly related to breast cancer in dense breasts (P=0.001; P=0.001; P=0.008; P<0.001; P=0.019; P=0.008; P=0.002, respectively). The sensitivity, specificity, PPV, NPV, accuracy and AUC of US BI-RADS category were 94.7%, 90.7%, 95.8%, 88.0%, 93.4% and 0.93 (95%CI, 0.88-0.97), respectively, while those of MRI BI-RADS category were 98.2%, 57.5%, 84.3%, 83.3%, 86.0% and 0.78 (95%CI, 0.71-0.85), respectively. MRI BI-RADS showed a significantly higher sensitivity than US BI-RADS (98.2% vs 94.7%, P=0.043), whereas US BI-RADS showed significantly higher specificity (90.7% vs 57.5%, P<0.001). US BI-RADS showed better diagnostic efficiency in differentiating nodules in dense breasts than MRI BI-RADS (AUC 0.93 vs 0.78, P<0.001).Conclusion: By combining the use of ES and conventional US, US BI-RADS had better diagnostic efficiency in differentiating nodules in dense breasts than MRI. For the diagnosis of malignant tumors in patients with dense breasts, MRI and US BI-RADS can be used as supplemental diagnostic tools to detect lesions, with US BI-RADS considered the preferred adjunctive resource.

Breast imaging reporting and data system (BI-RADS) provides standard descriptors but not detailed decision rules for characterizing breast lesions. Diffusion-weighted imaging (DWI) and T2-weighted imaging (T2WI) are also not incorporated in the BI-RADS. Several multiparametric magnetic resonance imaging (mpMRI)-based decision rules have been developed to differentiate breast lesions, but lack external validation. This study aims to externally validate several mpMRI-based decision rules for characterizing breast lesions and compare them with Kaiser score and BI-RADS category. There were 206 patients with 218 pathology-proven breast lesions (99 malignancies) included in this retrospective study from January 2018 to May 2018. Two radiologists blinded to pathology evaluated breast lesions according to the three mpMRI-based decision rules (Kim, Istomin, Zhong) and Kaiser score. BI-RADS category was extracted from radiology reports and also analysed. The diagnostic performances of the four decision rules and BI-RADS category were calculated and compared for different lesion types [mass and non-mass enhancement (NME)] and size (≤10 and >10 mm). The unnecessary biopsy rates for BI-RADS 4 lesions were calculated by the four decision rules. The three mpMRI-based decision rules showed area under the curve (AUC) of 0.81-0.87 for all lesions, 0.86-0.92 for mass lesions, 0.68-0.82 for NME, and 0.68-0.87 for lesion size ≤10 mm, 0.82-0.87 for lesion size >10 mm. Kaiser score showed the highest diagnostic performance for all subgroups except for lesion size ≤10 mm. No significant differences were found in AUC between Kaiser score and BI-RADS category. The mpMRI-based decision rules showed high sensitivity of 100% in all subgroups at the expense of low specificity (range, 2.9-41.2%). In contrast, Kaiser score demonstrated a significantly higher specificity of 73.5-92.9% than the three mpMRI-based decision rules at the cost of a decreased sensitivity (range, 60.0-93.6%) in different subgroups. The unnecessary biopsy rates for BI-RADS 4 lesions were 9.8% (Istomin), 12.2% (Zhong), 14.6% (Kim) and 70.7% (Kaiser score), respectively. The mpMRI-based decision rules showed high sensitivity but low specificity for characterizing breast lesions, and their diagnostic efficiencies were inferior to Kaiser score and BI-RADS category.

The Breast Imaging Reporting and Data System (BI-RADS) was introduced in 1993 to standardize the interpretation of mammograms. Though many studies have assessed the validity of the system, fewer have examined its reliability. Our objective is to identify predictors of reliability as measured by the kappa statistic. We identified studies conducted between 1993 and 2009 which reported kappa values for interpreting mammograms using any edition of BI-RADS. Bivariate and multivariate multilevel analyses were used to examine associations between potential predictors and kappa values. We identified ten eligible studies, which yielded 88 kappa values for the analysis. Potential predictors of kappa included: whether or not the study included negative cases, whether single- or two-view mammograms were used, whether or not mammograms were digital versus screen-film, whether or not the fourth edition of BI-RADS was utilized, the BI-RADS category being evaluated, whether or not readers were trained, whether or not there was an overlap in readers' professional activities, the number of cases in the study and the country in which the study was conducted. Our best multivariate model identified training, use of two-view mammograms and BI-RADS categories (masses, calcifications, and final assessments) as predictors of kappa. Training, use of two-view mammograms and focusing on mass description may be useful in increasing reliability in mammogram interpretation. Calcification and final assessment descriptors are areas for potential improvement. These findings are important for implementing policies in BI-RADS use before introducing the system in different settings and improving current implementations.

ABSTRACTObjectives: A comparison of mammographic breast densities of women living in London with those of women living in rural and suburban areas.Design and methods: Using the standard four American College of Radiology Breast Imaging Reporting and Data System (BIRADS) categories of mammographic density, 318 mammograms of women from London and 654 mammograms of women from outside the capital aged 27–87 years who had received mammography at the Princess Grace Hospital, London, were assessed for density. The association between having any dense tissue and area of residence was assessed using both ordered and standard logistic regression, giving odds ratio estimates of relative risk of dense tissue adjusting for age.Results: Adjusting for age, London residents had significantly higher levels of density (OR = 1.32, 95% CI 1.04–1.70, p = 0.02). The major difference occurred in the age group 45–54 years and was most strongly manifested as a higher rate in London for density of 25% or more (BIRADS categories 2–4) as compared to almost entirely fatty (BIRADS 1) (OR = 2.22, 95% CI 1.05–4.68, p = 0.035).Conclusion: The higher density is likely to be due to a different prevalence of risk factors in the London population. This study cannot ascertain the reason for the higher density in this urban population, but the result is a cause for concern given that screening uptake is lower in London. Increased attention to screening in urban areas and attention to screening quality for dense breast tissue might be prudent.

Objective: Objective: To evaluate the accuracy of radiology team recommendations compared with histopathology outcomes using Breast Imaging Reporting and Data System (BI-RADS) categories in breast ultrasound (US). Study Design: Retrospective cross-sectional study. Place and Duration: Radiology Department, Mian Meer Hospital, Lahore, in collaboration with Mayo Hospital, Lahore, Pathology Department, from 02 January 2021 to 03 June 2022. Methodology: Female patients referred from the Well Women Clinic (WWC), which provides breast health services to the local community, and who underwent breast ultrasound performed by consultant radiologists, were included. Patients had USS scans by non-consultants. Male and female patients referred from sources other than the Well Women Clinic were excluded. BI-RADS categories 1–5 were assigned and compared with American College of Radiology (ACR) BI-RADS guidelines and with histopathology reports where available. Category 0 was excluded due to incomplete data. SPSS v29 was used for statistical analysis. Results: Of 48,432 ultrasound referrals, 539 were from WWC; 361 met the inclusion criteria. Mean age was 36.23 ± 13.18 years. BI-RADS distribution: Category 1 = 56.5%, Categories 2–3 = 19.1%, Categories 4–5 = 24.4%. Agreement with ACR BI-RADS guideline recommendations was significant (χ²=475.3, p&lt;0.0001), with accuracies of 100% for discharge, 97% for follow-up, and 97% for referral. Histopathology confirmed malignancy in 41 patients (11.4%). Sensitivity and specificity for malignancy detection were 76% and 95% respectively, with overall diagnostic accuracy of 93%. Conclusion: BI-RADS-based breast ultrasound by consultant radiologists demonstrated high accuracy against histopathology and guideline recommendations and supporting its role as a reliable first-line diagnostic tool. Keywords: Breast Imaging Reporting and Data System, Breast ultrasound, Histopathology correlation, Benign breast disease, Breast cancer diagnosis. How to Cite This: Yousaf A, Zaidi NS, Rauf F, Ahmed R, Malik MR, Abbas S. Accuracy of Recommendations and Histopathology Results with Breast Imaging Reporting and Data System (BI-RADS) – A Retrospective Study. Isra Med J. 2025; 17(1):23-27. DOI: https://doi.org/10.55282/imj.oa1425

High mammographic density (MD) is a strong risk factor for breast cancer; however, data on the association between MD and breast cancer subtypes has been inconsistent and most studies have been conducted using data from Western women. Studies have shown that Asian women, who in general have denser breasts compared with Western women, were more likely to develop luminal B and HER2-enriched tumors. We previously reported that higher MD, assessed using Breast Imaging Reporting and Data System (BI-RADS) categories, was associated with HER2-enriched tumors among Chinese breast cancer patients, which may at least partially explain the higher incidence of HER2-enriched tumors observed in Asian women. Since BI-RADS MD measurement is categorical and known to be subjective, in this study, we aimed to re-evaluate the MD-subtype association using an automated quantitative density metric. The analysis included 1,549 women with invasive breast cancer who were diagnosed and treated in a tertiary hospital in Beijing, China. Subtypes were defined as Luminal A (n=606): ER+ and PR+, HER2-, and low Ki-67 (&amp;lt;30% positive staining)/grade (grades 1 or 2); luminal B/HER2- (n=446): ER+ and/or PR+, HER2-, and high Ki-67 (≥30%)/grade (grade 3); luminal B/HER2+ (n=182): ER+ and/or PR+, HER2+ (regardless of Ki-67/grade); HER2-enriched (n=121): ER-, PR-, and HER2+; Triple-negative (TN, n=194): ER-, PR-, and HER2-. Quantitative measures of MD for the contralateral unaffected breast, including total breast volume, fibroglandular volume, and Volpara Density Grade (VDG, based on predetermined cut-points from volumetric percent density that shows moderate agreement with visual BI-RADS categories), were obtained using VolparaDensity software. Polychotomous logistic regression was used to assess the association between MD (per 1-unit standard deviation increase) and breast cancer subtypes with the adjustment of age, menopausal status, parity, and body mass index (BMI). Compared with luminal A cases, cases with luminal B/HER2- (Odds ratio [OR], 1.15; 95% confidence interval [CI], 1.01-1.32; p=0.04), luminal B/HER2+ (OR, 1.22; 95% CI, 1.02-1.44; p=0.03), and HER2-enriched (OR, 1.32; 95% CI, 1.08-1.60; p=0.01) tumors had significantly higher fibroglandular dense volume, and these associations were stronger in women with normal BMI. In contrast, cases with TN tumors had lower non-dense volume compared with luminal A cases (OR, 0.83; 95% CI, 0.69-1.00; p=0.05). Percent density did not vary significantly by subtypes. When comparing dense versus non-dense breasts using VDG, we found that dense breasts were significantly associated with HER2-enriched tumors (OR, 2.49; 95% CI, 1.40-4.42; p=0.002). Our data suggest that increasing dense volume was associated with luminal B and HER2-enriched subtypes among Chinese breast cancer cases. These results may provide insight into the distinct distribution of breast cancer subtypes in Asian populations. Citation Format: Xiaohong R Yang, Yuan Tian, Jennifer Guida, Hela Koka, Ariane Chan, Changyuan Guo, Erni Li, Eric Tang, Hyuna Sung, Ning Lu, Nan Hu, Gretchen Gierach, Jing Li. Association between quantitative mammographic density and breast cancer subtypes among Chinese breast cancer patients [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-08-20.

Objective: To develop a nomogram model for the differential diagnosis of benign and malignant breast BI-RADS (Breast Imaging Reporting and Data System) category 4 nodules based on serum tumor specific protein 70 (SP70) and conventional laboratory indicators and validate its predictive efficacy. Methods: A case-control study design was used to retrospectively analyze the data of 429 female patients diagnosed with BI-RADS category 4 breast nodules by breast color doppler flow imaging at the First Affiliated Hospital of Nanjing Medical University from January 2021 to April 2022 with an age range of 16 to 91 years and a median age of 50 years, and the patients were divided into a training cohort (314 patients) and a validation cohort (115 patients) according to the inclusion time successively. Using postoperative pathological findings as the"gold standard", univariate and multivariate logistic regression analyses were used to identify the predictor variables used for the model. The nomogram, receiver operating characteristic (ROC) curves and calibration curves were drawn for the prediction model, and the discrimination and calibration of the model were evaluated using the consistency index (C-index) and calibration plots. Results: The postoperative pathological results showed that 286 (66.7%) were malignant nodules and 143 (33.3%) were benign nodules of 429 breast BI-RADS category 4 nodules. The serum SP70 (OR=1.227,95%CI: 1.033-1.458,P=0.020), NLR (OR=1.545,95%CI: 1.047-2.280,P=0.028), LDL-C (OR=2.215, 95%CI: 1.354-3.622, P=0.002), GLU (OR=2.050,95%CI:1.222-3.438,P=0.007), PT (OR=1.383,95%CI: 1.046-1.828,P=0.023), nodule diameter (OR=1.042, 95%CI: 1.008-1.076, P=0.015) and age (OR=1.062,95%CI: 1.011-1.116,P=0.016) were independent risk factors which could be used to distinguish benign and malignant breast BI-RADS category 4 nodules (P<0.05). The nomogram was plotted by the above seven independent variables, and the concordance index (C-index) for the training cohort and validation cohort were 0.842 (95%CI:0.786-0.898) and 0.787 (95%CI:0.687-0.886), respectively. The sensitivity and specificity of using this model to identify benign and malignant breast BI-RADS category 4 nodules in the training and validation cohort were 83.5%, 72.5% and 79.2%, 73.6%, respectively. The calibration curves showed good agreement between the predicted and actual values in the nomogram. Conclusions: This study combined serum SP70, conventional laboratory indicators and breast color doppler flow imaging to develop a nomogram model for the differential diagnosis of benign and malignant breast BI-RADS category 4 nodules. The model may have good predictive efficacy and may provide a basis for clinical treatment options, which is beneficial for guiding breast cancer screening and prevention.

Stereotactic vacuum-assisted breast biopsies in 500 women with microcalcifications: radiological and pathological correlations

Background: Breast cancer is a significant health issue affecting women worldwide, including Malaysia, where the incidence of breast cancer is increasing. Early detection of breast cancer plays a crucial role in improving treatment outcomes and reducing mortality rates. Aims and Objectives: This study examines the impact of adjunctive breast ultrasound (BUS) on screening for Malaysian women, offering evidence to enhance future protocols and decisions in the early detection. Materials and Methods: This study involved 5807 women aged 40 and above undergoing breast screening at Beacon Hospital from January to December 2022. Data collected included mammography and BUS results, categorized by breast imaging reporting and data system (BI-RADS) and diagnostic criteria. Suspicious cases on both modalities had histopathology examination (HPE) for confirmation. A subset of 1690 cases underwent both mammography and BUS screening and was compared to those who only had mammography using BI-RADS scores. Results: This study revealed that the majority of mammograms, 94.69% (5499 cases), were categorized as BI-RADS category 0, 1, 2, and 3, while 5.3% (307 cases) were classified as BI-RADS 4 and 5. Among the cases that underwent ultrasound (1688 cases), 23.43% of them were categorized as BI-RADS 4–5. Using ultrasound in addition to mammography increased the number of cases (BI-RADS 4–5) by 33.48%, resulting in a total of 460 cases identified compared to 307 cases, detected by mammography alone. Significantly, 39% (74 cases) of the 190 cases with confirmed malignancies (BI-RADS 6) by HPE were initially classified as negative mammography (BI-RADS 0–3) but were later identified as BI-RADS 5 on ultrasound. Conclusion: Ultrasound as a supplementary screening method holds the potential for boosting breast cancer detection sensitivity and reducing false negatives, thereby aiding in early detection and better treatment outcomes.

To develop a dual-modal neural network model to characterize ultrasound (US) images of breast masses. A combined US B-mode and color Doppler neural network model was developed to classify US images of the breast. Three datasets with breast masses were originally detected and interpreted by 20 experienced radiologists according to Breast Imaging-Reporting and Data System (BI-RADS) lexicon ((1) training set, 103212 masses from 45,433 + 12,519 patients. (2) held-out validation set, 2748 masses from 1197 + 395 patients. (3) test set, 605 masses from 337 + 78 patients). The neural network was first trained on training set. Then, the trained model was tested on a held-out validation set to evaluate agreement on BI-RADS category between the model and the radiologists. In addition, the model and a reader study of 10 radiologists were applied to the test set with biopsy-proven results. To evaluate the performance of the model in benign or malignant classifications, the receiver operating characteristic curve, sensitivities, and specificities were compared. The trained dual-modal model showed favorable agreement with the assessment performed by the radiologists (κ= 0.73; 95% confidence interval, 0.71-0.75) in classifying breast masses into four BI-RADS categories in the validation set. For the binary categorization of benign or malignant breast masses in the test set, the dual-modal model achieved the area under the ROC curve (AUC) of 0.982, while the readers scored an AUC of 0.948 in terms of the ROC convex hull. The dual-modal model can be used to assess breast masses at a level comparable to that of an experienced radiologist. • A neural network model based on ultrasonic imaging can classify breast masses into different Breast Imaging-Reporting and Data System categories according to the probability of malignancy. • A combined ultrasonic B-mode and color Doppler neural network model achieved a high level of agreement with the readings of an experienced radiologist and has the potential to automate the routine characterization of breast masses.

Women with elevated mammographic density have an increased risk of developing breast cancer. However, among women diagnosed with breast cancer, it is unclear whether higher density portends reduced survival, independent of other factors. We evaluated relationships between mammographic density and risk of death from breast cancer and all causes within the US Breast Cancer Surveillance Consortium. We studied 9232 women diagnosed with primary invasive breast carcinoma during 1996-2005, with a mean follow-up of 6.6 years. Mammographic density was assessed using the Breast Imaging Reporting and Data System (BI-RADS) density classification. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated by Cox proportional hazards regression; women with scattered fibroglandular densities (BI-RADS 2) were the referent group. All statistical tests were two-sided. A total of 1795 women died, of whom 889 died of breast cancer. In multivariable analyses (adjusted for site, age at and year of diagnosis, American Joint Committee on Cancer stage, body mass index, mode of detection, treatment, and income), high density (BI-RADS 4) was not related to risk of death from breast cancer (HR = 0.92, 95% CI = 0.71 to 1.19) or death from all causes (HR = 0.83, 95% CI = 0.68 to 1.02). Analyses stratified by stage and other prognostic factors yielded similar results, except for an increased risk of breast cancer death among women with low density (BI-RADS 1) who were either obese (HR = 2.02, 95% CI = 1.37 to 2.97) or had tumors of at least 2.0 cm (HR = 1.55, 95% CI = 1.14 to 2.09). High mammographic breast density was not associated with risk of death from breast cancer or death from any cause after accounting for other patient and tumor characteristics. Thus, risk factors for the development of breast cancer may not necessarily be the same as factors influencing the risk of death after breast cancer has developed.

Background: Radiologically dense area (DA) on a mammogram represents the stromal and epithelial tissue in the breast, while non-dense area (NDA) represents adipose tissue. Quantitative measurements of DA, NDA as well as the proportion of dense area in the breast (percent density, PD) adjusted for age and body mass index (BMI) have all been associated with breast cancer (BC) risk. Previous genome-wide association studies (GWAS) have identified genetic loci that show association with both mammographic density (MD) phenotypes and BC, but little is known about the genome-wide and local genetic correlation between MD phenotypes and BC. Methods: We estimated the genome-wide and local genetic correlation between MD phenotypes adjusted for age and BMI and BC, leveraging GWAS summary statistics for MD phenotypes based on 12,866 women (MODE consortium) and for BC based on 122,977 cases and 105,974 controls (BCAC consortium). We also conducted partitioned heritability analysis to identify specific functional annotations that contribute to MD heritability. Results: We observed a negative genetic correlation between DA and NDA (-0.62, p=0.002). Both DA and PD showed positive genetic correlations with overall BC (DA: rg=0.24, p=0.004; PD: rg=0.27, p=0.0001), estrogen receptor-positive (ER+) BC (DA: rg=0.23, p=0.005; PD: rg=0.23, p=0.001), and ER- BC (DA: rg=0.26, p=0.01; PD: rg=0.32, p=0.0006). In contrast, NDA showed a nonsignificant negative genetic correlation with BC (rg=-0.11, p=0.21). Local genetic correlation analyses between MD phenotypes and BC reflected known findings including positive correlation at 10q21 region between both PD and DA and BC (both p&amp;lt;5x10-7), negative correlation at 22q13.1 between DA and BC (p=4.6x10-6), and positive correlation at 8p11.23 between NDA and BC (p=2.8x10-6). Two known BC regions that have not previously been associated with MD showed suggestive genetic correlations between BC and DA: 1q21.2 (negative genetic correlation, p=1.3x10-4) and 12q21.31 (positive genetic correlation, p=6.7x10-4). In exploratory analyses estimating the genome-wide genetic correlation between MD and hormone-related phenotypes in UK Biobank (n=337,199), we observed nominally significant positive genetic correlations between PD and benign mammary dysplasia (rg=0.45, p=0.01), NDA and endometriosis (rg=0.45, p=0.02), and negative genetic correlation between NDA and menstrual cycle length (rg=-0.30, p=0.05). We identified multiple functional annotations that contribute to MD heritability including histone mark H3K27ac that marks active enhancers (~2,5-fold enrichment in heritability for both DA and NDA) and DNAsel Hypersensitive Sites (DHS, 9.1-fold enrichment in heritability for DA). Conclusion: Our results provide evidence of shared genetic basis between mammographic density and breast cancer. Citation Format: Hongjie Chen, Julie Douglas, William Gordon, Deborah Thompson, Fergus Couch, Peter Fasching, Per Hall, John Hopper, Gertraud Maskarenic, Marina Pollan, Jennifer Stone, Rulla Tamimi, Sara Lindstroem. Evidence for a shared genetic basis between breast cancer and mammographic density phenotypes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 224.