Reproducibility of BI-RADS breast density measures among community radiologists: a prospective cohort study

Current status and prospects of breast cancer imaging-based diagnosis using artificial intelligence

Breast imaging has several modalities, each unique in terms of its imaging position, evaluation index, and imaging method. Breast diagnosis is made by combining a large number of past imaging features with the clinical course and histological findings. Artificial intelligence (AI), which extracts the features from image data and evaluates them based on comprehensive analysis, has been making rapid progress in this regard. Many previous studies have demonstrated the usefulness and development potential of AI, such as machine learning and deep learning, in breast imaging. However, despite studies showing the good performance of AI models, their overall utilization remains low, since a large amount of diverse imaging data is required, and prospective verification is necessary to prove its high reproducibility and robustness. Sharing information and collaborating with multiple institutions to collect and verify images of different conditions and backgrounds are vital. If image diagnosis using AI can indeed ensure a more detailed diagnosis, such as breast cancer subtypes or prognosis, it can help develop personalized medicine, which is urgently required. The positive results of AI research, using such image information, can make each modality more valuable than ever. The current review summarized the results of previous studies using AI in each evaluation field and discussed the related future prospects.

Generalisable deep learning method for mammographic density prediction across imaging techniques and self-reported race

BACKGROUND

Breast density is an important risk factor for breast cancer complemented by a higher risk of cancers being missed during screening of dense breasts due to reduced sensitivity of mammography. Automated, deep learning-based prediction of breast density could provide subject-specific risk assessment and flag difficult cases during screening. However, there is a lack of evidence for generalisability across imaging techniques and, importantly, across race.

METHODS

This study used a large, racially diverse dataset with 69,697 mammographic studies comprising 451,642 individual images from 23,057 female participants. A deep learning model was developed for four-class BI-RADS density prediction. A comprehensive performance evaluation assessed the generalisability across two imaging techniques, full-field digital mammography (FFDM) and two-dimensional synthetic (2DS) mammography. A detailed subgroup performance and bias analysis assessed the generalisability across participants' race.

RESULTS

Here we show that a model trained on FFDM-only achieves a 4-class BI-RADS classification accuracy of 80.5% (79.7-81.4) on FFDM and 79.4% (78.5-80.2) on unseen 2DS data. When trained on both FFDM and 2DS images, the performance increases to 82.3% (81.4-83.0) and 82.3% (81.3-83.1). Racial subgroup analysis shows unbiased performance across Black, White, and Asian participants, despite a separate analysis confirming that race can be predicted from the images with a high accuracy of 86.7% (86.0-87.4).

CONCLUSIONS

Deep learning-based breast density prediction generalises across imaging techniques and race. No substantial disparities are found for any subgroup, including races that were never seen during model development, suggesting that density predictions are unbiased.

Disparities in Breast Cancer Care-How Factors Related to Prevention, Diagnosis, and Treatment Drive Inequity

Breast cancer survival has increased significantly over the last few decades due to more effective strategies for prevention and risk modification, advancements in imaging detection, screening, and multimodal treatment algorithms. However, many have observed disparities in benefits derived from such improvements across populations and demographic groups. This review summarizes published works that contextualize modern disparities in breast cancer prevention, diagnosis, and treatment and presents potential strategies for reducing disparities. We conducted searches for studies that directly investigated and/or reported disparities in breast cancer prevention, detection, or treatment. Demographic factors, social determinants of health, and inequitable healthcare delivery may impede the ability of individuals and communities to employ risk-mitigating behaviors and prevention strategies. The disparate access to quality screening and timely diagnosis experienced by various groups poses significant hurdles to optimal care and survival. Finally, barriers to access and inequitable healthcare delivery patterns reinforce inequitable application of standards of care. Cumulatively, these disparities underlie notable differences in the incidence, severity, and survival of breast cancers. Efforts toward mitigation will require collaborative approaches and partnerships between communities, governments, and healthcare organizations, which must be considered equal stakeholders in the fight for equity in breast cancer care and outcomes.

Development and evaluation of a method to assess breast cancer risk using a longitudinal history of mammographic density: a cohort study

BACKGROUND

Women with dense breasts have an increased risk of breast cancer. However, breast density is measured with variability, which may reduce the reliability and accuracy of its association with breast cancer risk. This is particularly relevant when visually assessing breast density due to variation in inter- and intra-reader assessments. To address this issue, we developed a longitudinal breast density measure which uses an individual woman's entire history of mammographic density, and we evaluated its association with breast cancer risk as well as its predictive ability.

METHODS

In total, 132,439 women, aged 40-73 yr, who were enrolled in Kaiser Permanente Washington and had multiple screening mammograms taken between 1996 and 2013 were followed up for invasive breast cancer through 2014. Breast Imaging Reporting and Data System (BI-RADS) density was assessed at each screen. Continuous and derived categorical longitudinal density measures were developed using a linear mixed model that allowed for longitudinal density to be updated at each screen. Predictive ability was assessed using (1) age and body mass index-adjusted hazard ratios (HR) for breast density (time-varying covariate), (2) likelihood-ratio statistics (ΔLR-χ2) and (3) concordance indices.

RESULTS

In total, 2704 invasive breast cancers were diagnosed during follow-up (median = 5.2 yr; median mammograms per woman = 3). When compared with an age- and body mass index-only model, the gain in statistical information provided by the continuous longitudinal density measure was 23% greater than that provided by BI-RADS density (follow-up after baseline mammogram: ΔLR-χ2 = 379.6 (degrees of freedom (df) = 2) vs. 307.7 (df = 3)), which increased to 35% (ΔLR-χ2 = 251.2 vs. 186.7) for follow-up after three mammograms (n = 76,313, 2169 cancers). There was a sixfold difference in observed risk between densest and fattiest eight-category longitudinal density (HR = 6.3, 95% CI 4.7-8.7), versus a fourfold difference with BI-RADS density (HR = 4.3, 95% CI 3.4-5.5). Discriminatory accuracy was marginally greater for longitudinal versus BI-RADS density (c-index = 0.64 vs. 0.63, mean difference = 0.008, 95% CI 0.003-0.012).

CONCLUSIONS

Estimating mammographic density using a woman's history of breast density is likely to be more reliable than using the most recent observation only, which may lead to more reliable and accurate estimates of individual breast cancer risk. Longitudinal breast density has the potential to improve personal breast cancer risk estimation in women attending mammography screening.

Diagnostic accuracy of automated ACR BI-RADS breast density classification using deep convolutional neural networks

OBJECTIVES

High breast density is a well-known risk factor for breast cancer. This study aimed to develop and adapt two (MLO, CC) deep convolutional neural networks (DCNN) for automatic breast density classification on synthetic 2D tomosynthesis reconstructions.

METHODS

In total, 4605 synthetic 2D images (1665 patients, age: 57 ± 37 years) were labeled according to the ACR (American College of Radiology) density (A-D). Two DCNNs with 11 convolutional layers and 3 fully connected layers each, were trained with 70% of the data, whereas 20% was used for validation. The remaining 10% were used as a separate test dataset with 460 images (380 patients). All mammograms in the test dataset were read blinded by two radiologists (reader 1 with two and reader 2 with 11 years of dedicated mammographic experience in breast imaging), and the consensus was formed as the reference standard. The inter- and intra-reader reliabilities were assessed by calculating Cohen's kappa coefficients, and diagnostic accuracy measures of automated classification were evaluated.

RESULTS

The two models for MLO and CC projections had a mean sensitivity of 80.4% (95%-CI 72.2-86.9), a specificity of 89.3% (95%-CI 85.4-92.3), and an accuracy of 89.6% (95%-CI 88.1-90.9) in the differentiation between ACR A/B and ACR C/D. DCNN versus human and inter-reader agreement were both "substantial" (Cohen's kappa: 0.61 versus 0.63).

CONCLUSION

The DCNN allows accurate, standardized, and observer-independent classification of breast density based on the ACR BI-RADS system.

KEY POINTS

• A DCNN performs on par with human experts in breast density assessment for synthetic 2D tomosynthesis reconstructions. • The proposed technique may be useful for accurate, standardized, and observer-independent breast density evaluation of tomosynthesis.

Impact of multi-source data augmentation on performance of convolutional neural networks for abnormality classification in mammography

Introduction

To date, most mammography-related AI models have been trained using either film or digital mammogram datasets with little overlap. We investigated whether or not combining film and digital mammography during training will help or hinder modern models designed for use on digital mammograms.

Methods

To this end, a total of six binary classifiers were trained for comparison. The first three classifiers were trained using images only from Emory Breast Imaging Dataset (EMBED) using ResNet50, ResNet101, and ResNet152 architectures. The next three classifiers were trained using images from EMBED, Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM), and Digital Database for Screening Mammography (DDSM) datasets. All six models were tested only on digital mammograms from EMBED.

Results

The results showed that performance degradation to the customized ResNet models was statistically significant overall when EMBED dataset was augmented with CBIS-DDSM/DDSM. While the performance degradation was observed in all racial subgroups, some races are subject to more severe performance drop as compared to other races.

Discussion

The degradation may potentially be due to ( 1) a mismatch in features between film-based and digital mammograms ( 2) a mismatch in pathologic and radiological information. In conclusion, use of both film and digital mammography during training may hinder modern models designed for breast cancer screening. Caution is required when combining film-based and digital mammograms or when utilizing pathologic and radiological information simultaneously.

Collaborative Privacy-preserving Approaches for Distributed Deep Learning Using Multi-Institutional Data

Deep learning (DL) algorithms have shown remarkable potential in automating various tasks in medical imaging and radiologic reporting. However, models trained on low quantities of data or only using data from a single institution often are not generalizable to other institutions, which may have different patient demographics or data acquisition characteristics. Therefore, training DL algorithms using data from multiple institutions is crucial to improving the robustness and generalizability of clinically useful DL models. In the context of medical data, simply pooling data from each institution to a central location to train a model poses several issues such as increased risk to patient privacy, increased costs for data storage and transfer, and regulatory challenges. These challenges of centrally hosting data have motivated the development of distributed machine learning techniques and frameworks for collaborative learning that facilitate the training of DL models without the need to explicitly share private medical data. The authors describe several popular methods for collaborative training and review the main considerations for deploying these models. They also highlight publicly available software frameworks for federated learning and showcase several real-world examples of collaborative learning. The authors conclude by discussing some key challenges and future research directions for distributed DL. They aim to introduce clinicians to the benefits, limitations, and risks of using distributed DL for the development of medical artificial intelligence algorithms. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material.

Association of Longitudinal Mammographic Breast Density Changes with Subsequent Breast Cancer Risk

Background Although Breast Imaging Reporting and Data System (BI-RADS) density classification has been used to assess future breast cancer risk, its reliability and validity are still debated in literature. Purpose To determine the association between overall longitudinal changes in mammographic breast density and breast cancer risk stratified by menopausal status. Materials and Methods In a retrospective cohort study using the Korean National Health Insurance Service database, women aged at least 40 years without a history of cancer who underwent three consecutive biennial mammographic screenings in 2009-2014 were followed up through December 2020. Participants were divided according to baseline breast density: fatty (BI-RADS categories a, b) versus dense (BI-RADS categories c, d) and then into subgroups on the basis of changes from the first to second and from second to third screenings. Women without change in breast density were used as the reference group. Main outcomes were incident breast cancer events, both invasive breast cancer and ductal carcinoma in situ. Cox proportion hazard regression was used to calculate the hazard ratio (HR) with adjustment for other covariables. Results Among 2 253 963 women (mean age, 59 years ± 9) there were 22 439 detected breast cancers. Premenopausal women with fatty breasts at the first screening had a higher risk of breast cancer as density increased in the second and third screenings (fatty-to-dense HR, 1.45 [95% CI: 1.27, 1.65]; dense-to-fatty HR, 1.53 [95% CI: 1.34, 1.74]; dense-to-dense HR, 1.93 [95% CI: 1.75, 2.13]). In premenopausal women with dense breasts at baseline, those in whom density continuously decreased had a 0.62-fold lower risk (95% CI: 0.56, 0.69). Similar results were observed in postmenopausal women, remaining significant after adjustment for baseline breast density or changes in body mass index (fatty-to-dense HR, 1.50 [95% CI: 1.39, 1.62]; dense-to-fatty HR, 1.42 [95% CI: 1.31, 1.53]; dense-to-dense HR, 1.62 [95% CI: 1.51, 1.75]). Conclusion In both premenopausal and postmenopausal women undergoing three consecutive biennial mammographic screenings, a consecutive increase in breast density augmented the future breast cancer risk whereas a continuous decrease was associated with a lower risk. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Kataoka et al in this issue.

Automatic mammographic breast density classification in Chinese women: clinical validation of a deep learning model

BACKGROUND

High breast density is a strong risk factor for breast cancer. As such, high consistency and accuracy in breast density assessment is necessary.

PURPOSE

To validate our proposed deep learning (DL) model and explore its impact on radiologists on density assessments.

MATERIAL AND METHODS

A total of 3732 mammographic cases were collected as a validated set: 1686 cases before the implementation of the DL model and 2046 cases after the DL model. Five radiologists were divided into two groups (junior and senior groups) to assess all mammograms using either two- or four-category evaluation. Linear-weighted kappa (K) and intraclass correlation coefficient (ICC) statistics were used to analyze the consistency between radiologists before and after implementation of the DL model.

RESULTS

The accuracy and clinical acceptance of the DL model for the junior group were 96.3% and 96.8% for two-category evaluation, and 85.6% and 89.6% for four-category evaluation, respectively. For the senior group, the accuracy and clinical acceptance were 95.5% and 98.0% for two-category evaluation, and 84.3% and 95.3% for four-category evaluation, respectively. The consistency within the junior group, the senior group, and among all radiologists improved with the help of the DL model. For two-category, their K and ICC values improved to 0.81, 0.81, and 0.80 from 0.73, 0.75, and 0.76. And for four-category, their K and ICC values improved to 0.81, 0.82, and 0.82 from 0.73, 0.79, and 0.78, respectively.

CONCLUSION

The DL model showed high accuracy and clinical acceptance in breast density categories. It is helpful to improve radiologists' consistency.

Breast Density: Current Knowledge, Assessment Methods, and Clinical Implications

Breast density is an accepted independent risk factor for the future development of breast cancer, and greater breast density has the potential to mask malignancies on mammography, thus lowering the sensitivity of screening mammography. The risk associated with dense breast tissue has been shown to be modifiable with changes in breast density. Numerous studies have sought to identify factors that influence breast density, including age, genetic, racial/ethnic, prepubertal, adolescent, lifestyle, environmental, hormonal, and reproductive history factors. Qualitative, semiquantitative, and quantitative methods of breast density assessment have been developed, but to date there is no consensus assessment method or reference standard for breast density. Breast density has been incorporated into breast cancer risk models, and there is growing consciousness of the clinical implications of dense breast tissue in both the medical community and public arena. Efforts to improve breast cancer screening sensitivity for women with dense breasts have led to increased attention to supplemental screening methods in recent years, prompting the American College of Radiology to publish Appropriateness Criteria for supplemental screening based on breast density.

A Model to Predict Upstaging to Invasive Carcinoma in Patients Preoperatively Diagnosed with Low-Grade Ductal Carcinoma In Situ of the Breast

Simple Summary

Surgical management is currently the main standard of care procedure used in order to treat ductal carcinoma in situ (DCIS) of the breast. Nevertheless, the survival benefit of surgical resection in patients with such lesions appears to be low, especially for low-grade DCIS. Low-grade DCIS typically exhibit a slow growth pattern and, in many cases, never fully develop into a clinically significant disease: discerning harmless lesions from potentially invasive ones could lead to avoid overtreatment in many patients. Nonetheless, up to 26% of patients with biopsy-proven DCIS can reveal a synchronous invasive carcinoma in surgical specimens. Here, we aimed to create a model of radiological and pathological criteria able to reduce the underestimation of vacuum assisted breast biopsy in DCIS, identifying patients at very low risk (e.g., <2%) of diagnostic underestimation.

Abstract

Background: We aimed to create a model of radiological and pathological criteria able to predict the upgrade rate of low-grade ductal carcinoma in situ (DCIS) to invasive carcinoma, in patients undergoing vacuum-assisted breast biopsy (VABB) and subsequent surgical excision. Methods: A total of 3100 VABBs were retrospectively reviewed, among which we reported 295 low-grade DCIS who subsequently underwent surgery. The association between patients’ features and the upgrade rate to invasive breast cancer (IBC) was evaluated by univariate and multivariate analysis. Finally, we developed a nomogram for predicting the upstage at surgery, according to the multivariate logistic regression model. Results: The overall upgrade rate to invasive carcinoma was 10.8%. At univariate analysis, the risk of upgrade was significantly lower in patients with greater age (p = 0.018), without post-biopsy residual lesion (p < 0.001), with a smaller post-biopsy residual lesion size (p < 0.001), and in the presence of low-grade DCIS only in specimens with microcalcifications (p = 0.002). According to the final multivariable model, the predicted probability of upstage at surgery was lower than 2% in 58 patients; among these 58 patients, only one (1.7%) upstage was observed, showing a good calibration of the model. Conclusions: An easy-to-use nomogram for predicting the upstage at surgery based on radiological and pathological criteria is able to identify patients with low-grade carcinoma in situ with low risk of upstaging to infiltrating carcinomas.

Mammographic Density Assessment by Artificial Intelligence-Based Computer-Assisted Diagnosis: A Comparison with Automated Volumetric Assessment

We evaluated and compared the mammographic density assessment of an artificial intelligence-based computer-assisted diagnosis (AI-CAD) program using inter-rater agreements between radiologists and an automated density assessment program. Between March and May 2020, 488 consecutive mammograms of 488 patients (56.2 ± 10.9 years) were collected from a single institution. We assigned four classes of mammographic density based on BI-RADS (Breast Imaging Reporting and Data System) using commercial AI-CAD (Lunit INSIGHT MMG), and compared inter-rater agreements between radiologists, AI-CAD, and another commercial automated density assessment program (Volpara®). The inter-rater agreement between AI-CAD and the reader consensus was 0.52 with a matched rate of 68.2% (333/488). The inter-rater agreement between Volpara® and the reader consensus was similar to AI-CAD at 0.50 with a matched rate of 62.7% (306/488). The inter-rater agreement between AI-CAD and Volpara® was 0.54 with a matched rate of 61.5% (300/488). In conclusion, density assessments by AI-CAD showed fair agreement with those of radiologists, similar to the agreement between the commercial automated density assessment program and radiologists.

Reporting and Perceptions of Breast Arterial Calcification on Mammography: A Survey of ACR Radiologists

RATIONALE AND OBJECTIVES

The ACR Breast Commission conducted a member survey to evaluate current practices of reporting breast arterial calcification (BAC) on mammography and to determine perceptions about the value of BAC communication and follow-up recommendations among radiologists.

MATERIALS AND METHODS

In September 2020, an 18-item online survey was emailed to radiologist members of the American College of Radiology (ACR). Questions included radiologist demographics, current BAC reporting practices, follow-up recommendations, and perceptions about BAC. Five-point Likert scales were used and multivariate analysis was performed.

RESULTS

Of 598 completed survey responses, up to 87% (522/598) of ACR radiologist members include BAC in mammogram reports. However, only 41% (212/522) of respondents report BAC 'always' or 'most of the time'. Radiologist factors significantly associated with BAC reporting include years in practice and fellowship training with those in practice longer more likely to report BAC (OR 1.10, 95% CI, [1.01-1.20], p = 0.023) and those with fellowship training less likely to report BAC (OR 0.63, 95% CI, [0.42-0.94], p = 0.024). When BAC is reported, 69% (360/522) simply indicate the presence of BAC, 23% (121/522) provide a subjective grading of BAC burden, and 1% (6/522) calculate a BAC score. Among the radiologists reporting BAC, 58% (301/522) make no subsequent recommendations, while the remainder recommend primary care follow-up (39%; 204/522), cardiology evaluation (13%; 68/522), and/or coronary calcium scoring CT (11%; 59/522). Overall, there was agreement from 66% (392/598) of respondents that BAC is a cardiovascular risk factor. However, there was no consensus on whether patients and/or providers should be informed about BAC or whether reporting of BAC should become a standardized practice in breast imaging. Older and more experienced radiologists are more likely to agree that BAC is a cardiovascular risk factor (p = 0.022), providers should be informed about BAC (p = 0.002 and 0.006), BAC reporting should be a standardized practice (p = 0.004 and 0.001), and feel more comfortable informing patients about BAC (p = 0.001 and 0.003).

CONCLUSION

Radiologists' reporting practices and perceptions regarding BAC are not homogeneous. Although many radiologists report BAC to varying degrees, it is not routinely reported or recommended for follow-up in mammogram reports. Experienced radiologists are more likely to include and value BAC in their breast imaging practice.

Quantitative Mammographic Density Measurements and Molecular Subtypes in Chinese Women With Breast Cancer

Background

Studies investigating associations between mammographic density (MD) and breast cancer subtypes have generated mixed results. We previously showed that having extremely dense breasts was associated with the human epidermal growth factor receptor-2 (HER2)-enriched subtype in Chinese breast cancer patients.

Methods

In this study, we reevaluated the MD-subtype association in 1549 Chinese breast cancer patients, using VolparaDensity software to obtain quantitative MD measures. All statistical tests were 2-sided.

Results

Compared with women with luminal A tumors, women with luminal B/HER2- (odds ratio [OR] = 1.20, 95% confidence interval [CI] = 1.04 to 1.38; P = .01), luminal B/HER2+ (OR = 1.22, 95% CI = 1.03 to 1.46; P = .03), and HER2-enriched tumors (OR = 1.30, 95% CI = 1.06 to 1.59; P = .01) had higher fibroglandular dense volume. These associations were stronger in patients with smaller tumors (<2 cm). In contrast, the triple-negative subtype was associated with lower nondense volume (OR = 0.82, 95% CI = 0.68 to 0.99; P = .04), and the association was only seen among older women (age 50 years or older).

Conclusion

Although biological mechanisms remain to be investigated, the associations for the HER2-enriched and luminal B subtypes with increasing MD may partially explain the higher prevalence of luminal B and HER2+ breast cancers previously reported in Asian women.

Deep Learning-Based Artificial Intelligence for Mammography

During the past decade, researchers have investigated the use of computer-aided mammography interpretation. With the application of deep learning technology, artificial intelligence (AI)-based algorithms for mammography have shown promising results in the quantitative assessment of parenchymal density, detection and diagnosis of breast cancer, and prediction of breast cancer risk, enabling more precise patient management. AI-based algorithms may also enhance the efficiency of the interpretation workflow by reducing both the workload and interpretation time. However, more in-depth investigation is required to conclusively prove the effectiveness of AI-based algorithms. This review article discusses how AI algorithms can be applied to mammography interpretation as well as the current challenges in its implementation in real-world practice.

External Validation of a Deep Learning Model for Predicting Mammographic Breast Density in Routine Clinical Practice

RATIONALE AND OBJECTIVES

Federal legislation requires patient notification of dense mammographic breast tissue because increased density is a marker of breast cancer risk and can limit the sensitivity of mammography. As previously described, we clinically implemented our deep learning model at the academic breast imaging practice where the model was developed with high clinical acceptance. Our objective was to externally validate our deep learning model on radiologist breast density assessments in a community breast imaging practice.

MATERIALS AND METHODS

Our deep learning model was implemented at a dedicated breast imaging practice staffed by both academic and community breast imaging radiologists in October 2018. Deep learning model assessment of mammographic breast density was presented to the radiologist during routine clinical practice at the time of mammogram interpretation. We identified 2174 consecutive screening mammograms after implementation of the deep learning model. Radiologist agreement with the model's assessment was measured and compared across radiologist groups.

RESULTS

Both academic and community radiologists had high clinical acceptance of the deep learning model's density prediction, with 94.9% (academic) and 90.7% (community) acceptance for dense versus nondense categories (p < 0.001). The proportion of mammograms assessed as dense by all radiologists decreased from 47.0% before deep learning model implementation to 41.0% after deep learning model implementation (p < 0.001).

CONCLUSION

Our deep learning model had a high clinical acceptance rate among both academic and community radiologists and reduced the proportion of mammograms assessed as dense. This is an important step to validating our deep learning model prior to potential widespread implementation.

Gut microbiome, body weight, and mammographic breast density in healthy postmenopausal women

PURPOSE

We examined gut microbiome (GM) profiles in relation to mammographic breast density (BD) and body mass index (BMI) in healthy postmenopausal women.

METHODS

Eligible women were postmenopausal, had a BMI ≤ 35 kg/m², and had not recently taken oral/IV antibiotics. All women provided a fecal sample and information on breast cancer risk factors. Mammographic BD was classified with the American College of Radiology's BI-RADS BD classification system. Bacterial DNA was isolated from fecal samples and the V1-V2 hypervariable regions of 16S rRNA were sequenced on the Illumina MiSeq platform. We examined associations of GM with indices of within-sample (alpha) diversity and the ratio of the two main phyla (Firmicutes and Bacteroidetes; F/B ratio) with BD and BMI.

RESULTS

Among 69 women with BD data, 39 had low BD (BI-RADS I/II) and 30 had high BD (BI-RADS III/IV). BMI was inversely associated with BD (mean BMI = 23.8 and 28.0 in women with high and low BD, respectively, p = 1.07 × 10^-5). Similar levels of GM diversity were found across weight groups according to Shannon (p = 0.83); Inverse Simpson (p = 0.97); and Chao1 (p = 0.31) indices. F/B ratio and microbiota diversity were suggestively greater in women with high vs. low BD (p = 0.35, 0.14, 0.15, and 0.17 for F/B ratio, Shannon, Inverse Simpson and Chao1, respectively).

CONCLUSION

Suggestive differences observed in women with high and low BD with respect to GM alpha diversity and prevalence of specific GM taxa need to be confirmed in larger studies.

A Multisite Study of a Breast Density Deep Learning Model for Full-Field Digital Mammography and Synthetic Mammography

PURPOSE

To develop a Breast Imaging Reporting and Data System (BI-RADS) breast density deep learning (DL) model in a multisite setting for synthetic two-dimensional mammographic (SM) images derived from digital breast tomosynthesis examinations by using full-field digital mammographic (FFDM) images and limited SM data.

MATERIALS AND METHODS

A DL model was trained to predict BI-RADS breast density by using FFDM images acquired from 2008 to 2017 (site 1: 57 492 patients, 187 627 examinations, 750 752 images) for this retrospective study. The FFDM model was evaluated by using SM datasets from two institutions (site 1: 3842 patients, 3866 examinations, 14 472 images, acquired from 2016 to 2017; site 2: 7557 patients, 16 283 examinations, 63 973 images, 2015 to 2019). Each of the three datasets were then split into training, validation, and test. Adaptation methods were investigated to improve performance on the SM datasets, and the effect of dataset size on each adaptation method was considered. Statistical significance was assessed by using CIs, which were estimated by bootstrapping.

RESULTS

Without adaptation, the model demonstrated substantial agreement with the original reporting radiologists for all three datasets (site 1 FFDM: linearly weighted Cohen κ [κw] = 0.75 [95% CI: 0.74, 0.76]; site 1 SM: κw = 0.71 [95% CI: 0.64, 0.78]; site 2 SM: κw = 0.72 [95% CI: 0.70, 0.75]). With adaptation, performance improved for site 2 (site 1: κw = 0.72 [95% CI: 0.66, 0.79], 0.71 vs 0.72, P = .80; site 2: κw = 0.79 [95% CI: 0.76, 0.81], 0.72 vs 0.79, P < .001) by using only 500 SM images from that site.

CONCLUSION

A BI-RADS breast density DL model demonstrated strong performance on FFDM and SM images from two institutions without training on SM images and improved by using few SM images.Supplemental material is available for this article.Published under a CC BY 4.0 license.

Multi-Institutional Assessment and Crowdsourcing Evaluation of Deep Learning for Automated Classification of Breast Density

OBJECTIVE

We developed deep learning algorithms to automatically assess BI-RADS breast density.

METHODS

Using a large multi-institution patient cohort of 108,230 digital screening mammograms from the Digital Mammographic Imaging Screening Trial, we investigated the effect of data, model, and training parameters on overall model performance and provided crowdsourcing evaluation from the attendees of the ACR 2019 Annual Meeting.

RESULTS

Our best-performing algorithm achieved good agreement with radiologists who were qualified interpreters of mammograms, with a four-class κ of 0.667. When training was performed with randomly sampled images from the data set versus sampling equal number of images from each density category, the model predictions were biased away from the low-prevalence categories such as extremely dense breasts. The net result was an increase in sensitivity and a decrease in specificity for predicting dense breasts for equal class compared with random sampling. We also found that the performance of the model degrades when we evaluate on digital mammography data formats that differ from the one that we trained on, emphasizing the importance of multi-institutional training sets. Lastly, we showed that crowdsourced annotations, including those from attendees who routinely read mammograms, had higher agreement with our algorithm than with the original interpreting radiologists.

CONCLUSION

We demonstrated the possible parameters that can influence the performance of the model and how crowdsourcing can be used for evaluation. This study was performed in tandem with the development of the ACR AI-LAB, a platform for democratizing artificial intelligence.

Fully Automatic Assessment of Background Parenchymal Enhancement on Breast MRI Using Machine-Learning Models

BACKGROUND

Automated measurement and classification models with objectivity and reproducibility are required for accurate evaluation of the breast cancer risk of fibroglandular tissue (FGT) and background parenchymal enhancement (BPE).

PURPOSE

To develop and evaluate a machine-learning algorithm for breast FGT segmentation and BPE classification.

STUDY TYPE

Retrospective.

POPULATION

A total of 794 patients with breast cancer, 594 patients assigned to the development set, and 200 patients to the test set.

FIELD STRENGTH/SEQUENCE

3T and 1.5T; T₂ -weighted, fat-saturated T₁ -weighted (T₁ W) with dynamic contrast enhancement (DCE).

ASSESSMENT

Manual segmentation was performed for the whole breast and FGT regions in the contralateral breast. The BPE region was determined by thresholding using the subtraction of the pre- and postcontrast T₁ W images and the segmented FGT mask. Two radiologists independently assessed the categories of FGT and BPE. A deep-learning-based algorithm was designed to segment and measure the volume of whole breast and FGT and classify the grade of BPE.

STATISTICAL TESTS

Dice similarity coefficients (DSC) and Spearman correlation analysis were used to compare the volumes from the manual and deep-learning-based segmentations. Kappa statistics were used for agreement analysis. Comparison of area under the receiver operating characteristic (ROC) curves (AUC) and F1 scores were calculated to evaluate the performance of BPE classification.

RESULTS

The mean (±SD) DSC for manual and deep-learning segmentations was 0.85 ± 0.11. The correlation coefficient for FGT volume from manual- and deep-learning-based segmentations was 0.93. Overall accuracy of manual segmentation and deep-learning segmentation in BPE classification task was 66% and 67%, respectively. For binary categorization of BPE grade (minimal/mild vs. moderate/marked), overall accuracy increased to 91.5% in manual segmentation and 90.5% in deep-learning segmentation; the AUC was 0.93 in both methods.

DATA CONCLUSION

This deep-learning-based algorithm can provide reliable segmentation and classification results for BPE.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.

Quantification of water and lipid density with dual-energy mammography: validation in postmortem breasts

OBJECTIVES

Breast cancer is the most common cancer in women and the second leading cause of cancer death. It is well known that breast density is an important risk factor for breast cancer and also can be used to personalize screening and for assessment of treatment response. Breast density has previously been correlated to volumetric water density. The purpose of this study is to validate the accuracy and precision of dual-energy mammography in measuring water density in postmortem breasts.

METHODS

Twenty pairs of postmortem breasts were imaged using dual-energy mammography with energy-sensitive photon-counting detectors. Chemical analysis was used as the reference standard to assess the accuracy of dual-energy mammography in measuring volumetric water and lipid density. Images from different views and contralateral breasts were used to assess estimate of precision for water and lipid volumetric density measurements.

RESULTS

The measured volumetric water and lipid density from dual-energy mammography and chemical analysis were in good agreement, where the standard errors of estimates (SEE) of both were calculated to be 2.1%. Volumetric water and lipid density measurements from different views were also in good agreement, with a SEE of 1.3% and 1.1%, respectively.

CONCLUSIONS

The results indicate that dual-energy mammography can be used to accurately measure volumetric water and lipid density in breast tissue. Accurate quantification of volumetric water density is expected to enhance its utility as a risk factor for breast cancer and for assessment of response to therapy.

KEY POINTS

• Dual-energy mammography can be used to accurately measure water and lipid volumetric density in breast tissue. • Improved quantification of volumetric water density is expected to enhance its utility for assessment of response to therapy and as a risk factor for breast cancer.

Breast density, MR imaging biomarkers, and breast cancer risk

Mammographic breast density and various breast MRI features are imaging biomarkers that can predict a woman's future risk of breast cancer. While mammographic density (MD) has been established as an independent risk factor for the development of breast cancer, MD assessment methods need to be accurate and reproducible for widespread clinical use in stratifying patients based on their risk. In addition, a number of breast MRI biomarkers using contrast-enhanced and noncontrast-enhanced techniques are also being investigated as risk predictors. The validation and standardization of these breast MRI biomarkers will be necessary for population-based clinical implementation of patient risk stratification, as well. This review provides an update on MD assessment methods, breast MRI biomarkers, and their ability to predict breast cancer risk.

Automated volumetric breast density measures: differential change between breasts in women with and without breast cancer

BACKGROUND

Given that breast cancer and normal dense fibroglandular tissue have similar radiographic attenuation, we examine whether automated volumetric density measures identify a differential change between breasts in women with cancer and compare to healthy controls.

METHODS

Eligible cases (n = 1160) had unilateral invasive breast cancer and bilateral full-field digital mammograms (FFDMs) at two time points: within 2 months and 1-5 years before diagnosis. Controls (n = 2360) were matched to cases on age and date of FFDMs. Dense volume (DV) and volumetric percent density (VPD) for each breast were assessed using Volpara™. Differences in DV and VPD between mammograms (median 3 years apart) were calculated per breast separately for cases and controls and their difference evaluated by using the Wilcoxon signed-rank test. To simulate clinical practice where cancer laterality is unknown, we examined whether the absolute difference between breasts can discriminate cases from controls using area under the ROC curve (AUC) analysis, adjusting for age, BMI, and time.

RESULTS

Among cases, the VPD and DV between mammograms of the cancerous breast decreased to a lesser degree (- 0.26% and - 2.10 cm³) than the normal breast (- 0.39% and - 2.74 cm³) for a difference of 0.13% (p value < 0.001) and 0.63 cm³ (p = 0.002), respectively. Among controls, the differences between breasts were nearly identical for VPD (- 0.02 [p = 0.92]) and DV (0.05 [p = 0.77]). The AUC for discriminating cases from controls using absolute difference between breasts was 0.54 (95% CI 0.52, 0.56) for VPD and 0.56 (95% CI, 0.54, 0.58) for DV.

CONCLUSION

There is a small relative increase in volumetric density measures over time in the breast with cancer which is not found in the normal breast. However, the magnitude of this difference is small, and this measure alone does not appear to be a good discriminator between women with and without breast cancer.

Relationship between Mammographic Density and Age in the United Arab Emirates Population

Objective

Higher breast density is a strong, independent risk factor for breast cancer. Breast density varies by age, ethnicity, and geographic area although dense breast tissue has been associated with younger age and premenopausal status. The relationship between breast density and age in women in the United Arab Emirates (UAE) has not been determined. This study evaluated breast density in the UAE population and its relationship with age.

Methods

Women participating in the national cancer screening program from August 2015 to May 2018 who underwent screening mammography were included. Breast parenchymal density was classified according to the American College of Radiology (ACR) Breast Imaging-Reporting and Data System (BI-RADS) from category a (almost entirely fatty) through d (extremely dense). Subjects were divided into six age groups, and the association between age and breast density was evaluated.

Results

Of the 4911 women included, 1604 (32.7%), 2149 (43.8%), 1055 (21.5%), and 103 (2.1%) were classified as having categories a–d breast density, respectively. A significant negative correlation was observed between age and breast density category (p < 0.001). Women of mean age 44 ± 7 years had the highest breast density, whereas those of mean age 56 ± 14 years had the lowest breast density. Comparisons of Emirati women with Lebanese and Western women showed that breast density was lower in Emirati women than in the other populations.

Conclusions

To our knowledge, this is the first study to evaluate the relationship between mammographic breast density and age in UAE women. As in other populations, age was inversely related to breast density, but the proportion of Emirati women with dense breasts was lower than in other populations. Because this study lacked demographic, clinical, and histopathological data, further evaluation is required.

Digital Analysis in Breast Imaging

Breast imaging is a multimodal approach that plays an essential role in the diagnosis of breast cancer. Mammography, sonography, magnetic resonance, and image-guided biopsy are imaging techniques used to search for malignant changes in the breast or precursors of malignant changes in, e.g., screening programs or follow-ups after breast cancer treatment. However, these methods still have some disadvantages such as interobserver variability and the mammography sensitivity in women with radiologically dense breasts. In order to overcome these difficulties and decrease the number of false positive findings, improvements in imaging analysis with the help of artificial intelligence are constantly being developed and tested. In addition, the extraction and correlation of imaging features with special tumor characteristics and genetics of the patients in order to get more information about treatment response, prognosis, and also cancer risk are coming more and more in focus. The aim of this review is to address recent developments in digital analysis of images and demonstrate their potential value in multimodal breast imaging.

The correlation between mammographic densities and molecular pathology in breast cancer

This study aimed to analyze the correlation between mammographic density obtained by density analysis software (DAS)/radiologists visual (RV) classification with molecular subtype, and the expression levels of estrogen receptor (ER), progesterone receptor (PR), Ki67 antigen (Ki-67), p53 gene (p53), and human epidermal growth factor receptor-2 (HER2). A total of 688 breast cancer patients with digital mammography and complete molecular pathological results in Tianjin Medical University Cancer Institute and Hospital between February 2015 and February 2016 were collected. The DAS-density grade (DASD) and the radiologists visually classified density grade (RVD) were evaluated by 3 radiologists. The correlation between density grade and the expression levels of ER, PR, Ki-67, p53, HER2 and breast cancer molecular subtype (PMS) were analyzed. The agreement between DASD and RVD was explored. ER, PR and HER-2 positive rate were significantly different among patients with different RVD grades (P< 0.05). HER2 positive rates showed an increasing trend following RVD upgrading (P𝑡𝑟𝑒𝑛𝑑< 0.05). HER-2 positive rate in RVD D1 + D2 was 7.69%, which was higher than that in D3 + D4 (P< 0.05). The ER and Ki-67 expressions in patients were markedly different among DASD (P= 0.009 and 0.002) and RVD (P= 0.012 and 0.036) with different grades. The kappa value of each DASD to RVD was 0.31 (P< 0.01). The RVD 3 proportion was 14.58% (63/432) in HER2 Over-expressing subtype, which was apparently higher than RVD1 (2.43%, 1/41) (P< 0.05). Breast density may be partial correlated with molecular pathology in breast cancer.

Mammographic Breast Density Assessment Using Deep Learning: Clinical Implementation

Purpose To develop a deep learning (DL) algorithm to assess mammographic breast density. Materials and Methods In this retrospective study, a deep convolutional neural network was trained to assess Breast Imaging Reporting and Data System (BI-RADS) breast density based on the original interpretation by an experienced radiologist of 41 479 digital screening mammograms obtained in 27 684 women from January 2009 to May 2011. The resulting algorithm was tested on a held-out test set of 8677 mammograms in 5741 women. In addition, five radiologists performed a reader study on 500 mammograms randomly selected from the test set. Finally, the algorithm was implemented in routine clinical practice, where eight radiologists reviewed 10 763 consecutive mammograms assessed with the model. Agreement on BI-RADS category for the DL model and for three sets of readings-(a) radiologists in the test set, (b) radiologists working in consensus in the reader study set, and (c) radiologists in the clinical implementation set-were estimated with linear-weighted κ statistics and were compared across 5000 bootstrap samples to assess significance. Results The DL model showed good agreement with radiologists in the test set (κ = 0.67; 95% confidence interval [CI]: 0.66, 0.68) and with radiologists in consensus in the reader study set (κ = 0.78; 95% CI: 0.73, 0.82). There was very good agreement (κ = 0.85; 95% CI: 0.84, 0.86) with radiologists in the clinical implementation set; for binary categorization of dense or nondense breasts, 10 149 of 10 763 (94%; 95% CI: 94%, 95%) DL assessments were accepted by the interpreting radiologist. Conclusion This DL model can be used to assess mammographic breast density at the level of an experienced mammographer. © RSNA, 2018 Online supplemental material is available for this article . See also the editorial by Chan and Helvie in this issue.

Evaluation of 2 breast cancer risk models in a benign breast disease cohort

BACKGROUND

More than 1.5 million women per year have a benign breast biopsy resulting in concern about their future breast cancer (BC) risk. This study examined the performance of 2 BC risk models that integrate clinical and histologic findings in this population.

METHODS

The BC risk at 5 and 10 years was estimated with the Breast Cancer Surveillance Consortium (BCSC) and Benign Breast Disease to Breast Cancer (BBD-BC) models for women diagnosed with benign breast disease (BBD) at the Mayo Clinic from 1997 to 2001. Women with BBD were eligible for the BBD-BC model, but the BCSC model also required a screening mammogram. Calibration and discrimination were assessed.

RESULTS

Fifty-six cases of BC were diagnosed among the 2142 women with BBD (median age, 50 years) within 5 years (118 were diagnosed within 10 years). The BBD-BC model had slightly better calibration at 5 years (0.89; 95% confidence interval [CI], 0.71-1.21) versus 10 years (0.81; 95% CI, 0.70-1.00) but similar discrimination in the 2 time periods: 0.68 (95% CI, 0.60-0.75) and 0.66 (95% CI, 0.60-0.71), respectively. In contrast, among the 1089 women with screening mammograms (98 cases of BC within 10 years), the BCSC model had better calibration (0.94; 95% CI, 0.85-1.43) and discrimination (0.63; 95% CI, 0.56-0.71) at 10 years versus 5 years (calibration, 1.31; 95% CI, 0.94-2.25; discrimination, 0.59; 95% CI, 0.46-0.71) where discrimination was not different from chance.

CONCLUSIONS

The BCSC and BBD-BC models were validated in the Mayo BBD cohort, although their performance differed by 5-year risk versus 10-year risk. Further enhancement of these models is needed to provide accurate BC risk estimates for women with BBD.

Automated and Clinical Breast Imaging Reporting and Data System Density Measures Predict Risk for Screen-Detected and Interval Cancers: A Case-Control Study

BACKGROUND

In 30 states, women who have had screening mammography are informed of their breast density on the basis of Breast Imaging Reporting and Data System (BI-RADS) density categories estimated subjectively by radiologists. Variation in these clinical categories across and within radiologists has led to discussion about whether automated BI-RADS density should be reported instead.

OBJECTIVE

To determine whether breast cancer risk and detection are similar for automated and clinical BI-RADS density measures.

DESIGN

Case-control.

SETTING

San Francisco Mammography Registry and Mayo Clinic.

PARTICIPANTS

1609 women with screen-detected cancer, 351 women with interval invasive cancer, and 4409 matched control participants.

MEASUREMENTS

Automated and clinical BI-RADS density assessed on digital mammography at 2 time points from September 2006 to October 2014, interval and screen-detected breast cancer risk, and mammography sensitivity.

RESULTS

Of women whose breast density was categorized by automated BI-RADS more than 6 months to 5 years before diagnosis, those with extremely dense breasts had a 5.65-fold higher interval cancer risk (95% CI, 3.33 to 9.60) and a 1.43-fold higher screen-detected risk (CI, 1.14 to 1.79) than those with scattered fibroglandular densities. Associations of interval and screen-detected cancer with clinical BI-RADS density were similar to those with automated BI-RADS density, regardless of whether density was measured more than 6 months to less than 2 years or 2 to 5 years before diagnosis. Automated and clinical BI-RADS density measures had similar discriminatory accuracy, which was higher for interval than screen-detected cancer (c-statistics: 0.70 vs. 0.62 [P < 0.001] and 0.72 vs. 0.62 [P < 0.001], respectively). Mammography sensitivity was similar for automated and clinical BI-RADS categories: fatty, 93% versus 92%; scattered fibroglandular densities, 90% versus 90%; heterogeneously dense, 82% versus 78%; and extremely dense, 63% versus 64%, respectively.

LIMITATION

Neither automated nor clinical BI-RADS density was assessed on tomosynthesis, an emerging breast screening method.

CONCLUSION

Automated and clinical BI-RADS density similarly predict interval and screen-detected cancer risk, suggesting that either measure may be used to inform women of their breast density.

PRIMARY FUNDING SOURCE

National Cancer Institute.

Comparison of variability in breast density assessment by BI-RADS category according to the level of experience

Background Only few studies have assessed variability in the results obtained by the readers with different experience levels in comparison with automated volumetric breast density measurements. Purpose To examine the variations in breast density assessment according to BI-RADS categories among readers with different experience levels and to compare it with the results of automated quantitative measurements. Material and Methods Density assignment was done for 1000 screening mammograms by six readers with three different experience levels (breast-imaging experts, general radiologists, and students). Agreement level between the results obtained by the readers and the Volpara automated volumetric breast density measurements was assessed. The agreement analysis using two categories-non-dense and dense breast tissue-was also performed. Results Intra-reader agreement for experts, general radiologists, and students were almost perfect or substantial (k = 0.74-0.95). The agreement between visual assessments of the breast-imaging experts and volumetric assessments by Volpara was substantial (k = 0.77). The agreement was moderate between the experts and general radiologists (k = 0.67) and slight between the students and Volpara (k = 0.01). The agreement for the two category groups (nondense and dense) was almost perfect between the experts and Volpara (k = 0.83). The agreement was substantial between the experts and general radiologists (k = 0.78). Conclusion We observed similar high agreement levels between visual assessments of breast density performed by radiologists and the volumetric assessments. However, agreement levels were substantially lower for the untrained readers.

InforMD: a new initiative to raise public awareness about breast density

On a mammogram, breast density (also known as mammographic density) is shown as white and bright regions and is associated with reduced sensitivity in cancer detection and increased breast cancer risk. However, many Australian women are unaware of the significance of breast density as it is not routinely reported or discussed. In order to address this lack of knowledge, Australian breast cancer researchers with expertise in mammographic density formed the InforMD alliance (INformation FORum on Mammographic Density) in 2016. The alliance is working to raise awareness of breast density with the goal of improving breast cancer diagnosis and health outcomes for women. The InforMD website (www.InforMD.org.au) was launched in October 2016, coinciding with a major nationwide public awareness campaign by the alliance during breast cancer awareness month. The website contains unbiased, accurate, updated information on breast density. The website also provides summaries of major research articles in layperson language, recent news items related to breast density, links to relevant information for health professionals, events, and feature articles. Members of the public and health professionals can also subscribe for news updates. The interactive online Forum section facilitates discussion between health professionals, scientists and members of the public. To increase online traffic to the website, Facebook (www.facebook.com/BeInforMD) and Twitter (https://twitter.com/BeInforMD_) pages were launched in December 2016. Since its launch, InforMD has generated considerable interest. The public awareness campaign reached over 7 million Australians through a combination of newspaper, TV, radio, and online news. The website has attracted 13,058 unique visitors and 30,353 page views (data as of 19/12/2017). Breast cancer researchers have a significant role to play in disseminating information to the public on breast density. A combination of mainstream and social media, together with a well-informed and updated website, has laid the groundwork for the InforMD alliance to reach a wide audience.

Breast density: why all the fuss?

The term "breast density" or mammographic density (MD) denotes those components of breast parenchyma visualised at mammography that are denser than adipose tissue. MD is composed of a mixture of epithelial and stromal components, notably collagen, in variable proportions. MD is most commonly assessed in clinical practice with the time-honoured method of visual estimation of area-based percent density (PMD) on a mammogram, with categorisation into quartiles. The computerised semi-automated thresholding method, Cumulus, also yielding area-based percent density, is widely used for research purposes; however, the advent of fully automated volumetric methods developed as a consequence of the widespread use of digital mammography (DM) and yielding both absolute and percent dense volumes, has resulted in an explosion of interest in MD recently. Broadly, the importance of MD is twofold: firstly, the presence of marked MD significantly reduces mammographic sensitivity for breast cancer, even with state-of-the-art DM. Recognition of this led to the formation of a powerful lobby group ('Are You Dense') in the US, as a consequence of which 32 states have legislated for mandatory disclosure of MD to women undergoing mammography. Secondly, it is now widely accepted that MD is in itself a risk factor for breast cancer, with a four-to sixfold increased relative risk in women with PMD in the highest quintile compared to those with PMD in the lowest quintile. Consequently, major research efforts are underway to assess whether use of MD could provide a major step forward towards risk-adapted, personalised breast cancer prevention, imaging, and treatment.

Association between mammographic breast density and histologic features of benign breast disease

BACKGROUND

Over 40% of women undergoing breast screening have mammographically dense breasts. Elevated mammographic breast density (MBD) is an established breast cancer risk factor and is known to mask tumors within the dense tissue. However, the association of MBD with high risk benign breast disease (BBD) is unknown.

METHOD

We analyzed data for 3400 women diagnosed with pathologically confirmed BBD in the Mayo Clinic BBD cohort from 1985-2001, with a clinical MBD measure (either parenchymal pattern (PP) or Breast Imaging Reporting and Data Systems (BI-RADS) density) and expert pathology review. Risk factor information was collected from medical records and questionnaires. MBD was dichotomized as dense (PP classification P2 or DY, or BI-RADS classification c or d) or non-dense (PP classification N1 or P1, or BI-RADS classification a or b). Associations of clinical and histologic characteristics with MBD were examined using logistic regression analysis to estimate odds ratios (ORs) with 95% confidence intervals (CIs).

RESULTS

Of 3400 women in the study, 2163 (64%) had dense breasts. Adjusting for age and body mass index (BMI), there were positive associations of dense breasts with use of hormone therapy (HT), lack of lobular involution, presence of atypical lobular hyperplasia (ALH), histologic fibrosis, columnar cell hyperplasia/flat epithelia atypia (CCH/FEA), sclerosing adenosis (SA), cyst, usual ductal hyperplasia, and calcifications. In fully adjusted multivariate models, HT (1.3, 95% CI 1.1-1.5), ALH (1.5, 95% CI 1.0-2.2), lack of lobular involution (OR 1.6, 95% CI 1.2-2.1, compared to complete involution), fibrosis (OR 2.2, 95% CI 1.9-2.6) and CCH/FEA (OR 1.3, 95% CI 1.0-1.6) remained significantly associated with high MBD.

CONCLUSION

Our findings support an association between high risk BBD and high MBD, suggesting that risks associated with the latter may act early in breast carcinogenesis.

Population-Attributable Risk Proportion of Clinical Risk Factors for Breast Cancer

IMPORTANCE

Many established breast cancer risk factors are used in clinical risk prediction models, although the proportion of breast cancers explained by these factors is unknown.

OBJECTIVE

To determine the population-attributable risk proportion (PARP) for breast cancer associated with clinical breast cancer risk factors among premenopausal and postmenopausal women.

DESIGN, SETTING, AND PARTICIPANTS

Case-control study with 1:10 matching on age, year of risk factor assessment, and Breast Cancer Surveillance Consortium (BCSC) registry. Risk factor data were collected prospectively from January 1, 1996, through October 31, 2012, from BCSC community-based breast imaging facilities. A total of 18 437 women with invasive breast cancer or ductal carcinoma in situ were enrolled as cases and matched to 184 309 women without breast cancer, with a total of 58 146 premenopausal and 144 600 postmenopausal women enrolled in the study.

EXPOSURES

Breast Imaging Reporting and Data System (BI-RADS) breast density (heterogeneously or extremely dense vs scattered fibroglandular densities), first-degree family history of breast cancer, body mass index (>25 vs 18.5-25), history of benign breast biopsy, and nulliparity or age at first birth (≥30 years vs <30 years).

MAIN OUTCOMES AND MEASURES

Population-attributable risk proportion of breast cancer.

RESULTS

Of the 18 437 women with breast cancer, the mean (SD) age was 46.3 (3.7) years among premenopausal women and 61.7 (7.2) years among the postmenopausal women. Overall, 4747 (89.8%) premenopausal and 12 502 (95.1%) postmenopausal women with breast cancer had at least 1 breast cancer risk factor. The combined PARP of all risk factors was 52.7% (95% CI, 49.1%-56.3%) among premenopausal women and 54.7% (95% CI, 46.5%-54.7%) among postmenopausal women. Breast density was the most prevalent risk factor for both premenopausal and postmenopausal women and had the largest effect on the PARP; 39.3% (95% CI, 36.6%-42.0%) of premenopausal and 26.2% (95% CI, 24.4%-28.0%) of postmenopausal breast cancers could potentially be averted if all women with heterogeneously or extremely dense breasts shifted to scattered fibroglandular breast density. Among postmenopausal women, 22.8% (95% CI, 18.3%-27.3%) of breast cancers could potentially be averted if all overweight and obese women attained a body mass index of less than 25.

CONCLUSIONS AND RELEVANCE

Most women with breast cancer have at least 1 breast cancer risk factor routinely documented at the time of mammography, and more than half of premenopausal and postmenopausal breast cancers are explained by these factors. These easily assessed risk factors should be incorporated into risk prediction models to stratify breast cancer risk and promote risk-based screening and targeted prevention efforts.

Combining quantitative and qualitative breast density measures to assess breast cancer risk

BACKGROUND

Accurately identifying women with dense breasts (Breast Imaging Reporting and Data System [BI-RADS] heterogeneously or extremely dense) who are at high breast cancer risk will facilitate discussions of supplemental imaging and primary prevention. We examined the independent contribution of dense breast volume and BI-RADS breast density to predict invasive breast cancer and whether dense breast volume combined with Breast Cancer Surveillance Consortium (BCSC) risk model factors (age, race/ethnicity, family history of breast cancer, history of breast biopsy, and BI-RADS breast density) improves identifying women with dense breasts at high breast cancer risk.

METHODS

We conducted a case-control study of 1720 women with invasive cancer and 3686 control subjects. We calculated ORs and 95% CIs for the effect of BI-RADS breast density and Volpara™ automated dense breast volume on invasive cancer risk, adjusting for other BCSC risk model factors plus body mass index (BMI), and we compared C-statistics between models. We calculated BCSC 5-year breast cancer risk, incorporating the adjusted ORs associated with dense breast volume.

RESULTS

Compared with women with BI-RADS scattered fibroglandular densities and second-quartile dense breast volume, women with BI-RADS extremely dense breasts and third- or fourth-quartile dense breast volume (75% of women with extremely dense breasts) had high breast cancer risk (OR 2.87, 95% CI 1.84-4.47, and OR 2.56, 95% CI 1.87-3.52, respectively), whereas women with extremely dense breasts and first- or second-quartile dense breast volume were not at significantly increased breast cancer risk (OR 1.53, 95% CI 0.75-3.09, and OR 1.50, 95% CI 0.82-2.73, respectively). Adding continuous dense breast volume to a model with BCSC risk model factors and BMI increased discriminatory accuracy compared with a model with only BCSC risk model factors (C-statistic 0.639, 95% CI 0.623-0.654, vs. C-statistic 0.614, 95% CI 0.598-0.630, respectively; P < 0.001). Women with dense breasts and fourth-quartile dense breast volume had a BCSC 5-year risk of 2.5%, whereas women with dense breasts and first-quartile dense breast volume had a 5-year risk ≤ 1.8%.

CONCLUSIONS

Risk models with automated dense breast volume combined with BI-RADS breast density may better identify women with dense breasts at high breast cancer risk than risk models with either measure alone.

Risk-based Breast Cancer Screening: Implications of Breast Density

The approach to breast cancer screening has changed over time from a general approach to a more personalized, risk-based approach. Women with dense breasts, one of the most prevalent risk factors, are now being informed that they are at increased risk of developing breast cancer and should consider supplemental screening beyond mammography. This article reviews the current evidence regarding the impact of breast density relative to other known risk factors, the evidence regarding supplemental screening for women with dense breasts, supplemental screening options, and recommendations for physicians having shared decision-making discussions with women who have dense breasts.

Inter-reader Variability in the Use of BI-RADS Descriptors for Suspicious Findings on Diagnostic Mammography: A Multi-institution Study of 10 Academic Radiologists

RATIONALE AND OBJECTIVES

The study aimed to determine the inter-observer agreement among academic breast radiologists when using the Breast Imaging Reporting and Data System (BI-RADS) lesion descriptors for suspicious findings on diagnostic mammography.

MATERIALS AND METHODS

Ten experienced academic breast radiologists across five medical centers independently reviewed 250 de-identified diagnostic mammographic cases that were previously assessed as BI-RADS 4 or 5 with subsequent pathologic diagnosis by percutaneous or surgical biopsy. Each radiologist assessed the presence of the following suspicious mammographic findings: mass, asymmetry (one view), focal asymmetry (two views), architectural distortion, and calcifications. For any identified calcifications, the radiologist also described the morphology and distribution. Inter-observer agreement was determined with Fleiss kappa statistic. Agreement was also calculated by years of experience.

RESULTS

Of the 250 lesions, 156 (62%) were benign and 94 (38%) were malignant. Agreement among the 10 readers was strongest for recognizing the presence of calcifications (k = 0.82). There was substantial agreement among the readers for the identification of a mass (k = 0.67), whereas agreement was fair for the presence of a focal asymmetry (k = 0.21) or architectural distortion (k = 0.28). Agreement for asymmetries (one view) was slight (k = 0.09). Among the categories of calcification morphology and distribution, reader agreement was moderate (k = 0.51 and k = 0.60, respectively). Readers with more experience (10 or more years in clinical practice) did not demonstrate higher levels of agreement compared to those with less experience.

CONCLUSIONS

Strength of agreement varies widely for different types of mammographic findings, even among dedicated academic breast radiologists. More subtle findings such as asymmetries and architectural distortion demonstrated the weakest agreement. Studies that seek to evaluate the predictive value of certain mammographic features for malignancy should take into consideration the inherent interpretive variability for these findings.

Influences of race and breast density on related cognitive and emotion outcomes before mandated breast density notification

RATIONALE

Many states have adopted laws mandating breast density (BD) notification for applicable women; however, very little is known about what women knew or felt about BD and related breast cancer (BC) risk before implementation of BD notification laws.

OBJECTIVE

We examined between-race differences in the extent to which having dense breasts was associated with women's related BD cognition and emotion, and with health care providers' communication about BD.

METHODS

We received surveys between May and October of 2015 assessing health care provider (HCP) communication about BD, BD-related knowledge, BD-related anxiety and BC worry from 182 African American (AA) and 113 European American (EA) women in the state of Michigan for whom we had radiologists' assessments of BD.

RESULTS

Whereas having dense breasts was not associated with any BD-related cognition or emotion, there were robust effects of race as follows: EA women were more likely to have been told about BD by a HCP, more likely to know their BD status, had greater knowledge of BD and of BC risk, and had greater perceptions of BC risk and worry; AA women had greater BD-related anxieties. EA women's greater knowledge of their own BD status was directly related to the increased likelihood of HCP communication about BD. However, HCP communication about BD attenuated anxiety for AA women only.

CONCLUSION

We present the only data of which we are aware that examines between-race differences in the associations between actual BD, HCP communication and BD related cognition and emotion before the implementation of BD notification laws. Our findings suggest that the BD notification laws could yield positive benefits for disparities in BD-related knowledge and anxiety when the notifications are followed by discussions with health care providers.

Increased Risk of Developing Breast Cancer after a False-Positive Screening Mammogram

BACKGROUND

Women with a history of a false-positive mammogram result may be at increased risk of developing subsequent breast cancer.

METHODS

Using 1994 to 2009 Breast Cancer Surveillance Consortium data, we included women ages 40 to 74 years with a screening mammogram that resulted in a false-positive with recommendation for additional imaging, false-positive with recommendation for biopsy, or true-negative with no cancer within one year following the examination. We used partly conditional Cox proportional hazards survival models to assess the association between a false-positive mammogram result and subsequent breast cancer, adjusting for potential confounders. Adjusted survival curves stratified by breast density and false-positive result were used to evaluate changes in risk over time.

RESULTS

During 12,022,560 person-years of follow-up, 48,735 cancers were diagnosed. Compared with women with a true-negative examination, women with a false-positive with additional imaging recommendation had increased risk of developing breast cancer [adjusted HR (aHR) = 1.39; 95% confidence interval (CI), 1.35-1.44] as did women with a false-positive with a biopsy recommendation (aHR = 1.76; 95% CI,1.65-1.88). Results stratifying by breast density were similar to overall results except among women with almost entirely fatty breasts in which aHRs were similar for both the false-positive groups. Women with a false-positive result had persistently increased risk of developing breast cancer 10 years after the false-positive examination.

CONCLUSION/IMPACT

Women with a history of a false-positive screening mammogram or biopsy recommendation were at increased risk of developing breast cancer for at least a decade, suggesting that prior false-positive screening may be useful in risk prediction models.

Variation in Mammographic Breast Density Assessments Among Radiologists in Clinical Practice: A Multicenter Observational Study

BACKGROUND

About half of the United States has legislation requiring radiology facilities to disclose mammographic breast density information to women, often with language recommending discussion of supplemental screening options for women with dense breasts.

OBJECTIVE

To examine variation in breast density assessment across radiologists in clinical practice.

DESIGN

Cross-sectional and longitudinal analyses of prospectively collected observational data.

SETTING

30 radiology facilities within the 3 breast cancer screening research centers of the Population-based Research Optimizing Screening through Personalized Regimens (PROSPR) consortium.

PARTICIPANTS

Radiologists who interpreted at least 500 screening mammograms during 2011 to 2013 (n = 83). Data on 216 783 screening mammograms from 145 123 women aged 40 to 89 years were included.

MEASUREMENTS

Mammographic breast density, as clinically recorded using the 4 Breast Imaging Reporting and Data System categories (heterogeneously dense and extremely dense categories were considered "dense" for analyses), and patient age, race, and body mass index (BMI).

RESULTS

Overall, 36.9% of mammograms were rated as showing dense breasts. Across radiologists, this percentage ranged from 6.3% to 84.5% (median, 38.7% [interquartile range, 28.9% to 50.9%]), with multivariable adjustment for patient characteristics having little effect (interquartile range, 29.9% to 50.8%). Examination of patient subgroups revealed that variation in density assessment across radiologists was pervasive in all but the most extreme patient age and BMI combinations. Among women with consecutive mammograms interpreted by different radiologists, 17.2% (5909 of 34 271) had discordant assessments of dense versus nondense status.

LIMITATION

Quantitative measures of mammographic breast density were not available for comparison.

CONCLUSION

There is wide variation in density assessment across radiologists that should be carefully considered by providers and policymakers when considering supplemental screening strategies. The likelihood of a woman being told she has dense breasts varies substantially according to which radiologist interprets her mammogram.

PRIMARY FUNDING SOURCE

National Institutes of Health.

Racial Differences in Quantitative Measures of Area and Volumetric Breast Density

BACKGROUND

Increased breast density is a strong risk factor for breast cancer and also decreases the sensitivity of mammographic screening. The purpose of our study was to compare breast density for black and white women using quantitative measures.

METHODS

Breast density was assessed among 5282 black and 4216 white women screened using digital mammography. Breast Imaging-Reporting and Data System (BI-RADS) density was obtained from radiologists' reports. Quantitative measures for dense area, area percent density (PD), dense volume, and volume percent density were estimated using validated, automated software. Breast density was categorized as dense or nondense based on BI-RADS categories or based on values above and below the median for quantitative measures. Logistic regression was used to estimate the odds of having dense breasts by race, adjusted for age, body mass index (BMI), age at menarche, menopause status, family history of breast or ovarian cancer, parity and age at first birth, and current hormone replacement therapy (HRT) use. All statistical tests were two-sided.

RESULTS

There was a statistically significant interaction of race and BMI on breast density. After accounting for age, BMI, and breast cancer risk factors, black women had statistically significantly greater odds of high breast density across all quantitative measures (eg, PD nonobese odds ratio [OR] = 1.18, 95% confidence interval [CI] = 1.02 to 1.37, P = .03, PD obese OR = 1.26, 95% CI = 1.04 to 1.53, P = .02). There was no statistically significant difference in BI-RADS density by race.

CONCLUSIONS

After accounting for age, BMI, and other risk factors, black women had higher breast density than white women across all quantitative measures previously associated with breast cancer risk. These results may have implications for risk assessment and screening.

Classification of fatty and dense breast parenchyma: comparison of automatic volumetric density measurement and radiologists' classification and their inter-observer variation

BACKGROUND

Automatically calculated breast density is a promising alternative to subjective BI-RADS density assessment. However, such software needs a cutoff value for density classification.

PURPOSE

To determine the volumetric density threshold which classifies fatty and dense breasts with highest accuracy compared to average BI-RADS density assessment, and to analyze radiologists' inter-observer variation.

MATERIAL AND METHODS

A total of 537 full field digital mammography examinations were randomly selected from a population based screening program. Five radiologists assessed density using the BI-RADS density scale, where BI-RADS I-II were classified as fatty and III-IV as dense. A commercially available software (Quantra) calculated volumetric breast density. We calculated the cutoff (threshold) values in volumetric density that yielded highest accuracy compared to median and individual radiologists' classification. Inter-observer variation was analyzed using the kappa statistic.

RESULTS

The threshold that best matched the median radiologists' classification was 10%, which resulted in 87% accuracy. Thresholds that best matched individual radiologist's classification had a range of 8-15%. A total of 191 (35.6 %) cases were scored both dense and fatty by at least one radiologist. Fourteen (2.6 %) cases were unanimously scored by the radiologists, yet differently using automatic assessment. The agreement (kappa) between reader's median classification and individual radiologists was 0.624 to 0.902, and agreement between median classification and Quantra was 0.731.

CONCLUSION

The optimal volumetric threshold of 10% using automatic assessment would classify breast parenchyma as fatty or dense with substantial accuracy and consistency compared to radiologists' BI-RADS categorization, which suffers from high inter-observer variation.

The Use of Ultrasound in Breast Cancer Screening of Asymptomatic Women with Dense Breast Tissue: A Narrative Review

INTRODUCTION

Mammography is the standard screening modality for breast cancer; however, sensitivity reduces with increasing breast density, resulting in the potential for masking of cancer. Ultrasound is a potential supplemental screening tool, but its routine use is controversial.

METHODS

A database search was performed with keywords "ultrasound" and "breast density and screening", including variations. Articles were included if they assessed the use of hand-held ultrasound as a supplemental screening modality in women with dense breasts.

DISCUSSION

Twelve articles were identified. No high-level evidence articles were identified. Cancer detection rates increased with the addition of ultrasound-to-mammography screening protocols. However, this was associated with increased costs per cancer detected, an increased biopsy rate, and a low positive predictive value. The survival benefit, cost versus benefit, and psychological impact of the addition of ultrasound is unknown.

CONCLUSIONS

The addition of ultrasound to a screening program in an asymptomatic population of women with dense breast tissue detects additional cancers compared with mammography alone. Knowledge regarding a survival or cost benefit associated with increased cancer detection, and the psychological impact of the addition of ultrasound is unknown. Further research is needed to assess whether the addition of ultrasound is cost-effective with respect to clinical outcome and survival.

Consistency of breast density categories in serial screening mammograms: A comparison between automated and human assessment

Reliable breast density measurement is needed to personalize screening by using density as a risk factor and offering supplemental screening to women with dense breasts. We investigated the categorization of pairs of subsequent screening mammograms into density classes by human readers and by an automated system. With software (VDG) and by four readers, including three specialized breast radiologists, 1000 mammograms belonging to 500 pairs of subsequent screening exams were categorized into either two or four density classes. We calculated percent agreement and the percentage of women that changed from dense to non-dense and vice versa. Inter-exam agreement (IEA) was calculated with kappa statistics. Results were computed for each reader individually and for the case that each mammogram was classified by one of the four readers by random assignment (group reading). Higher percent agreement was found with VDG (90.4%, CI 87.9-92.9%) than with readers (86.2-89.2%), while less plausible changes from non-dense to dense occur less often with VDG (2.8%, CI 1.4-4.2%) than with group reading (4.2%, CI 2.4-6.0%). We found an IEA of 0.68-0.77 for the readers using two classes and an IEA of 0.76-0.82 using four classes. IEA is significantly higher with VDG compared to group reading. The categorization of serial mammograms in density classes is more consistent with automated software than with a mixed group of human readers. When using breast density to personalize screening protocols, assessment with software may be preferred over assessment by radiologists.