OBJECTIVES:To evaluate patient acceptance and feedback regarding supplemental imaging modalities: automated whole-breast ultrasound (ABUS), contrast-enhanced mammography (CEM), and abbreviated breast MRI (AB-MRI) within the BRAID (Breast Screening: Risk Adaptive Imaging for Density) trial. MATERIALS AND METHODS:An adapted Testing Morbidities Index questionnaire was utilised to capture participant experiences and perceptions (January-April 2024) related to AB-MRI, ABUS and CEM. Likert-scale questions assessed discomfort, anxiety, and overall satisfaction for each imaging modality, while thematic analysis was applied to free-text patient feedback. Additionally, reasons for withdrawal were recorded for each modality. RESULTS:Among 159 women providing feedback, 57/159 (35.8%) underwent ABUS, 52/159 (32.7%) CEM, and 50/159 (31.5%) AB-MRI. Acceptability of ABUS, CEM and AB-MRI was rated similarly to mammography by 71/159 (64.8%) of these respondents, with 72/159 (45.3%) considering them superior. Mild-to-moderate discomfort due to breast compression was reported for ABUS and CEM, whereas AB-MRI resulted in the least discomfort. Pre-procedural anxiety was observed across all imaging modalities, particularly with contrast-enhanced techniques; however, experiences were generally well-tolerated. Effective communication and pre-test information reduced anxiety levels, with most participants willing to repeat the procedures. 151/984 (15.3%) withdrawals in BRAID were due to adverse patient experiences, with contrast-enhanced techniques accounting for most of these withdrawals (CEM: 69/151, 45.7%; AB-MRI: 66/151, 43.7%; ABUS: 12/151, 7.9%). The main reasons for withdrawal were unhappiness with the allocated imaging arm and discomfort or anxiety during the procedure. CONCLUSION:Supplemental imaging modalities are generally well-accepted by patients with benefit throughout gained by clear communication and preparedness. CRITICAL RELEVANCE STATEMENT:Feedback from a subgroup of women participating in the BRAID trial shows that supplemental imaging alongside routine screening is well-accepted. Clear communication and empathetic care further improve acceptance, supporting a shift toward personalised breast cancer screening for women with dense breasts. KEY POINTS:Understanding women's imaging experiences is essential for optimising breast screening practices. Acceptability of supplemental imaging was rated similar to or better than mammography by most participants. Clear, empathetic communication reduced anxiety and improved experience with contrast-enhanced imaging.

The impact of incorporating artificial intelligence (AI) into a double-read breast-screening workflow, including arbitration, is unclear. This retrospective study included 50,000 representative women from two NHS breast-screening centers. All the women had long-term follow-up, allowing us to determine whether use of AI leads to earlier cancer detection. Cases requiring arbitration (8,732 cases) were read by 22 readers in a reader study, following their normal arbitration workflow. Overall, after arbitration, replacing the second reader with AI was noninferior (5% margin) to two human readers in terms of sensitivity and specificity (P < 0.001) while offering a workload benefit. Arbitration improved the specificity of the AI arm by overruling cases incorrectly recalled by the AI tool; however, it also overruled the AI tool recall decision for some interval and next-round cancers. Further development of the AI tool alongside improvement in its explainability could lead to the earlier detection of cancers.

To compare tumour characteristics of screen-detected and interval cancers and their relationship with breast density and AJCC prognostic stage. In this retrospective study, women screened with mammography between 1/4/2017 and 30/3/2020 at one site were included. Tumour characteristics were compared for screen-detected and interval cancers with breast density (Volpara) and AJCC (8th edition) prognostic stage for interval cancers. Categorical variables were compared using Pearson’s χ², Fisher’s exact, or binomial tests. From 55,010 attendees, 723 cancers were diagnosed (463 screen-detected; 260 interval). Time to interval cancer diagnosis was longer in women with dense breasts compared with non-dense breasts (median: 767 vs 624 days; p = 0.04). Earlier stage interval cancers ( < IIA) were more likely diagnosed in the second and third year compared to year 1 ((33/78) 48.7

Since the discovery of x-rays, the role of women in imaging has become increasingly important. Reflecting the changes in the roles of women in society, and the increasing acceptance that women make an effective and complimentary contribution to the workforce, the number of women working in imaging has increased in the past few decades. This is due in part by the change from male only medical schools early last century to having equal numbers of male and female medical students in many countries by the seventies. Radiology is an attractive career option for women and over the past two decades the proportion of women has increased and is now around 41% in radiology. This has resulted an increase in the number of women in leadership roles. This article examines the evidence for women in various roles, explores why having women in positions of influence is important, and looks at trends in the data. The recent move away from Equality, Diversity and Inclusion (EDI) is disappointing as this affirmative action did help greater numbers of women into senior roles. It is clear there are many opportunities for professional women to step up and we should encourage and support them to do so.

Breast cancer is a major cause of cancer death among women, emphasising the importance of early detection for improved treatment outcomes and quality of life. Mammography, the primary diagnostic imaging test, poses challenges due to the high variability and patterns in mammograms. Double reading of mammograms is recommended in many screening programmes to improve diagnostic accuracy but increases the workload for radiologists. Therefore, researchers have explored Machine Learning models to support expert decision-making. Stand-alone models have shown comparable or superior performance to radiologists, but some studies observe decreased sensitivity when facing multiple datasets and sites, indicating the need for highly generalisable and robust models. This work devises MammoDG, a novel deep-learning framework for generalisable, robust, and reliable analysis of cross-domain multi-center mammography data. MammoDG introduces a cross-channel cross-view enhancement module to conduct information interaction among multi-view mammograms. Additionally, an instance-level contrastive regularisation is designed to mitigate the distribution gap between multi-center data and thus enhance generalisation capabilities. Extensive validation demonstrates the superiority of MammoDG over existing models, highlighting the critical importance of domain generalisation for trustworthy mammography analysis in the presence of imaging protocol variations.

Artificial intelligence (AI) promises to enhance breast cancer screening. Here we evaluated Google's mammography AI system (version 1.2) across two phases: a retrospective study using 115,973 mammograms from five National Health Service screening services with 39-month follow-up and prospective noninterventional feasibility deployment at 12 sites (9,266 cases). The primary endpoint was AI sensitivity and specificity versus first reader using a 5% noninferiority margin. The secondary endpoints were performance versus second or consensus readers and breast-level analyses. Retrospectively, AI achieved superior sensitivity (0.541 versus 0.437 for first reader, P < 0.001) and noninferior specificity (0.943 versus 0.952, P < 0.001). Cancer detection rate increased from 7.54 to 9.33 per 1,000 women, with AI detecting 25.0% of interval cancers. Performance was particularly strong for first screens (39.3% fewer recalls, 8.8% higher detection) and invasive cancers. No systematic demographic disparities were observed. Simulated second-reader replacement reduced reading time by 32% while increasing detection by 17.7%. Prospective deployment confirmed technical feasibility but revealed a distribution shift requiring threshold recalibration. Implementation requires adaptive calibration and continuous monitoring to ensure safety and equity.

The current practice of pre-operative imaging in breast cancer is highly varied throughout Europe. Therefore, the European Society of Breast Imaging (EUSOBI) launched a call to all other European scientific societies involved in breast care to provide expert advice for pre-operative imaging in breast cancer and come to a common understanding of the types of evidence required for clinical practice guidelines for diagnostic tests. A panel comprising 13 experts (invited based on their level of expertise and representation in medical societies) voted on statements and questions encompassing all aspects of pre-operative staging. Consensus was reached in 67.4

To investigate the surgical impact of preoperative breast MRI in patients diagnosed with invasive lobular breast cancer (ILC) in a prospective observational study. The prospective MIPA observational study database was queried for patients aged 18–80 with newly diagnosed unilateral ILC at needle biopsy referred for primary surgery. Patients who underwent preoperative MRI (MRI group) were matched (1:1) with those who did not (noMRI group) according to nine confounding covariates. Surgical outcomes were compared between the matched groups with nonparametric statistics after calculating odds ratios (ORs). A total of 547 women with unilateral needle biopsy-diagnosed ILC were identified (158 noMRI group, 389 MRI group). After patient matching, each group retained 103 patients, for a total of 206 matched patients. For the rate of women having a first-line mastectomy, there was no significant difference between the MRI group (21.4

BACKGROUND:It is not known which supplemental imaging technique is most beneficial for women with dense breasts attending breast screening. This study compares abbreviated MRI, automated whole breast ultrasound (ABUS), and contrast-enhanced mammography versus standard of care in women with dense breasts and a negative mammogram. We report on interim results from the first round of supplemental imaging. METHODS:In this UK randomised controlled trial, at ten breast screening sites, women (aged 50-70 years) were independently allocated by batches (day/mobile screening van) to either abbreviated MRI, ABUS, or contrast-enhanced mammography or standard of care (full-field digital mammography) varied by modality availability at each centre. Women were invited if their mammogram was negative and they had dense breasts. Primary outcome was detection rate, defined as the percentage of women with a positive result on supplemental imaging that resulted in histologically confirmed breast cancer. Analysis was by imaging received (intention to treat) using network meta-analysis, treating each site as a study in the meta-analysis, with two analyses carried out: one using only the three active intervention arms (primary analysis) that compared the three supplemental imaging techniques with respect to cancer detection, recall, and biopsy rates in addition to those resulting from full-field digital mammography alone; and one with the addition of the observational data from Cambridge on full-field digital mammography alone. This trial is closed for recruitment and is registered with ClinicalTrials.gov, NCT04097366. FINDINGS:From October 18, 2019, to March 30, 2024, 9361 eligible women were recruited and randomly assigned (2318 to abbreviated MRI, 2240 to ABUS, 2235 to contrast-enhanced mammography, and 2568 to standard of care). Of those, 6305 completed supplementary imaging (2130 in the abbreviated MRI, 2141 in the ABUS, and 2035 in the contrast-enhanced mammography) and were included in the outcome analysis. The cancer detection rate was 17·4 (95% CI 12·2-23·9, n=37) per 1000 examinations for abbreviated MRI, 4·2 (1·9-8·0, n=9) per 1000 examinations for ABUS, and 19·2 (13·7-26·1, n=39) per 1000 examinations for contrast-enhanced mammography, of which 15·0 (10·3-21·1, n=32) per 1000 women for abbreviated MRI, 4·2 (1·9-8·0, n=9) per 1000 examinations for ABUS, and 15·7 (10·8-22·1, n=32) per 1000 examinations for contrast-enhanced mammography were invasive cancers. The detection rates for abbreviated MRI were significantly higher than for ABUS (p=0·047) and non-significantly higher than for contrast-enhanced mammography (p=0·62). There was one case of extravasation in the abbreviated MRI arm (0·5 events per 1000 examinations), no adverse events in the ABUS arm, and 24 iodinated contrast reactions (17 minor [8·4 events per 1000 examinations], six moderate [2·9 events per 1000 examinations], and one severe [0·5 events per 1000 examinations]) and three extravasations (1·5 extravasations per 1000 examinations) in the contrast-enhanced mammography arm. INTERPRETATION:Abbreviated MRI and contrast-enhanced mammography detected three times as many invasive cancers compared with ABUS, with cancers being half the size. This study shows that supplemental imaging could lead to earlier detection of cancer in women with dense breasts but does not estimate the level of overdiagnosis. FUNDING:Cancer Research UK, GE Healthcare, and Bayer Healthcare.

Background Deep learning risk algorithms for personalized breast cancer screening outperform traditional methods in retrospective evaluations, but triennial screening assessments are lacking. Purpose To evaluate the predictive ability of 3-year risk scores generated by a deep learning algorithm (Mirai) to identify women who developed interval cancers (ICs) in the UK breast screening program, which invites women aged 50-70 years for triennial mammography. Materials and Methods For this retrospective study, Mirai processed digital screening mammograms with negative results collected from a 3-year cohort (January 2014 to December 2016) across two sites and two primary mammography systems. Exclusions included screen-detected cancers (baseline and next round), implants, and nonstandard views. The reference standard was no cancer diagnosis within 40 months of negative screening, confirmed histopathologically. The primary objective was predicting ICs at 1-, 2-, and 3-year time points after baseline screening. Secondary objectives were assessing predictions across age quartiles and Breast Imaging Reporting and Data System (BI-RADS) breast densities. Areas under the receiver operating characteristic curve (AUCs) and true positives (ICs) were calculated across operating thresholds. Risk score distributions were compared with the Mann-Whitney U test, and AUCs were compared with the DeLong test. Results Analysis included 134 217 examinations from the same number of women (mean age, 59.1 years ± 7.9 [SD]), including 524 ICs. There was no evidence of performance differences among 1-, 2-, and 3-year IC predictions (P ≥ .63), age quartiles (P ≥ .73), or breast densities (P ≥ .99). Overall AUCs were 0.72 (95% CI: 0.65, 0.78), 0.67 (95% CI: 0.64, 0.70), and 0.67 (95% CI: 0.65, 0.70) for 1-, 2-, and 3-year IC predictions, respectively. C indexes for age quartiles were 0.67 (95% CI: 0.62, 0.71) for age younger than 52 years, 0.70 (95% CI: 0.65, 0.75) for age 52-58 years, 0.71 (95% CI: 0.67, 0.75) for age 59-65 years, and 0.71 (95% CI: 0.67, 0.75) for age of 66 years and older. C indexes for BI-RADS categories a, b, c, and d were 0.70 (95% CI: 0.62, 0.78), 0.69 (95% CI: 0.65, 0.73), 0.68 (95% CI: 0.64, 0.71), and 0.67 (95% CI: 0.62, 0.73), respectively. Three-year risk scores retrospectively predicted 3.6% (19 of 524), 14.5% (76 of 524), 26.1% (137 of 524), and 42.4% (222 of 524) of ICs for women assigned the highest 1%, 5%, 10%, and 20% of scores. Conclusion Mirai could identify women for more frequent screening or additional imaging, detecting ICs earlier. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Philpotts in this issue.

Early detection of breast cancer is critical for improving patient outcomes. While mammography remains the primary screening modality, magnetic resonance imaging (MRI) is increasingly recommended as a supplemental tool for women with dense breast tissue and those at elevated risk. However, the acquisition and interpretation of multiparametric breast MRI are time-consuming and require specialized expertise, limiting scalability in clinical practice. Artificial intelligence (AI) methods have shown promise in supporting breast MRI interpretation, but their development is hindered by the limited availability of large, diverse, and publicly accessible datasets. To address this gap, we present a publicly available, multi-center breast MRI dataset collected across six clinical institutions in five European countries. The dataset comprises 741 examinations from women undergoing screening or diagnostic breast MRI and includes malignant, benign, and non-lesion cases. Data were acquired using heterogeneous scanners, field strengths, and acquisition protocols, reflecting real-world clinical variability. In addition, we report baseline benchmark experiments using a transformer-based model to illustrate potential use cases of the dataset and to provide reference performance for future methodological comparisons.

To estimate tumour volume doubling time (TVDT) of interval cancers (ICs). Two radiologists retrospectively reviewed prior screening and diagnostic mammograms and measured mean diameter on “visible” ICs. Univariate analyses of clinicopathological variables (ER, HER2, grade, age at diagnosis, and breast density) were undertaken, and those with p < 0.1 were included in a generalised linear model to estimate TVDT, cancer size at screening, and time of cancer visibility for “non-visible” tumours. From 2011 to 2018, 476 ICs were diagnosed, almost half in the third year after screening with 86

BACKGROUND:To evaluate the acceptability of a risk-based breast cancer screening (BCS) strategy among professionals involved in MyPeBS study in 6 countries. METHODS:After qualitative interviews, a questionnaire was built with a Delphi method: to evaluate professionals' basic understanding, satisfaction and reactions to each stage of the trial, opinions on BCS and its future. The questionnaire was distributed by emailing 698 investigators, who forwarded it to all categories of professionals involved in trial recruitment (physicians, medical secretaries, nurses, and mammography technicians). Descriptive statistics were used to summarize views on acceptability. RESULTS:Among the 198 respondents, most declared being at ease with the trial design and the concept of breast cancer risk estimation. They were mostly comfortable explaining the different trial steps, communicating risk estimation, and answering women's questions. Some professionals were not comfortable explaining high (7.1%) and low-risk categories (9%) and did not feel sufficiently trained (26.5%). Although professionals were mostly confident about risk-based approaches and the potential of this to improve breast cancer screening (93.5%), 58% called for further validation of the risk-models to predict risk before implementation in population-based programs. They expressed concerns about the complexity of this screening strategy, stressing the need to properly inform the public and to train professionals in delivering risk assessment. CONCLUSION:This first study assessing the perspectives of professionals delivering risk-based BCS. As professional acceptability is key for successful implementation, training for all professionals and tools to help them communicate risk to women will be necessary to develop risk assessment in BCS. TRIAL REGISTRATION:Study sponsor: Unicancer. My personalised breast screening (MyPeBS). CLINICALTRIALS:gov (2018) available at https://www. CLINICALTRIALS:gov/ct2/show/NCT03672331 .

Artificial intelligence (AI) projects in healthcare research and practice require approval from information governance (IG) teams within relevant healthcare providers. Navigating this approval process has been highlighted as a key challenge for AI innovation in healthcare by many stakeholders focused on the development and adoption of AI. Data privacy and impact assessments are a part of the approval process which is often identified as the focal point for these challenges. This perspective reports insights from a multidisciplinary workshop aiming to characterise challenges and explore potential solutions collaboratively. Themes around the variation in AI technologies, governance processes and stakeholder perspectives arose, highlighting the need for training initiatives, communities of practice and the standardization of governance processes and structures across NHS Trusts.

Purpose: A survey conducted by the European Society of Breast Imaging (EUSOBI) in 2023 revealed significant variations in Quality Assurance (QA) practices across Europe. The UK encourages regular performance monitoring for screen readers. This study aimed to assess the variability in diagnostic performance among readers participating in a wider prospective randomised trial across multiple countries. Method: In this retrospective multinational study, breast imaging readers from the MyPeBS clinical trial examined a test set of 40 challenging breast screening cases using the PERFORMS software, from March 2021 to February 2022. The challenging set, enriched with biopsy-proven cancers, aimed to differentiate readers by their level of diagnostic performance. Cancer detection and correct return to screen rates were calculated for each participant. Results: A total of 110 readers from 6 countries completed the PERFORMS test set, while 88 also completed an accompanying questionnaire collecting information about their breast screening work and experience. The study revealed variability in cancer detection rates (M = 73.6 %, SD = 19.7 %, range 0.0 %-100.0 %) and correct return to screen rates (M = 79.7 %, SD = 10.5 %, range 46.4 %-100.0 %). Outliers with extremely low cancer detection (2.7 % of participants) and correct return to screen rates (1.8 % of participants) were also identified. Conclusions: Breast imaging readers' performance in test set-based assessments like PERFORMS can reflect realworld screening proficiency. The presence of outlier readers with low diagnostic performance on the test highlights the need for double reading and for standardised QA protocols to ensure patient safety and service efficiency.

Breast cancer is the most prevalent cancer in women in Europe, and while all European countries have some form of screening for breast cancer, disparities in organization and implementation exist. Breast density is a well-established risk factor for breast cancer; however, most countries in Europe do not have recommendations in place for notification of breast density or additional supplementary imaging for women with dense breasts. Various supplemental screening modalities have been investigated in Europe, and when comparing modalities, MRI has been shown to be superior in cancer detection rate and in detecting small invasive disease that may impact long-term survival, as demonstrated in the Dense Tissue and Early Breast Neoplasm Screening (DENSE) trial in the Netherlands. Based on convincing evidence, the European Society of Breast Imaging issued recommendations that women with category D density undergo breast MRI from ages 50 to 70 years at least every 4 years and preferably every 2 to 3 years. However, currently no countries in Europe routinely offer women with BI-RADS category D density breasts MRI as supplemental imaging. The reasons for lack of implementation of MRI screening are multifactorial. Concerns regarding increased recalls have been cited, as have cost and lack of resources. However, studies have demonstrated breast MRI in women with BI-RADS category D density breasts to be cost-effective compared with the current breast cancer screening standard of biannual mammography. Furthermore, abbreviated MRI protocols could facilitate more widespread use of affordable MRI screening. Women's perception on breast density notification and supplemental imaging is key to successful implementation.

The aim of this work is to evaluate the performance of deep learning (DL) models for breast cancer diagnosis with MRI. A literature search was conducted on Web of Science, PubMed, and IEEE Xplore for relevant studies published from January 2015 to February 2024. The study was registered with the PROSPERO International Prospective Register of Systematic Reviews (protocol no. CRD42024485371). The quality assessment of diagnostic accuracy studies-2 (QUADAS2) tool and the Must AI Criteria-10 (MAIC-10) checklist were used to assess quality and risk of bias. The meta-analysis included studies reporting DL for breast cancer diagnosis and their performance, from which pooled summary estimates for the area under the curve (AUC), sensitivity, and specificity were calculated. A total of 40 studies were included, of which only 21 were eligible for quantitative analysis. Convolutional neural networks (CNNs) were used in 62.5

Over the next 5 years, new breast cancer screening guidelines recommending magnetic resonance imaging (MRI) for certain patients will significantly increase the volume of imaging data to be analyzed. While this increase poses challenges for radiologists, artificial intelligence (AI) offers potential solutions to manage this workload. However, the development of AI models is often hindered by manual annotation requirements and strict data-sharing regulations between institutions. In this study, we present an integrated pipeline combining weakly supervised learning—reducing the need for detailed annotations—with local AI model training via swarm learning (SL), which circumvents centralized data sharing. We utilized three datasets comprising 1372 female bilateral breast MRI exams from institutions in three countries: the United States (US), Switzerland, and the United Kingdom (UK) to train models. These models were then validated on two external datasets consisting of 649 bilateral breast MRI exams from Germany and Greece. Upon systematically benchmarking various weakly supervised two-dimensional (2D) and three-dimensional (3D) deep learning (DL) methods, we find that the 3D-ResNet-101 demonstrates superior performance. By implementing a real-world SL setup across three international centers, we observe that these collaboratively trained models outperform those trained locally. Even with a smaller dataset, we demonstrate the practical feasibility of deploying SL internationally with on-site data processing, addressing challenges such as data privacy and annotation variability. Combining weakly supervised learning with SL enhances inter-institutional collaboration, improving the utility of distributed datasets for medical AI training without requiring detailed annotations or centralized data sharing. Breast cancer screening guidelines are expanding to include more MRI scans, increasing the amount of imaging data doctors must analyze. This study explored how artificial intelligence (AI) can help manage this increased workload while overcoming challenges such as limited data sharing between hospitals and the need for detailed annotations on each image. Researchers used MRI scans from five hospitals in the US, Switzerland, the UK, Germany, and Greece to train and test AI models. They found that a specific type of AI model performed the best, and that training AI collaboratively across hospitals improved results compared to training at individual sites. This approach could make AI tools more effective and secure for use in healthcare, potentially improving breast cancer detection and patient outcomes. Saldanha, Zhu et al. present an integrated pipeline combining weakly supervised learning with local artificial intelligence (AI) model training via swarm learning (SL) to circumvent a need for centralized data sharing. Deploying SL internationally with on-site data processing addresses challenges such as data privacy and annotation variability enabling AI training across international datasets while preserving data privacy.