Differentiation Between Phyllodes Tumor and Fibroadenoma of the Breast: A Radiomics Prediction Model Based on Full-Field Digital Mammography & Digital Tomosynthesis.

OBJECTIVE: To assess the diagnostic performance of FFDM-based and DBT-based radiomics models to differentiate breast phyllodes tumors from fibroadenomas. METHODS: 192 patients (93 phyllodes tumors and 99 fibroadenomas) who underwent mammography were retrospectively enrolled. Radiomic features were respectively extracted from FFDM and the clearest slice of DBT images. A least absolute shrinkage and selection operator (LASSO) regression was used to select radiomics features. A combined model was constructed by radiomics and radiological signatures. Machine learning classification was done using logistic regression based on radiomics or radiological signatures (clinical model). Four radiologists were tested on phyllodes tumors and fibroadenomas with and without optimal model assistance. The area under the receiver operating characteristic (ROC) curve (AUC) was computed to assess the performance of each model or radiologist. The Delong test and McNemar's test were performed to compare the performance. RESULTS: The combined model yielded the highest performance with an AUC of 0.948 (95%CI: 0.889-1.000) in the testing set, slightly higher than the FFDM-radiomics model (AUC of 0.937, 95%CI: 0.841-0.984) and the DBT-radiomics model (AUC of 0.860, 95%CI: 0.742-0.936) and significantly superior to the clinical model (AUC of 0.719, 95%CI: 0.585-0.829). With the combined model aid, the AUCs of four radiologists were improved from 0.808 to 0.914 (p=0.079), 0.759 to 0.888 (p=0.015), 0.717 to 0.846 (p=0.004), and 0.629 to 0.803 (p=0.001). CONCLUSION: Radiomics analysis based on FFDM and DBT shows promise in differentiating phyllodes tumors from fibroadenomas.

Enhancing Radiologist Reading Performance by Ordering Screening Mammograms Based on Characteristics That Promote Visual Adaptation.

Background Mammographic background characteristics may stimulate human visual adaptation, allowing radiologists to detect abnormalities more effectively. However, it is unclear whether density, or another image characteristic, drives visual adaptation. Purpose To investigate whether screening performance improves when screening mammography examinations are ordered for batch reading according to mammographic characteristics that may promote visual adaptation. Materials and Methods This retrospective multireader multicase study was performed with mammograms obtained between September 2016 and May 2019. The screening examinations, each consisting of four mammograms, were interpreted by 13 radiologists in three distinct orders: randomly, by increasing volumetric breast density (VBD), and based on a self-supervised learning (SSL) encoding (examinations automatically grouped as "looking similar"). An eye tracker recorded radiologists' eye movements during interpretation. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity of random-ordered readings were compared with those of VBD- and SSL-ordered readings using mixed-model analysis of variance. Reading time, fixation metrics, and perceived density were compared using Wilcoxon signed-rank tests. Results Mammography examinations (75 with breast cancer, 75 without breast cancer) from 150 women (median age, 55 years [IQR, 50-63]) were read. The examinations ordered by increasing VBD versus randomly had an increased AUC (0.93 [95% CI: 0.91, 0.96] vs 0.92 [95% CI: 0.89, 0.95]; P = .009), without evidence of a difference in specificity (89% [871 of 975] vs 86% [837 of 975], P = .04) and sensitivity (both 81% [794 of 975 vs 788 of 975], P = .78), and a reduced reading time (24.3 vs 27.9 seconds, P < .001), fixation count (47 vs 52, P < .001), and fixation time in malignant regions (3.7 vs 4.6 seconds, P < .001). For SSL-ordered readings, there was no evidence of differences in AUC (0.92 [95% CI: 0.89, 0.95]; P = .70), specificity (84% [820 of 975], P = .37), sensitivity (80% [784 of 975], P = .79), fixation count (54, P = .05), or fixation time in malignant regions (4.6 seconds, P > .99) compared with random-ordered readings. Reading times were significantly higher for SSL-ordered readings compared with random-ordered readings (28.4 seconds, P = .02). Conclusion Screening mammography examinations ordered from low to high VBD improved screening performance while reducing reading and fixation times. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Grimm in this issue.

Diagnostic Accuracy of Screening Contrast-enhanced Mammography for Women with Extremely Dense Breasts at Increased Risk of Breast Cancer.

Background Mammogram interpretation is challenging in female patients with extremely dense breasts (Breast Imaging Reporting and Data System [BI-RADS] category D), who have a higher breast cancer risk. Contrast-enhanced mammography (CEM) has recently emerged as a potential alternative; however, data regarding CEM utility in this subpopulation are limited. Purpose To evaluate the diagnostic performance of CEM for breast cancer screening in female patients with extremely dense breasts. Materials and Methods This retrospective single-institution study included consecutive CEM examinations in asymptomatic female patients with extremely dense breasts performed from December 2012 to March 2022. From CEM examinations, low-energy (LE) images were the equivalent of a two-dimensional full-field digital mammogram. Recombined images highlighting areas of contrast enhancement were constructed using a postprocessing algorithm. The sensitivity and specificity of LE images and CEM images (ie, including both LE and recombined images) were calculated and compared using the McNemar test. Results This study included 1299 screening CEM examinations (609 female patients; mean age, 50 years ± 9 [SD]). Sixteen screen-detected cancers were diagnosed, and two interval cancers occured. Five cancers were depicted at LE imaging and an additional 11 cancers were depicted at CEM (incremental cancer detection rate, 8.7 cancers per 1000 examinations). CEM sensitivity was 88.9% (16 of 18; 95% CI: 65.3, 98.6), which was higher than the LE examination sensitivity of 27.8% (five of 18; 95% CI: 9.7, 53.5) (P = .003). However, there was decreased CEM specificity (88.9%; 1108 of 1246; 95% CI: 87.0, 90.6) compared with LE imaging (specificity, 96.2%; 1199 of 1246; 95% CI: 95.0, 97.2) (P < .001). Compared with specificity at baseline, CEM specificity at follow-up improved to 90.7% (705 of 777; 95% CI: 88.5, 92.7; P = .01). Conclusion Compared with LE imaging, CEM showed higher sensitivity but lower specificity in female patients with extremely dense breasts, although specificity improved at follow-up. © RSNA, 2024 See also the editorial by Lobbes in this issue.

Artificial Intelligence Algorithm for Subclinical Breast Cancer Detection.

Importance: Early breast cancer detection is associated with lower morbidity and mortality. Objective: To examine whether a commercial artificial intelligence (AI) algorithm for breast cancer detection could estimate the development of future cancer. Design, Setting, and Participants: This retrospective cohort study of 116 495 women aged 50 to 69 years with no prior history of breast cancer before they underwent at least 3 consecutive biennial screening examinations used scores from an AI algorithm (INSIGHT MMG, version 1.1.7.2; Lunit Inc; used September 28, 2022, to April 5, 2023) for breast cancer detection and screening data from multiple, consecutive rounds of mammography performed from September 13, 2004, to December 21, 2018, at 9 breast centers in Norway. The statistical analyses were performed from September 2023 to August 2024. Exposure: Artificial intelligence algorithm score indicating suspicion for the presence of breast cancer. The algorithm provided a continuous cancer detection score for each examination ranging from 0 to 100, with increasing values indicating a higher likelihood of cancer being present on the current mammogram. Main Outcomes and Measures: Maximum AI algorithm score for cancer detection and absolute difference in score among breasts of women developing screening-detected cancer, women with interval cancer, and women who screened negative. Results: The mean (SD) age at the first study round was 58.5 (4.5) years for 1265 women with screening-detected cancer in the third round, 57.4 (4.6) years for 342 women with interval cancer after 3 negative screening rounds, and 56.4 (4.9) years for 116 495 women without breast cancer all 3 screening rounds. The mean (SD) absolute differences in AI scores among breasts of women developing screening-detected cancer were 21.3 (28.1) at the first study round, 30.7 (32.5) at the second study round, and 79.0 (28.9) at the third study round. The mean (SD) differences prior to interval cancer were 19.7 (27.0) at the first study round, 21.0 (27.7) at the second study round, and 34.0 (33.6) at the third study round. The mean (SD) differences among women who did not develop breast cancer were 9.9 (17.5) at the first study round, 9.6 (17.4) at the second study round, and 9.3 (17.3) at the third study round. Areas under the receiver operating characteristic curve for the absolute difference were 0.63 (95% CI, 0.61-0.65) at the first study round, 0.72 (95% CI, 0.71-0.74) at the second study round, and 0.96 (95% CI, 0.95-0.96) at the third study round for screening-detected cancer and 0.64 (95% CI, 0.61-0.67) at the first study round, 0.65 (95% CI, 0.62-0.68) at the second study round, and 0.77 (95% CI, 0.74-0.79) at the third study round for interval cancers. Conclusions and Relevance: In this retrospective cohort study of women undergoing screening mammography, mean absolute AI scores were higher for breasts developing vs not developing cancer 4 to 6 years before their eventual detection. These findings suggest that commercial AI algorithms developed for breast cancer detection may identify women at high risk of a future breast cancer, offering a pathway for personalized screening approaches that can lead to earlier cancer diagnosis.

Breast tumor segmentation using neural cellular automata and shape guided segmentation in mammography images.

PURPOSE: Using computer-aided design (CAD) systems, this research endeavors to enhance breast cancer segmentation by addressing data insufficiency and data complexity during model training. As perceived by computer vision models, the inherent symmetry and complexity of mammography images make segmentation difficult. The objective is to optimize the precision and effectiveness of medical imaging. METHODS: The study introduces a hybrid strategy combining shape-guided segmentation (SGS) and M3D-neural cellular automata (M3D-NCA), resulting in improved computational efficiency and performance. The implementation of Shape-guided segmentation (SGS) during the initialization phase, coupled with the elimination of convolutional layers, enables the model to effectively reduce computation time. The research proposes a novel loss function that combines segmentation losses from both components for effective training. RESULTS: The robust technique provided aims to improve the accuracy and consistency of breast tumor segmentation, leading to significant improvements in medical imaging and breast cancer detection and treatment. CONCLUSION: This study enhances breast cancer segmentation in medical imaging using CAD systems. Combining shape-guided segmentation (SGS) and M3D-neural cellular automata (M3D-NCA) is a hybrid approach that improves performance and computational efficiency by dealing with complex data and not having enough training data. The approach also reduces computing time and improves training efficiency. The study aims to improve breast cancer detection and treatment methods in medical imaging technology.

Reproductive factors and mammographic density within the International Consortium of Mammographic Density: A cross-sectional study.

BACKGROUND: Elevated mammographic density (MD) for a woman's age and body mass index (BMI) is an established breast cancer risk factor. The relationship of parity, age at first birth, and breastfeeding with MD is less clear. We examined the associations of these factors with MD within the International Consortium of Mammographic Density (ICMD). METHODS: ICMD is a consortium of 27 studies with pooled individual-level epidemiological and MD data from 11,755 women without breast cancer aged 35-85 years from 22 countries, capturing 40 country-& ethnicity-specific population groups. MD was measured using the area-based tool Cumulus. Meta-analyses across population groups and pooled analyses were used to examine linear regression associations of square-root (√) transformed MD measures (percent MD (PMD), dense area (DA), and non-dense area (NDA)) with parity, age at first birth, ever/never breastfed and lifetime breastfeeding duration. Models were adjusted for age at mammogram, age at menarche, BMI, menopausal status, use of hormone replacement therapy, calibration method, mammogram view and reader, and parity and age at first birth when not the association of interest. RESULTS: Among 10,988 women included in these analyses, 90.1% (n = 9,895) were parous, of whom 13% (n = 1,286) had ≥ five births. The mean age at first birth was 24.3 years (Standard deviation = 5.1). Increasing parity (per birth) was inversely associated with √PMD (ß: - 0.05, 95% confidence interval (CI): - 0.07, - 0.03) and √DA (ß: - 0.08, 95% CI: - 0.12, - 0.05) with this trend evident until at least nine births. Women who were older at first birth (per five-year increase) had higher √PMD (ß:0.06, 95% CI:0.03, 0.10) and √DA (ß:0.06, 95% CI:0.02, 0.10), and lower √NDA (ß: - 0.06, 95% CI: - 0.11, - 0.01). In stratified analyses, this association was only evident in women who were post-menopausal at MD assessment. Among parous women, no associations were found between ever/never breastfed or lifetime breastfeeding duration (per six-month increase) and √MD. CONCLUSIONS: Associations with higher parity and older age at first birth with √MD were consistent with the direction of their respective associations with breast cancer risk. Further research is needed to understand reproductive factor-related differences in the composition of breast tissue and their associations with breast cancer risk.

Acceptability of de-intensified screening for women at low risk of breast cancer: a randomised online experimental survey.

BACKGROUND: Risk-stratified approaches to breast screening show promise for increasing benefits and reducing harms. But the successful implementation of such an approach will rely on public acceptability. To date, research suggests that while increased screening for women at high risk will be acceptable, any de-intensification of screening for low-risk groups may be met with less enthusiasm. We report findings from a population-based survey of women in England, approaching the age of eligibility for breast screening, to compare the acceptability of current age-based screening with two hypothetical risk-adapted approaches for women at low risk of breast cancer. METHODS: An online survey of 1,579 women aged 40-49 with no personal experience of breast cancer or mammography. Participants were recruited via a market research panel, using target quotas for educational attainment and ethnic group, and were randomised to view information about (1) standard NHS age-based screening; (2) a later screening start age for low-risk women; or (3) a longer screening interval for low-risk women. Primary outcomes were cognitive, emotional, and global acceptability. ANOVAs and multiple regression were used to compare acceptability between groups and explore demographic and psychosocial factors associated with acceptability. RESULTS: All three screening approaches were judged to be acceptable on the single-item measure of global acceptability (mean score > 3 on a 5-point scale). Scores for all three measures of acceptability were significantly lower for the risk-adapted scenarios than for age-based screening. There were no differences between the two risk-adapted scenarios. In multivariable analysis, higher breast cancer knowledge was positively associated with cognitive and emotional acceptability of screening approach. Willingness to undergo personal risk assessment was not associated with experimental group. CONCLUSION: We found no difference in the acceptability of later start age vs. longer screening intervals for women at low risk of breast cancer in a large sample of women who were screening naïve. Although acceptability of both risk-adapted scenarios was lower than for standard age-based screening, overall acceptability was reasonable. The positive associations between knowledge and both cognitive and emotional acceptability suggests clear and reassuring communication about the rationale for de-intensified screening may enhance acceptability.

Breast Cancers Detected during a Decade of Screening with Digital Breast Tomosynthesis: Comparison with Digital Mammography.

Background Digital breast tomosynthesis (DBT) has been shown to help increase cancer detection compared with two-dimensional digital mammography (DM). However, it is unclear whether additional tumor detection will improve outcomes or lead to overdiagnosis of breast cancer. Purpose This study aimed to compare cancer types and stages over 3 years of DM screening and 10 years of DBT screening to determine the effect of DBT. Materials and Methods A retrospective search identified breast cancers detected by using screening mammography from August 2008 through July 2021. Data collected included demographic, imaging, and pathologic information. Invasive cancers 2 cm or larger, human epidermal growth factor 2-positive or triple-negative tumors greater than 10 mm, axillary nodes positive for cancer, and distant organ spread were considered advanced cancers. The DBT and DM cohorts were compared and further analyzed by prevalent versus incident examinations. False-negative findings were also assessed. Results A total of 1407 breast cancers were analyzed (142 with DM, 1265 with DBT). DBT showed a higher rate of cancer depiction than DM (5.3 vs four cancers per 1000, respectively; P = .001), with a similar ratio of invasive cancers to ductal carcinomas in situ (76.5%:23.5% [968 and 297 of 1265, respectively] vs 71.1%:28.9% [101 and 41 of 142, respectively]). Mean invasive cancer size did not differ between DM and DBT (1.44 cm ± 0.93 [SD] vs 1.36 cm ± 1.14, respectively; P = .49), but incident DBT cases were smaller than prevalent cases (1.2 cm ± 1.0 vs 1.6 cm ± 1.4, respectively; P < .001). DBT and DM had similar rates of invasive cancer subtypes: low grade (26.5% [243 of 912] vs 29% [28 of 96], respectively), moderate grade (57.2% [522 of 912] vs 51% [49 of 96], respectively), and high grade (16.1% [147 of 912] vs 20% [19 of 96], respectively) (P = .65). The proportion of advanced cancers was lower with DBT than DM (32.6% [316 of 968] vs 43.6% [44 of 101], respectively; P = .04) and between DBT prevalent and incident screening (39.1% [133 of 340] vs 29.1% [183 of 628], respectively; P = .003). There was no difference in interval cancer rates (0.14 per 1000 with DM and 0.2 per 1000 with DBT; P = .42) for both groups. Conclusion DBT helped to increase breast cancer detection rate and depicted invasive cancers with a lower rate of advanced cancers compared with DM, with further improvement observed at incident rounds of screening. © RSNA, 2024 See also the editorial by Kim and Woo in this issue.

WHO-recommended levels of physical activity in relation to mammographic breast density, mammographic tumor appearance, and mode of detection of breast cancer.

BACKGROUND: Despite known benefits of physical activity in reducing breast cancer risk, its impact on mammographic characteristics remain unclear and understudied. This study aimed to investigate associations between pre-diagnostic physical activity and mammographic features at breast cancer diagnosis, specifically mammographic breast density (MBD) and mammographic tumor appearance (MA), as well as mode of cancer detection (MoD). METHODS: Physical activity levels from study baseline (1991-1996) and mammographic information from the time of invasive breast cancer diagnosis (1991-2014) of 1116 women enrolled in the Malmö Diet and Cancer Study cohort were used. Duration and intensity of physical activity were assessed according to metabolic equivalent of task hours (MET-h) per week, or World Health Organization (WHO) guideline recommendations. MBD was dichotomized into low-moderate or high, MA into spiculated or non-spiculated tumors, and MoD into clinical or screening detection. Associations were investigated through logistic regression analyses providing odds ratios (OR) with 95% confidence intervals (CI) in crude and multivariable-adjusted models. RESULTS: In total, 32% of participants had high MBD at diagnosis, 37% had non-spiculated MA and 50% had clinical MoD. Overall, no association between physical activity and MBD was found with increasing MET-h/week or when comparing women who exceeded WHO guidelines to those subceeding recommendations (ORadj 1.24, 95% CI 0.78-1.98). Likewise, no differences in MA or MoD were observed across categories of physical activity. CONCLUSIONS: No associations were observed between pre-diagnostic physical activity and MBD, MA, or MoD at breast cancer diagnosis. While physical activity is an established breast cancer prevention strategy, it does not appear to modify mammographic characteristics or screening detection.

Implications of breast density for breast cancer screening.

ABSTRACT: Extremely dense breasts can be an independent risk factor for breast cancer. A new FDA rule requires that patients be notified of their breast density and the possible benefits of additional imaging to screen for breast cancer. Clinicians should be cognizant of the data about breast cancer risk, breast density, and recommendations to change screening techniques if patients, particularly premenopausal females, have extremely dense breasts but no other known risk factors.

Meta-analysis and systematic review of the diagnostic value of contrast-enhanced spectral mammography for the detection of breast cancer.

OBJECTIVE: The objective is to evaluate the diagnostic effectiveness of contrast-enhanced spectral mammography (CESM) in the diagnosis of breast cancer. DESIGN: DATA SOURCES: PubMed, Embase and Cochrane libraries up to 18 June 2022. ELIGIBILITY CRITERIA FOR SELECTING STUDIES: We included trials studies, compared the results of different researchers on CESM in the diagnosis of breast cancer, and calculated the diagnostic value of CESM for breast cancer. DATA EXTRACTION AND SYNTHESIS: Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) evaluated the methodological quality of all the included studies. The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses specification. In addition to sensitivity and specificity, other important parameters were explored in an analysis of CESM accuracy for breast cancer diagnosis. For overall accuracy estimation, summary receiver operating characteristic curves were calculated. STATA V.14.0 was used for all analyses. RESULTS: This meta-analysis included a total of 12 studies. According to the summary estimates for CESM in the diagnosis of breast cancer, the pooled sensitivity and specificity were 0.97 (95% CI 0.92 to 0.98) and 0.76 (95% CI 0.64 to 0.85), respectively. Positive likelihood ratio was 4.03 (95% CI 2.65 to 6.11), negative likelihood ratio was 0.05 (95% CI 0.02 to 0.09) and the diagnostic odds ratio was 89.49 (95% CI 45.78 to 174.92). Moreover, there was a 0.95 area under the curve. CONCLUSIONS: The CESM has high sensitivity and good specificity when it comes to evaluating breast cancer, particularly in women with dense breasts. Thus, provide more information for clinical diagnosis and treatment.

Evaluating the Impact of a Clinical Breast Examination Screening Program within the IraPEN Project.

PURPOSE: Breast cancer is a prevalent global cancer and a leading cause of mortality in developed countries. In 2015, Iran introduced the Package of Essential Noncommunicable Diseases (IraPEN) as a pilot project to tackle prevalent noncommunicable diseases, including breast cancer. However, there is limited research evaluating the implementation, costs, and outcomes of breast cancer screening within IraPEN. Therefore, this study aims to investigate the costs and outcomes of the clinical breast examination screening program in Isfahan from 2017 to 2020. METHOD: This descriptive cost-outcome study utilized data from 450,876 individuals aged 30 to 69 who participated in clinical breast examination screening. The outcomes assessed in this program encompassed the number of participants, the number of individuals identified with symptoms, referrals to the next level of examination, the number of individuals undergoing mammography, recorded mammography results, and the number of cases of breast cancer identified. Direct costs were estimated, including personnel, infrastructure, equipment, and other related expenses. RESULTS: |The findings revealed that the direct costs of the breast cancer screening program in Isfahan between 2017 and 2020 were 310,514,608,558 Rials, equivalent to approximately 15,470,633 PPP$. These expenses led to the identification of 134,508 individuals with symptoms, referrals of 258,599 individuals to the subsequent level of examination, and approximately 55,974 individuals undergoing mammography tests. CONCLUSION: This study demonstrates that the breast cancer screening program provides a significant number of women in the target age group with breast self-examination education while raising public awareness about this disease.

Detection of breast cancer in digital breast tomosynthesis with vision transformers.

Digital Breast Tomosynthesis (DBT) has revolutionized more traditional breast imaging through its three-dimensional (3D) visualization capability that significantly enhances lesion discernibility, reduces tissue overlap, and improves diagnostic precision as compared to conventional two-dimensional (2D) mammography. In this study, we propose an advanced Computer-Aided Detection (CAD) system that harnesses the power of vision transformers to augment DBT's diagnostic efficiency. This scheme uses a neural network to glean attributes from the 2D slices of DBT followed by post-processing that considers features from neighboring slices to categorize the entire 3D scan. By leveraging a transfer learning technique, we trained and validated our CAD framework on a unique dataset consisting of 3,831 DBT scans and subsequently tested it on 685 scans. Of the architectures tested, the Swin Transformer outperformed the ResNet101 and vanilla Vision Transformer. It achieved an impressive AUC score of 0.934 ± 0.026 at a resolution of 384 × 384. Increasing the image resolution from 224 to 384 not only maintained vital image attributes but also led to a marked improvement in performance (p-value = 0.0003). The Mean Teacher algorithm, a semi-supervised method using both labeled and unlabeled DBT slices, showed no significant improvement over the supervised approach. Comprehensive analyses across different lesion types, sizes, and patient ages revealed consistent performance. The integration of attention mechanisms yielded a visual narrative of the model's decision-making process that highlighted the prioritized regions during assessments. These findings should significantly propel the methodologies employed in DBT image analysis by setting a new benchmark for breast cancer diagnostic precision.

AI-integrated Screening to Replace Double Reading of Mammograms: A Population-wide Accuracy and Feasibility Study.

Mammography screening supported by deep learning-based artificial intelligence (AI) solutions can potentially reduce workload without compromising breast cancer detection accuracy, but the site of deployment in the workflow might be crucial. This retrospective study compared three simulated AI-integrated screening scenarios with standard double reading with arbitration in a sample of 249 402 mammograms from a representative screening population. A commercial AI system replaced the first reader (scenario 1: integrated AIfirst), the second reader (scenario 2: integrated AIsecond), or both readers for triaging of low- and high-risk cases (scenario 3: integrated AItriage). AI threshold values were chosen based partly on previous validation and setting the screen-read volume reduction at approximately 50% across scenarios. Detection accuracy measures were calculated. Compared with standard double reading, integrated AIfirst showed no evidence of a difference in accuracy metrics except for a higher arbitration rate (+0.99%, P < .001). Integrated AIsecond had lower sensitivity (-1.58%, P < .001), negative predictive value (NPV) (-0.01%, P < .001), and recall rate (-0.06%, P = .04) but a higher positive predictive value (PPV) (+0.03%, P < .001) and arbitration rate (+1.22%, P < .001). Integrated AItriage achieved higher sensitivity (+1.33%, P < .001), PPV (+0.36%, P = .03), and NPV (+0.01%, P < .001) but lower arbitration rate (-0.88%, P < .001). Replacing one or both readers with AI seems feasible; however, the site of application in the workflow can have clinically relevant effects on accuracy and workload. Keywords: Mammography, Breast, Neoplasms-Primary, Screening, Epidemiology, Diagnosis, Convolutional Neural Network (CNN) Supplemental material is available for this article. Published under a CC BY 4.0 license.