Digital Mammography Screening Research Articles

Abstract PO2-28-09: Long-term Breast Cancer Risk Prediction in Black Women: External Validation of a Mammography-driven AI Model

Abstract Despite the demonstrated potential of artificial intelligence (AI) in breast cancer risk assessment for personalizing screening recommendations, further validation is required regarding AI model bias and generalizability. We performed external validation in Black women of a mammography-driven AI breast cancer risk model (Mirai) originally developed on screening cohorts primarily consisting of White women. In this institutional review board-approved, Health Insurance Portability and Protection Act (HIPAA)-compliant study under a waiver of consent, we retrospectively analyzed a case–cohort sample nested within the core academic breast cancer screening practice of BJC Healthcare, the hospital partner of Washington University in St. Louis. For the purposes of this validation study, relying on 2D digital mammography (DM) images, we focused on Black women presenting for annual DM screening (Selenia or Selenia Dimensions; Hologic) between 2008 and 2018. Eligible breast cancer cases were derived from all women with a breast cancer diagnosis (with associated biopsy-confirmed tumor pathology via institutional cancer registry) after negative (BI-RADS 1 or 2) DM screening 1 to 5 years prior to cancer diagnosis. We also identified a random sample of controls, defined as women who had negative (BI-RADS 1 or 2) DM screening, with 1 to 5 years of screening follow-up without a cancer diagnosis. Risk scores for all DM exams were calculated via the Mirai model. Performance was evaluated using concordance-index (C) analyses and associated 95% confidence intervals (CIs) for the entire cohort, as well as for study subgroups of invasive versus in-situ cancer and cancer molecular subtypes. We analyzed 1368 DM screening exams, including 672 DM exams from 391 women diagnosed with breast cancer (mean age, 58 years; standard deviation, 10 years) and 696 DM exams from 406 controls (mean age, 55 years; standard deviation, 10 years). The overall C-index was 0.62 [95% CIs 0.60–0.64] for all Black women, which was lower compared to previously reported validation results for Mirai in studies of similar design on predominantly White and racially diverse screening cohorts (C-index = 0.67-0.78). There was no evidence of a significant difference between invasive and in-situ cancer (C-index = 0.64 [95% CIs 0.61–0.66] vs. 0.64 [95% CIs 0.61–0.67]). Compared to other cancer molecular subtypes, performance was significantly higher among triple-negative (C-index = 0.67 [95% CIs 0.62–0.71]) and estrogen receptor (ER) negative cancer (C-index = 0.66 [95% CIs 0.61–0.70]). A previously developed mammography-driven AI model showed overall good performance in long-term breast cancer risk assessment in a dataset of Black women only, particularly for triple-negative and ER- cancer types. However, performance was lower compared to previously reported validation results from similar studies on predominantly White and racially diverse screening cohorts. Our results suggest that further refinements are needed towards more accurate breast cancer risk assessment in Black women. Long-term breast cancer risk prediction performance (C-indices) of Mirai in the full cohort of Black women and in study subgroups. Citation Format: Juanita Hernandez Lopez, Zhixin Sun, Shirin Shoushtari, Ulugbek Kamilov, Debbie Bennett, Aimilia Gastounioti. Long-term Breast Cancer Risk Prediction in Black Women: External Validation of a Mammography-driven AI Model [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-28-09.

Screening mammography performance according to breast density: a comparison between radiologists versus standalone intelligence detection

BackgroundArtificial intelligence (AI) algorithms for the independent assessment of screening mammograms have not been well established in a large screening cohort of Asian women. We compared the performance of screening digital mammography considering breast density, between radiologists and AI standalone detection among Korean women.MethodsWe retrospectively included 89,855 Korean women who underwent their initial screening digital mammography from 2009 to 2020. Breast cancer within 12 months of the screening mammography was the reference standard, according to the National Cancer Registry. Lunit software was used to determine the probability of malignancy scores, with a cutoff of 10% for breast cancer detection. The AI’s performance was compared with that of the final Breast Imaging Reporting and Data System category, as recorded by breast radiologists. Breast density was classified into four categories (A–D) based on the radiologist and AI-based assessments. The performance metrics (cancer detection rate [CDR], sensitivity, specificity, positive predictive value [PPV], recall rate, and area under the receiver operating characteristic curve [AUC]) were compared across breast density categories.ResultsMean participant age was 43.5 ± 8.7 years; 143 breast cancer cases were identified within 12 months. The CDRs (1.1/1000 examination) and sensitivity values showed no significant differences between radiologist and AI-based results (69.9% [95% confidence interval [CI], 61.7–77.3] vs. 67.1% [95% CI, 58.8–74.8]). However, the AI algorithm showed better specificity (93.0% [95% CI, 92.9–93.2] vs. 77.6% [95% CI, 61.7–77.9]), PPV (1.5% [95% CI, 1.2–1.9] vs. 0.5% [95% CI, 0.4–0.6]), recall rate (7.1% [95% CI, 6.9–7.2] vs. 22.5% [95% CI, 22.2–22.7]), and AUC values (0.8 [95% CI, 0.76–0.84] vs. 0.74 [95% CI, 0.7–0.78]) (all P < 0.05). Radiologist and AI-based results showed the best performance in the non-dense category; the CDR and sensitivity were higher for radiologists in the heterogeneously dense category (P = 0.059). However, the specificity, PPV, and recall rate consistently favored AI-based results across all categories, including the extremely dense category.ConclusionsAI-based software showed slightly lower sensitivity, although the difference was not statistically significant. However, it outperformed radiologists in recall rate, specificity, PPV, and AUC, with disparities most prominent in extremely dense breast tissue.

Abstract 6619: Dynamic prediction with repeated mammogram images improves 5-year breast cancer risk performance

Abstract Background: Increasing use of digital screening mammography opens potential to add individualized dynamic measures of breast image history to predict future breast cancer risk. Current prediction models use a single mammogram measure, usually breast density, at one point in time, to add to prediction models based on established risk factors. Our previous work has shown that change in each breast over time is related to risk of diagnosis with breast cancer. Methods: We apply a dynamic prediction model that includes full field digital mammogram in each breast to predict future 5-year risk. Data from Joanne Knight Breast Health Cohort (JKBHC) serve as training population with 290 pathology-confirmed breast cancer cases and 658 controls sampled with a total number of 8,712 craniocaudal (CC) view mammograms for analysis. Nurses’ Health Study 2 (NHS2) serves as the external validation with 188 cases and 462 controls sampled with a total number of 3,712 CC-views. The dynamic risk prediction model aims to estimate the probability of future breast cancer risk conditional on the past history of the longitudinal mammograms. We train the model on the JKBH cohort, retaining 20% for testing and 15% for validation. We report a time dependent AUC when evaluating model performance. Results: The mean age of women was 56.6 years at entry to cohort; 81.3% were white. Adding trajectory of mammogram images in each breast to a model improved performance. The mean number of mammograms conditional on 5 years of observation in JKBHC was 3.36 (1.49) and in NHS2 2.36 (1.53). We obtained a 5-year dynamic AUC of 0.81 (0.05) in the JKBHC internal validation and 0.77 (0.06) in the NHS2 external validation. Model performance was comparable in pre- and postmenopausal women. Because breasts with high mammographic density hinder detection of tumors we stratify analyses by dense and non-dense at baseline. Among women with dense breasts at baseline (BIRADS C and D), we observe a dynamic 5-year risk AUC = 0.77. In the subgroup with less dense breasts (BIRADS A and B) we observe AUC = 0.74. Conclusions: With this approach we can accommodate the irregularly observed time-varying images and project only partial information prior to doing the prediction. Adding trajectory of mammogram history in each breast improves 5-year breast cancer risk prediction beyond models used in clinic today. Citation Format: Graham A. Colditz, Rulla M. Tamimi, Bernard Rosner, Shu Jiang. Dynamic prediction with repeated mammogram images improves 5-year breast cancer risk performance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6619.

Abstract 3449: Is AI-enhanced breast ultrasound ready for breast cancer screening in low-resource environments? A systematic review

Abstract Purpose. Screening mammography is unavailable in many low-resource areas. We ask if the state-of-the-art in artificial intelligence (AI)-enhanced breast ultrasound (BUS) is sufficiently accurate to be used for primary breast cancer screening in low-resource regions. Background. Since the 1980s, high-income countries have implemented mammographic screening programs, leading to breast cancer mortality reduction in screened women.1 Mammography is unavailable in many low-resource regions, such as the USAPI. Furthermore, travel difficulties and lack of radiologists hinder implementation. AI combined with portable BUS may address limitations of the high-income paradigm. In this systematic review, we ask if AI-enhanced BUS can detect/segment lesions (Objective 1) and classify lesions as cancerous (Objective 2). Methods. Two reviewers independently assessed articles from 1/1/2016 to 8/6/2023 from PubMed, Google Scholar, and citation searching. Studies which report on AI development and report performance on a patient-wise, held-out test set met the inclusion criteria. Studies were characterized by AI task and clinical application time. Dataset composition and performance were examined via narrative data synthesis. QUADAS-2 bias assessment was performed using criteria for each AI task. Success in (2) is defined by meeting minimum screening performance guidelines.2,3 Results. PubMed yielded 281 studies, Google Scholar yielded 225 studies, and a manual citation search yielded 41 studies. From 382 unique full texts evaluated, 52 articles met all inclusion criteria: 3 frame selection, 2 real-time detection, 2 combination, 14 segmentation-only, and 31 classification-only. Lesion segmentation-only models achieved a 90th percentile Dice similarity coefficient of 0.913 on generally small datasets. The best evidence for lesion cancer classification reported 0.976 area under the curve. All studies faced elevated bias and applicability concerns under QUADAS-2. Conclusion. Reported performance for (1) is insufficient to introduce AI-enhanced BUS for breast cancer screening. Evidence supporting AI-enhanced BUS for (2) is dependent on few studies relying on internal datasets, limiting generalizability. Geographically diverse clinical trials are needed to confirm and improve robustness of performance of AI-enhanced BUS for (1) and (2). References. 1. Marmot MG, et al. The benefits and harms of breast cancer screening: an independent review. British journal of cancer. 2013;108(11):2205-2240. 2. Lehman CD, et al. National Performance Benchmarks for Modern Screening Digital Mammography: Update from the Breast Cancer Surveillance Consortium. Radiology. 2017-04-01 2017;283(1):49-58. doi:10.1148/radiol.2016161174 3. Rosenberg RD, et al. Performance Benchmarks for Screening Mammography. Radiology. 2006-10-01 2006;241(1):55-66. doi:10.1148/radiol.2411051504 Citation Format: Arianna Bunnell, Dustin Valdez, Fredrik Strand, Yannik Glaser, Peter Sadowski, John A. Shepherd. Is AI-enhanced breast ultrasound ready for breast cancer screening in low-resource environments? A systematic review [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3449.

The risk-based breast screening (RIBBS) study protocol: a personalized screening model for young women.

The optimal mammography screening strategy for women aged 45-49years is a matter of debate. We present the RIBBS study protocol, a quasi-experimental, prospective, population-based study comparing a risk- and breast density-stratified screening model (interventional cohort) with annual digital mammography (DM) screening (observational control cohort) in a real-world setting. The interventional cohort consists of 10,269 women aged 45years enrolled between 2020 and 2021 from two provinces of the Veneto Region(northen Italy). At baseline, participants underwent two-view digital breast tomosynthesis (DBT) and completed the Tyrer-Cuzick risk prediction model. Volumetric breast density (VBD) was calculated from DBT and the lifetime risk (LTR) was estimated by including VBD among the risk factors. Based on VBD and LTR, women were classified into five subgroups with specific screening protocols for subsequent screening rounds: (1) LTR ≤ 17% and nondense breast: biennial DBT; (2) LTR ≤ 17% and dense breast: biennial DBT and ultrasound; (3) LTR 17-30% or LTR > 30% without family history of BC, and nondense breast: annual DBT; (4) LTR 17-30% or > 30% without family history of BC, and dense breast: annual DBT and ultrasound; and (5) LTR > 30% and family history of BC: annual DBT and breast MRI. The interventional cohort is still ongoing. An observational, nonequivalent control cohort of 43,000 women aged 45years participating in an annual DM screening programme was recruited in three provinces of the neighbouring Emilia-Romagna Region. Cumulative incidence rates of advanced BC at three, five, and ten years between the two cohorts will be compared, adjusting for the incidence difference at baseline.Trial registration This study is registered on Clinicaltrials.gov (NCT05675085).

Assessment of salivary cadmium levels and breast density in the Marin Women's Study.

We aimed to determine if salivary cadmium (Cd) levels had any association with breast density, hoping to establish a less invasive cost-effective method of stratifying Cd burden as an environmental breast cancer risk factor. Salivary Cd levels were quantified from the Marin Women's Study, a Marin County, California population composite. Volumetric compositional breast density (BDsxa ) data were measured by single x-ray absorptiometry techniques. Digital screening mammography was performed by the San Francisco Mammography Registry. Radiologists reviewed mammograms and assigned a Breast Imaging-Reporting and Data System score. Early morning salivary Cd samples were assayed. Association analyses were then performed. Cd was quantifiable in over 90% of saliva samples (mean = 55.7 pg/L, SD = 29). Women with higher saliva Cd levels had a non-significant odds ratio of 1.34 with BI-RAD scores (3 or 4) (95% CI 0.75-2.39, p = 0.329). Cd levels were higher in current smokers (mean = 61.4 pg/L, SD = 34.8) than former smokers or non-smokers. These results were non-significant. Pilot data revealed that higher age and higher BMI were associated with higher BI-RAD scores (p < 0.001). Salivary Cd is a viable quantification source in large epidemiologic studies. Association analyses between Cd levels and breast density may provide additional information for breast cancer risk assessment, risk reduction plans, and future research directions. Further work is needed to demonstrate a more robust testing protocol before the extent of its usefulness can be established.

Eye radiation dose from breast cancer screening using full field digital mammography and digital breast tomosynthesis: A phantom study

IntroductionThe eye lens is recently classified as one of the most radiosensitive tissues by the International Commission on Radiological Protection (ICRP), and it has been suggested that the eye lens receives radiation dose during mammography due to scatter radiation. The aim of this study was to investigate the radiation dose received by the lens of the clients’ eye from Full Field Digital Mammography (FFDM) and Digital Breast Tomosynthesis (DBT) screening. MethodsThe eye radiation dose received by ATOM dosimetry phantom was estimated with thermo-luminescent dosemeters (TLDs). One TLD was utilised for each eye. A breast phantom was exposed for four-view screening mammography using 16 FFDM machines and one DBT machine. The breast phantom was exposed three times for each mammographic position and an average TLD dose reading was considered to minimise random error. ResultsFor four-view FFDM screening, the phantom eye radiation dose ranges from 0.013 mGy to 0.029 mGy with a mean±sd of 0.019 ± 0.005 mGy. A higher eye radiation dose of 0.041 mGy was recorded from four-view DBT screening. The statistical analysis demonstrated that the eye lens radiation dose is strongly and significantly correlated to breast organ dose and X-ray tube voltage. ConclusionThe phantom eye lens was exposed to scatter radiation from FFDM and DBT screening. The measured dose via the four-view DBT screening was higher than the four-view FFDM screening, but sits below the internationally acceptable ranges. If the findings of our paper hold true in practice, then the risk to the lens of the eyes for women attending breast screening is acceptable. Implications for practiceThe new lens radiation dose levels recommended by the ICRP necessitate the reevaluation of eye radiation dose from different radiographic examinations, especially those used for screening purpose where healthy individuals involved.

Trends in patient dose and compression force for digital (DR) mammography systems over an eleven-year period.

This study evaluated trends in patient dose and compression force for screening digital (DR) mammography systems. The results of five audits (carried out in 2011, 2014, 2018, 2020 and 2022) were compared. For every audit, anonymised screening examinations from each system consisting of the standard craniocaudal (CC) and mediolateral oblique (MLO) views of both breasts were analysed. Exposure parameters were extracted from the Digital Imaging and Communications in Medicine (DICOM) header and the mean glandular dose (MGD) for each image was calculated. Trends in the distribution of MGD, compressed breast thickness, compression force and compression force per radiographer were investigated. The mean MGD per image (and mean compressed breast thickness) was 1.20 mGy (58mm), 1.53 mGy (59mm), 1.83 mGy (61mm), 1.94 mGy (60mm) and 2.11 mGy (61mm) for 2011, 2014, 2018, 2020 and 2022 respectively. The mean (and standard deviation) compression force was 114 (32) N, 112 (29) N, 108 (27) N, 104 (24) N and 100 (23) N for 2011, 2014, 2018, 2020 and 2022 respectively. The mean MGD per image has increased over time but remains below internationally established Diagnostic Reference Levels (DRLs). This increase is primarily due to a change in the distribution of the different manufacturers and digital detector technologies, rather than an increase in the dose of the individual systems over time. The mean compression force has decreased over time in response to client feedback surveys. The standard deviation has also reduced, indicating more consistent application of force.

Discordant and false-negative interpretations at digital breast tomosynthesis in the prospective Oslo Tomosynthesis Screening Trial (OTST) using independent double reading.

To analyze discordant and false-negatives of double reading digital breast tomosynthesis (DBT) versus digital mammography (DM) including reading times in the Oslo Tomosynthesis Screening Trial (OTST), and reclassify these in a retrospective reader study as missed, minimal sign, or true-negatives. The prospective OTST comparing double reading DBT vs. DM had paired design with four parallel arms: DM, DM + computer aided detection, DBT + DM, and DBT + synthetic mammography. Eight radiologists interpreted images in batches using a 5-point scale. Reading time was automatically recorded. A retrospective reader study including four radiologists classified screen-detected cancers with at least one false-negative score and screening examinations of interval cancers as negative, non-specific minimal sign, significant minimal sign, and missed; the two latter groups are defined "actionable." Statistics included chi-square, Fisher's exact, McNemar's, and Mann-Whitney U tests. Discordant rate (cancer missed by one reader) for screen-detected cancers was overall comparable (DBT (31% [71/227]) and DM (30% [52/175]), p= .81), significantly lower at DBT for spiculated cancers (DBT, 19% [20/106] vs. DM, 36% [38/106], p= .003), but high (28/49 = 57%, p= 0.001) for DBT-only detected spiculated cancers. Reading time and sensitivity varied among readers. False-negative DBT-only detected spiculated cancers had shorter reading time than true-negatives in 46% (13/28). Retrospective evaluation classified the following DBT exams "actionable": three missed by both readers, 95% (39/41) of discordant cancers detected by both modes, all 30 discordant DBT-only cancers, 25% (13/51) of interval cancers. Discordant rate was overall comparable for DBT and DM, significantly lower at DBT for spiculated cancers, but high for DBT-only detected spiculated lesions. Most false-negative screen-detected DBT were classified as "actionable." Retrospective evaluation of false-negative interpretations from the Oslo Tomosynthesis Screening Trial shows that most discordant and several interval cancers could have been detected at screening. This underlines the potential for modern AI-based reading aids and triage, as high-volume screening is a demanding task. • Digital breast tomosynthesis (DBT) screening is more sensitive and has higher specificity compared to digital mammography screening, but high-volume DBT screening is a demanding task which can result in high discordance rate among readers. • Independent double reading DBT screening had overall comparable discordance rate as digital mammography, lower for spiculated masses seen on both modalities, and higher for small spiculated cancer seen only on DBT. • Almost all discordant digital breast tomosynthesis-detected cancers (72 of 74) and 25% (13 of 51) of the interval cancers in the Oslo Tomosynthesis Screening Trial were retrospectively classified as actionable and could have been detected by the readers.

Use of novel artificial intelligence computer-assisted detection (AI-CAD) for screening mammography: an analysis of 17,884 consecutive two-view full-field digital mammography screening exams.

Novel artificial intelligence computer-assisted detection (AI-CAD) systems based on deep learning (DL) promise to support screen reading. To test a DL-AI-CAD system compared to human reading on consecutive screening mammograms. In this retrospective study, 17,884 consecutive anonymized screening mammograms, double-read from January to November 2018, were processed by the DL-AI-CAD system. AI-CAD reading was considered positive if the AI-CAD case scores exceeded 30 (range = 1-100) and the lesion was correctly marked. Likewise, human reading (R1 or R2, respectively) was considered positive if the lesion was correctly identified and called. Receiver operating characteristic (ROC) analysis was performed and accuracy data were calculated. Ground truth for benign lesions: absence of malignancy after cancer registry matching (2022); for malignancy: histopathologic proof; evaluation was patient-based. In total, 114 screen-detected and 17 interval cancers (ICA) occurred. ROC analysis of screen-detected cancers yielded an AUC of 89% for AI-CAD. Sensitivity/specificity was 81.7%/80.2% for AI-CAD; 77.1%/91.7% for R1; 78.6/91.6% for R2. Combining each human reading with AI-CAD was as sensitive as human double-reading (all approximately 88%), but less specific (approximately 75%) compared to human double-reading (approximately 87%). These AI-CAD combinations required consensus readings for twice as many cases as the human combination. Four of 17 ICA exceeded a case score of 30; two of four CAD correctly marked the quadrant of the subsequent ICA. Including ICA cases, this AI-CAD achieved comparable sensitivity to human reading at lower specificity. Combining human reading and AI-CAD allows increasing sensitivity compared to single-reading.

Digital Breast Tomosynthesis versus Digital Mammography Screening Performance on Successive Screening Rounds from the Breast Cancer Surveillance Consortium.

Background Prior cross-sectional studies have observed that breast cancer screening with digital breast tomosynthesis (DBT) has a lower recall rate and higher cancer detection rate compared with digital mammography (DM). Purpose To evaluate breast cancer screening outcomes with DBT versus DM on successive screening rounds. Materials and Methods In this retrospective cohort study, data from 58 breast imaging facilities in the Breast Cancer Surveillance Consortium were collected. Analysis included women aged 40-79 years undergoing DBT or DM screening from 2011 to 2020. Absolute differences in screening outcomes by modality and screening round were estimated during the study period by using generalized estimating equations with marginal standardization to adjust for differences in women's risk characteristics across modality and round. Results A total of 523 485 DBT examinations (mean age of women, 58.7 years ± 9.7 [SD]) and 1 008 123 DM examinations (mean age, 58.4 years ± 9.8) among 504 863 women were evaluated. DBT and DM recall rates decreased with successive screening round, but absolute recall rates in each round were significantly lower with DBT versus DM (round 1 difference, -3.3% [95% CI: -4.6, -2.1] [P < .001]; round 2 difference, -1.8% [95% CI: -2.9, -0.7] [P = .003]; round 3 or above difference, -1.2% [95% CI: -2.4, -0.1] [P = .03]). DBT had significantly higher cancer detection (difference, 0.6 per 1000 examinations [95% CI: 0.2, 1.1]; P = .009) compared with DM only for round 3 and above. There were no significant differences in interval cancer rate (round 1 difference, 0.00 per 1000 examinations [95% CI: -0.24, 0.30] [P = .96]; round 2 or above difference, 0.04 [95% CI: -0.19, 0.31] [P = .76]) or total advanced cancer rate (round 1 difference, 0.00 per 1000 examinations [95% CI: -0.15, 0.19] [P = .94]; round 2 or above difference, -0.06 [95% CI: -0.18, 0.11] [P = .43]). Conclusion DBT had lower recall rates and could help detect more cancers than DM across three screening rounds, with no difference in interval or advanced cancer rates. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Skaane in this issue.

Standalone AI for Breast Cancer Detection at Screening Digital Mammography and Digital Breast Tomosynthesis: A Systematic Review and Meta-Analysis.

Background There is considerable interest in the potential use of artificial intelligence (AI) systems in mammographic screening. However, it is essential to critically evaluate the performance of AI before it can become a modality used for independent mammographic interpretation. Purpose To evaluate the reported standalone performances of AI for interpretation of digital mammography and digital breast tomosynthesis (DBT). Materials and Methods A systematic search was conducted in PubMed, Google Scholar, Embase (Ovid), and Web of Science databases for studies published from January 2017 to June 2022. Sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) values were reviewed. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 and Comparative (QUADAS-2 and QUADAS-C, respectively). A random effects meta-analysis and meta-regression analysis were performed for overall studies and for different study types (reader studies vs historic cohort studies) and imaging techniques (digital mammography vs DBT). Results In total, 16 studies that include 1 108 328 examinations in 497 091 women were analyzed (six reader studies, seven historic cohort studies on digital mammography, and four studies on DBT). Pooled AUCs were significantly higher for standalone AI than radiologists in the six reader studies on digital mammography (0.87 vs 0.81, P = .002), but not for historic cohort studies (0.89 vs 0.96, P = .152). Four studies on DBT showed significantly higher AUCs in AI compared with radiologists (0.90 vs 0.79, P < .001). Higher sensitivity and lower specificity were seen for standalone AI compared with radiologists. Conclusion Standalone AI for screening digital mammography performed as well as or better than radiologists. Compared with digital mammography, there is an insufficient number of studies to assess the performance of AI systems in the interpretation of DBT screening examinations. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Scaranelo in this issue.

Effect of artificial intelligence-based computer-aided diagnosis on the screening outcomes of digital mammography: a matched cohort study.

To investigate whether artificial intelligence-based computer-aided diagnosis (AI-CAD) can improve radiologists' performance when used to support radiologists' interpretation of digital mammography (DM) in breast cancer screening. A retrospective database search identified 3158 asymptomatic Korean women who consecutively underwent screening DM between January and December 2019 without AI-CAD support, and screening DM between February and July 2020 with image interpretation aided by AI-CAD in a tertiary referral hospital using single reading. Propensity score matching was used to match the DM with AI-CAD group in a 1:1 ratio with the DM without AI-CAD group according to age, breast density, experience level of the interpreting radiologist, and screening round. Performance measures were compared with the McNemar test and generalized estimating equations. A total of 1579 women who underwent DM with AI-CAD were matched with 1579 women who underwent DM without AI-CAD. Radiologists showed higher specificity (96% [1500 of 1563] vs 91.6% [1430 of 1561]; p < 0.001) and lower abnormal interpretation rates (AIR) (4.9% [77 of 1579] vs 9.2% [145 of 1579]; p < 0.001) with AI-CAD than without. There was no significant difference in the cancer detection rate (CDR) (AI-CAD vs no AI-CAD, 8.9 vs 8.9 per 1000 examinations; p = 0.999), sensitivity (87.5% vs 77.8%; p = 0.999), and positive predictive value for biopsy (PPV3) (35.0% vs 35.0%; p = 0.999) according to AI-CAD support. AI-CAD increases the specificity for radiologists without decreasing sensitivity as a supportive tool in the single reading of DM for breast cancer screening. This study shows that AI-CAD could improve the specificity of radiologists' DM interpretation in the single reading system without decreasing sensitivity, suggesting that it can benefit patients by reducing false positive and recall rates. • In this retrospective-matched cohort study (DM without AI-CAD vs DM with AI-CAD), radiologists showed higher specificity and lower AIR when AI-CAD was used to support decision-making in DM screening. • CDR, sensitivity, and PPV for biopsy did not differ with and without AI-CAD support.

Screening Outcomes of Supplemental Automated Breast US in Asian Women with Dense and Nondense Breasts.

Background Automated breast (AB) US effectively depicts mammographically occult breast cancers in Western women. However, few studies have focused on the outcome of supplemental AB US in Asian women who have denser breasts than Western women. Purpose To evaluate the performance of supplemental AB US on mammography-based breast cancer screening in Asian women with dense breasts and those with nondense breasts. Materials and Methods A retrospective database search identified asymptomatic Korean women who underwent digital mammography (DM) and supplemental AB US screening for breast cancer between January 2018 and December 2019. We excluded women without sufficient follow-up, established final diagnosis, or histopathologic results. Performance measures of DM alone and AB US combined with DM (hereafter AB US plus DM) were compared. The primary outcome was cancer detection rate (CDR), and the secondary outcomes were sensitivity and specificity. Subgroup analyses were performed based on mammography density. Results From 2785 screening examinations in 2301 women (mean age, 52 years ± 9 [SD]), 28 cancers were diagnosed (26 screening-detected cancers, two interval cancers). When compared with DM alone, AB US plus DM resulted in a higher CDR of 9.3 per 1000 examinations (95% CI: 7.7, 10.3) versus 6.5 per 1000 examinations (95% CI: 5.2, 7.2; P < .001) and a higher sensitivity of 90.9% (95% CI: 77.3, 100.0) versus 63.6% (95% CI: 40.9, 81.8; P < .001) but a lower specificity of 86.8% (95% CI: 85.2, 88.2) versus 94.6% (95% CI: 93.6, 95.5; P < .001) in women with dense breasts. In women with nondense breasts, AB US plus DM resulted in a higher CDR of 9.5 per 1000 examinations (95% CI: 7.1, 10.6) versus 6.3 per 1000 examinations (95% CI: 3.5, 7.1; P < .001), whereas specificity was lower at 95.2% (95% CI: 93.4, 96.8) versus 97.1% (95% CI: 95.8, 98.4; P < .001). Conclusion In Asian women, the addition of automated breast US to digital mammography showed higher cancer detection rates but lower specificities in both dense and nondense breasts. © RSNA, 2023 Supplemental material is available for this article.

Mammographic Screening in Routine Practice: Multisite Study of Digital Breast Tomosynthesis and Digital Mammography Screenings.

Background The use of digital breast tomosynthesis (DBT) is increasing over digital mammography (DM) following studies demonstrating lower recall rates (RRs) and higher cancer detection rates (CDRs). However, inconsistent interpretation of evidence on the risks and benefits of mammography has resulted in varying screening mammography recommendations. Purpose To evaluate screening outcomes among women in the United States who underwent routine DM or DBT mammographic screening. Materials and Methods This retrospective cohort study included women aged 40-79 years who underwent DM or DBT screening mammograms between January 2014 and December 2020. Outcomes of RR, CDR, positive predictive value of recall (PPV1), biopsy rate, and positive predictive value of biopsy (PPV3) were compared between DM and DBT with use of adjusted multivariable logistic regression models. Results A total of 2 528 063 screening mammograms from 1 100 447 women (mean age, 57 years ± 10 [SD]) were included. In crude analyses, DBT (1 693 727 screening mammograms vs 834 336 DM screening mammograms) demonstrated lower RR (10.3% [95% CI: 10.3, 10.4] for DM vs 8.9% [95% CI: 8.9, 9.0] for DBT; P < .001) and higher CDR (4.5 of 1000 screening mammograms [95% CI: 4.3, 4.6] vs 5.3 of 1000 [95% CI: 5.2, 5.5]; P < .001), PPV1 (4.3% [95% CI: 4.2, 4.5] vs 5.9% [95% CI: 5.7, 6.0]; P < .001), and biopsy rates (14.5 of 1000 screening mammograms [95% CI: 14.2, 14.7] vs 17.6 of 1000 [95% CI: 17.4, 17.8]; P < .001). PPV3 was similar between cohorts (30.0% [95% CI: 29.2, 30.9] for DM vs 29.3% [95% CI: 28.7, 29.9] for DBT; P = .16). After adjustment for age, breast density, site, and index year, associations remained stable with respect to statistical significance. Conclusion Women undergoing digital breast tomosynthesis had improved screening mammography outcomes compared with women who underwent digital mammography. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Bae and Seo in this issue.

Detection and evaluation of breast pathology based on fuzzy clustering and discrete wavelet transform

The aim of the work is to develop a method for extracting a breast tumor and an uninfected area of the breast within the framework of medical imaging using a combination of fuzzy clustering with discrete wavelet transform (FCMDWT) tools. Two sets of mammographic breast image data were used in the study: a subset of breast images of a digital base mammography screening data (CBIS-DDSM) and private data set. Contrast enhancement techniques were used to improve image quality before using FCMDWT for tumor segmentation and analysis. To assess the accuracy of tumor isolation metrics such as the Dice coefficient (F1 score) and intersection by union (IoU) were used. The mean IoU was 98.41 and F1 score was 96.47 using the proposed FCMDWT.

Abstract P070: Volumetric parenchymal pattern analysis for breast cancer risk estimation

Abstract Introduction: Mammographic breast density is among the strongest risk factors for breast cancer. However, breast density is typically assessed subjectively by the radiologist according to the Breast Imaging Reporting and Data System (BI-RADS) based on 2 dimensional (2D) digital mammography (DM) images. Digital breast tomosynthesis (DBT) is quickly replacing DM and allows more detailed volumetric imaging of the breast. Advances in radiomics, the high-throughput extraction of radiologic features, has enabled characterization of breast parenchymal complexity beyond breast density alone. The purpose of this study was to compare the performance of volumetric parenchymal pattern analysis from DBT and DM with conventional breast density measurement with respect to breast cancer risk estimation. Methods: We performed a case control study among women with concurrent DM and DBT screening (Selenia Dimensions, Hologic Inc.) at our institution between 3/2011-12/2014. Cases were diagnosed with breast cancer within 1 year of screening; controls were confirmed negative or benign at 1 year follow-up, matched on race (Black, White, other/unknown) and age (5-year bins). After exclusions for imaging artifacts, craniocaudal (CC) and mediolateral oblique (MLO) views for 187 cases and 737 controls, in six image formats were assessed: 1) raw (“FOR PROCESSING”) DM; 2) processed (“FOR PRESENTATION”) DM; 3) raw DBT central projection; 4) processed DBT central projection; 5) DBT central reconstructed slice; and 6) DBT reconstructed stack. For cases, we analyzed the breast contralateral to cancer diagnosis; for controls the same breast as the matched case. We extracted radiomic features using a lattice-based approach with the publicly available CaPTk software, averaging features for each breast over CC and MLO views. We examined 3 lattice window sizes (6.4, 12.8, and 25.6 mm) and 23 resolutions for image resampling (0.075 - 2mm). We performed PCA on the resulting 487 features for each combination of window size and resolution and built conditional logistic regression models to assess the association of the first 7 principal components with breast cancer, with models including age, BMI, and BI-RADS density. For each image type we calculated the model C-statistic at all window sizes and resolutions, for a total of 2304 experimental conditions. Results: Features from reconstructed DBT scans had on average higher C-statistics across all experimental conditions. A model using only age, BMI, and BI-RADS density had a C-statistic of 0.61. Models using radiomic features plus age, BMI, and BI-RADS density had mean C-statistic of 0.68 (IQR 0.68, 0.69) for reconstructed DBT scans; for all other image types, the mean C-statistic ranged from 0.64 to 0.66. Conclusions: Incorporating volumetric breast parenchymal patterns from DBT improves breast cancer risk estimation beyond markers derived from DM and beyond conventional BI-RADS density. Citation Format: Eric A. Cohen, Omid Haji Maghsoudi, Raymond Acciavatti, Lauren Pantalone, Walter Mankowski, Alex A Nguyen, Christopher G. Scott, Stacey Winham, Andrew D. Maidment, Anne Marie McCarthy, Celine M Vachon, Emily F Conant, Despina Kontos. Volumetric parenchymal pattern analysis for breast cancer risk estimation. [abstract]. In: Proceedings of the AACR Special Conference: Precision Prevention, Early Detection, and Interception of Cancer; 2022 Nov 17-19; Austin, TX. Philadelphia (PA): AACR; Can Prev Res 2023;16(1 Suppl): Abstract nr P070.

Digital Breast Tomosynthesis Plus Ultrasound Versus Digital Mammography Plus Ultrasound for Screening Breast Cancer in Women With Dense Breasts.

To compare the outcomes of digital breast tomosynthesis (DBT) screening combined with ultrasound (US) with those of digital mammography (DM) combined with US in women with dense breasts. A retrospective database search identified consecutive asymptomatic women with dense breasts who underwent breast cancer screening with DBT or DM and whole-breast US simultaneously between June 2016 and July 2019. Women who underwent DBT + US (DBT cohort) and DM + US (DM cohort) were matched using 1:2 ratio according to mammographic density, age, menopausal status, hormone replacement therapy, and a family history of breast cancer. The cancer detection rate (CDR) per 1000 screening examinations, abnormal interpretation rate (AIR), sensitivity, and specificity were compared. A total of 863 women in the DBT cohort were matched with 1726 women in the DM cohort (median age, 53 years; interquartile range, 40-78 years) and 26 breast cancers (9 in the DBT cohort and 17 in the DM cohort) were identified. The DBT and DM cohorts showed comparable CDR (10.4 [9 of 863; 95% confidence interval {CI}: 4.8-19.7] vs. 9.8 [17 of 1726; 95% CI: 5.7-15.7] per 1000 examinations, respectively; P = 0.889). DBT cohort showed a higher AIR than the DM cohort (31.6% [273 of 863; 95% CI: 28.5%-34.9%] vs. 22.4% [387 of 1726; 95% CI: 20.5%-24.5%]; P < 0.001). The sensitivity for both cohorts was 100%. In women with negative findings on DBT or DM, supplemental US yielded similar CDRs in both DBT and DM cohorts (4.0 vs. 3.3 per 1000 examinations, respectively; P = 0.803) and higher AIR in the DBT cohort (24.8% [188 of 758; 95% CI: 21.8%-28.0%] vs. 16.9% [257 of 1516; 95% CI: 15.1%-18.9%; P < 0.001). DBT screening combined with US showed comparable CDR but lower specificity than DM screening combined with US in women with dense breasts.

Use of Full-quality DICOM Images Compared to Minimally Compressed Mammograms in JPEG Format for Radiology Training: A Study From Radiologist and Radiographer Perspectives

Running online training in mammography interpretation poses a challenge to radiologists and reporting radiographers due to the large size of digital mammograms in DICOM format and limited bandwidth capabilities of the users for image transmission. This study aims to compare image quality between the full-quality with minimal compressed JPEG and DICOM format of mammograms on a diagnostic monitor through the evaluation of radiologists and radiographers. Twelve participants including six radiologists and six radiographers participated as observers in this study. The observers viewed 60 2D digital mammography screening cases (22 cancer and 38 normal cases) in DICOM and minimal compressed JPEG formats on a 5MP diagnostic monitor. A 5-point Likert scale was provided for observers to compare the quality of mammograms between the two formats, with text anchors indicating to one image being significantly better, slightly better or of equal quality in terms of technical and diagnostic aspects. Nonparametric descriptive statistics were used to evaluate the ratings of radiologists and radiographers in different characteristics of mammograms of two image formats. The DICOM and JPEG images were statistically equivalent through ratings from radiographers in brightness, contrast, dynamic range, sharpness, no significant distortion, no significant noise, and background homogeneity in all mammograms. Similarly, most radiologists rated DICOM and JPEG images clinically and statistically equivalent with respect to difficulty of interpretation, brightness, contrast, dynamic range, sharpness, the appearance of Cooper's ligaments, visibility of subtle microcalcifications, visibility of structures at the margins of the breast. Normal cases were marginally favored by radiologists in DICOM format (ranging from 0.4% to 5.3%) while cancer cases in JPEG (ranging from 0.8% to 7.6%) received slightly higher rating. Findings showed that baseline full-quality with minimal compression JPEG was equivalent to the DICOM format of full-field digital mammograms which suggests that this type of JPEG could be used for online training and education in radiology.

Breast cancer screening with digital breast tomosynthesis: comparison of different reading strategies implementing artificial intelligence

ObjectivesDigital breast tomosynthesis (DBT) can detect more cancers than the current standard breast screening method, digital mammography (DM); however, it can substantially increase the reading workload and thus hinder implementation in screening. Artificial intelligence (AI) might be a solution. The aim of this study was to retrospectively test different ways of using AI in a screening workflow.MethodsAn AI system was used to analyse 14,772 double-read single-view DBT examinations from a screening trial with paired DM double reading. Three scenarios were studied: if AI can identify normal cases that can be excluded from human reading; if AI can replace the second reader; if AI can replace both readers. The number of detected cancers and false positives was compared with DM or DBT double reading.ResultsBy excluding normal cases and only reading 50.5% (7460/14,772) of all examinations, 95% (121/127) of the DBT double reading detected cancers could be detected. Compared to DM screening, 27% (26/95) more cancers could be detected (p < 0.001) while keeping recall rates at the same level. With AI replacing the second reader, 95% (120/127) of the DBT double reading detected cancers could be detected—26% (25/95) more than DM screening (p < 0.001)—while increasing recall rates by 53%. AI alone with DBT has a sensitivity similar to DM double reading (p = 0.689).ConclusionAI can open up possibilities for implementing DBT screening and detecting more cancers with the total reading workload unchanged. Considering the potential legal and psychological implications, replacing the second reader with AI would probably be most the feasible approach.Key Points• Breast cancer screening with digital breast tomosynthesis and artificial intelligence can detect more cancers than mammography screening without increasing screen-reading workload.• Artificial intelligence can either exclude low-risk cases from double reading or replace the second reader.• Retrospective study based on paired mammography and digital breast tomosynthesis screening data.