Mammography For Detection Of Breast Cancer Research Articles

Surgical marker navigation system in breast conserving surgery and excision of non-palpable lesions: first Spanish single-center experience.

Due to the establishment of screening mammography for breast cancer detection, the number of non-palpable lesions has increased. Thus, an optimal localization system is mandatory for the excision of non-palpable breast tumors. The aim of the study is to report the feasibility Surgical Marker Navigation (SMN) system Sirius Pintuition® for the excision of non-palpable breast tumors and non-palpable axillary lymph nodes. A retrospective observational study of patients undergoing breast-conserving surgery and lymph node excision guided by SMN between December 2022 and May 2023 was performed. A total of 84 patients underwent excision of non-palpable breast tumors (77; 91.7%) or non-palpable axillary lymph-nodes (7; 8.3%) using SMN. In total, 94 markers were placed, in 74 patients (88.1%) only one marker was placed, whereas in 10 patients (11.9%) two markers were placed to correctly localize the lesion in the operating room. Most markers were placed using ultrasonographic guidance (69; 82.1%). Seventy-seven patients underwent breast-conserving surgery (91.7%) and 7 (8.3%) lymph node excision. In 10 cases (11.9%), the marker was accidentally displaced during surgery due to the use of magnetized instruments, although the specimen could be removed. In sum, all the markers were removed from the patients, although the marker retrieval rate, as we defined it (percentage of patients in whom the initial excised specimen contained the marker divided by the total number of patients), was 88.1%. The use of Sirius Pintuition® SMN for non-palpable breast tumors and non-palpable lymph nodes is feasible, with a retrieval rate of 88.1%.

Breast Cancer Detection in Mammography using Faster Region Convolutional Neural Networks and Group Convolution

Breast Cancer Detection in Mammography using Faster Region Convolutional Neural Networks and Group Convolution

Abstract PO2-13-12: Serum Thioredoxin 1 Enhances Breast Cancer Detection in Mammography and Breast Ultrasound, Including Dense Breasts

Abstract Background: Breast cancer (BC), when detected in its early stages, generally carries a more favorable prognosis than other types of cancer. Currently, the screening and early detection of breast cancer rely predominantly on image-oriented diagnostic methods. However, these methods inherently fall short in capturing the dynamic nature of cancer cells. Mammography has known limitations in clinical sensitivity, particularly in cases with high breast density. While breast ultrasound can mitigate some of these limitations, there remain unmet needs in this field. Previously, we reported the promising clinical performance (Sensitivity: 96.43%, Specificity: 97.32%) of serum thioredoxin 1 (Trx1) as a potential novel means for detecting breast cancer. In this study, we aim to evaluate the extent to which serum Trx1 can enhance the accuracy of breast cancer diagnosis, including in dense breasts, within the imaging-based diagnostic system powered by mammography and ultrasonography. Methods: We have generated monoclonal antibodies against Trx1 and developed an ELISA kit (DxMe® BC Kit, E&S Healthcare, Korea) that quantifies Trx1 in serum. A total of 308 biopsy-confirmed breast cancer patients, who had undergone mammography prior to their final diagnosis, were recruited. The clinical information of these 308 patients was retrospectively examined and collected. The data from 254 individuals, who also had results from breast ultrasound, and 261 individuals, who had records of breast density, were analyzed collectively. Freshly collected blood was prepared as previously indicated, and then tested with the kit to determine the level of Trx1. All serum samples were measured twice, and statistical analyses were conducted using ROC analysis, one-way ANOVA, and unpaired t-tests. Results: Clinical sensitivity for BC detection was analyzed using the Trx1 levels of the 254 subjects who underwent both mammography and breast ultrasound. The respective sensitivities of mammography, breast ultrasound, and the DxMe® BC Kit were 72.83%, 92.91%, and 96.46%. With combined analysis of mammography, breast ultrasound, and Trx1, the sensitivity increased to 98.43% for mammography + breast ultrasound, and 98.82% for mammography + Trx1. When breast ultrasound + Trx1 was considered, the sensitivity rose to 99.61%. Likewise, the sensitivity of mammography + breast ultrasound + Trx1 was also measured at 99.61%. Among the 261 breast cancer patients with available information about breast density, 186 patients (71.3%) had dense breasts. The clinical sensitivity of mammography and Trx1 was specifically analyzed for these dense breast cases. Mammography demonstrated a sensitivity of 64.3%, while the DxMe® BC kit exhibited a sensitivity of 97.2%. When mammography and Trx1 were analyzed in parallel, sensitivity increased to 99.7%. Conclusion: This study demonstrated that the DxMe® BC kit, which quantifies serum Trx1 levels, can enhance the accuracy of conventional image-based diagnostics for breast cancer screening and early detection by more than 25%. This improvement in sensitivity was observed when mammography and breast ultrasound were each used in conjunction with Trx1, with no significant difference noted between these two combinations. Notably, in the case of dense breasts, the Trx1 test improved sensitivity by more than 30% compared to mammography. Given that sensitivity increases when mammography and Trx1 tests are used together, it showed that incorporating the Trx1 test alongside traditional mammography could significantly enhance diagnostic performance in dense breasts up to 35%. This approach has the potential to streamline the breast cancer diagnostic process, reduce time and cost, and improve the reliability of current breast cancer screening and early detection methods, regardless of breast density. Citation Format: Young Kim, Hyemi Ko, Yeri Lee, Keunkoo Shin, Jongam Song, Jeryoung Kim, Kyounghoon Suh, Jinsun Lee. Serum Thioredoxin 1 Enhances Breast Cancer Detection in Mammography and Breast Ultrasound, Including Dense Breasts [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 PO2-13-12.

Mammography with deep learning for breast cancer detection.

X-ray mammography is currently considered the golden standard method for breast cancer screening, however, it has limitations in terms of sensitivity and specificity. With the rapid advancements in deep learning techniques, it is possible to customize mammography for each patient, providing more accurate information for risk assessment, prognosis, and treatment planning. This paper aims to study the recent achievements of deep learning-based mammography for breast cancer detection and classification. This review paper highlights the potential of deep learning-assisted X-ray mammography in improving the accuracy of breast cancer screening. While the potential benefits are clear, it is essential to address the challenges associated with implementing this technology in clinical settings. Future research should focus on refining deep learning algorithms, ensuring data privacy, improving model interpretability, and establishing generalizability to successfully integrate deep learning-assisted mammography into routine breast cancer screening programs. It is hoped that the research findings will assist investigators, engineers, and clinicians in developing more effective breast imaging tools that provide accurate diagnosis, sensitivity, and specificity for breast cancer.

Application of Artificial Intelligence in the Mammographic Detection of Breast Cancer in Saudi Arabian Women

Background: Breast cancer has a 14.8% incidence rate and an 8.5% fatality rate in Saudi Arabia. Mammography is useful for the early detection of breast cancer. Researchers have been developing artificial intelligence (AI) algorithms for early breast cancer diagnosis and reducing false-positive mammography results. The aim of this study was to examine the performance and accuracy of an AI system in breast cancer screening among Saudi women. Materials and Methods: This is a retrospective cross-sectional study that included 378 mammograms collected from 2017 to 2021 from government hospitals in Jeddah, Saudi Arabia. The patients’ demographic and clinical information were collected from files and electronic medical records. The radiologists’ assessments of the mammograms were based on Breast Imaging Reporting and Data System (BIRADS) scores. Follow-up or biopsy reports verified the radiologists’ findings. The MammoScreen system was the AI tool used in this study. Data were analyzed using SPSS Version 25. Results: The patients’ mean age was 50.31 years. Most patients had breast density B (42.3%) followed by A (27.2%) and C (25.9%). Most malignant cases were invasive ductal carcinomas (37.3%). Of the 181 cancer cases, 36.9% were BIRADS category V. The area under the curve for the AI detection (0.923; 95% confidence interval [CI], 0.893–0.954) was greater than that for the radiologists’ interpretation (0.838; 95% CI, 0.796–0.881). The AI detection agreed with the histopathological result in 167 positive (91.3%) and 182 negative cases (93.3%). The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the AI system were 92.8%, 91.9%, 91.3%, 93.3%, and 92.3%, respectively. The radiologist’s interpretation agreed with the pathology report in 180 positive (73.8%) and 134 negative cases (100%). Its sensitivity, specificity, PPV, NPV, and accuracy were 100%, 67.7%, 73.8%, 100%, and 83.1%, respectively. Conclusions: The AI system tested in this study had better accuracy and diagnostic performance than the radiologists and thus could be used as a support diagnostic tool for breast cancer detection in clinical practice and to reduce false-positive recalls.

Reproducibility and Explainability of Deep Learning in Mammography: A Systematic Review of Literature

Abstract Background Although abundant literature is currently available on the use of deep learning for breast cancer detection in mammography, the quality of such literature is widely variable. Purpose To evaluate published literature on breast cancer detection in mammography for reproducibility and to ascertain best practices for model design. Methods The PubMed and Scopus databases were searched to identify records that described the use of deep learning to detect lesions or classify images into cancer or noncancer. A modification of Quality Assessment of Diagnostic Accuracy Studies (mQUADAS-2) tool was developed for this review and was applied to the included studies. Results of reported studies (area under curve [AUC] of receiver operator curve [ROC] curve, sensitivity, specificity) were recorded. Results A total of 12,123 records were screened, of which 107 fit the inclusion criteria. Training and test datasets, key idea behind model architecture, and results were recorded for these studies. Based on mQUADAS-2 assessment, 103 studies had high risk of bias due to nonrepresentative patient selection. Four studies were of adequate quality, of which three trained their own model, and one used a commercial network. Ensemble models were used in two of these. Common strategies used for model training included patch classifiers, image classification networks (ResNet in 67%), and object detection networks (RetinaNet in 67%). The highest reported AUC was 0.927 ± 0.008 on a screening dataset, while it reached 0.945 (0.919–0.968) on an enriched subset. Higher values of AUC (0.955) and specificity (98.5%) were reached when combined radiologist and Artificial Intelligence readings were used than either of them alone. None of the studies provided explainability beyond localization accuracy. None of the studies have studied interaction between AI and radiologist in a real world setting. Conclusion While deep learning holds much promise in mammography interpretation, evaluation in a reproducible clinical setting and explainable networks are the need of the hour.

Automated breast cancer detection in mammography using ensemble classifier and feature weighting algorithms

Automated breast cancer detection in mammography using ensemble classifier and feature weighting algorithms

Diagnostic performance of screening mammography according to menstrual cycle among Asian women.

To investigate the performance metrics of screening mammography according to menstrual cycle week in premenopausal Asian women. This retrospective study included 69,556 premenopausal Asian women who underwent their first screening mammography between 2011 and 2019. The presence or absence of a breast cancer diagnosis within 12 months after the index screening mammography served as the reference standard, determined by linking the study data to the national cancer registry data. Menstrual cycles were calculated, and participants were assigned to groups according to weeks 1-4. The performance metrics included cancer detection rate (CDR), sensitivity, specificity, and positive predictive value (PPV), with comparisons across menstrual cycles. Among menstrual cycles, the lowest CDR at 4.7 per 1000 women (95% confidence interval [CI], 3.8-5.8 per 1000 women) was observed in week 4 (all P < 0.05). The highest sensitivity of 72.7% (95% CI, 61.4-82.3) was observed in week 1, although the results failed to reach statistical significance. The highest specificity of 80.4% (95% CI, 79.5-81.3%) was observed in week 1 (P = 0.01). The lowest PPV of 2.2% (95% CI, 1.8-2.7) was observed in week 4 (all P < 0.05). Screening mammography tended to show a higher performance during week 1 and a lower performance during week 4 of the menstrual cycle among Asian women. These results emphasize the importance of timing recommendations that consider menstrual cycles to optimize the effectiveness of screening mammography for breast cancer detection.

High-resolution synthesis of high-density breast mammograms: Application to improved fairness in deep learning based mass detection.

Computer-aided detection systems based on deep learning have shown good performance in breast cancer detection. However, high-density breasts show poorer detection performance since dense tissues can mask or even simulate masses. Therefore, the sensitivity of mammography for breast cancer detection can be reduced by more than 20% in dense breasts. Additionally, extremely dense cases reported an increased risk of cancer compared to low-density breasts. This study aims to improve the mass detection performance in high-density breasts using synthetic high-density full-field digital mammograms (FFDM) as data augmentation during breast mass detection model training. To this end, a total of five cycle-consistent GAN (CycleGAN) models using three FFDM datasets were trained for low-to-high-density image translation in high-resolution mammograms. The training images were split by breast density BI-RADS categories, being BI-RADS A almost entirely fatty and BI-RADS D extremely dense breasts. Our results showed that the proposed data augmentation technique improved the sensitivity and precision of mass detection in models trained with small datasets and improved the domain generalization of the models trained with large databases. In addition, the clinical realism of the synthetic images was evaluated in a reader study involving two expert radiologists and one surgical oncologist.

Fast Watershed Segmentation for Breast Cancer Detection

Within engineering and computer specializations, image processing is the most important study topic. It is one of today's fastest-growing technologies, with applications in a variety of biological sectors, including cancer sickness. According to the latest data from throughout the world, breast cancer is the most lethal of all cancer kinds. It is the most frequent cancer in women and the second leading cause of cancer mortality in women. In this study, we advocate using a watershed transformation to create a fast segmentation technique. This allows for the blending of updated information about picture objects, extending the partitioning of the dividing waterline and therefore the standard watershed technique. The method requires a mechanism to express the test picture in terms of the amount of change around every given pixel before it can begin the watershed modification. Each pixel in the greyscale representation of the original picture is subjected to the Sobel operator. According to the form, the tumors identified are round or semicircular and the light of the tumor dims as we travel away from its core. The complement for this previous data may be seen as a local minimum that necessitated the start of the watershed process. As a result, each tumor picture may be represented as a lake, with the center in the complement tumor picture being the least value. The identification of tumor percentage gets more reliable after using the approach. As a consequence, our computer-aided diagnostic method for mammographic breast cancer detection has improved significantly thanks to the novel methodology. The method was written in MATLAB and tested on a Windows computer. The strategy was put to the test using photos from MIAS (Mammogram Image Analysis Society, UK), which offers a consistent categorization system for mammographic examinations. In this study, we advocate using a watershed transformation to create a rapid segmentation technique. This allows for the blending of data about picture objects, extending the partitioning of the dividing waterline and therefore the standard watershed technique.

Abstract 181: High-accuracy breast cancer detection in mammography using EfficientNet and end-to-end training

Abstract Background:Breast cancer (BC) is the second most common cancer among women. BC screening is usually based on mammography interpreted by radiologists. Recently, some researchers have used deep learning to automatically diagnose BC in mammography and so assist radiologists. The progress of BC detection algorithms can be measured by their performance on public datasets. The CBIS-DDSM is a widely used public dataset composed of scanned mammographies, equally divided into malignant and non-malignant (benign) images. Each image is accompanied by the segmentation of the lesion. Shen et al. (Nature Sci. Rep., 2019) presented a BC detection algorithm using an “end-to-end” approach to train deep neural networks. In this algorithm, a patch classifier is first trained to classify local image patches. The patch classifier's weights are then used to initialize the whole image classifier, that is refined using datasets with the cancer status of the whole image. They achieved an AUC of 0.87 [0.84, 0.90] in classifying CBIS-DDSM images, using their best single-model, single-view breast classifier. They used ResNet (He et al., CVPR 2016) as the basis of their algorithm. Our hypothesis was that replacing the old ResNet with the modern EfficientNet (Tan et al., arXiv 2019) and MobileNetV2 (Sandler et al.,CVPR 2018) would result in greater accuracy. Methods:We tested many different models, to conclude that the best model is obtained using EfficientNet-B4 as the base model, with a MobileNetV2 block at the top, followed by a dense layer with two output categories. We trained the patch classifier using 52,528 patches with 224x224 pixels extracted from CBIS-DDSM. From each image, we extracted 20 patches: 10 patches containing the lesion and 10 from the background (without lesion). The patch classifier weights were then used to initialize the whole image classifier, that was trained using the end-to-end approach with CBIS-DDSM images resized to 1152x896 pixels, with data augmentation. The training was performed using a step learning rate of 1e-4 for the first 20 epochs then 1e-5 for the remaining 10 and batch size of 4, using 10-fold cross-validation. We used 81% of the dataset for training, 9% for validation and 10% for testing. Results:We obtained an AUC of 0.8963±0.06, using a single-model, single-view classifier and without test-time data augmentation. Conclusions:Using EfficientNet and MobileNetV2 as the basis of the BC detection algorithm (instead of ResNet), we obtained an improvement in classifying CBIS-DDSM images into malignant/non-malignant: AUC has increased from 0.87 to 0.896. Our AUC is also larger than other recent papers in the literature, such as Shu et al. (IEEE Trans Med. Image, 2020) that achieved an AUC of 0.838 in the same CBIS-DDSM dataset. Citation Format: Daniel G. Petrini, Carlos Shimizu, Gabriel V. Valente, Guilherme Folgueira, Guilherme A. Novaes, Maria L. Katayama, Pedro Serio, Rosimeire A. Roela, Tatiana C. Tucunduva, Maria Aparecida A. Folgueira, Hae Y. Kim. High-accuracy breast cancer detection in mammography using EfficientNet and end-to-end training [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 181.

DeepCAT: Deep Computer-Aided Triage of Screening Mammography.

Although much deep learning research has focused on mammographic detection of breast cancer, relatively little attention has been paid to mammography triage for radiologist review. The purpose of this study was to develop and test DeepCAT, a deep learning system for mammography triage based on suspicion of cancer. Specifically, we evaluate DeepCAT's ability to provide two augmentations to radiologists: (1) discarding images unlikely to have cancer from radiologist review and (2) prioritization of images likely to contain cancer. We used 1878 2D-mammographic images (CC & MLO) from the Digital Database for Screening Mammography to develop DeepCAT, a deep learning triage system composed of 2 components: (1) mammogram classifier cascade and (2) mass detector, which are combined to generate an overall priority score. This priority score is used to order images for radiologist review. Of 595 testing images, DeepCAT recommended low priority for 315 images (53%), of which none contained a malignant mass. In evaluation of prioritizing images according to likelihood of containing cancer, DeepCAT's study ordering required an average of 26 adjacent swaps to obtain perfect review order. Our results suggest that DeepCAT could substantially increase efficiency for breast imagers and effectively triage review of mammograms with malignant masses.

Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach.

Breast cancer remains a global challenge, causing over 600,000 deaths in 2018 (ref. 1). To achieve earlier cancer detection, health organizations worldwide recommend screening mammography, which is estimated to decrease breast cancer mortality by 20-40% (refs. 2,3). Despite the clear value of screening mammography, significant false positive and false negative rates along with non-uniformities in expert reader availability leave opportunities for improving quality and access4,5. To address these limitations, there has been much recent interest in applying deep learning to mammography6-18, and these efforts have highlighted two key difficulties: obtaining large amounts of annotated training data and ensuring generalization across populations, acquisition equipment and modalities. Here we present an annotation-efficient deep learning approach that (1) achieves state-of-the-art performance in mammogram classification, (2) successfully extends to digital breast tomosynthesis (DBT; '3D mammography'), (3) detects cancers in clinically negative prior mammograms of patients with cancer, (4) generalizes well to a population with low screening rates and (5) outperforms five out of five full-time breast-imaging specialists with an average increase in sensitivity of 14%. By creating new 'maximum suspicion projection' (MSP) images from DBT data, our progressively trained, multiple-instance learning approach effectively trains on DBT exams using only breast-level labels while maintaining localization-based interpretability. Altogether, our results demonstrate promise towards software that can improve the accuracy of and access to screening mammography worldwide.

DIGITAL BREAST TOMOSYNTHESIS AND FULL-FIELD DIGITAL MAMMOGRAPHY IN BREAST CANCER DETECTION ASSOCIATED WITH FOUR ASYMMETRY TYPES

The aim: Comparing sensitivity and specificity of digital breast tomosynthesis and full-field digital mammography in breast cancer detection associated with four different types of asymmetries according to BI-RADS Atlas. Materials and methods: Study included 201 patients with four types of asymmetries according BI-RADS atlas (asymmetry - 81 (40,3%), focal asymmetry - 82 (40,8%), global asymmetry - 36 (17,9%) and developing asymmetry - 2 (1,0%)) who underwent full-field digital mammography, digital breast tomosynthesis and hand-held full breast ultrasound from January 2017 to June 2018. The general rate of breast cancer for the 201 patients with asymmetries was 8 cases (4,0%) (IBC, n=6 (3,0%); DCIS, n=2 (1,0%) other findings associated with asymmetries were non-malignant, n=10 (5,0%) (sclerosing adenosis, n=5 (2,5%); fibroadenomatosis, n=3 (1,5%); simple cyst, n=1 (0,5%); radial scar associated with papilloma, typical ductal hyperplasia and sclerosing adenosis, n=1 (0,5%). Results: Analysis of the results showed that sensitivity of digital breast tomosynthesis was 75.0% [95% CI, 34.91% to 96.81%] and specificity was 94.8% [95% CI, 90.68% to 97.49%] which was superior to full-field digital mammography sensitivity 50.0% [95% CI, 15.70% to 84.30%] and specificity 91.19% [95% CI, 86.27% to 94.78%] for breast cancer detection associated with different types of asymmetries. Conclusions: Using of digital breast tomosynthesis in assessment of breast asymmetries can improve sensitivity and specificity in breast cancer detection and reduce number of unnecessary biopsies and short-interval follow-up examinations.

Shotgun proteomics coupled to nanoparticle-based biomarker enrichment reveals a novel panel of extracellular matrix proteins as candidate serum protein biomarkers for early-stage breast cancer detection

BackgroundThe lack of specificity and high degree of false positive and false negative rates when using mammographic screening for detecting early-stage breast cancer is a critical issue. Blood-based molecular assays that could be used in adjunct with mammography for increased specificity and sensitivity could have profound clinical impact. Our objective was to discover and independently verify a panel of candidate blood-based biomarkers that could identify the earliest stages of breast cancer and complement current mammographic screening approaches.MethodsWe used affinity hydrogel nanoparticles coupled with LC-MS/MS analysis to enrich and analyze low-abundance proteins in serum samples from 20 patients with invasive ductal carcinoma (IDC) breast cancer and 20 female control individuals with positive mammograms and benign pathology at biopsy. We compared these results to those obtained from five cohorts of individuals diagnosed with cancer in organs other than breast (ovarian, lung, prostate, and colon cancer, as well as melanoma) to establish IDC-specific protein signatures. Twenty-four IDC candidate biomarkers were then verified by multiple reaction monitoring (LC-MRM) in an independent validation cohort of 60 serum samples specifically including earliest-stage breast cancer and benign controls (19 early-stage (T1a) IDC and 41 controls).ResultsIn our discovery set, 56 proteins were increased in the serum samples from IDC patients, and 32 of these proteins were specific to IDC. Verification of a subset of these proteins in an independent cohort of early-stage T1a breast cancer yielded a panel of 4 proteins, ITGA2B (integrin subunit alpha IIb), FLNA (Filamin A), RAP1A (Ras-associated protein-1A), and TLN-1 (Talin-1), which classified breast cancer patients with 100% sensitivity and 85% specificity (AUC of 0.93).ConclusionsUsing a nanoparticle-based protein enrichment technology, we identified and verified a highly specific and sensitive protein signature indicative of early-stage breast cancer with no false positives when assessing benign and inflammatory controls. These markers have been previously reported in cell-ECM interaction and tumor microenvironment biology. Further studies with larger cohorts are needed to evaluate whether this biomarker panel improves the positive predictive value of mammography for breast cancer detection.

Repeat Screening Outcomes with Digital Breast Tomosynthesis Plus Synthetic Mammography for Breast Cancer Detection: Results from the Prospective Verona Pilot Study.

Background Few results are available about subsequent outcomes after screening with digital breast tomosynthesis (DBT). Purpose To compare the diagnostic accuracy of a screening round with DBT plus synthetic mammography (SM) (hereafter, DBT+SM) and the repeat screening round with DBT with SM (hereafter, DBT+SM) or full-field digital mammography (FFDM) with FFDM screening. Materials and Methods This prospective study (Verona Pilot Study, clinical trial identification: 2015/1238) included women screened with DBT+SM between April 2015 and March 2017 and rescreened with DBT+SM or FFDM between April 2017 and March 2019. Screening performance (recall rate, cancer detection rate [CDR], and positive predictive value of recall [PPV1]) was compared with that obtained from 28 680 women screened with FFDM between 2013 and 2014 (control group). Cancer stages were compared between modalities and screening rounds. A χ2 test was used to evaluate differences. P < .05 was indicative of a statistically significant difference. Results Of 34 638 women enrolled, 32 870 (median age, 58 years; age range, 52-71 years) underwent repeat screening-16 198 with DBT+SM and 16 672 with FFDM. The CDR was higher for repeat screening with DBT+SM than for the control group with FFDM (8.1 per 1000 women screened vs 4.5 per 1000 women screened, respectively; P < .01) and was not significantly lower for repeat screening with FFDM (3.5 per 1000 women screened vs 4.5 per 1000 women screened, respectively; P = .11). Compared with the control group, there was no difference in the recall rate at repeat screening with both DBT+SM (3.71% vs 3.40%, respectively; P = .10) and FFDM (3.71% vs 3.69%, P = .92), whereas PPV1 was higher only when repeat screening was performed with DBT+SM (23.8% vs 12.0%, P < .01). At repeat screening, the proportion of cancers stage II or higher was 14.5% (19 of 131 cancers) with DBT+SM and 8.5% (five of 59 cancers) with FFDM, both of which were lower than the proportion in the control group with FFDM (30 of 110 cancers, 27.3%) (P ≤ .01). Conclusion At repeat screening, digital breast tomosynthesis plus synthetic mammography depicted more cancers than full-field digital mammography (FFDM) and found a lower number of stage II cancers compared with FFDM. © RSNA, 2020 See also the editorial by Bae in this issue.

Artificial intelligence for breast cancer detection in mammography and digital breast tomosynthesis: State of the art.

Screening for breast cancer with mammography has been introduced in various countries over the last 30 years, initially using analog screen-film-based systems and, over the last 20 years, transitioning to the use of fully digital systems. With the introduction of digitization, the computer interpretation of images has been a subject of intense interest, resulting in the introduction of computer-aided detection (CADe) and diagnosis (CADx) algorithms in the early 2000's. Although they were introduced with high expectations, the potential improvement in the clinical realm failed to materialize, mostly due to the high number of false positive marks per analyzed image. In the last five years, the artificial intelligence (AI) revolution in computing, driven mostly by deep learning and convolutional neural networks, has also pervaded the field of automated breast cancer detection in digital mammography and digital breast tomosynthesis. Research in this area first involved comparison of its capabilities to that of conventional CADe/CADx methods, which quickly demonstrated the potential of this new technology. In the last couple of years, more mature and some commercial products have been developed, and studies of their performance compared to that of experienced breast radiologists are showing that these algorithms are on par with human-performance levels in retrospective data sets. Although additional studies, especially prospective evaluations performed in the real screening environment, are needed, it is becoming clear that AI will have an important role in the future breast cancer screening realm. Exactly how this new player will shape this field remains to be determined, but recent studies are already evaluating different options for implementation of this technology. The aim of this review is to provide an overview of the basic concepts and developments in the field AI for breast cancer detection in digital mammography and digital breast tomosynthesis. The pitfalls of conventional methods, and how these are, for the most part, avoided by this new technology, will be discussed. Importantly, studies that have evaluated the current capabilities of AI and proposals for how these capabilities should be leveraged in the clinical realm will be reviewed, while the questions that need to be answered before this vision becomes a reality are posed.

Breast Cancer Detection via Mammographic Images : A Survey

Breast cancer is a top killer disease for women globally. The long term survival rate of women can be improved through early and effective screening of breast cancer cells. Currently, a mammogram is the recommended tool for breast cancer screening since it can identify breast cancer cells several years before physical signs appear and it is cost effective. This paper analyzes mammographic detection of breast cancer by providing an explanation on development and classification of Breast Cancer, Image representation models for breast tumor, mammography technologies, a discussion on various mammographic signs of breast cancer, breast cancer feature extraction techniques, popular breast cancer classification techniques, comparative analysis of existing mammogram breast cancer databases, and a review of mammographic breast cancer detection studies are presented. Finally, a highlight on future work is given.

Comparative analysis between synthetic mammography reconstructed from digital breast tomosynthesis and full-field digital mammography for breast cancer detection and visibility

PurposeTo compare observer performance between synthetic mammography (2DSM) and full-field digital mammography (FFDM) for breast cancer detection and visibility. MethodA retrospective analysis was conducted on 136 histopathologically proven cases of breast cancer in patients who underwent FFDM and digital breast tomosynthesis (DBT). 2DSM images were reconstructed from DBT data, and 2DSM and FFDM images were reviewed and evaluated for mammographic features, probability of malignancy (BI-RADS classification), and lesion conspicuity. DBT images were not reviewed. Statistical differences in cancer detection rates between 2DSM and FFDM images were analyzed using the McNemar test, agreement on BI-RADS assessment between 2DSM and FFDM was assessed using Cohen’s kappa test, and the Wilcoxon’s signed rank test was used to compare visibility scores. ResultsMean cancer detection rates with 2DSM and FFDM images were 84.6 % and 87.8 %, respectively. In subgroup analyses, differences in breast density, tumor size, and presence of calcifications were not statistically significant. Agreement between 2DSM and FFDM images for BI-RADS classification was graded as good with Cohen’s k-coefficient of 0.78 ± 0.05. Visibility scores in both modalities of images were similar for all lesions combined; however, 2DSM had significantly better visibility scores for calcified cancers (p < 0.01), and in dense breast tissue (p < 0.01). ConclusionsDiagnostic performances of 2DSM and FFDM images were comparable for detecting breast cancers, and it is possible that 2DSM may eliminate the need for additional FFDM during DBT-based imaging due to advances in image reconstruction methods.

Stand-Alone Artificial Intelligence for Breast Cancer Detection in Mammography: Comparison With 101 Radiologists.

Artificial intelligence (AI) systems performing at radiologist-like levels in the evaluation of digital mammography (DM) would improve breast cancer screening accuracy and efficiency. We aimed to compare the stand-alone performance of an AI system to that of radiologists in detecting breast cancer in DM. Nine multi-reader, multi-case study datasets previously used for different research purposes in seven countries were collected. Each dataset consisted of DM exams acquired with systems from four different vendors, multiple radiologists' assessments per exam, and ground truth verified by histopathological analysis or follow-up, yielding a total of 2652 exams (653 malignant) and interpretations by 101 radiologists (28 296 independent interpretations). An AI system analyzed these exams yielding a level of suspicion of cancer present between 1 and 10. The detection performance between the radiologists and the AI system was compared using a noninferiority null hypothesis at a margin of 0.05. The performance of the AI system was statistically noninferior to that of the average of the 101 radiologists. The AI system had a 0.840 (95% confidence interval [CI] = 0.820 to 0.860) area under the ROC curve and the average of the radiologists was 0.814 (95% CI = 0.787 to 0.841) (difference 95% CI = -0.003 to 0.055). The AI system had an AUC higher than 61.4% of the radiologists. The evaluated AI system achieved a cancer detection accuracy comparable to an average breast radiologist in this retrospective setting. Although promising, the performance and impact of such a system in a screening setting needs further investigation.