Showing papers on "Digital mammography published in 2020"
TL;DR: This synopsis of the European Breast Guidelines provides recommendations regarding organized screening programs for women aged 40 to 75 years who are at average risk and the addition of hand-held ultrasonography, automated breast ultr Masonography, or magnetic resonance imaging compared with mammography alone.
Abstract: Description The European Commission Initiative for Breast Cancer Screening and Diagnosis guidelines (European Breast Guidelines) are coordinated by the European Commission's Joint Research Centre. The target audience for the guidelines includes women, health professionals, and policymakers. Methods An international guideline panel of 28 multidisciplinary members, including patients, developed questions and corresponding recommendations that were informed by systematic reviews of the evidence conducted between March 2016 and December 2018. GRADE (Grading of Recommendations Assessment, Development and Evaluation) Evidence to Decision frameworks were used to structure the process and minimize the influence of competing interests by enhancing transparency. Questions and recommendations, expressed as strong or conditional, focused on outcomes that matter to women and provided a rating of the certainty of evidence. Recommendations This synopsis of the European Breast Guidelines provides recommendations regarding organized screening programs for women aged 40 to 75 years who are at average risk. The recommendations address digital mammography screening and the addition of hand-held ultrasonography, automated breast ultrasonography, or magnetic resonance imaging compared with mammography alone. The recommendations also discuss the frequency of screening and inform decision making for women at average risk who are recalled for suspicious lesions or who have high breast density.
130 citations
TL;DR: The most common breast cancer imaging modalities, public, most cited and recently updated breast cancer databases, histopathological based breast cancer image analysis, and DL application types in medical image analysis are reviewed.
Abstract: Computer-aided image analysis for better understanding of images has been time-honored approaches in the medical computing field. In the conventional machine learning approach, the domain experts in medical images are mandatory for image annotation that subsequently to be used for feature engineering. However, in deep learning, a big jump has been made to help the researchers do segmentation, feature extraction, classification, and detection from raw medical images obtained using digital breast tomosynthesis, digital mammography, magnetic resonance imaging, and ultrasound imaging modalities. As a result, deep learning (DL) has gained a state-of-the-art in many application areas, for example, breast cancer image analysis. In this survey paper, we reviewed the most common breast cancer imaging modalities, public, most cited and recently updated breast cancer databases, histopathological based breast cancer image analysis, and DL application types in medical image analysis. We finally conclude by pointing out the research gaps to be addressed in the future.
90 citations
TL;DR: Experimental results show that MA-CNN is a powerful tool for diagnosing breast cancer by means of classifying the mammogram images with overall sensitivity of 96% and 0.99 AUC.
Abstract: Breast cancer is one of the leading causes of cancer death among women in worldwide. Early diagnosis of breast cancer improves the chance of survival by aiding proper clinical treatments. The digital mammography examination helps in diagnosing the breast cancer at its earlier stage. In this paper, Multiscale All Convolutional Neural Network (MA-CNN) is developed to assist the radiologist in diagnosing the breast cancer effectively. MA-CNN is a convolutional neural network-based approach that classifies mammogram images accurately. Convolutional neural networks are excellent in extracting the task specific features, since the feature learning is associated with classification task in order to attain the improved performance. The proposed approach automatically categorizes the mammographic images on mini-MIAS dataset into normal, malignant and benign classes. This model improves the accuracy of the classification system by fusing the wider context of information using multiscale filters without negotiating the computation speed. Experimental results show that MA-CNN is a powerful tool for diagnosing breast cancer by means of classifying the mammogram images with overall sensitivity of 96% and 0.99 AUC.
78 citations
TL;DR: Digital breast tomosynthesis screening was associated with detection of a higher proportion of poor-prognosis cancers than was digital mammography and improved compared with DM for 5 years of DBT at the population level.
Abstract: The sensitivity of digital breast tomosynthesis (DBT) was higher than that of digital mammography (DM) in each of the first 5 years after implementation of DBT. DBT also helped detect a higher prop...
71 citations
TL;DR: The proposed deep learning modeling approach can predict short-term breast cancer risk using normal screening mammogram images, and CC view was consistently more predictive than MLO view in both deep learning models, regardless the input sub-regions.
Abstract: PURPOSE To investigate two deep learning-based modeling schemes for predicting short-term risk of developing breast cancer using prior normal screening digital mammograms in a case-control setting. METHODS We conducted a retrospective Institutional Review Board-approved study on a case-control cohort of 226 patients (including 113 women diagnosed with breast cancer and 113 controls) who underwent general population breast cancer screening. For each patient, a prior normal (i.e., with negative or benign findings) digital mammogram examination [including mediolateral oblique (MLO) view and craniocaudal (CC) view two images] was collected. Thus, a total of 452 normal images (226 MLO view images and 226 CC view images) of this case-control cohort were analyzed to predict the outcome, i.e., developing breast cancer (cancer cases) or remaining breast cancer-free (controls) within the follow-up period. We implemented an end-to-end deep learning model and a GoogLeNet-LDA model and compared their effects in several experimental settings using two mammographic view images and inputting two different subregions of the images to the models. The proposed models were also compared to logistic regression modeling of mammographic breast density. Area under the receiver operating characteristic curve (AUC) was used as the model performance metric. RESULTS The highest AUC was 0.73 [95% Confidence Interval (CI): 0.68-0.78; GoogLeNet-LDA model on CC view] when using the whole-breast and was 0.72 (95% CI: 0.67-0.76; GoogLeNet-LDA model on MLO + CC view) when using the dense tissue, respectively, as the model input. The GoogleNet-LDA model significantly (all P < 0.05) outperformed the end-to-end GoogLeNet model in all experiments. CC view was consistently more predictive than MLO view in both deep learning models, regardless of the input subregions. Both models exhibited superior performance than the percent breast density (AUC = 0.54; 95% CI: 0.49-0.59). CONCLUSIONS The proposed deep learning modeling approach can predict short-term breast cancer risk using normal screening mammogram images. Larger studies are needed to further reveal the promise of deep learning in enhancing imaging-based breast cancer risk assessment.
64 citations
University of Washington1, Kaiser Permanente2, University of California, Davis3, University of California, San Francisco4, University of North Carolina at Chapel Hill5, Dartmouth College6, University of Vermont7, The Dartmouth Institute for Health Policy and Clinical Practice8, University of Illinois at Chicago9
TL;DR: Women undergoing baseline screening may benefit most from digital breast tomosynthesis, whereas on subsequent screens, the benefits of digital breasttomosynthesis may vary by age and density category.
Abstract: Importance Digital mammography (DM) and digital breast tomosynthesis (DBT) are used for routine breast cancer screening. There is minimal evidence on performance outcomes by age, screening round, and breast density in community practice. Objective To compare DM vs DBT performance by age, baseline vs subsequent screening round, and breast density category. Design, Setting, and Participants This comparative effectiveness study assessed 1 584 079 screening examinations of women aged 40 to 79 years without prior history of breast cancer, mastectomy, or breast augmentation undergoing screening mammography at 46 participating Breast Cancer Surveillance Consortium facilities from January 2010 to April 2018. Exposures Age, Breast Imaging Reporting and Data System breast density category, screening round, and modality. Main Outcomes and Measures Absolute rates and relative risks (RRs) of screening recall and cancer detection. Results Of 1 273 492 DM and 310 587 DBT examinations analyzed, 1 028 891 examinations (65.0%) were of white non-Hispanic women; 399 952 women (25.2%) were younger than 50 years; and 671 136 women (42.4%) had heterogeneously dense or extremely dense breasts. Adjusted differences in DM vs DBT performance were largest on baseline examinations: for example, per 1000 baseline examinations in women ages 50 to 59, recall rates decreased from 241 examinations for DM to 204 examinations for DBT (RR, 0.84; 95% CI, 0.73-0.98), and cancer detection rates increased from 5.9 with DM to 8.8 with DBT (RR, 1.50; 95% CI, 1.10-2.08). On subsequent examinations, women aged 40 to 79 years with heterogeneously dense breasts had improved recall rates and improved cancer detection with DBT. For example, per 1000 examinations in women aged 50 to 59 years, the number of recall examinations decreased from 102 with DM to 93 with DBT (RR, 0.91; 95% CI, 0.84-0.98), and cancer detection increased from 3.7 with DM to 5.3 with DBT (RR, 1.42; 95% CI, 1.23-1.64). Women aged 50 to 79 years with scattered fibroglandular density also had improved recall and cancer detection rates with DBT. Women aged 40 to 49 years with scattered fibroglandular density and women aged 50 to 79 years with almost entirely fatty breasts benefited from improved recall rates without change in cancer detection rates. No improvements in recall or cancer detection rates were observed in women with extremely dense breasts on subsequent examinations for any age group. Conclusions and Relevance This study found that improvements in recall and cancer detection rates with DBT were greatest on baseline mammograms. On subsequent screening mammograms, the benefits of DBT varied by age and breast density. Women with extremely dense breasts did not benefit from improved recall or cancer detection with DBT on subsequent screening rounds.
60 citations
TL;DR: Deep learning would work efficiently in screening breast cancer in digital mammograms of various densities, which could be maximized in breasts with lower parenchyma density.
Abstract: Mammography plays an important role in screening breast cancer among females, and artificial intelligence has enabled the automated detection of diseases on medical images. This study aimed to develop a deep learning model detecting breast cancer in digital mammograms of various densities and to evaluate the model performance compared to previous studies. From 1501 subjects who underwent digital mammography between February 2007 and May 2015, craniocaudal and mediolateral view mammograms were included and concatenated for each breast, ultimately producing 3002 merged images. Two convolutional neural networks were trained to detect any malignant lesion on the merged images. The performances were tested using 301 merged images from 284 subjects and compared to a meta-analysis including 12 previous deep learning studies. The mean area under the receiver-operating characteristic curve (AUC) for detecting breast cancer in each merged mammogram was 0.952 ± 0.005 by DenseNet-169 and 0.954 ± 0.020 by EfficientNet-B5, respectively. The performance for malignancy detection decreased as breast density increased (density A, mean AUC = 0.984 vs. density D, mean AUC = 0.902 by DenseNet-169). When patients’ age was used as a covariate for malignancy detection, the performance showed little change (mean AUC, 0.953 ± 0.005). The mean sensitivity and specificity of the DenseNet-169 (87 and 88%, respectively) surpassed the mean values (81 and 82%, respectively) obtained in a meta-analysis. Deep learning would work efficiently in screening breast cancer in digital mammograms of various densities, which could be maximized in breasts with lower parenchyma density.
48 citations
TL;DR: Screening high risk women 30-39 with annual MRI only may be sufficient for cancer detection and should be evaluated further, particularly for mutation carriers; among women 50-69, detection is most effective when mammography is included with annualMRI.
Abstract: BACKGROUND The Ontario Breast Screening Program expanded in July 2011 to screen high-risk women age 30-69 years with annual magnetic resonance imaging (MRI) and digital mammography. This study examined the benefits of screening with mammography and MRI by age and risk criteria. METHODS This prospective cohort study included 8782 women age 30-69 years referred to the High Risk Ontario Breast Screening Program from July 2011 to June 2015, with final results to December 2016. Cancer detection rates, sensitivity, and specificity of MRI and mammography combined were compared with each modality individually within risk groups stratified by age using generalized estimating equation models. Prognostic features of screen-detected breast cancers were compared by modality using Fisher exact test. All P values are two-sided. RESULTS Among 20 053 screening episodes, there were 280 screen-detected breast cancers (cancer detection rate = 14.0 per 1000, 95% confidence interval [CI] = 12.4 to 15.7). The sensitivity of mammography was statistically significantly lower than that of MRI plus mammography (40.8%, 95% CI = 29.3% to 53.5% vs 96.0%, 95% CI = 92.2% to 98.0%, P < .001). In mutation carriers age 30-39 years, sensitivity of the combination was comparable with MRI alone (100.0% vs 96.8%, 95% CI = 79.2% to 100.0%, P = .99) but with statistically significantly decreased specificity (78.0%, 95% CI = 74.7% to 80.9% vs 86.2%, 95% CI = 83.5% to 88.5%, P < .001). In women age 50-69 years, combining MRI and mammography statistically significantly increased sensitivity compared with MRI alone (96.3%, 95% CI = 90.6% to 98.6% vs 90.9%, 95% CI = 83.6% to 95.1%, P = .02), with a small but statistically significant decrease in specificity (84.2%, 95% CI = 83.1% to 85.2% vs 90.0%, 95% CI = 89.2% to 90.9%, P < .001). CONCLUSIONS Screening high risk women age 30-39 years with annual MRI only may be sufficient for cancer detection and should be evaluated further, particularly for mutation carriers. Among women age 50-69 years, detection is most effective when mammography is included with annual MRI.
43 citations
TL;DR: DBT is more sensitive than DM, while the addition of DM to DBT provides no additional diagnostic benefit in detecting breast cancer compared to digital breast tomosynthesis alone.
Abstract: No consensus exists on digital breast tomosynthesis (DBT) utilization for breast cancer detection. We performed a diagnostic test accuracy systematic review and meta-analysis comparing DBT, combined DBT and digital mammography (DM), and DM alone for breast cancer detection in average-risk women. MEDLINE and EMBASE were searched until September 2018. Comparative design studies reporting on the diagnostic accuracy of DBT and/or DM for breast cancer detection were included. Demographic, methodologic, and diagnostic accuracy data were extracted. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 tool. Accuracy metrics were pooled using bivariate random-effects meta-analysis. The impact of multiple covariates was assessed using meta-regression. PROSPERO ID: CRD 42018111287. Thirty-eight studies reporting on 488,099 patients (13,923 with breast cancer) were included. Eleven studies were at low risk of bias. DBT alone, combined DBT and DM, and DM alone demonstrated sensitivities of 88% (95% confidence interval [CI] 83–92), 88% (CI 83–92), and 79% (CI 75–82), as well as specificities of 84% (CI 76–89), 81% (CI 73–88), and 79% (CI 71–85), respectively. The greater sensitivities of DBT alone and combined DBT and DM compared to DM alone were preserved in the combined meta-regression models accounting for other covariates (p = 0.003–0.006). No significant difference in diagnostic accuracy between DBT alone and combined DBT and DM was identified (p = 0.175–0.581). DBT is more sensitive than DM, while the addition of DM to DBT provides no additional diagnostic benefit. Consideration of these findings in breast cancer imaging guidelines is recommended. • Digital breast tomosynthesis with or without additional digital mammography is more sensitive in detecting breast cancer than digital mammography alone in women at average risk for breast cancer.
• The addition of digital mammography to digital breast tomosynthesis provides no additional diagnostic benefit in detecting breast cancer compared to digital breast tomosynthesis alone.
• The specificity of digital breast tomosynthesis with or without additional digital mammography is no different than digital mammography alone in the detection of breast cancer.
39 citations
23 Nov 2020
TL;DR: This study’s findings show that the proposed deep learning-based techniques can produce a clinically useful computer-aided tool for breast density analysis by digital mammography.
Abstract: Breast density estimation with visual evaluation is still challenging due to low contrast and significant fluctuations in the mammograms' fatty tissue background. The primary key to breast density classification is to detect the dense tissues in the mammographic images correctly. Many methods have been proposed for breast density estimation; nevertheless, most of them are not fully automated. Besides, they have been badly affected by low signal-to-noise ratio and variability of density in appearance and texture. This study intends to develop a fully automated and digitalized breast tissue segmentation and classification using advanced deep learning techniques. The conditional Generative Adversarial Networks (cGAN) network is applied to segment the dense tissues in mammograms. To have a complete system for breast density classification, we propose a Convolutional Neural Network (CNN) to classify mammograms based on the standardization of Breast Imaging-Reporting and Data System (BI-RADS). The classification network is fed by the segmented masks of dense tissues generated by the cGAN network. For screening mammography, 410 images of 115 patients from the INbreast dataset were used. The proposed framework can segment the dense regions with an accuracy, Dice coefficient, Jaccard index of 98%, 88%, and 78%, respectively. Furthermore, we obtained precision, sensitivity, and specificity of 97.85%, 97.85%, and 99.28%, respectively, for breast density classification. This study's findings are promising and show that the proposed deep learning-based techniques can produce a clinically useful computer-aided tool for breast density analysis by digital mammography.
39 citations
TL;DR: Deep learning algorithms to automatically assess BI-RADS breast density were developed and crowdsourced annotations, including those from attendees who routinely read mammograms, had higher agreement with the algorithm than with the original interpreting radiologists.
Abstract: Objective We developed deep learning algorithms to automatically assess BI-RADS breast density Methods Using a large multi-institution patient cohort of 108,230 digital screening mammograms from the Digital Mammographic Imaging Screening Trial, we investigated the effect of data, model, and training parameters on overall model performance and provided crowdsourcing evaluation from the attendees of the ACR 2019 Annual Meeting Results Our best-performing algorithm achieved good agreement with radiologists who were qualified interpreters of mammograms, with a four-class κ of 0667 When training was performed with randomly sampled images from the data set versus sampling equal number of images from each density category, the model predictions were biased away from the low-prevalence categories such as extremely dense breasts The net result was an increase in sensitivity and a decrease in specificity for predicting dense breasts for equal class compared with random sampling We also found that the performance of the model degrades when we evaluate on digital mammography data formats that differ from the one that we trained on, emphasizing the importance of multi-institutional training sets Lastly, we showed that crowdsourced annotations, including those from attendees who routinely read mammograms, had higher agreement with our algorithm than with the original interpreting radiologists Conclusion We demonstrated the possible parameters that can influence the performance of the model and how crowdsourcing can be used for evaluation This study was performed in tandem with the development of the ACR AI-LAB, a platform for democratizing artificial intelligence
TL;DR: There is no evidence with high or moderate quality showing that DBT compared with digital mammography decreases recall rates, as well as false positive and false negative rates, for women attending population-based breast cancer screenings.
Abstract: We proposed to compare the accuracy and effectiveness of digital breast tomosynthesis (DBT), plus digital or synthetic mammography, with digital mammography alone in women attending population-based breast cancer screenings. We performed a systematic review and included controlled studies comparing DBT with digital mammography for breast cancer screening. Search strategies were applied to the MEDLINE, Embase, LILACS, and CENTRAL databases. With moderate quality of evidence, in 1,000 screens, DBT plus digital mammography increased the overall and invasive breast cancer rates by 3 and 2 (RR 1.36, 95% CI 1.18 to 1.58 and RR 1.51, 95% CI 1.27 to 1.79, respectively). DBT plus synthetic mammography increased both overall and invasive breast cancer rates by 2 (RR 1.38, 95% CI 1.24 to 1.54 and RR 1.37, 95% CI 1.22 to 1.55, respectively). DBT did not improve recall, false positive and false negative rates. However due to heterogeneity the quality of evidence was low. For women attending population-based breast cancer screenings, DBT increases rates of overall and invasive breast cancer. There is no evidence with high or moderate quality showing that DBT compared with digital mammography decreases recall rates, as well as false positive and false negative rates.
TL;DR: This study presents the first review which focuses on the detection of architectural distortion (AD) from mammographic images and discovered that using a deep learning approach, such as the convolution neural network (CNN) method, can yield a significant increase in performance for the task of Detection of architectural distortions.
Abstract: Breast cancer is a type of cancer that has risen to be the second cause of death among women. Classification of breast tissues into normal, benign, or malignant depends on the presence of abnormalities like microcalcifications, masses, architectural distortions, and asymmetries. Architectural distortion (AD) is subtle in detection with no association with masses but shows the abnormal arrangement of tissue strands, often in a radial, spiculation, or random pattern. It is widely rated as the third symptom of breast cancer which is the most commonly missed abnormality. Most computational approaches characterizing abnormalities in breast images often concentrate on the detection of microcalcification and masses with architectural distortions appearing as a secondary finding. The subtle nature and a minimal occurrence of architectural distortions may seem to complicate computational approaches for its detection. As a result, little research interest has been recorded in this area. It is widely reported that some cases of recent breast cancer are wrongly diagnosed due to the omission in detecting the presence of architectural distortion at the early stage of the disease. However, we discovered that most computational solutions to early detection of breast cancer are focused mainly on detecting other abnormalities such as masses and microcalcification, which are some evidence of the advanced stage of the disease. To emphasise the little efforts channeled towards detection of AD compared to other abnormalities, this article aims to detail the review of such studies in the last decade. To the best of our knowledge, this study presents the first review which focuses on the detection of architectural distortion (AD) from mammographic images. Furthermore, this article presents a comprehensive review of approaches, advances, and challenges on the computational methods for detecting AD, with the sole aim of advancing the use of deep learning models in detecting AD. Moreover, a comparative study of performance analyses of articles surveyed in this article is investigated. Our findings revealed that about 70% of the existing literature adopted Gabor Filters, while just less than 10% leveraged on the state-of-the-art performances recorded in computer vision and deep learning, in building outstanding computational models for the detection of AD. The current study also discovered that using a deep learning approach, such as the convolution neural network (CNN) method, can yield a significant increase in performance for the task of detection of architectural distortions. This assertion is based on literature results obtained using the CNN, which generates an accuracy of 99.4% compared to the use of Gabor filters method, which accounts for 95% accuracy.
TL;DR: Addition of CEDM for evaluation of low-moderate suspicion soft tissue breast lesions can substantially reduce biopsy of benign lesions without compromising cancer detection.
TL;DR: The MG-based radiomics nomogram could be used as a non-invasive and reliable tool in predicting ALN metastasis and may facilitate to assist clinicians for pre-operative decision-making.
Abstract: Objective:To establish a radiomics nomogram by integrating clinical risk factors and radiomics features extracted from digital mammography (MG) images for pre-operative prediction of axillary lymph...
TL;DR: With the transition to digital breast tomosynthesis-guided biopsy, more architectural distortions were biopsied, more radial sclerosing lesions were identified, and more discordance existed in radiologic and pathologic examinations, with a similar percentage of carcinomas diagnosed.
Abstract: Background Digital breast tomosynthesis (DBT)-guided biopsy is increasingly used in practice. It is important to know expected changes in biopsy targets, pathologic results, and discordance rates. Purpose To compare biopsy target types, pathologic results, and discordance rates for 2 years preceding and 2 years following implementation of DBT-guided biopsy. Materials and Methods All 9-gauge vacuum-assisted core biopsies from a single tertiary breast center that used digital mammography (DM) stereotactic guidance from 2013 to 2015 and DBT-guided biopsy from 2015 to 2017 were retrospectively reviewed. All mammographic examinations were performed with DBT. Patient demographics, biopsy target type, pathologic reports, surgical excision specimens when available, breast density, and imaging follow-up results were recorded. Biopsy targets and discordance rates between radiologic and pathologic examinations were compared between the two biopsy groups. Generalized mixed modeling was used to examine results before and after DBT-guided biopsy. Results A total of 1313 women underwent 1405 breast biopsies: 643 by using DM (August 2013 to July 2015) (median age, 56 years; interquartile range, 49-66 years) and 762 by using DBT (August 2015 to July 2017) (median age, 58 years; interquartile range, 50-67 years), (P = .58). Calcifications were the most common biopsy target for both groups, constituting 89.9% (578 of 643) of DM-guided biopsies and 71.1% (542 of 762) of DBT-guided biopsies (P = .03). The rate of architectural distortion biopsies was 2.0% (13 of 643) with DM-guided biopsy and 17.7% (135 of 762) with DBT-guided biopsy (P = .01). Although overall malignancy rate was similar for DM-guided biopsy (27.8% [179 of 643]) and DBT-guided biopsy (24.8% [191 of 762], P = .54), DBT-guided biopsy helped identify a similar percentage of invasive malignancies (37.4% [72 of 191] vs 29.0% [52 of 179] at DM P = .35), but more radial sclerosing lesions (8.3% [95% confidence interval {CI}: 6.0, 10.0] vs 1.7% [95% CI: 1.0, 3.1]) (P = .01). The discordance rate was 1.4% (95% CI: 1.0, 2.7) with DM-guided biopsy and 4.5% (95% CI: 3.2, 6.3) with DBT-guided biopsy (P = .01). Of the 34 discordant DBT-guided biopsies, 30 were architectural distortions. Conclusion With the transition to digital breast tomosynthesis-guided biopsy, more architectural distortions were biopsied, more radial sclerosing lesions were identified, and more discordance existed in radiologic and pathologic examinations, with a similar percentage of carcinomas diagnosed. © RSNA, 2020 Online supplemental material is available for this article.
TL;DR: Despite few reported adverse screening outcomes with SM-DBT, radiologists have concerns about image quality, specifically calcification characterization, and most radiologists screening with DBT have SM.
Abstract: Synthesized digital mammography (SM) was developed to replace digital mammography (DM) in digital breast tomosynthesis (DBT) imaging to reduce radiation dose. This survey assessed utilization and attitudes regarding SM in DBT screening. The study was institutional review board exempt. An online survey was sent to members of the Society of Breast Imaging in June 2018. Questions included practice information, utilization of DBT and SM, perception of change in recall rates (RRs) and cancer detection rates (CDRs) with SM–DBT versus DM–DBT, and attitudes regarding SM versus DM in DBT screening. χ2 Tests were used to compare response frequencies across groups. In all, 312 of 2,600 Society of Breast Imaging members responded to the survey (12%). Of respondents, 96% reported DBT capability, and 83% reported SM capability. Of those without SM, the most cited reasons were cost or administration and image quality concerns (both 32%). In addition, 40% reported combined SM and DM use in DBT screens, and 52% reported SM use without DM in the majority of DBT screens. The overall satisfaction with SM was 3.4 of 5 (1-5 scale). Most cited SM advantages were decreased dose (85%) and increased lesion conspicuity (27%). The most cited SM disadvantages were calcification characterization (61%) and decreased image quality (31%). Most respondents were unsure if CDRs changed (44%) and RR changed (30%) with few reporting adverse outcomes (6% RR increase, 1% CDR decrease). Most radiologists screening with DBT have SM, but only one-half have replaced DM with SM. Despite few reported adverse screening outcomes with SM–DBT, radiologists have concerns about image quality, specifically calcification characterization.
TL;DR: DBT and C EDM have superior diagnostic accuracy of 2D synthetized MX to identify and classify breast lesions, and CEDM combined with D BT has better diagnostic performance compared with DBT alone.
Abstract: To compare diagnostic performance of contrast-enhanced dual-energy digital mammography (CEDM) and digital breast tomosynthesis (DBT) alone and in combination compared to 2D digital mammography (MX) and dynamic contrast-enhanced MRI (DCE-MRI) in women with breast lesions. We enrolled 100 consecutive patients with breast lesions (BIRADS 3-5 at imaging or clinically suspicious). CEDM, DBT, and DCE-MRI 2D were acquired. Synthetized MX was obtained by DBT. A total of 134 lesions were investigated on 111 breasts of 100 enrolled patients: 53 were histopathologically proven as benign and 81 as malignant. Nonparametric statistics and receiver operating characteristic (ROC) curve were performed. Two-dimensional synthetized MX showed an area under ROC curve (AUC) of 0.764 (sensitivity 65%, specificity 80%), while AUC was of 0.845 (sensitivity 80%, specificity 82%) for DBT, of 0.879 (sensitivity 82%, specificity 80%) for CEDM, and of 0.892 (sensitivity 91%, specificity 84%) for CE-MRI. DCE-MRI determined an AUC of 0.934 (sensitivity 96%, specificity 88%). Combined CEDM with DBT findings, we obtained an AUC of 0.890 (sensitivity 89%, specificity 74%). A difference statistically significant was observed only between DCE-MRI and CEDM (P = .03). DBT, CEDM, CEDM combined to tomosynthesis, and DCE-MRI had a high ability to identify multifocal and bilateral lesions with a detection rate of 77%, 85%, 91%, and 95% respectively, while 2D synthetized MX had a detection rate for multifocal lesions of 56%. DBT and CEDM have superior diagnostic accuracy of 2D synthetized MX to identify and classify breast lesions, and CEDM combined with DBT has better diagnostic performance compared with DBT alone. The best results in terms of diagnostic performance were obtained by DCE-MRI. Dynamic information obtained by time-intensity curve including entire phase of contrast agent uptake allows a better detection and classification of breast lesions.
TL;DR: TheEvidence around the relative benefits and harms of breast cancer screening using a single radiologist to examine each female's mammograms for signs of cancer, or two radiologists (double reading) is explored.
Abstract: In this article, we explore the evidence around the relative benefits and harms of breast cancer screening using a single radiologist to examine each female's mammograms for signs of cancer (single reading), or two radiologists (double reading). First, we briefly explore the historical evidence using film-screen mammography, before providing an in-depth description of evidence using digital mammography. We classify studies according to which exact version of double reading they use, because the evidence suggests that effectiveness of double reading is contingent on whether the two radiologists are blinded to one another's decisions, and how the decisions of the two radiologists are integrated. Finally, we explore the implications for future mammography, including using artificial intelligence as the second reader, and applications to more complex three-dimensional imaging techniques such as tomosynthesis.
TL;DR: An automated classification method for diagnosing breast lesions on digital breast tomosynthesis images using radiomics to comprehensively analyze the radiological images is developed and indicates that the proposed method may help in more accurately diagnosing cases that are difficult to classify on images.
Abstract: In digital mammography, which is used for the early detection of breast tumors, oversight may occur due to overlap between normal tissues and lesions. However, since digital breast tomosynthesis can acquire three-dimensional images, tissue overlapping is reduced, and, therefore, the shape and distribution of the lesions can be easily identified. However, it is often difficult to distinguish between benign and malignant breast lesions on images, and the diagnostic accuracy can be reduced due to complications from radiological interpretations, owing to acquisition of a higher number of images. In this study, we developed an automated classification method for diagnosing breast lesions on digital breast tomosynthesis images using radiomics to comprehensively analyze the radiological images. We extracted an analysis area centered on the lesion and calculated 70 radiomic features, including the shape of the lesion, existence of spicula, and texture information. The accuracy was compared by inputting the obtained radiomic features to four classifiers (support vector machine, random forest, naive Bayes, and multi-layer perceptron), and the final classification result was obtained as an output using a classifier with high accuracy. To confirm the effectiveness of the proposed method, we used 24 cases with confirmed pathological diagnosis on biopsy. We also compared the classification results based on the presence or absence of dimension reduction using least absolute shrinkage and a selection operator (LASSO). As a result, when the support vector machine was used as a classifier, the correct identification rate of the benign tumors was 55% and that of malignant tumors was 84%, with best results. These results indicate that the proposed method may help in more accurately diagnosing cases that are difficult to classify on images.
TL;DR: Full-field digital mammography allows better discrimination or prediction of breast cancer in the general female population than screen-film mammography or computed radiography, regardless of age, breast density, or screening round.
Abstract: Full-field digital mammography had higher sensitivity and superior screening accuracy than screen-film mammography and computed radiography, despite its slightly lower specificity.
Posted Content•
TL;DR: A one-stage object detection architecture, called Breast Mass Detection Network (BMassDNet), based on anchor-free and feature pyramid which makes the detection of breast masses of different sizes well adapted and introduces a truncation normalization method and combines it with adaptive histogram equalization to enhance the contrast between the breast mass and the surrounding environment.
Abstract: Background and Objective: Accurate detection of breast masses in mammography images is critical to diagnose early breast cancer, which can greatly improve the patients survival rate. However, it is still a big challenge due to the heterogeneity of breast masses and the complexity of their surrounding environment.Methods: To address these problems, we propose a one-stage object detection architecture, called Breast Mass Detection Network (BMassDNet), based on anchor-free and feature pyramid which makes the detection of breast masses of different sizes well adapted. We introduce a truncation normalization method and combine it with adaptive histogram equalization to enhance the contrast between the breast mass and the surrounding environment. Meanwhile, to solve the overfitting problem caused by small data size, we propose a natural deformation data augmentation method and mend the train data dynamic updating method based on the data complexity to effectively utilize the limited data. Finally, we use transfer learning to assist the training process and to improve the robustness of the model ulteriorly.Results: On the INbreast dataset, each image has an average of 0.495 false positives whilst the recall rate is 0.930; On the DDSM dataset, when each image has 0.599 false positives, the recall rate reaches 0.943.Conclusions: The experimental results on datasets INbreast and DDSM show that the proposed BMassDNet can obtain competitive detection performance over the current top ranked methods.
TL;DR: The VCT platform will provide a framework for scanner design optimization, comparison between different scanner designs and between different modalities or protocols on computational breast models, without the need for scanning actual patients as in conventional clinical trials.
TL;DR: The accuracy and the effectiveness of the proposed technique for the detection and classification of breast Micro-Calcifications (MCs), which are diagnostically significant but difficult to detect findings, is evaluated on a unique dataset.
Abstract: Radiologists worldwide use mammography as a reliable tool for breast cancer screening. However, mammography assessment is challenging even for well-trained radiologists, leading to a pressing need for Computer Aided Diagnosis (CAD) systems. In this work, a novel technique for the detection and classification of breast Micro-Calcifications (MCs), which are diagnostically significant but difficult to detect findings, is presented. The proposed method is based on the subtraction of temporally sequential mammogram pairs, after pre-processing and image registration, followed by machine-learning. The classification was performed using several features extracted from the subtracted mammograms and selected during training to optimize the accuracy of the results. Six classifiers were tested in a leave-one-patient-out, 4, 5 and 10 fold cross-validation process. This technique was evaluated on a unique dataset, consisting of temporal sequences of mammograms from 80 patients taken between 1 to 6 years apart. The resulting 320 mammograms were reviewed by 2 radiologists who precisely marked each MC location. The accuracy of classifying MCs as benign or suspicious improved from 91.42% without temporal subtraction and an Ensemble of Decision Trees (EDT), to 99.55% with the use of sequential mammograms and Support Vector Machines (SVMs) with leave-one-patient-out validation. The improvement was statistically significant (p-value <; 0.005). These results verify the accuracy and the effectiveness of the proposed technique should to be further evaluated on a larger dataset.
TL;DR: Screening with digital breast tomosynthesis combined with synthetic two-dimensional mammograms (DBT+SM) versus digital mammography (DM) yielded lower recall rates for women with Volpara Density Grade (VDG) 1 and VDG 2, and adjusted relative risk of recall and screen-detected breast cancer increased with denser breasts for DBT+SM but not for DM.
Abstract: The relative risks of recall and screen-detected breast cancer increased by automated breast density category 1–4 for digital breast tomosynthesis combined with synthetic mammograms but not for dig...
TL;DR: In this article, artefacts that are commonly encountered in clinical practice are described to ease the recognition and help troubleshoot solutions to prevent or minimise them.
Abstract: Contrast-enhanced digital mammography (CEDM) is a diagnostic tool for breast cancer detection. Artefacts are observed in about 10% of CEDM examinations. Understanding CEDM artefacts is important to prevent diagnostic misinterpretation. In this article, we have described the artefacts that we have commonly encountered in clinical practice; we hope to ease the recognition and help troubleshoot solutions to prevent or minimise them.
TL;DR: The range of human first-reader performance measures within a population-based screening cohort of 1 million screening mammograms to gauge the performance of emerging AI CAD systems was determined to show a wide range of performance differences between high-volume readers.
Abstract: Overall sensitivity of breast cancer detection in a screening cohort ranged from 63% to 84% for the least sensitive quartile to the most sensitive quartile for high-volume radiologists; the largest...
TL;DR: SM is noninferior to FFDM in the detection of microcalcifications and the subjective diagnostic confidence in the two modalities is comparable, according to radiologists' preference between the two imaging modalities.
Abstract: OBJECTIVE. The purpose of this study is to compare the performance of 2D synthetic mammography (SM) to that of full-field digital mammography (FFDM) in the detection of microcalcifications and to e...
TL;DR: In this paper, a novel model that trains both texton representation and deep CNN representation for mass classification tasks is proposed, which can be used to assist in the classification of benign and malignant breast masses.
Abstract: Prompt diagnosis of benign and malignant breast masses is essential for early breast cancer screening. Convolutional neural networks (CNNs) can be used to assist in the classification of benign and malignant breast masses. A persistent problem in current mammography mass classification via CNN is the lack of local-invariant features, which cannot effectively respond to geometric image transformations or changes caused by imaging angles. In this study, a novel model that trains both texton representation and deep CNN representation for mass classification tasks is proposed. Rotation-invariant features provided by the maximum response filter bank are incorporated with the CNN-based classification. The fusion after implementing the reduction approach is used to address the deficiencies of CNN in extracting mass features. This model is tested on public datasets, CBIS-DDSM, and a combined dataset, namely, mini-MIAS and INbreast. The fusion after implementing the reduction approach on the CBIS-DDSM dataset outperforms that of the other models in terms of area under the receiver operating curve (0.97), accuracy (94.30%), and specificity (97.19%). Therefore, our proposed method can be integrated with computer-aided diagnosis systems to achieve precise screening of breast masses.
TL;DR: Switching from digital mammography to biennial digital breast tomosynthesis is not cost-effective at a willingness-to-pay threshold of €20 000 per life-year gained, but digital Breast Tomosynthesis has a higher probability of being more cost- effective than digital mammographers at a threshold of€35’000 perLife-year gain.
Abstract: Breast cancer screening with digital breast tomosynthesis compared with digital mammography was predicted to increase life-years gained by 7% and decrease false-positive results by 2% and was cost-...