Digital Breast Tomosynthesis: State of the Art

Detection of breast cancer in digital breast tomosynthesis with vision transformers

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

AI-based strategies in breast mass ≤ 2 cm classification with mammography and tomosynthesis

PURPOSE

To evaluate the diagnosis performance of digital mammography (DM) and digital breast tomosynthesis (DBT), DM combined DBT with AI-based strategies for breast mass ≤ 2 cm.

MATERIALS AND METHODS

DM and DBT images in 483 patients including 512 breast masses were acquired from November 2018 to November 2019. Malignant and benign tumours were determined by biopsies using histological analysis and follow-up within 24 months. The radiomics and deep learning methods were employed to extract the breast mass features in images and finally for benign and malignant classification. The DM, DBT and DM combined DBT (DM + DBT) images were fed into radiomics and deep learning models to construct corresponding models, respectively. The area under the receiver operating characteristic curve (AUC) was employed to estimate model performance. An external dataset of 146 patients from March 2021 to December 2022 from another center was enrolled for external validation.

RESULTS

In the internal testing dataset, compared with the DM model and the DBT model, the DM + DBT models based on radiomics and deep learning both showed statistically significant higher AUCs [0.810 (RA-DM), 0.823 (RA-DBT) and 0.869 (RA-DM + DBT), P ≤ 0.001; 0.867 (DL-DM), 0.871 (DL-DBT) and 0.908 (DL-DM + DBT), P = 0.001]. The deep learning models present superior to the radiomics models in the experiments with only DM (0.867 vs 0.810, P = 0.001), only DBT (0.871 vs 0.823, P = 0.001) and DM + DBT (0.908 vs 0.869, P = 0.003).

CONCLUSIONS

DBT has a clear additional value for diagnosing breast mass less than 2 cm compared with only DM. AI-based methods, especially deep learning, can help achieve excellent performance.

Deep learning of mammogram images to reduce unnecessary breast biopsies: a preliminary study

BACKGROUND

Patients with a Breast Imaging Reporting and Data System (BI-RADS) 4 mammogram are currently recommended for biopsy. However, 70-80% of the biopsies are negative/benign. In this study, we developed a deep learning classification algorithm on mammogram images to classify BI-RADS 4 suspicious lesions aiming to reduce unnecessary breast biopsies.

MATERIALS AND METHODS

This retrospective study included 847 patients with a BI-RADS 4 breast lesion that underwent biopsy at a single institution and included 200 invasive breast cancers, 200 ductal carcinoma in-situ (DCIS), 198 pure atypias, 194 benign, and 55 atypias upstaged to malignancy after excisional biopsy. We employed convolutional neural networks to perform 4 binary classification tasks: (I) benign vs. all atypia + invasive + DCIS, aiming to identify the benign cases for whom biopsy may be avoided; (II) benign + pure atypia vs. atypia-upstaged + invasive + DCIS, aiming to reduce excision of atypia that is not upgraded to cancer at surgery; (III) benign vs. each of the other 3 classes individually (atypia, DCIS, invasive), aiming for a precise diagnosis; and (IV) pure atypia vs. atypia-upstaged, aiming to reduce unnecessary excisional biopsies on atypia patients.

RESULTS

A 95% sensitivity for the "higher stage disease" class was ensured for all tasks. The specificity value was 33% in Task I, and 25% in Task II, respectively. In Task III, the respective specificity value was 30% (vs. atypia), 30% (vs. DCIS), and 46% (vs. invasive tumor). In Task IV, the specificity was 35%. The AUC values for the 4 tasks were 0.72, 0.67, 0.70/0.73/0.72, and 0.67, respectively.

CONCLUSION

Deep learning of digital mammograms containing BI-RADS 4 findings can identify lesions that may not need breast biopsy, leading to potential reduction of unnecessary procedures and the attendant costs and stress.

Evaluating the Margins of Breast Cancer Tumors by Using Digital Breast Tomosynthesis with Deep Learning: A Preliminary Assessment

BACKGROUND

The assessment information of tumor margins is extremely important for the success of the breast cancer surgery and whether the patient undergoes a second operation. However, conducting surgical margin assessments is a time-consuming task that requires pathology-related skills and equipment, and often cannot be provided in a timely manner. To address this challenge, digital breast tomosynthesis technology was utilized to generate detailed cross-sectional images of the breast tissue and integrate deep learning algorithms for image segmentation, achieving an assessment of tumor margins during surgery.

METHODS

this study utilized post-operative tissue samples from 46 patients who underwent breast-conserving treatment, and generated image sets using digital breast tomosynthesis for the training and evaluation of deep learning models.

RESULTS

Deep learning algorithms effectively identifying the tumor area. They achieved a Mean Intersection over Union (MIoU) of 0.91, global accuracy of 99%, weighted IoU of 44%, precision of 98%, recall of 83%, F1 score of 89%, and dice coefficient of 93% on the training dataset; for the testing dataset, MIoU was at 83%, global accuracy at 97%, weighted IoU at 38%, precision at 87%, recall rate at 69%, F1 score at 76%, dice coefficient at 86%.

CONCLUSIONS

The initial evaluation suggests that the deep learning-based image segmentation method is highly accurate in measuring breast tumor margins. This helps provide information related to tumor margins during surgery, and by using different datasets, this research method can also be applied to the surgical margin assessment of various types of tumors.

Comparing accuracy of tomosynthesis plus digital mammography or synthetic 2D mammography in breast cancer screening: baseline results of the MAITA RCT consortium

AIM

The analyses here reported aim to compare the screening performance of digital tomosynthesis (DBT) versus mammography (DM).

METHODS

MAITA is a consortium of four Italian trials, REtomo, Proteus, Impeto, and MAITA trial. The trials adopted a two-arm randomised design comparing DBT plus DM (REtomo and Proteus) or synthetic-2D (Impeto and MAITA trial) versus DM; multiple vendors were included. Women aged 45 to 69 years were individually randomised to one round of DBT or DM.

FINDINGS

From March 2014 to February 2022, 50,856 and 63,295 women were randomised to the DBT and DM arm, respectively. In the DBT arm, 6656 women were screened with DBT plus synthetic-2D. Recall was higher in the DBT arm (5·84% versus 4·96%), with differences between centres. With DBT, 0·8/1000 (95% CI 0·3 to 1·3) more women received surgical treatment for a benign lesion. The detection rate was 51% higher with DBT, ie. 2·6/1000 (95% CI 1·7 to 3·6) more cancers detected, with a similar relative increase for invasive cancers and ductal carcinoma in situ. The results were similar below and over the age of 50, at first and subsequent rounds, and with DBT plus DM and DBT plus synthetic-2D. No learning curve was appreciable. Detection of cancers >= 20 mm, with 2 or more positive lymph nodes, grade III, HER2-positive, or triple-negative was similar in the two arms.

INTERPRETATION

Results from MAITA confirm that DBT is superior to DM for the detection of cancers, with a possible increase in recall rate. DBT performance in screening should be assessed locally while waiting for long-term follow-up results on the impact of advanced cancer incidence.

An unusual artifact observed on screening mammography in a patient with an LVAD

PURPOSE

Screening mammography and digital breast tomosynthesis consist of high-resolution x-ray images to identify findings that are potentially indicative of breast cancer, enabling early detection and reduction of breast cancer mortality. Imaging artifacts can occasionally occur, sometimes due to patient-related medical devices. Because of continuous evolution of new technologies, there is potential for novel artifacts to be encountered. In this technical note, we report an unusual artifact in the screening mammogram of a patient with an Abbott HeartMate 3 left ventricular assist device (LVAD).

METHODS

A 72-year-old patient with a HeartMate 3 LVAD presented to our breast imaging facility for a standard screening exam with digital breast tomosynthesis (Selenia Dimensions, Hologic Inc., Bedford, MA) and synthetic 2D images (C-view, Hologic Inc., Bedford, MA).

RESULTS

Linear artifacts oriented in the anteroposterior dimension demonstrating a spatial periodicity of ∼1.4 mm were seen on all left breast images, whereas concurrent right breast images did not demonstrate any artifacts. Repeat attempts using two identical digital breast tomosynthesis units demonstrated the same artifacts. No other exam at our imaging center that day demonstrated any such artifacts. Mammogram exams performed on this patient prior to her LVAD placement did not exhibit any similar artifacts.

CONCLUSION

Findings support the patient's LVAD as the underlying source of linear artifacts observed on left breast images, particularly given the proximity of the LVAD to the left breast. With the number of patients receiving LVAD placement on the rise, as well as increasing median survival rates status post LVAD implantation, recognition of this LVAD related artifact on mammography may be important.

Tomosynthesis performance compared to radiography and computed tomography for sacroiliac joint structural damage detection in patients with suspected axial spondyloarthritis

Purpose

To compare tomosynthesis performance to radiography for the differentiation of sacroiliitis versus normal or degenerative changes in sacroiliac joints in patients with suspected axial spondyloarthritis (SpA).

Materials and methods

Radiography, tomosynthesis and CT of sacroiliac joints (29 patients) were performed on the same day in consecutive patients with suspected SpA. The examinations were retrospectively read independently, blinded by two radiologists (one junior and one senior, and twice by one junior). Interobserver and intraobserver agreement was evaluated using the kappa coefficient. Effective doses for each imaging sensitivity, specificity and accuracy were assessed and compared with CT as gold standard.

Results

CT detected 15/58 joints with sacroiliitis. The imaging sensitivity, specificity and accuracy were 60%, 84% and 44%, respectively, for radiography and 87%, 91% and 77% for tomosynthesis. The mean effective dose for tomosynthesis was significantly lower than that of CT (5-fold less) and significantly higher than that of radiography (8-fold more).

Conclusion

Tomosynthesis is superior to radiography for sacroiliitis detection in patients with suspected SpA, with 5-fold less radiation exposure than CT.

[Digital Breast Tomosynthesis Quality Control Manual Overview]

Recently, mammography systems equipped with digital breast tomosynthesis (DBT) have become widely used in Japan. Therefore, it is urgently necessary to establish a quality control method for DBTs. So far, we have been studying acceptance tests for DBTs with reference to EUREF. In 2020, IEC 61223-3-6 was published, which provides not only acceptance tests but also constancy test methods. Therefore, we conducted data collection using DBTs sold in Japan and examined the feasibility of conducting constancy tests. Although there were some items that were difficult to implement in each device, we were able to confirm quality control items that could be implemented in many devices. In addition, we were able to confirm routine tests that enable rapid evaluation. Based on these results, we have developed a "Digital Breast Tomosynthesis Quality Control Manual". In this paper, we report an overview of the manual and the results of routine tests.

A cadaveric breast cancer tissue phantom for phase-contrast X-ray imaging applications

BACKGROUND

As mammography X-ray imaging technologies advance and provide elevated contrast in soft tissues, a need has developed for reliable imaging phantoms for use in system design and component calibration. In advanced imaging modalities such as refraction-based methods, it is critical that developed phantoms capture the biological details seen in clinical precancerous and cancerous cases while minimizing artifacts that may be caused due to phantom production. This work presents the fabrication of a breast tissue imaging phantom from cadaveric breast tissue suitable for use in both transmission and refraction-enhanced imaging systems.

METHODS

Human cancer cell tumors were grown orthotopically in nude athymic mice and implanted into the fixed tissue while maintaining the native tumor/adipose tissue interface.

RESULTS

The resulting human-murine tissue hybrid phantom was mounted on a clear acrylic housing for absorption and refraction X-ray imaging. Digital breast tomosynthesis was also performed.

CONCLUSION

Both attenuation-based imaging and refraction-based imaging of the phantom are presented to confirm the suitability of this phantom's use in both imaging modalities.

Deep learning model for predicting the presence of stromal invasion of breast cancer on digital breast tomosynthesis

To develop a deep learning (DL)-based algorithm to predict the presence of stromal invasion in breast cancer using digital breast tomosynthesis (DBT). Our institutional review board approved this retrospective study and waived the requirement for informed consent from the patients. Initially, 499 patients (mean age 50.5 years, age range, 29-90 years) who were referred to our hospital under the suspicion of breast cancer and who underwent DBT between March 1 and August 31, 2019, were enrolled in this study. Among the 499 patients, 140 who underwent surgery after being diagnosed with breast cancer were selected for the analysis. Based on the pathological reports, the 140 patients were classified into two groups: those with non-invasive cancer (n = 20) and those with invasive cancer (n = 120). VGG16, Resnet50, DenseNet121, and Xception architectures were used as DL models to differentiate non-invasive from invasive cancer. The diagnostic performance of the DL models was assessed based on the area under the receiver operating characteristic curve (AUC). The AUC for the four models were 0.56 [95% confidence intervals (95% CI) 0.49-0.62], 0.67 (95% CI 0.62-0.74), 0.71 (95% CI 0.65-0.75), and 0.75 (95% CI 0.69-0.81), respectively. Our proposed DL model trained on DBT images is useful for predicting the presence of stromal invasion in breast cancer.

Impact of loss functions on the performance of a deep neural network designed to restore low-dose digital mammography

Digital mammography is currently the most common imaging tool for breast cancer screening. Although the benefits of using digital mammography for cancer screening outweigh the risks associated with the x-ray exposure, the radiation dose must be kept as low as possible while maintaining the diagnostic utility of the generated images, thus minimizing patient risks. Many studies investigated the feasibility of dose reduction by restoring low-dose images using deep neural networks. In these cases, choosing the appropriate training database and loss function is crucial and impacts the quality of the results. In this work, we used a standard residual network (ResNet) to restore low-dose digital mammography images and evaluated the performance of several loss functions. For training purposes, we extracted 256,000 image patches from a dataset of 400 images of retrospective clinical mammography exams, where dose reduction factors of 75% and 50% were simulated to generate low and standard-dose pairs. We validated the network in a real scenario by using a physical anthropomorphic breast phantom to acquire real low-dose and standard full-dose images in a commercially available mammography system, which were then processed through our trained model. We benchmarked our results against an analytical restoration model for low-dose digital mammography. Objective assessment was performed through the signal-to-noise ratio (SNR) and the mean normalized squared error (MNSE), decomposed into residual noise and bias. Statistical tests revealed that the use of the perceptual loss (PL4) resulted in statistically significant differences when compared to all other loss functions. Additionally, images restored using the PL4 achieved the closest residual noise to the standard dose. On the other hand, perceptual loss PL3, structural similarity index (SSIM) and one of the adversarial losses achieved the lowest bias for both dose reduction factors. The source code of our deep neural network is available at https://github.com/WANG-AXIS/LdDMDenoising.

Stepwise Implementation of 2D Synthesized Screening Mammography and Its Effect on Stereotactic Biopsy of Microcalcifications

RATIONALE AND OBJECTIVES

Information evaluating the efficacy of 2D synthesized mammography (2Ds) reconstructions in microcalcification detection is limited. This study used stereotactic biopsy data for microcalcifications to evaluate the stepwise implementation of 2Ds in screening mammography. The study aim was to identify whether 2Ds + digital breast tomosynthesis (DBT) is non-inferior to 2D digital mammography (2DM) + 2Ds + DBT, 2DM + DBT, and 2DM in identifying microcalcifications undergoing further diagnostic imaging and stereotactic biopsy.

MATERIALS AND METHODS

Retrospective stereotactic biopsy data were extracted following 151,736 screening mammograms of healthy women (average age, 56.3 years; range, 30-89 years), performed between 2012 and 2019. The stereotactic biopsy data were separated into 2DM, 2DM + DBT, 2DM + 2Ds + DBT, and 2Ds + DBT arms and examined using Fisher's exact test to compare the detection rates of all cancers, invasive cancers, DCIS, and ADH between modalities for patients undergoing stereotactic biopsy of microcalcifications.

RESULTS

No statistical significance in cancer detection was seen for 2Ds + DBT among those calcifications that underwent stereotactic biopsy when comparing the 2Ds + DBT to 2DM, 2DM + DBT, and 2DM + 2Ds + DBT imaging combinations.

CONCLUSION

These data suggest that 2Ds + DBT is non-inferior to 2DM + DBT in detecting microcalcifications that will undergo stereotactic biopsy.

Digital Breast Tomosynthesis: Towards Dose Reduction through Image Quality Improvement

Currently, breast cancer is the most commonly diagnosed type of cancer worldwide. Digital Breast Tomosynthesis (DBT) has been widely accepted as a stand-alone modality to replace Digital Mammography, particularly in denser breasts. However, the image quality improvement provided by DBT is accompanied by an increase in the radiation dose for the patient. Here, a method based on 2D Total Variation (2D TV) minimization to improve image quality without the need to increase the dose was proposed. Two phantoms were used to acquire data at different dose ranges (0.88-2.19 mGy for Gammex 156 and 0.65-1.71 mGy for our phantom). A 2D TV minimization filter was applied to the data, and the image quality was assessed through contrast-to-noise ratio (CNR) and the detectability index of lesions before and after filtering. The results showed a decrease in 2D TV values after filtering, with variations of up to 31%, increasing image quality. The increase in CNR values after filtering showed that it is possible to use lower doses (-26%, on average) without compromising on image quality. The detectability index had substantial increases (up to 14%), especially in smaller lesions. So, not only did the proposed approach allow for the enhancement of image quality without increasing the dose, but it also improved the chances of detecting small lesions that could be overlooked.

Supplemental Breast Cancer Screening in Women with Dense Breasts and Negative Mammography: A Systematic Review and Meta-Analysis

Background The best supplemental breast cancer screening modality in women at average risk or intermediate risk for breast cancer with dense breast and negative mammogram remains to be determined. Purpose To conduct systematic review and meta-analysis comparing clinical outcomes of the most common available supplemental screening modalities in women at average risk or intermediate risk for breast cancer in patients with dense breasts and mammography with negative findings. Materials and Methods A comprehensive search was conducted until March 12, 2020, in Medline, Epub Ahead of Print and In-Process and Other Non-Indexed Citations; Embase Classic and Embase; Cochrane Central Register of Controlled Trials; and Cochrane Database of Systematic Reviews, for Randomized Controlled Trials and Prospective Observational Studies. Incremental cancer detection rate (CDR); positive predictive value of recall (PPV1); positive predictive value of biopsies performed (PPV3); and interval CDRs of supplemental imaging modalities, digital breast tomosynthesis, handheld US, automated breast US, and MRI in non-high-risk patients with dense breasts and mammography negative for cancer were reviewed. Data metrics and risk of bias were assessed. Random-effects meta-analysis and two-sided metaregression analyses comparing each imaging modality metrics were performed (PROSPERO; CRD42018080402). Results Twenty-two studies reporting 261 233 screened patients were included. Of 132 166 screened patients with dense breast and mammography negative for cancer who met inclusion criteria, a total of 541 cancers missed at mammography were detected with these supplemental modalities. Metaregression models showed that MRI was superior to other supplemental modalities in CDR (incremental CDR, 1.52 per 1000 screenings; 95% CI: 0.74, 2.33; P < .001), including invasive CDR (invasive CDR, 1.31 per 1000 screenings; 95% CI: 0.57, 2.06; P < .001), and in situ disease (rate of ductal carcinoma in situ, 1.91 per 1000 screenings; 95% CI: 0.10, 3.72; P < .04). No differences in PPV1 and PPV3 were identified. The limited number of studies prevented assessment of interval cancer metrics. Excluding MRI, no statistically significant difference in any metrics were identified among the remaining imaging modalities. Conclusion The pooled data showed that MRI was the best supplemental imaging modality in women at average risk or intermediate risk for breast cancer with dense breasts and mammography negative for cancer. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Hooley and Butler in this issue.

Visual adaptation to medical images: a comparison of digital mammography and tomosynthesis

Purpose

Radiologists and other image readers spend prolonged periods inspecting medical images. The visual system can rapidly adapt or adjust sensitivity to the images that an observer is currently viewing, and previous studies have demonstrated that this can lead to pronounced changes in the perception of mammogram images. We compared these adaptation effects for images from different imaging modalities to explore both general and modality-specific consequences of adaptation in medical image perception.

Approach

We measured perceptual changes induced by adaptation to images acquired by digital mammography (DM) or digital breast tomosynthesis (DBT), which have both similar and distinct textural properties. Participants (nonradiologists) adapted to images from the same patient acquired from each modality or for different patients with American College of Radiology-Breast Imaging Reporting and Data System (BI-RADS) classification of dense or fatty tissue. The participants then judged the appearance of composite images formed by blending the two adapting images (i.e., DM versus DBT or dense versus fatty in each modality).

Results

Adaptation to either modality produced similar significant shifts in the perception of dense and fatty textures, reducing the salience of the adapted component in the test images. In side-by-side judgments, a modality-specific adaptation effect was not observed. However, when the images were directly fixated during adaptation and testing, so that the textural differences between the modalities were more visible, significantly different changes in the sensitivity to the noise in the images were observed.

Conclusions

These results confirm that observers can readily adapt to the visual properties or spatial textures of medical images in ways that can bias their perception of the images, and that adaptation can also be selective for the distinctive visual features of images acquired by different modalities.

Audit of Prior Screening Mammograms of Screen-Detected Cancers: Implications for the Delay in Breast Cancer Detection

When cancer is detected in a screening mammogram, on occasion retrospective review of prior screening (pre-index) mammograms indicates a likely presence of cancer. These missed cancers during pre-index screens constitute a delay in detection and diagnosis. This study was undertaken to quantify the missed cancer rate by auditing pre-index screens to improve the quality of mammography screening practice. From a cohort of 135 screen-detected cancers, 120 pre-index screening mammograms could be retrieved and served as the study sample. A consensus read by 2 radiologists who interpreted the pre-index screens in an unblinded manner with full knowledge of cancer location, cancer type, lesion type, and pathology served as the truth or reference standard. Five radiologists interpreted the pre-index screens in a blinded manner. Established performance metrics such as sensitivity and specificity were quantified for each reader in interpreting these pre-index screens in a blinded manner. All five radiologists detected lesions in 8/120 (6.7%) screens. Excluding the 2 readers whose performance was close to random, all the 3 remaining readers detected lesions in 13 pre-index screens. This indicates that there is a delay in diagnosis by at least one cycle from 8/120 (6.7%) to 13/120 (10.8%). There were no observable trends in terms of either the cancer type or the lesion type. Auditing prior screening mammograms in screen-detected cancers can help in identifying the proportion of cases that were missed during interpretation and help in quantifying the delay in breast cancer detection.

Artificial Intelligence in Breast X-Ray Imaging

This topical review is focused on the clinical breast x-ray imaging applications of the rapidly evolving field of artificial intelligence (AI). The range of AI applications is broad. AI can be used for breast cancer risk estimation that could allow for tailoring the screening interval and the protocol that are woman-specific and for triaging the screening exams. It also can serve as a tool to aid in the detection and diagnosis for improved sensitivity and specificity and as a tool to reduce radiologists' reading time. AI can also serve as a potential second 'reader' during screening interpretation. During the last decade, numerous studies have shown the potential of AI-assisted interpretation of mammography and to a lesser extent digital breast tomosynthesis; however, most of these studies are retrospective in nature. There is a need for prospective clinical studies to evaluate these technologies to better understand their real-world efficacy. Further, there are ethical, medicolegal, and liability concerns that need to be considered prior to the routine use of AI in the breast imaging clinic.

[Breast Imaging Reporting and Data System (BI-RADS): Advantages and Limitations]

Breast Imaging Reporting and Data System (BI-RADS) is a communication and data tracking system that standardizes and controls the quality of reporting by presenting lexicon descriptors, assessment categories, and recommendations for managing breast lesions. Using standardized terminology recommended by BI-RADS, radiologists can concisely and reproducibly communicate breast imaging results to clinicians. They can also provide the estimated malignant probability of the lesions found and guide management for them by determining the final assessment category. The limitations of BI-RADS 5th edition currently in use are that there are some areas for which standardized terminologies still need to be established, and that the diagnostic criteria of MRI assessment categories 3 and 4 are ambiguous compared to those for mammography or ultrasound. The next revision of BI-RADS is expected to include solutions for overcoming current limitations.

Artificial intelligence for digital breast tomosynthesis: Impact on diagnostic performance, reading times, and workload in the era of personalized screening

The ultimate goals of the application of artificial intelligence (AI) to digital breast tomosynthesis (DBT) are the reduction of reading times, the increase of diagnostic performance, and the reduction of interval cancer rates. In this review, after outlining the journey from computer-aided detection/diagnosis systems to AI applied to digital mammography (DM), we summarize the results of studies where AI was applied to DBT, noting that long-term advantages of DBT screening and its crucial ability to decrease the interval cancer rate are still under scrutiny. AI has shown the capability to overcome some shortcomings of DBT in the screening setting by improving diagnostic performance and by reducing recall rates (from -2 % to -27 %) and reading times (up to -53 %, with an average 20 % reduction), but the ability of AI to reduce interval cancer rates has not yet been clearly investigated. Prospective validation is needed to assess the cost-effectiveness and real-world impact of AI models assisting DBT interpretation, especially in large-scale studies with low breast cancer prevalence. Finally, we focus on the incoming era of personalized and risk-stratified screening that will first see the application of contrast-enhanced breast imaging to screen women with extremely dense breasts. As the diagnostic advantage of DBT over DM was concentrated in this category, we try to understand if the application of AI to DM in the remaining cohorts of women with heterogeneously dense or non-dense breast could close the gap in diagnostic performance between DM and DBT, thus neutralizing the usefulness of AI application to DBT.

Developing breast lesion detection algorithms for digital breast tomosynthesis: Leveraging false positive findings

BACKGROUND

Due to the complex nature of digital breast tomosynthesis (DBT) in imaging techniques, reading times are longer than 2D mammograms. A robust computer-aided diagnosis system in DBT could help radiologists reduce their workload and reading times.

PURPOSE

The purpose of this study was to develop algorithms for detecting biopsy-proven breast lesions on DBT using multi-depth level convolutional models and leveraging non-biopsied samples. As biopsied positive samples in a lesion dataset are limited, we hypothesized that false positive (FP) findings by detection algorithms from non-biopsied benign lesions could improve detection algorithms by using them as data augmentation.

APPROACH

We first extracted 2D slices from DBT volumes with biopsy-proven breast lesions (cancer and benign), with non-biopsied benign lesions (actionable), and for controls. Then, to provide lesion continuity along the z-direction, we combined a lesion slice with its immediate adjacent slices to synthesize 2.5-dimensional (2.5D) images of the lesion by assigning them into R, G, and B color channels. We used 224 biopsy-proven lesions from 39 cancer and 62 benign patients from a DBTex challenge dataset of 1000 scans. We included the 2.5D images of immediate neighboring slices from the lesion's center to increase the number of training samples. For lesion detection, we used the YOLOv5 algorithm as our base network. We trained a baseline algorithm (medium-depth level) using biopsied samples to detect actionable FPs in non-biopsied images. Afterward, we fine-tuned the baseline model on the augmented image set (actionable FPs added). For lesion inferencing, we processed the DBT volume slice-by-slice to estimate bounding boxes in each slice, and then combined them by connecting bounding boxes along the depth via volumetric morphological closing. We trained an additional model (large) with deeper-depth levels by repeating the above process. Finally, we developed an ensemble algorithm by combining the medium and large detection models. We used the free-response operating characteristic curve to evaluate our algorithms. We reported mean sensitivity per FPs per DBT volume only for biopsied views and sensitivity at 2-false positives per image (2FPI) for all views. However, due to the limited accessibility to the truth of the challenge validation and test datasets, we used sensitivity at 2FPI for statistical evaluation.

RESULTS

For the DBTex independent validation set, the medium baseline model achieved a mean sensitivity of 0.627 FPs per DBT volume, and a sensitivity of 0.640 at 2FPI. After adding actionable FP lesions, the model had an improved 2FPI of 0.769 over the baseline (p-value = 0.013). Our ensemble algorithm with multi-depth levels (medium + large) achieved a mean sensitivity of 0.815 FPs per DBT volume and an improved sensitivity at 2FPI of 0.80 over the baseline (p-value < 0.001) on the validation set. Finally, our ensemble model achieved a mean sensitivity of 0.786 FPs per DBT volume and a sensitivity of 0.743 at 2FPI on the DBTex independent test set.

CONCLUSIONS

Our results show that actionable FP findings hold useful information for lesion detection algorithms, and our ensemble detection model with multi-depth levels improves lesion detection performance.

Performance evaluation of digital breast tomosynthesis systems: physical methods and experimental data

Digital breast tomosynthesis (DBT) has become a well-established breast imaging technique, whose performance has been investigated in many clinical studies, including a number of prospective clinical trials. Results from these studies generally point to non-inferiority in terms of microcalcification detection and superior mass-lesion detection for DBT imaging compared to digital mammography (DM). This modality has become an essential tool in the clinic for assessment and ad-hoc screening but is not yet implemented in most breast screening programmes at a state or national level. While evidence on the clinical utility of DBT has been accumulating, there has also been progress in the development of methods for technical performance assessment and quality control of these imaging systems. DBT is a relatively complicated ‘pseudo-3D’ modality whose technical assessment poses a number of difficulties. This paper reviews methods for the technical performance assessment of DBT devices, starting at the component level in part one and leading up to discussion of system evaluation with physical test objects in part two. We provide some historical and basic theoretical perspective, often starting from methods developed for DM imaging. Data from a multi-vendor comparison are also included, acquired under the medical physics quality control protocol developed by EUREF and currently being consolidated by a European Federation of Organisations for Medical Physics working group. These data and associated methods can serve as a reference for the development of reference data and provide some context for clinical studies.

[Patterns in the Use and Perception of Digital Breast Tomosynthesis: A Survey of Korean Breast Radiologists]

Purpose

To evaluate the pattern of use and the perception of digital breast tomosynthesis (DBT) among Korean breast radiologists.

Materials and Methods

From March 22 to 29, 2021, an online survey comprising 27 questions was sent to members of the Korean Society of Breast Imaging. Questions related to practice characteristics, utilization and perception of DBT, and research interests. Results were analyzed based on factors using logistic regression.

Results

Overall, 120 of 257 members responded to the survey (response rate, 46.7%), 67 (55.8%) of whom reported using DBT. The overall satisfaction with DBT was 3.31 (1-5 scale). The most-cited DBT advantages were decreased recall rate (55.8%), increased lesion conspicuity (48.3%), and increased cancer detection (45.8%). The most-cited DBT disadvantages were extra cost for patients (46.7%), insufficient calcification characterization (43.3%), insufficient improvement in diagnostic performance (39.2%), and radiation dose (35.8%). Radiologists reported increased storage requirements and interpretation time for barriers to implementing DBT.

Conclusion

Further improvement of DBT techniques reflecting feedback from the user's perspective will help increase the acceptance of DBT in Korea.

Online Calibration of a Linear Micro Tomosynthesis Scanner

In a linear tomosynthesis scanner designed for imaging histologic samples of several centimeters size at 10 µm resolution, the mechanical instability of the scanning stage (±10 µm) exceeded the resolution of the image system, making it necessary to determine the trajectory of the stage for each scan to avoid blurring and artifacts in the images that would arise from the errors in the geometric information used in 3D reconstruction. We present a method for online calibration by attaching a layer of randomly dispersed micro glass beads or calcium particles to the bottom of the sample stage. The method was based on a parametric representation of the rigid body motion of the sample stage-marker layer assembly. The marker layer was easy to produce and proven effective in the calibration procedure.

Online Geometric Calibration of a Hybrid CT System for Ultrahigh-Resolution Imaging

A hybrid imaging system consisting of a standard computed tomography (CT) scanner and a low-profile photon-counting detector insert in contact with the patient's body has been used to produce ultrahigh-resolution images in a limited volume in chest scans of patients. The detector insert is placed on the patient bed as needed and not attached. Thus, its position and orientation in the scanner is dependent on the patient's position and scan settings. To allow accurate image reconstruction, we devised a method of determining the relative geometry of the detector insert and the CT scanner for each scan using fiducial markers. This method uses an iterative registration algorithm to align the markers in the reconstructed volume from the detector insert to that of the concurrent CT scan. After obtaining precise geometric information of the detector insert relative to the CT scanner, the two complementary sets of images are summed together to create a detailed image with reduced artifacts.

Comparison of Abbreviated Breast MRI vs Digital Breast Tomosynthesis for Breast Cancer Detection among Women with a History of Breast Cancer

RATIONALE AND OBJECTIVES

To compare the diagnostic performance of abbreviated breast MRI (AB-MRI) and digital breast tomosynthesis (DBT) in women with a personal history (PH) of breast cancer as a postoperative screening tool.

MATERIALS AND METHODS

A total of 471 patients who completed both DBT and AB-MRI examinations were included in this study (median age, 54.5 years). The detected cancer characteristics were analyzed. The cancer detection rate (CDR), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and area under the curve (AUC) were calculated by receiver operating characteristic (ROC) curve analysis.

RESULTS

Eleven malignancies were diagnosed, and most of the detected cancers were stage I (7 of 11, 63.6%). Eight were invasive ductal carcinomas (IDC), and 3 were ductal carcinoma in situ (DCIS). Of the 11 recurrences, 6 malignancies were detected by DBT, and 11 were detected by AB-MRI. AB-MRI detected all 8 IDC and 3 DCIS lesions, and DBT detected 6 of 8 IDC lesions. The CDRs for DBT and AB-MRI screenings were 12.7 and 23.4 per 1,000 women, respectively. The sensitivity, specificity, PPV, NPV, and accuracy of DBT versus AB-MRI were 54.6% versus 100%, 97.6% versus 96.5%, 35.3% versus 40.7%, 98.9% versus 100%, and 96.6% versus 96.6%, respectively. AB-MRI showed a higher AUC value (0.983) than DBT (0.761) (p = 0.0049).

CONCLUSION

AB-MRI showed an improved CDR, especially for invasive cancer. The diagnostic performance of AB-MRI was superior to that of DBT with high sensitivity and PPV without sacrificing specificity in women with a PH of breast cancer.

Breast dosimetry in alternative X-ray-based imaging modalities used in current clinical practices

In X-ray breast imaging, Digital Mammography (DM) and Digital Breast Tomosynthesis (DBT), are the standard and largely used techniques, both for diagnostic and screening purposes. Other techniques, such as dedicated Breast Computed Tomography (BCT) and Contrast Enhanced Mammography (CEM) have been developed as an alternative or a complementary technique to the established ones. The performance of these imaging techniques is being continuously assessed to improve the image quality and to reduce the radiation dose. These imaging modalities are predominantly used in the diagnostic setting to resolve incomplete or indeterminate findings detected with conventional screening examinations and could potentially be used either as an adjunct or as a primary screening tool in select populations, such as for women with dense breasts. The aim of this review is to describe the radiation dosimetry for these imaging techniques, and to compare the mean glandular dose with standard breast imaging modalities, such as DM and DBT.

Applying artificial intelligence technology to assist with breast cancer diagnosis and prognosis prediction

Breast cancer remains the most diagnosed cancer in women. Advances in medical imaging modalities and technologies have greatly aided in the early detection of breast cancer and the decline of patient mortality rates. However, reading and interpreting breast images remains difficult due to the high heterogeneity of breast tumors and fibro-glandular tissue, which results in lower cancer detection sensitivity and specificity and large inter-reader variability. In order to help overcome these clinical challenges, researchers have made great efforts to develop computer-aided detection and/or diagnosis (CAD) schemes of breast images to provide radiologists with decision-making support tools. Recent rapid advances in high throughput data analysis methods and artificial intelligence (AI) technologies, particularly radiomics and deep learning techniques, have led to an exponential increase in the development of new AI-based models of breast images that cover a broad range of application topics. In this review paper, we focus on reviewing recent advances in better understanding the association between radiomics features and tumor microenvironment and the progress in developing new AI-based quantitative image feature analysis models in three realms of breast cancer: predicting breast cancer risk, the likelihood of tumor malignancy, and tumor response to treatment. The outlook and three major challenges of applying new AI-based models of breast images to clinical practice are also discussed. Through this review we conclude that although developing new AI-based models of breast images has achieved significant progress and promising results, several obstacles to applying these new AI-based models to clinical practice remain. Therefore, more research effort is needed in future studies.

What eye tracking can tell us about how radiologists use automated breast ultrasound

Purpose: Automated breast ultrasound (ABUS) presents three-dimensional (3D) representations of the breast in the form of stacks of coronal and transverse plane images. ABUS is especially useful for the assessment of dense breasts. Here, we present the first eye tracking data showing how radiologists search and evaluate ABUS cases. Approach: Twelve readers evaluated single-breast cases in 20-min sessions. Positive findings were present in 56% of the evaluated cases. Eye position and the currently visible coronal and transverse slice were tracked, allowing for reconstruction of 3D "scanpaths." Results: Individual readers had consistent search strategies. Most readers had strategies that involved examination of all available images. Overall accuracy was 0.74 (sensitivity = 0.66 and specificity = 0.84). The 20 false negative errors across all readers can be classified using Kundel's (1978) taxonomy: 17 are "decision" errors (readers found the target but misclassified it as normal or benign). There was one recognition error and two "search" errors. This is an unusually high proportion of decision errors. Readers spent essentially the same proportion of time viewing coronal and transverse images, regardless of whether the case was positive or negative, correct or incorrect. Readers tended to use a "scanner" strategy when viewing coronal images and a "driller" strategy when viewing transverse images. Conclusions: These results suggest that ABUS errors are more likely to be errors of interpretation than of search. Further research could determine if readers' exploration of all images is useful or if, in some negative cases, search of transverse images is redundant following a search of coronal images.

A randomised controlled trial of digital breast tomosynthesis vs digital mammography as primary screening tests: Screening results over subsequent episodes of the Proteus Donna study

Proteus Donna is a randomised controlled trial aimed at prospectively evaluating screening with digital breast tomosynthesis (DBT), including interval cancer detection (ICD) and cancer detection (CD) in the analysis as a cumulative measure over subsequent screening episodes. Consenting women aged 46 to 68 attending the regional Breast Screening Service were randomly assigned to conventional digital mammography (DM, control arm) or DBT in addition to DM (DBT, study arm). At the subsequent round all participants underwent DM. Thirty-six months follow-up allowed for the identification of cancers detected in the subsequent screening and interscreening interval. Relative risk (RR) and 95% confidence interval (95% CI) were computed. Cumulative CD and Nelson-Aalen incidence were analysed over the follow-up period. Between 31 December 2014 and 31 December 2017, 43 022 women were randomised to DM and 30 844 to DBT. At baseline, CD was significantly higher (RR: 1.44, 95% CI: 1.21-1.71) in the study arm. ICD did not differ significantly between the two arms (RR: 0.92, 95% CI: 0.62-1.35). At subsequent screening with DM, the CD was lower (nearly significant) in the study arm (RR: 0.83, 95% CI: 0.65-1.06). Over the follow-up period, the cumulative CD (comprehensive of ICD) was slightly higher in the study arm (RR: 1.15, 95% CI: 1.01-1.31). The Nelson-Aalen cumulative incidence over time remained significantly higher in the study arm for approximately 24 months. Benign lesions detection was higher in the study arm at baseline and lower at subsequent tests. Outcomes are consistent with a lead time gain of DBT compared to DM, with an increase in false positives and moderate overdiagnosis.

Power Spectrum Analysis of Breast Parenchyma with Digital Breast Tomosynthesis Images in a Longitudinal Screening Cohort from Two Vendors

RATIONALE AND OBJECTIVES

To quantitatively compare breast parenchymal texture between two Digital Breast Tomosynthesis (DBT) vendors using images from the same patients.

MATERIALS AND METHODS

This retrospective study included consecutive patients who had normal screening DBT exams performed in January 2018 from GE and normal screening DBT exams in adjacent years from Hologic. Power spectrum analysis was performed within the breast tissue region. The slope of a linear function between log-frequency and log-power, β, was derived as a quantitative measure of breast texture and compared within and across vendors along with secondary parameters (laterality, view, year, image format, and breast density) with correlation tests and t-tests.

RESULTS

A total of 24,339 DBT slices or synthetic 2D images from 85 exams in 25 women were analyzed. Strong power-law behavior was verified from all images. Values of β d did not differ significantly for laterality, view, or year. Significant differences of β were observed across vendors for DBT images (Hologic: 3.4±0.2 vs GE: 3.1±0.2, 95% CI on difference: 0.27 to 0.30) and synthetic 2D images (Hologic: 2.7±0.3 vs GE: 3.0±0.2, 95% CI on difference: -0.36 to -0.27), and density groups with each vendor: scattered (GE: 3.0±0.3, Hologic: 3.3±0.3) vs. heterogeneous (GE: 3.2±0.2, Hologic: 3.4±0.1), 95% CI (-0.27, -0.08) and (-0.21, -0.05), respectively.

CONCLUSION

There are quantitative differences in the presentation of breast imaging texture between DBT vendors and across breast density categories. Our findings have relevance and importance for development and optimization of AI algorithms related to breast density assessment and cancer detection.

A Randomized Trial Comparing Breast Cancer Incidence and Interval Cancers after Tomosynthesis Plus Mammography versus Mammography Alone

Background Adding digital breast tomosynthesis (DBT) to digital mammography (DM) improves breast cancer screening sensitivity, but how this impacts mortality and other end points is unknown. Purpose To compare interval and overall breast cancer incidence after screening with DBT plus DM versus DM alone. Materials and Methods In this prospective trial (RETomo), women attending screening were randomized to one round of DBT plus DM (experimental arm) or to DM (control arm). All were then rescreened with DM after 12 months (women aged 45-49 years) or after 24 months (50-69 years). The primary outcome was interval cancer incidence. Cumulative incidence up to the subsequent screening round plus 9 months (21- and 33-month follow-up for women aged 45-49 and 50-69, respectively) was also reported. Ductal carcinomas in situ are included. Subgroup analyses by age and breast density were conducted; 95% CIs computed according to binomial distribution are reported. Results Baseline cancer detection was higher in the DBT plu DM arm than DM arm (101 of 13 356 women vs 61 of 13 521 women; relative detection, 1.7 [95% CI: 1.2, 2.3]). The mean age ± standard deviation for the women in both arms was 55 years ± 7. Interval cancer incidence was similar in the two arms (21 vs 22 cancers; relative incidence, 0.97 [95% CI: 0.53, 1.8]). Cumulative incidence remained higher in the DBT plus DM arm in women over 50 (153 vs 124 cancers; relative incidence, 1.2 [95% CI: 0.99, 1.6]), while it was similar in the two arms in women aged 45-49 (36 vs 41 cancers; relative incidence, 0.89 [95% CI: 0.57, 1.4]). Conclusion In women younger than 50 years, the benefit of early diagnosis seemed to be appreciable, while for women over age 50, the higher sensitivity of tomosynthesis plus mammography was not matched by a subsequent reduction in cancers at the next screening examination or in the intervening interval. Clinical trial registration no. NCT02698202 © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by Lee and Ray in this issue.

Cone-beam breast CT using an offset detector: effect of detector offset and image reconstruction algorithm

Objective.A dedicated cone-beam breast computed tomography (BCT) using a high-resolution, low-noise detector operating in offset-detector geometry has been developed. This study investigates the effects of varying detector offsets and image reconstruction algorithms to determine the appropriate combination of detector offset and reconstruction algorithm.Approach.Projection datasets (300 projections in 360°) of 30 breasts containing calcified lesions that were acquired using a prototype cone-beam BCT system comprising a 40 × 30 cm flat-panel detector with 1024 × 768 detector pixels were reconstructed using Feldkamp-Davis-Kress (FDK) algorithm and served as the reference. The projection datasets were retrospectively truncated to emulate cone-beam datasets with sinograms of 768×768 and 640×768 detector pixels, corresponding to 5 cm and 7.5 cm lateral offsets, respectively. These datasets were reconstructed using the FDK algorithm with appropriate weights and an ASD-POCS-based Fast, total variation-Regularized, Iterative, Statistical reconstruction Technique (FRIST), resulting in a total of 4 offset-detector reconstructions (2 detector offsets × 2 reconstruction methods). Signal difference-to-noise ratio (SDNR), variance, and full-width at half-maximum (FWHM) of calcifications in two orthogonal directions were determined from all reconstructions. All quantitative measurements were performed on images in units of linear attenuation coefficient (1/cm).Results.The FWHM of calcifications did not differ (P > 0.262) among reconstruction algorithms and detector formats, implying comparable spatial resolution. For a chosen detector offset, the FRIST algorithm outperformed FDK in terms of variance and SDNR (P < 0.0001). For a given reconstruction method, the 5 cm offset provided better results.Significance.This study indicates the feasibility of using the compressed sensing-based, FRIST algorithm to reconstruct sinograms from offset-detectors. Among the reconstruction methods and detector offsets studied, FRIST reconstructions corresponding to a 30 cm × 30 cm with 5 cm lateral offset, achieved the best performance. A clinical prototype using such an offset geometry has been developed and installed for clinical trials.

Breast Radiation Exposure of 3D Digital Breast Tomosynthesis Compared to Full-Field Digital Mammography in a Clinical Follow-Up Setting

According to a position paper of the European Commission Initiative on Breast Cancer (ECIBC), DBT is close to being introduced in European breast cancer screening programmes. Our study aimed to examine radiation dose delivered by digital breast tomosynthesis (DBT) and digital mammography (FFDM) in comparison to sole FFDM in a clinical follow-up setting and in an identical patient cohort. Retrospectively, 768 breast examinations of 96 patients were included. Patients received both DBT and FFDM between May 2015 and July 2019: (I) FFDM in cranio-caudal (CC) and DBT in mediolateral oblique (MLO) view, as well as a (II) follow-up examination with FFDM in CC and MLO view. The mean glandular dose (MGD) was determined by the mammography system according to Dance’s model. The MGD (standard deviation (SD), interquartile range (IQR)) was distributed as follows: (I) (CC_FFDM+MLO_DBT) (a) left FFDM_CC 1.40 mGy (0.36 mGy, 1.13–1.59 mGy), left DBT_MLO 1.62 mGy (0.51 mGy, 1.27–1.82 mGy); (b) right FFDM_CC 1.36 mGy (0.34 mGy, 1.14–1.51 mGy), right DBT_MLO 1.59 mGy (0.52 mGy, 1.27–1.62 mGy). (II) (CC_FFDM+MLO_FFDM) (a) left FFDM_CC 1.35 mGy (0.35 mGy, 1.10–1.60 mGy), left FFDM_MLO 1.40 mGy (0.39 mGy, 1.12–1.59 mGy), (b) right FFDM_CC 1.35 mGy (0.33 mGy, 1.12–1.48 mGy), right FFDM_MLO 1.40 mGy (0.36 mGy, 1.14–1.58 mGy). MGD was significantly higher for DBT mlo views compared to FFDM (p < 0.001). Radiation dose was significantly higher for DBT in MLO views compared to FFDM. However, the MGD of DBT MLO lies below the national diagnostic reference level of 2 mGy for an FFDM view. Hence, our results support the use of either DBT or FFDM as suggested in the ECIBC’s Guidelines.

Value of BI-RADS 3 Audits

Objectives

BI-RADS 3 is an established assessment category in which the probability of malignancy is equal to or less than 2%. However, monitoring adherence to imaging criteria can be challenging and there are few established benchmarks for auditing BI-RADS 3 assignments. In this study, we explore some parameters that could serve as useful tools for quality control and clinical practice management.

Materials and Methods

This retrospective study covered a 4-year period (Jan 2014-Dec 2017) and included all women over 40 years who were recalled from a screening exam and had an initial assignment of BI-RADS 3 (probably benign) category after diagnostic workup. A follow-up period of 2 years following the assignment of BI-RADS 3 was used for quantitative quality control metrics.

Results

Among 135,765 screening exams, 13,453 were recalled and 1,037 BI-RADS 3 cases met inclusion criteria. The follow-up rate at 24 months was 86.7%. The upgrade rate was 7.4% (77/1,037) [CI: 5.9-9.2%] and the PPV3 was 33.8% (26/77) [CI: 23.4-45.5%]. The cancer yield was 2.51% (26/1,037) [CI: 1.64-3.65%] and did not differ (p=0. 243) from the 2% probability of malignancy. The initial BI-RADS3 per screening exam and per recall from screening were 0.76% (1,037/135,765) [CI: 0.72-0.81%] and 7.7% (1,037/13,453) [CI: 7.26-8.17%], respectively.

Conclusion

Regular audit of BIRADS 3 metrics has the potential to provide additional insights for clinical practice management. Data from varied clinical settings with input from an expert committee could help establish benchmarks for these metrics.

Leveraging Multi-Task Learning to Cope With Poor and Missing Labels of Mammograms

In breast cancer screening, binary classification of mammograms is a common task aiming to determine whether a case is malignant or benign. A Computer-Aided Diagnosis (CADx) system based on a trainable classifier requires clean data and labels coming from a confirmed diagnosis. Unfortunately, such labels are not easy to obtain in clinical practice, since the histopathological reports of biopsy may not be available alongside mammograms, while normal cases may not have an explicit follow-up confirmation. Such ambiguities result either in reducing the number of samples eligible for training or in a label uncertainty that may decrease the performances. In this work, we maximize the number of samples for training relying on multi-task learning. We design a deep-neural-network-based classifier yielding multiple outputs in one forward pass. The predicted classes include binary malignancy, cancer probability estimation, breast density, and image laterality. Since few samples have all classes available and confirmed, we propose to introduce the uncertainty related to the classes as a per-sample weight during training. Such weighting prevents updating the network's parameters when training on uncertain or missing labels. We evaluate our approach on the public INBreast and private datasets, showing statistically significant improvements compared to baseline and independent state-of-the-art approaches. Moreover, we use mammograms from Susan G. Komen Tissue Bank for fine-tuning, further demonstrating the ability to improve the performances in our multi-task learning setup from raw clinical data. We achieved the binary classification performance of AUC = 80.46 on our private dataset and AUC = 85.23 on the INBreast dataset.

PERFORMANCE COMPARISON OF SYSTEMS WITH FULL-FIELD DIGITAL MAMMOGRAPHY, DIGITAL BREAST TOMOSYNTHESIS AND CONTRAST-ENHANCED SPECTRAL MAMMOGRAPHY

The purpose is to compare full-field digital mammography (FFDM), digital breast tomosynthesis (DBT) and contrast-enhanced spectral mammography (CESM) technologies on three mammography systems in terms of image quality and patient dose. Two Senographe Essential with DBT and CESM (denoted S1 and S2) and one Selenia Dimensions (S3) with FFDM and DBT were considered. Dosimetry methods recommended in the European protocol were used. Image quality was tested with CDMAM in FFDM and DBT and with ideal observer method in FFDM. Mean values of mean glandular dose (MGD) from whole patient samples on S1, S2 and S3 were as follows: FFDM 1.65, 1.84 and 2.23 mGy; DBT 2.03, 1.96 and 2.87 mGy; CESM 2.65 and 3.16 mGy, respectively. S3 exhibited better low-contrast detectability for the smallest sized discs of CDMAM and ideal observer in FFDM, and for the largest sized discs in DBT, at similar dose levels.

Screening Digital Breast Tomosynthesis: Radiation Dose Among Patients With Breast Implants

OBJECTIVE

To compare the mean glandular dose (MGD), cancer detection rate (CDR), and recall rate (RR) among screening examinations of patients with breast implants utilizing various digital breast tomosynthesis (DBT)-based imaging protocols.

METHODS

This IRB-approved retrospective study included 1998 women with breast implants who presented for screening mammography between December 10, 2013, and May 29, 2020. Images were obtained using various protocol combinations of DBT and 2D digital mammography. Data collected included MGD, implant type and position, breast density, BI-RADS final assessment category, CDR, and RR. Statistical analysis utilized type II analysis of variance and the chi-square test.

RESULTS

The highest MGD was observed in the DBT only protocol, while the 2D only protocol had the lowest (10.29 mGy vs 5.88 mGy, respectively). Statistically significant difference in MGD was observed across protocols (P < 0.0001). The highest per-view MGD was among DBT full-field (FF) views in both craniocaudal and mediolateral oblique projections (P < 0.0001). No significant difference was observed in RR among protocols (P = 0.17). The combined 2D (FF only) + DBT implant-displaced (ID) views protocol detected the highest number of cancers (CDR, 7.2 per 1000), but this was not significantly different across protocols (P = 0.48).

CONCLUSION

The combination of 2D FF views and DBT ID views should be considered for women with breast implants in a DBT-based screening practice when aiming to minimize radiation exposure without compromising the sensitivity of cancer detection. Avoidance of DBT FF in this patient population is recommended to minimize radiation dose.

Effect of glandularity on the detection of simulated cancers in planar, tomosynthesis, and synthetic 2D imaging of the breast using a hybrid virtual clinical trial

PURPOSE

The purpose of this study was to measure the threshold diameter of calcifications and masses for 2D imaging, digital breast tomosynthesis (DBT), and synthetic 2D images, for a range of breast glandularities. This study shows the limits of detection for each of the technologies and the strengths and weaknesses of each in terms of visualizing the radiological features of small cancers.

METHODS

Mathematical voxel breast phantoms with glandularities by volume of 9%, 18%, and 30% with a thickness of 53 mm were created. Simulated ill-defined masses and calcification clusters with a range of diameters were inserted into some of these breast models. The imaging characteristics of a Siemens Inspiration X-ray system were measured for a 29 kV, tungsten/rhodium anode/filter combination. Ray tracing through the breast models was undertaken to create simulated 2D and DBT projection images. These were then modified to adjust the image sharpness, and to add scatter and noise. The mean glandular doses for the images were 1.43, 1.47, and 1.47 mGy for 2D and 1.92, 1.97, and 1.98 mGy for DBT for the three glandularities. The resultant images were processed to create 2D, DBT planes and synthetic 2D images. Patches of the images with or without a simulated lesion were extracted, and used in a four-alternative forced choice study to measure the threshold diameters for each imaging mode, lesion type, and glandularity. The study was undertaken by six physicists.

RESULTS

The threshold diameters of the lesions were 6.2, 4.9, and 6.7 mm (masses) and 225, 370, and 399 μm, (calcifications) for 2D, DBT, and synthetic 2D, respectively, for a breast glandularity of 18%. The threshold diameter of ill-defined masses is significantly smaller for DBT than for both 2D (p≤0.006) and synthetic 2D (p≤0.012) for all glandularities. Glandularity has a significant effect on the threshold diameter of masses, even for DBT where there is reduced background structure in the images. The calcification threshold diameters for 2D images were significantly smaller than for DBT and synthetic 2D for all glandularities. There were few significant differences for the threshold diameter of calcifications between glandularities, indicating that the background structure has little effect on the detection of calcifications. We measured larger but nonsignificant differences in the threshold diameters for synthetic 2D imaging than for 2D imaging for masses in the 9% (p = 0.059) and 18% (p = 0.19) glandularities. The threshold diameters for synthetic 2D imaging were larger than for 2D imaging for calcifications (p < 0.001) for all glandularities.

CONCLUSIONS

We have shown that glandularity has only a small effect on the detection of calcifications, but the threshold diameter of masses was significantly larger for higher glandularity for all of the modalities tested. We measured nonsignificantly larger threshold diameters for synthetic 2D imaging than for 2D imaging for masses at the 9% (p = 0.059) and 18% (p = 0.19) glandularities and significantly larger diameters for calcifications (p < 0.001) for all glandularities. The lesions simulated were very subtle and further work is required to examine the clinical effect of not seeing the smallest calcifications in clusters.

A phase sensitive x-ray breast tomosynthesis system: Preliminary patient images with cancer lesions

Phase-sensitive x-ray imaging continues to attract research for its ability to visualize weakly absorbing details like those often encountered in biology and medicine. We have developed and assembled the first inline-based high-energy phase sensitive breast tomosynthesis (PBT) system, which is currently undergoing patient imaging testing at a clinical site. The PBT system consists of a microfocus polychromatic x-ray source and a direct conversion-based flat panel detector coated with a 1 mm thick amorphous selenium layer allowing a high detective quantum efficiency at high energies. The PBT system scans a compressed breast over 15° with 9 angular projection views. The high-energy scan parameters are carefully selected to ensure similar or lower mean glandular dose levels to the clinical standard of care systems. Phase retrieval and data binning are applied to the phase contrast angular projection views and a filtered back-projection algorithm is used to reconstruct the final images. This article reports the distributions of radiation dose versus thickness of the compressed breasts at 59 and 89 kV and sample PBT images acquired from 3 patients. Preliminary PBT images demonstrate the feasibility of this new imaging modality to acquire breast images at lower radiation dose as compared to the clinical digital breast tomosynthesis system with enhanced lesion characteristics (i.e. lesion spiculation and margins).

Assessment of the uterine dose in digital mammography and digital breast tomosynthesis

INTRODUCTION

Digital Mammography (DM-2D) and more recently Digital Breast Tomosynthesis (DBT), are two of the most effective imaging modalities for breast cancer detection, often used in screening programmes. It may happen that exams using these two imaging modalities are inadvertently performed to pregnant women. The objective of this study is to assess the dose in the uterus due to DM-2D and DBT exams, according to two main irradiation scenarios: in the 1st scenario the exposure parameters were pre-selected directly by the imaging system, while in the 2nd scenario, the maximum exposure parameters were chosen.

METHODS

The mammography equipment used was a Siemens Mammomat Inspiration. A physical anthropomorphic phantom, PMMA plates (simulating a breast thickness of 6 cm) and thermoluminescent dosimeters (TLDs) were used to measure entrance air kerma values on the phantom's breast and abdomen in order to successively estimate the mean glandular dose (MGD) and the dose in the uterus. For the two irradiation scenarios chosen, two-breast imaging modalities were selected: 1) DBT in Cranio-Caudal (CC) view (with 28 kV and 160 mAs as exposure parameters), 2) DBT and DM in Medio Lateral-Oblique (MLO) and CC views (with 34 kV and 250 mAs as exposure parameters).

RESULTS

In the 1st scenario, the TLD measurements did not detect significant dose values in the abdomen whereas the MGD estimated using the D.R. Dance model was in close agreement with data available in the literature. In the 2nd scenario, there was no significant difference in MGD estimation between the different views, whereas the air kerma values in the abdomen (in DBT mode, CC and MLO) were 0.049 mGy and 0.004 mGy respectively. In CC DM-2D mode the abdomen air kerma value was 0.026 mGy, with no significant detected value in MLO view.

CONCLUSIONS

For the dose in the uterus, the obtained values seem to indicate that DM-2D and DBT examinations inadvertently performed during pregnancy do not pose a significant radiological risk, even considering the case of overexposure in both breasts.

IMPLICATIONS FOR PRACTICE

The accurate knowledge of the doses in DM-2D and DBT will contribute to raise the awareness among medical practitioners involved in breast imaging empowering them to provide accurate information about dose levels in the uterus, improving their radiation risk communication skills and consequently helping to reduce the anxiety of pregnant women undergoing this type of examinations.

Dedicated breast CT: state of the art-Part II. Clinical application and future outlook

Dedicated breast CT is being increasingly used for breast imaging. This technique provides images with no compression, removal of tissue overlap, rapid acquisition, and available simultaneous assessment of microcalcifications and contrast enhancement. In this second installment in a 2-part review, the current status of clinical applications and ongoing efforts to develop new imaging systems are discussed, with particular emphasis on how to achieve optimized practice including lesion detection and characterization, response to therapy monitoring, density assessment, intervention, and implant evaluation. The potential for future screening with breast CT is also addressed. KEY POINTS: • Dedicated breast CT is an emerging modality with enormous potential in the future of breast imaging by addressing numerous clinical needs from diagnosis to treatment. • Breast CT shows either noninferiority or superiority with mammography and numerical comparability to MRI after contrast administration in diagnostic statistics, demonstrates excellent performance in lesion characterization, density assessment, and intervention, and exhibits promise in implant evaluation, while potential application to breast cancer screening is still controversial. • New imaging modalities such as phase-contrast breast CT, spectral breast CT, and hybrid imaging are in the progress of R & D.

A Data Set and Deep Learning Algorithm for the Detection of Masses and Architectural Distortions in Digital Breast Tomosynthesis Images

IMPORTANCE

Breast cancer screening is among the most common radiological tasks, with more than 39 million examinations performed each year. While it has been among the most studied medical imaging applications of artificial intelligence, the development and evaluation of algorithms are hindered by the lack of well-annotated, large-scale publicly available data sets.

OBJECTIVES

To curate, annotate, and make publicly available a large-scale data set of digital breast tomosynthesis (DBT) images to facilitate the development and evaluation of artificial intelligence algorithms for breast cancer screening; to develop a baseline deep learning model for breast cancer detection; and to test this model using the data set to serve as a baseline for future research.

DESIGN, SETTING, AND PARTICIPANTS

In this diagnostic study, 16 802 DBT examinations with at least 1 reconstruction view available, performed between August 26, 2014, and January 29, 2018, were obtained from Duke Health System and analyzed. From the initial cohort, examinations were divided into 4 groups and split into training and test sets for the development and evaluation of a deep learning model. Images with foreign objects or spot compression views were excluded. Data analysis was conducted from January 2018 to October 2020.

EXPOSURES

Screening DBT.

MAIN OUTCOMES AND MEASURES

The detection algorithm was evaluated with breast-based free-response receiver operating characteristic curve and sensitivity at 2 false positives per volume.

RESULTS

The curated data set contained 22 032 reconstructed DBT volumes that belonged to 5610 studies from 5060 patients with a mean (SD) age of 55 (11) years and 5059 (100.0%) women. This included 4 groups of studies: (1) 5129 (91.4%) normal studies; (2) 280 (5.0%) actionable studies, for which where additional imaging was needed but no biopsy was performed; (3) 112 (2.0%) benign biopsied studies; and (4) 89 studies (1.6%) with cancer. Our data set included masses and architectural distortions that were annotated by 2 experienced radiologists. Our deep learning model reached breast-based sensitivity of 65% (39 of 60; 95% CI, 56%-74%) at 2 false positives per DBT volume on a test set of 460 examinations from 418 patients.

CONCLUSIONS AND RELEVANCE

The large, diverse, and curated data set presented in this study could facilitate the development and evaluation of artificial intelligence algorithms for breast cancer screening by providing data for training as well as a common set of cases for model validation. The performance of the model developed in this study showed that the task remains challenging; its performance could serve as a baseline for future model development.

Theoretical investigation of the signal performance of HgI2x-ray converters incorporating a Frisch grid structure at mammographic energies

Active matrix, flat-panel imagers (AMFPIs) suffer from decreased detective quantum efficiency under conditions of low dose per image frame (such as for digital breast tomosynthesis, fluoroscopy and cone-beam CT) due to low signal compared to the additive electronic noise. One way to address this challenge is to introduce a high-gain x-ray converter called particle-in-binder mercuric iodide (PIB HgI₂) which exhibits 3-10 times higher x-ray sensitivity compared to that of a-Se and CsI:Tl converters employed in commercial AMFPI systems. However, a remaining challenge for practical implementation of PIB HgI₂is the high level of image lag, which is believed to largely originate from the trapping of holes. Towards addressing this challenge, this paper reports a theoretical investigation of the use of a Frisch grid structure embedded in the converter to suppress hole signal-which would be expected to reduce image lag. The grid acts as a third electrode sandwiched between a continuous top electrode and pixelated bottom electrodes having a 100μm pitch. Signal properties of such a detector are investigated as a function of VDR (the ratio of the voltage difference between the electrodes in the region below the grid to that above the grid), grid pitch (the center-to-center distance between two neighboring grid wires) andR_GRID(the ratio of grid wire width to grid pitch) for mammographic x-ray energies. The results show that smaller grid pitch suppresses hole signal to a higher degree (up to ∼96%) while a larger gap between grid wires and higher VDR provide minimally impeded electron transport. Examination of the tradeoff between maximizing electron signal and minimizing hole signal indicates that a grid design having a grid pitch of 20μm withR_GRIDof 50% and 65% provides hole signal suppression of ∼93% and ∼95% for VDR of 1 and 3, respectively.

Digital Breast Tomosynthesis: Potential Benefits in Routine Clinical Practice

Digital breast tomosynthesis (DBT) is gradually being implemented in routine clinical breast imaging practice. The technique of image acquisition reduces the confounding effect of overlapping breast tissue, which substantially affects cancer detection, abnormal recall, and interval cancer rates in a screening/ surveillance setting. In a diagnostic setting, tomosynthesis also allows for improved lesion localization and characterization over conventional imaging, which potentially improves the accuracy and improved workflow efficiency. To optimize the utility of tomosynthesis, imagers should be aware of the pertinent aspects of image acquisition as it relates to interpretation, the appearance of benign and malignant pathologies, and sources of possible misinterpretation. This article aims to provide a practical knowledge base of DBT and demonstrate its potential benefits when incorporated into routine clinical practice.

Radiomics: A Primer for Breast Radiologists

Radiomics has a long-standing history in breast imaging with computer-aided detection (CAD) for screening mammography developed in the late 20th century. Although conventional CAD had widespread adoption, the clinical benefits for experienced breast radiologists were debatable due to high false-positive marks and subsequent increased recall rates. The dramatic growth in recent years of artificial intelligence-based analysis, including machine learning and deep learning, has provided numerous opportunities for improved modern radiomics work in breast imaging. There has been extensive radiomics work in mammography, digital breast tomosynthesis, MRI, ultrasound, PET-CT, and combined multimodality imaging. Specific radiomics outcomes of interest have been diverse, including CAD, prediction of response to neoadjuvant therapy, lesion classification, and survival, among other outcomes. Additionally, the radiogenomics subfield that correlates radiomics features with genetics has been very proliferative, in parallel with the clinical validation of breast cancer molecular subtypes and gene expression assays. Despite the promise of radiomics, there are important challenges related to image normalization, limited large unbiased data sets, and lack of external validation. Much of the radiomics work to date has been exploratory using single-institution retrospective series for analysis, but several promising lines of investigation have made the leap to clinical practice with commercially available products. As a result, breast radiologists will increasingly be incorporating radiomics-based tools into their daily practice in the near future. Therefore, breast radiologists must have a broad understanding of the scope, applications, and limitations of radiomics work.

Digital Breast Tomosynthesis: an Overview

Breast cancer is emerging as the most common malignancy in Indian women. Mammography is one of the few screening modalities available to the modern world that has proved itself of much use by aiding early detection and treatment of non-palpable, node-negative breast cancers. However, due to its two-dimensional nature, many cases of malignancies are still missed, to be detected at a later date or by an alternate modality. In 2011, FDA approved the supplemental use of digital breast tomosynthesis (DBT) in screening and diagnostic set ups. The acquisition of multiple low-dose projection images of the compressed parenchyma provided a 'third' dimension to the mammogram whereby the breast tissue could be seen layer by layer on the workstation. It improves cancer detection rate, and reduces recall rate and false-positive findings by improving lesion characterization. The current review discusses the principle of DBT with a comprehensive study of the literature.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13193-021-01310-y.