Subtopic Deep Dive
Digital Breast Tomosynthesis
Research Guide
What is Digital Breast Tomosynthesis?
Digital Breast Tomosynthesis (DBT) is a limited-angle tomographic imaging technique that acquires multiple low-dose projection images of the breast to reconstruct pseudo-3D slices, reducing tissue superposition artifacts compared to 2D mammography.
DBT improves lesion conspicuity and cancer detection rates while lowering recall rates in screening. Over 10 key papers since 2006, including Sechopoulos (2013, 417 citations) on image acquisition and Lång et al. (2015, 283 citations) on standalone screening performance. Clinical trials show DBT + mammography combo reduces false positives by 15-40%.
Why It Matters
DBT lowers false-positive recalls by 37% and increases cancer detection by 1.2 per 1000 screens versus digital mammography alone (Skaane et al., 2013). It enhances detection in dense breasts, critical for 40-50% of women (Mann et al., 2022). AI integration boosts mass detection sensitivity to 0.91 AUC (Samala et al., 2016). Combo-mode with synthetic 2D cuts radiation dose while maintaining efficacy (Comstock et al., 2020).
Key Research Challenges
Limited Angular Range Reconstruction
DBT uses 15-50° arcs, causing out-of-plane artifacts in reconstructions. Zhang et al. (2006, 280 citations) compared cone-beam methods, finding simultaneous algebraic reconstruction most robust but computationally intensive. Balancing resolution and artifact reduction remains unresolved.
Radiation Dose Optimization
DBT delivers 1.5-2.5 times dose of 2-view mammography despite low-dose projections. Svahn et al. (2014, 223 citations) reviewed dose estimates, noting variability by vendor. Minimizing dose without sensitivity loss challenges screening protocols.
Reading Workflow Efficiency
Interpreting 50-100 slices per breast increases radiologist time by 20-30%. Chong et al. (2019) outlined clinical practice, highlighting combo-mode needs. AI prioritization could help but requires validation (Geras et al., 2019).
Essential Papers
Screening for Breast Cancer: U.S. Preventive Services Task Force Recommendation Statement
Albert L. Siu, on behalf of the U.S. Preventive Services Task Force · 2016 · Annals of Internal Medicine · 2.1K citations
The USPSTF recommends biennial screening mammography for women aged 50 to 74 years. (B recommendation) The decision to start screening mammography in women prior to age 50 years should be an indivi...
Comparison of Abbreviated Breast MRI vs Digital Breast Tomosynthesis for Breast Cancer Detection Among Women With Dense Breasts Undergoing Screening
Christopher Comstock, Constantine Gatsonis, Gillian M. Newstead et al. · 2020 · JAMA · 430 citations
ClinicalTrials.gov Identifier: NCT02933489.
A review of breast tomosynthesis. Part I. The image acquisition process
Ioannis Sechopoulos · 2013 · Medical Physics · 417 citations
Mammography is a very well‐established imaging modality for the early detection and diagnosis of breast cancer. However, since the introduction of digital imaging to the realm of radiology, more ad...
Breast cancer screening in women with extremely dense breasts recommendations of the European Society of Breast Imaging (EUSOBI)
Ritse M. Mann, Alexandra Athanasiou, Pascal Baltzer et al. · 2022 · European Radiology · 363 citations
Mass detection in digital breast tomosynthesis: Deep convolutional neural network with transfer learning from mammography
Ravi K. Samala, Heang‐Ping Chan, Lubomir M. Hadjiiski et al. · 2016 · Medical Physics · 287 citations
Purpose: Develop a computer‐aided detection (CAD) system for masses in digital breast tomosynthesis (DBT) volume using a deep convolutional neural network (DCNN) with transfer learning from mammogr...
An overview of mammographic density and its association with breast cancer
Shayan Nazari, Pinku Mukherjee · 2018 · Breast Cancer · 286 citations
Artificial Intelligence for Mammography and Digital Breast Tomosynthesis: Current Concepts and Future Perspectives
Krzysztof J. Geras, Ritse M. Mann, Linda Moy · 2019 · Radiology · 284 citations
Although computer-aided diagnosis (CAD) is widely used in mammography, conventional CAD programs that use prompts to indicate potential cancers on the mammograms have not led to an improvement in d...
Reading Guide
Foundational Papers
Start with Sechopoulos (2013) for acquisition principles (417 citations), then Zhang et al. (2006) for reconstruction algorithms (280 citations), followed by Svahn (2014) for dose metrics (223 citations) to build technical base.
Recent Advances
Study Comstock (2020, 430 citations) for MRI vs DBT comparison in dense breasts, Mann (2022, 363 citations) for EUSOBI guidelines, Chong (2019, 232 citations) for clinical practice.
Core Methods
Projection acquisition (9-25 arcs, 15-50°), simultaneous algebraic reconstruction (SART), filtered back-projection (FBP), AI CAD with DCNN transfer learning (Samala 2016), combo-mode with synthetic 2D.
How PapersFlow Helps You Research Digital Breast Tomosynthesis
Discover & Search
Research Agent uses citationGraph on Sechopoulos (2013) to map 400+ tomosynthesis papers, revealing reconstruction (Zhang 2006) and AI detection (Samala 2016) clusters. exaSearch queries 'DBT dose reduction trials' yield 50+ results with filters for >200 citations. findSimilarPapers on Lång (2015) surfaces 20 population trials like Skaane (2013).
Analyze & Verify
Analysis Agent runs readPaperContent on Comstock (2020) to extract AUC metrics from dense breast trial tables. verifyResponse (CoVe) cross-checks claims against Mann (2022) for EUSOBI guidelines. runPythonAnalysis loads DBT sensitivity data from Samala (2016) CSV extracts, computes GRADE B evidence via meta-analysis with pandas (mean sensitivity 0.88, p<0.01).
Synthesize & Write
Synthesis Agent detects gaps like 'standalone DBT vs combo in Asian cohorts' from 20 papers, flags contradictions between Svahn (2014) dose data and vendor claims. Writing Agent uses latexSyncCitations to compile 15-paper review with auto-numbered DBT workflow diagrams via latexGenerateFigure. exportMermaid visualizes trial comparisons (FFDM vs DBT recall rates).
Use Cases
"Compare mass detection AUC of DCNN models in DBT vs mammography from recent papers"
Research Agent → searchPapers('DBT DCNN mass detection') → Analysis Agent → runPythonAnalysis(pandas meta-analysis of Samala 2016 + Geras 2019 AUCs) → outputs CSV with pooled 0.91 AUC and forest plot
"Draft LaTeX section on DBT reconstruction artifacts citing Zhang 2006 and Sechopoulos 2013"
Synthesis Agent → gap detection → Writing Agent → latexEditText + latexSyncCitations(10 papers) → latexCompile → researcher gets PDF-ready subsection with equations and cited figures
"Find GitHub repos implementing DBT cone-beam reconstruction from Medical Physics papers"
Research Agent → paperExtractUrls(Zhang 2006) → Code Discovery → paperFindGithubRepo → githubRepoInspect → outputs 3 repos with ALRT recon code, README summaries, and runPythonAnalysis benchmarks
Automated Workflows
Deep Research workflow conducts systematic review: searchPapers(100 DBT screening papers) → citationGraph → DeepScan(7-step GRADE grading on sensitivity/recall) → structured report ranking Lång (2015) highest evidence. Theorizer generates hypotheses like 'AI-synthesized 2D optimizes combo DBT' from Comstock (2020) + Geras (2019). DeepScan verifies dose claims across Svahn (2014) and 15 trials via CoVe chain.
Frequently Asked Questions
What defines Digital Breast Tomosynthesis?
DBT acquires 15-25 projections over 15-50° arc, reconstructing 1mm slices to eliminate summation artifacts (Sechopoulos, 2013).
What are core DBT reconstruction methods?
Limited-angle cone-beam methods include FBP, SART, and ALRT; Zhang et al. (2006) showed ALRT best for artifact reduction.
What are key papers on DBT screening performance?
Lång et al. (2015, 283 citations) proved one-view DBT viable standalone; Skaane et al. (2013) showed combo reduces recalls 37%.
What open problems exist in DBT research?
Dose standardization across vendors (Svahn 2014), AI workflow integration (Geras 2019), and standalone vs combo validation in diverse populations.
Research Digital Radiography and Breast Imaging with AI
PapersFlow provides specialized AI tools for Medicine researchers. Here are the most relevant for this topic:
Systematic Review
AI-powered evidence synthesis with documented search strategies
AI Literature Review
Automate paper discovery and synthesis across 474M+ papers
Find Disagreement
Discover conflicting findings and counter-evidence
Paper Summarizer
Get structured summaries of any paper in seconds
See how researchers in Health & Medicine use PapersFlow
Field-specific workflows, example queries, and use cases.
Start Researching Digital Breast Tomosynthesis with AI
Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.
See how PapersFlow works for Medicine researchers