Introduction
This session (Part 2) focuses on practical, clinic-facing aspects of full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT). It addresses frequently asked questions on dose, disadvantages, accreditation, whether DBT will replace FFDM, artifacts, quality control (QC), and radiation safety. Emphasis is placed on protocol choices that affect dose and workflow, artifact recognition and mitigation, accreditation/QC expectations, and the operational realities of implementing and sustaining DBT in clinical practice.
Reframing “3D”: Tomosynthesis Is Not Volumetric CT
DBT provides limited-depth resolution and should not be equated with true volumetric CT. Using the ACR accreditation phantom and high-resolution CT demonstrates that DBT sharpens a target plane while blurring above and below it, leading to out-of-plane persistence of high-contrast objects.
Key Points
- DBT has limited z-axis resolution with plane-dependent sharpness; objects persist as blur above/below the focal plane.
- High-contrast structures (e.g., clips, discs) can produce extensive through-plane artifacts.
- CT resolves masses, microcalcifications, and fibers volumetrically; DBT cannot be reformatted into orthogonal planes by the user on most systems.
- The ACR phantom features appear sharply at ~7 mm; DBT typically visualizes 4–5 masses, 4 calcification groups (rarely 5), and 4–6 fibers depending on windowing.
DBT Image Formation and Z-Resolution Artifacts
DBT reconstruction creates tomographic planes with effective slice width that broadens away from the focal plane, generating apparent “ghosts” for structures outside the plane.
Key Points
- Out-of-plane persistence reflects reconstruction/back-projection behavior, not true anatomy.
- Effective slice thickness increases with distance from the in-focus plane.
- Vertical persistence of a high-contrast object (e.g., aluminum sphere, acrylic disc) is an artifact of limited angular sampling.
- Recognize that DBT’s strengths are reduced tissue overlap; depth resolution remains constrained.
Mean Glandular Dose: DBT vs FFDM and the Role of Protocol
Modern DBT systems deliver per-view mean glandular dose (MGD) comparable to FFDM. The key determinant of total exam dose is the clinical protocol (e.g., combo mode vs synthetic 2D).
Key Points
- Typical phantom MGD: FFDM ~1.2 mGy/view; DBT ~1.4–1.5 mGy/view; historic “3 mGy/view” is an outdated film-era cap.
- Patient-specific factors and beam quality matter; tungsten target with appropriate filtration and higher kVp can reduce MGD in some breasts.
- Removal of the anti-scatter grid during DBT can lower dose relative to FFDM for some vendors.
- Protocol choice (2D only, 2D+DBT combo, DBT with synthetic 2D) drives total dose and compression time.
Vendor Configurations, Beam Quality, and Evidence on Dose
Marketing materials highlight system-specific advantages (target/filter, detector type). Independent analyses show overall comparability across vendors under clinical conditions.
Key Points
- Vendor claims (e.g., tungsten/rhodium vs molybdenum-based spectra) are condition-dependent and often use acrylic proxies.
- AAPM TG-223 (2014) concluded relative glandular doses are broadly comparable across vendors over typical thicknesses/compositions and scan angles.
- Breast composition is heterogeneous; acrylic or “50/50” breast assumptions are imperfect surrogates.
Acquisition Protocols, Synthetic 2D, and Workflow-Dose Trade-offs
The mix of 2D and DBT acquisitions influences dose, time under compression, and reading workflow.
Key Points
- 2D-only: ~1.2 mGy/view; 1–2 seconds.
- Combo (2D+DBT same view): additive dose (~2.65 mGy phantom example) and ~10 s compression; still within legacy limits but higher than single-acquisition approaches.
- DBT with synthetic 2D: single DBT acquisition (~1.4–1.5 mGy/view) with software-generated 2D; reduces dose vs combo.
- Availability: Hologic synthetic 2D FDA-cleared; other vendors’ synthetic approaches may be for guidance only (check current regulatory status).
- Mammography and DBT remain low-dose relative to many radiologic exams; effective dose is negligible in terms of long-term risk.
Practical Disadvantages and Operational Impacts of DBT
While clinically advantageous, DBT introduces complexity, cost, and workflow considerations.
Key Points
- Longer acquisition and interpretation time; more images and larger datasets.
- Increased equipment and maintenance costs; specialized diagnostic workstations required.
- PACS/storage/network demands escalate; plan for bandwidth and archive growth.
- QC is more complex and time-intensive; synthetic 2D may underestimate breast density vs FFDM.
- Microcalcifications may span multiple planes; interpretive adaptation needed.
Common Digital Mammography and DBT Artifacts
Artifacts arise from detector nonuniformity, post-processing, environment, motion, and display issues; these affect both FFDM and DBT.
Key Points
- Detector tiling/uniformity: tile gain variations, gaps; corrected by flat-fielding; delamination requires detector replacement.
- Dead pixels/columns: may be software-mapped; extensive failures necessitate service and potential detector swap.
- Environmental effects: temperature extremes can cause severe artifacts; stabilize room and system before imaging.
- Ghosting (a-Se): short-term “burn-in” and long-term nonuniformity; exacerbated by cold rooms and repeated exposures; may require detector replacement.
- Collimator alignment: verify light vs x-ray field; misalignment may follow service—requires prompt check and physics verification within 30 days.
- Motion blur: more evident on radiologist’s diagnostic workstation; retake if suspected.
- Monitor calibration: acquisition and diagnostic displays require periodic QC (grayscale, contrast patches).
- Grid artifacts: non-reciprocating bucky produces gridline shadows (FFDM); DBT uses linear grids, not cellular.
DBT-Specific Artifacts and Reconstruction Solutions
Certain artifacts are unique to DBT’s angular acquisition and reconstruction.
Key Points
- Zipper artifact: out-of-plane high-density objects (e.g., clips, coarse calcifications) create directional streaks along tube motion; reduced by iterative reconstruction.
- Truncation: incomplete breast coverage across projections produces edge artifacts; improved with better positioning/centering.
- Plane persistence: masses may appear across many planes (more in narrow-angle systems); wide-angle improves mass conspicuity but can slightly soften calcification sharpness.
- Synthetic 2D can inherit out-of-plane artifacts unless vendor algorithms suppress them (e.g., advanced filtering in newer systems).
Accreditation, Training, and Quality Control for DBT
DBT accreditation is distinct from FFDM but follows analogous processes; in the absence of universal DBT QC standards, use vendor protocols.
Key Points
- Accrediting bodies (FDA, ACR, states) now uniformly accredit DBT; DBT and FFDM are accredited separately.
- No universal DBT QC/dose standard; follow manufacturer QC manuals and data forms.
- Synthetic 2D acceptance varies by vendor/regulator; not all synthetic products are accreditation substitutes for 2D.
- Training: Radiologists—≥8 hours (including ≥6 hours hands-on for DBT reading); Physicists—≥8 hours (hands-on/online). Vendors commonly provide accredited training.
- ACR’s newer digital mammography phantom/manual enhances FFDM QC sensitivity but does not apply to DBT.
Phantom Scoring, CNR/SNR, and Interpreting QC Outcomes
QC scoring and thresholds differ between FFDM and DBT, reflecting DBT’s different acquisition and resolution characteristics.
Key Points
- FFDM ACR phantom passing: ≥5 fibers, ≥4 speck groups, ≥4 masses (max score 16).
- DBT QC includes SNR/CNR on the disc; typical thresholds (e.g., CNR > 40; ≤15% CNR variability) are rarely failed on modern systems.
- DBT phantoms are scored by scrolling to the sharp plane (~7 mm) and counting features; pixel binning can reduce in-plane resolution vs FFDM.
- High-resolution DBT modes (reduced binning) improve planar sharpness but remain subject to motion and z-resolution limits.
Corrective Actions and When to Involve the Medical Physicist
Know which failures halt clinical use immediately vs those with a 30-day correction window.
Key Points
- Immediate repair/hold exams: failure of system resolution (line pairs), excessive entrance exposure/MGD, failed phantom image quality (tech or physicist), failed SNR/CNR, flat-field calibration failure, compression force out of limits, critical display/printer failures.
- 30-day correction: unit assembly (collimation, angulation, displays), artifact evaluations (including ghosting), beam quality/HVL, AEC performance and reproducibility, output rate, visual checks, compression thickness indicator, repeat analysis review.
- Annual physics survey required; additional physics evaluation after repairs/replacements affecting dose or image quality.
- Common MQSA inspection failures: personnel qualifications/CMEs, report content, non-adherence to vendor QC manuals, infection control, and documented cases of falsified QC logs—allocate sufficient time and resources for QC.
Will Tomosynthesis Replace FFDM? Clinical, Operational, and Market Factors
DBT adoption is driven by clinical performance and patient demand, but FFDM retains roles, especially for calcifications and resource-constrained settings.
Key Points
- DBT improves cancer detection vs FFDM by reducing tissue overlap; can also acquire standard 2D.
- Calcification assessment may be more straightforward on FFDM; DBT slices distribute calc clusters across planes.
- DBT increases dose/time when added to 2D; synthetic 2D mitigates this and is increasingly adopted.
- Operational costs (storage, network, maintenance) and reading time are higher with DBT.
- Market adoption is substantial; a large and growing fraction of accredited units are DBT-capable.
Radiation Safety and Shielding for DBT Suites
Radiation protection principles are largely shared with FFDM, with attention to DBT’s angular exposures and workload growth.
Key Points
- Scatter originates from the patient; position the operator behind a protective barrier between the gantry and workstation; maximize distance.
- Use face shields and proper collimation; avoid under/over-collimation.
- Shielding design (including adjacent rooms, above/below floors) must be reviewed by a qualified medical physicist—required for new, replacement, or relocated units.
- DBT uses higher kVp and produces a broader scatter distribution than FFDM, but additional room shielding is rarely required; typical walls (e.g., two layers of gypsum board) are often adequate with sufficient distance.
- Accurately project workload; rapid volume growth can alter shielding adequacy assumptions and occupancy considerations.
Conclusion
Implementing DBT delivers meaningful clinical gains but requires informed protocol selection, rigorous QC, and thoughtful operational planning. Per-view dose is broadly comparable to FFDM; total exam dose hinges on acquisition strategy and the use of synthetic 2D. Recognizing DBT’s depth-resolution limits and characteristic artifacts is essential for accurate interpretation and efficient troubleshooting. Separate accreditation and vendor-driven QC remain the standard for DBT, with defined thresholds for corrective action and physics involvement. Finally, routine radiation safety practices, combined with realistic workload assessments, ensure safe and compliant DBT suite operations.
