[Syllabus]: DBT: Challenges and Pitfalls

Steven Poplack, MD, FACR, FSBI

Steven Poplack, MD, FACR, FSBI(bio)

  • Professor of Radiology
  • Department of Radiology, Stanford University School of Medicine

Introduction

This syllabus presents a case-based, practice-oriented review of digital breast tomosynthesis (DBT) with emphasis on interpretation challenges, pitfalls, and practical management. Through real-world examples, it addresses how to leverage DBT to reduce unnecessary recalls, improve specificity, and increase cancer detection, while avoiding common errors. Topics include distinguishing superimposition from true findings, recognizing architectural distortion, integrating synthetic mammography with DBT, optimizing diagnostic workflows, selecting and executing biopsy strategies, ensuring radiology–pathology concordance, managing false negatives, and using adjunct MRI and ultrasound intelligently.

Principles and Benefits of Digital Breast Tomosynthesis (DBT)

DBT enhances lesion conspicuity by reducing summation artifacts, supporting both lower recall rates and higher cancer detection. Its value is maximized when readers confidently dismiss superimposed parenchyma while recognizing subtle true abnormalities that are missed or equivocal on planar images.

Key Points

  • Expected benefits: decreased recall rate, increased specificity, increased sensitivity/cancer detection.
  • DBT introduces new and different information beyond the planar (digital or synthetic) mammogram.
  • Reader training is essential to avoid increasing recalls for non-genuine findings.

Screening Workflow: Binary Decisions at Screening

Screening DBT management should remain binary—either pass (BI-RADS 1–2) or recall (BI-RADS 0). Probably benign (BI-RADS 3) and biopsy recommendations (BI-RADS 4–5) should not originate from the screening setting.

Key Points

  • Screening decisions: pass vs recall; do not assign BI-RADS 3 or recommend biopsy at screening.
  • Reserve risk stratification and intervention planning for the diagnostic setting.
  • Establish a structured approach to track recalls and correlate outcomes for continuous learning.

The Learning Curve: Superimposition vs Genuine Abnormality

The central interpretive challenge is distinguishing tissue overlap from true pathology (asymmetry/mass/distortion). Failure to use DBT to confidently override summation artifacts negates specificity gains.

Key Points

  • Superimposition typically shows alternating dense and fatty slices, curvilinear streaming, and non-reproduction of planar size/shape (“wait for it and it never arrives”).
  • Genuine asymmetry persists across multiple contiguous slices with convex outward margins and size parity with the planar finding.
  • Practice reconstructing the planar composite from component DBT slices (mass + ligaments/vessels).

Integrating Planar and Tomosynthesis Findings

Correlating synthetic or digital mammography with DBT requires mental reconstruction of the planar appearance from slice-by-slice anatomy to determine whether an apparent irregularity is summation or a true lesion.

Key Points

  • Identify structures in focus across slices (e.g., ligaments, ducts, vessels) and the mass core.
  • Confirm whether DBT reproduces the planar finding’s morphology and dimensions.
  • Use both CC and MLO DBT datasets to validate two-view findings when possible.

Synthetic Mammography vs Digital Mammography: Visibility Pitfalls

Low-density asymmetries may be less conspicuous on synthetic mammograms than on digital mammograms; however, DBT typically reveals a conspicuous correlate.

Key Points

  • Do not rely on synthetic images alone when they appear normal.
  • Always review DBT even when planar images are unremarkable.
  • Expect DBT to “bail out” synthetic limitations but maintain vigilance for subtle planar cues.

Identifying and Managing Focal Asymmetry

Focal asymmetry on DBT is defined by two-view presence or multi-slice persistence with features exceeding simple overlap. Subtle distortion may coexist.

Key Points

  • Multi-slice, multi-view reproduction supports a genuine focal asymmetry.
  • Associated distortion increases suspicion and typically mandates recall.
  • Ultrasound correlation is pursued after completing diagnostic mammographic problem-solving.

Architectural Distortion: Recognition and Differential Diagnosis

DBT excels at unveiling architectural distortion—straight dense and lucent lines converging on a central point—often occult or equivocal on planar views.

Key Points

  • True distortion: converging straight lines with central focal point; planar “tenting” at the fat–glandular junction is a valuable sign.
  • Superimposition: curvilinear lines that traverse uninterrupted, with admixture of fatty slices.
  • Differential: invasive carcinoma (IDC/ILC), DCIS, radial scar/complex sclerosing lesion, sclerosing adenosis; malignancy rates for DBT-only distortions range ~20–50% in series.

Distortion vs Tissue Disorganization

Some early cancers present as subtle tissue disorganization lacking a discrete mass or classic linear convergence. These require a high index of suspicion and longitudinal comparison.

Key Points

  • Curvier, less orderly convergence can still represent early invasive cancer.
  • Compare with priors for interval development or persistent disorganization.
  • Small sub-centimeter distortions may be difficult to reproduce on diagnostic planar views—do not dismiss if convincing on screening DBT.

Completing Diagnostic Mammography Before Ultrasound

Diagnostic workup should localize and characterize the finding mammographically before using ultrasound, which may be negative and cannot “rescue” an incomplete mammographic assessment.

Key Points

  • Establish location and suspicion level and plan management assuming ultrasound might be negative.
  • If ultrasound is negative but DBT findings remain suspicious, proceed with stereotactic/DBT-guided sampling or alternative localization.
  • Do not allow a negative ultrasound to negate a convincing DBT abnormality.

Rolled CC Views: Technique and Utility

Rolled CC views, combined with DBT, alter tissue paths within the CC orientation and are particularly useful to confirm genuineness and localize lesions in the mediolateral plane.

Key Points

  • Roll upper breast medially or laterally; name based on direction upper breast is positioned.
  • Reproduction on both rolled views indicates a genuine lesion; directional motion localizes to the upper breast.
  • Often more informative than simply repeating CC spot compression, which may reproduce initial positioning.

Ultrasound Correlates: When and How to Use US

Use ultrasound after completing diagnostic mammography to characterize correlates and guide biopsy. Some DBT-only distortions will lack US correlates, which informs biopsy route selection.

Key Points

  • Targeted US is first-line for palpable or localized mammographic findings once site is defined.
  • Absence of a US correlate in the setting of suspicious DBT distortion favors stereotactic/DBT-guided sampling.
  • Recognize lesion-specific sonographic appearances (e.g., mucinous carcinoma: oval, circumscribed, nearly isoechoic to fat, lacks internal septa, remains discrete when turned).

Biopsy Strategy for DBT-Detected Findings

Biopsy method selection depends on visibility across modalities and available equipment. DBT-guided systems streamline accurate targeting of DBT-only abnormalities.

Key Points

  • Preferred: DBT-guided or stereotactic biopsy for DBT-only distortion/masses without clear US correlates.
  • If DBT-guided biopsy is unavailable, stereotactic biopsy may be performed using reliable landmarks (e.g., calcifications) with post-biopsy DBT confirmation.
  • Reserve ultrasound-guided biopsy for confident US correlates; avoid sampling likely non-correlates.

Radiology–Pathology Concordance for Distortion

Concordance assessment is critical. Certain benign pathologies are discordant with true architectural distortion and require additional action.

Key Points

  • Concordant benign diagnoses for true distortion: radial scar/complex sclerosing lesion, sclerosing adenosis, occasionally sclerosing papilloma.
  • Discordant benign diagnoses: stromal fibrosis, nonspecific fibrocystic change for a DBT-demonstrated true distortion.
  • Discordant outcomes necessitate repeat sampling (preferably image-guided at the original target) or surgical excision; avoid short-interval follow-up.

Intramammary Lymph Node Mimics

Do not overcall fat-containing reniform lesions in the axillary tail as intramammary lymph nodes without confirming a fatty hilum and appropriate location.

Key Points

  • Atypical location (too medial) or lack of definite fat should prompt targeted ultrasound.
  • Adjacent fat can mimic a hilum; paired cysts or other masses may simulate nodal morphology.
  • Confirm with US morphology (cortex/hilum) before assigning nodal etiology.

Fat-Containing Lesions on DBT: Not Always Benign

DBT may depict fat within or adjacent to a lesion; fat alone does not confer benignity if margins are suspicious.

Key Points

  • “No-touch” fat-containing lesions should also have circumscribed margins (e.g., hamartoma, intramammary lymph node, fat necrosis).
  • Indistinct or spiculated margins despite fat signal malignancy until proven otherwise (e.g., ILC engulfing fat).
  • Pursue full diagnostic workup and biopsy if morphology is suspicious.

False-Negative DBT: Within-Slice Density Obscuration

Cancers fully embedded within dense parenchyma may be occult on DBT if the fat–glandular interface is undisturbed on the slice.

Key Points

  • Palpable or focal clinical abnormalities require targeted ultrasound even when DBT is negative.
  • Expect occult DBT presentation for large intraparenchymal masses lacking calcifications/distortion.
  • Maintain a low threshold for adjunct imaging when clinical suspicion persists.

MRI in Problem Solving and Staging

MRI is seldom used as a problem-solver but can be valuable when biopsy is indicated but not technically feasible, and for staging to define disease extent. Second-look ultrasound should target MRI-detected lesions for potential biopsy.

Key Points

  • Use MRI to facilitate biopsy planning/localization when DBT-only abnormalities are not amenable to stereotactic/DBT-guided biopsy.
  • Biopsy additional MRI-detected lesions that may alter management; do not proceed directly to mastectomy without tissue confirmation when feasible.
  • Second-look ultrasound can identify correlates for US-guided biopsy; avoid overcalling MR-only non-mass enhancement (e.g., apocrine metaplasia/fibrocystic change).

Clip Placement and Post-Biopsy Verification

Post-biopsy imaging must confirm accurate targeting and clip placement relative to the original DBT finding to avoid sampling error.

Key Points

  • Compare post-biopsy views (or DBT) with initial scout to match landmarks and ensure same slice/location.
  • If the clip does not correspond to the DBT target, treat as discordant and re-biopsy/relocate as indicated.
  • Documentation of concordance supports appropriate follow-up or definitive management.

Practical Case-Based Pearls

A set of high-yield operational insights informed by case experience.

Key Points

  • Use DBT deliberately to reduce recalls for superimposition; otherwise, specificity gains are lost.
  • Low-density asymmetries may be occult on synthetic images—DBT review is mandatory.
  • Rolled CC DBT views help confirm genuineness and localize upper-breast lesions.
  • Small DBT-detected distortions can be hard to reproduce on diagnostic planar images—do not exclude based on planar negativity.
  • DBT-only distortions carry meaningful malignancy risk; sample when suspicious even without US or MRI correlates.

Conclusion

DBT transforms screening and diagnostic breast imaging by revealing subtle, clinically significant findings while enabling recall reduction—provided readers master the distinction between superimposition and true pathology. Optimal performance requires structured workflows: complete diagnostic mammography before ultrasound, use rolled views judiciously, select biopsy techniques aligned with imaging visibility, and enforce rigorous radiology–pathology concordance. Recognizing pitfalls—synthetic mammography limitations, false negatives from within-slice density, nodal mimics, and the fallacy of “fat equals benign”—ensures accurate, efficient care. Thoughtful integration of adjunct ultrasound and MRI, with targeted tissue sampling when management may change, underpins safe, evidence-based practice in the era of tomosynthesis.

Ready to earn credit? Sign up for the Course.