BI-RADS Assessment Categories

The BI-RADS assessment system assigns a category (0–6) to each breast MRI examination, paired with a concordant management recommendation and an estimated likelihood of cancer.

Assessment Category Table

CategoryNameManagementCancer Likelihood
0IncompleteAdditional imaging or prior comparisonN/A
1NegativeRoutine screening (including MRI per guidelines)~0%
2BenignRoutine screening~0%
3Probably Benign6-month follow-up or continued surveillance>0% to ≤2%
4SuspiciousTissue diagnosis (biopsy)>2% to <95%
4ALow suspicionTissue diagnosis>2% to ≤10%
4BModerate suspicionTissue diagnosis>10% to ≤50%
4CHigh suspicionTissue diagnosis>50% to <95%
5Highly Suggestive of MalignancyTissue diagnosis≥95%
6Known Biopsy-Proven MalignancyClinical follow-up; definitive local therapyN/A

Assessment Hierarchy

When multiple findings exist, use the highest-order actionable assessment:

5 > 4 > 0 > 6 > 3 > 2/1

Board Pearl

A category 4 finding supersedes a category 6 finding at the examination level. If a patient has biopsy-proven cancer (cat 6) in the right breast and a suspicious new finding (cat 4) in the left breast, the overall exam assessment is category 4 because it requires immediate action.

Category 0 — Incomplete

Use of Category 0 should be rare on MRI. Limited to three scenarios:

  1. Additional imaging needed for characterization (e.g., US to confirm likely lymph node)
  2. Prior imaging needed for comparison
  3. Technically inadequate exam (e.g., contrast not administered)

Category 0 should NOT be used when findings are sufficiently characterized by MRI and additional imaging is recommended for detection/guidance only.

Board Pearl

v2025 split Category 0 into two sub-types: “Need Additional Imaging Evaluation” and “Need Prior Imaging for Comparison” — matching the 2024 MQSA regulation update.

Category 3 — Probably Benign

Criteria

  • Cancer likelihood: >0% to ≤2%
  • Goal: decrease false-positive biopsy recommendations
  • Desirable frequency: ≤5% of examinations

Imaging Features Supporting Category 3

Masses: Oval + circumscribed + homogeneous/dark septations + T2 hyperintense → cancer ≤2%

NME: Focal or regional distribution + homogeneous enhancement → frequently benign (but no single feature alone is sufficient)

No individual feature alone is sufficient for category 3 — combinations of features are required.

Follow-Up Timing

  • 6 months12 months24 months
  • Stable at 24 months → upgrade to Category 2
  • Increase/progression → upgrade to Category 4 or 5

Category 3 in Current Breast Cancer

Avoid category 3 in the setting of current breast cancer. Findings that would otherwise be category 3 should be assigned category 4A to enable near-term tissue diagnosis for treatment planning.

Board Pearl

Caution with category 3 on non-baseline MRI: one study showed category 3 malignancy rate of 2% on baseline exams but 9% on non-baseline exams. New findings on follow-up MRI warrant more suspicion.

Category 4 — Suspicious

  • Subdivisions (4A, 4B, 4C) are now optional for MRI (previously only for mammography/US)
  • Management: tissue diagnosis
  • Report should include method of recommended tissue diagnosis and contingencies

Category 5 — Highly Suggestive of Malignancy

  • No individual feature alone confers ≥95% cancer likelihood
  • Highest PPV features: spiculated margin, linear/segmental NME
  • Combinations of suspicious features required
  • Percutaneous tissue diagnosis remains standard of care (not excisional biopsy)

Category 6 — Known Biopsy-Proven Malignancy

When to Use

  • Biopsy-proven malignancy before definitive local therapy
  • Enhancing findings that are definitively the biopsy-proven cancer
  • More extensive findings with high certainty of malignancy (same morphology, contiguous)
  • No enhancement at known malignancy site (e.g., post-NAC complete MRI response)

Additional Close Findings (ACFs)

Category 6 may also include ACFs when ALL criteria met:

  1. Within 2 cm of biopsy-proven malignancy
  2. Do not increase total extent by >2 cm
  3. Would not change clinical management

Other separate suspicious findings → assign separate category 0, 4, or 5.

Board Pearl

The term “satellite” should NOT be used for additional findings near a known cancer — it is not sufficiently defined regarding proximity and suspicion level. Use “additional close finding (ACF)” instead.

Targeted Ultrasound for MRI Findings

  • Use “targeted,” “MRI-targeted,” or “MRI-directed” — not “second-look” ultrasound
  • Overall US detection of MRI findings: pooled rate 58%
  • Masses: 66% US detection rate
  • NME: only 29% US detection rate → lower threshold for MRI-guided biopsy