BI-RADS Calculator: A Complete Guide to Breast Imaging Structured Reporting
Why Structured Breast Imaging Reports Matter
Breast cancer is the most commonly diagnosed cancer in women worldwide. Imaging plays a central role at every stage — screening, diagnosis, staging, and treatment monitoring. That central role carries a correspondingly heavy documentation burden.
The challenge is not just finding findings. It is communicating them accurately, consistently, and in a form that clinicians can act on. An unstructured narrative report describing a "somewhat irregular, slightly hypoechoic lesion in the right breast" leaves too much room for interpretation. The referring physician does not know what follow-up is appropriate. The patient does not know how worried to be. If the case is reviewed later — by a different radiologist, a tumor board, or a medicolegal reviewer — the original assessment is difficult to reconstruct.
Structured reporting with the BI-RADS lexicon solves this problem. By applying standardized descriptors and mapping them to defined assessment categories with explicit management recommendations, BI-RADS transforms a subjective narrative into an actionable clinical document. This is not bureaucratic formalism. It is how modern, high-volume breast imaging practice maintains quality at scale.
What BI-RADS Is and Where It Comes From
BI-RADS stands for Breast Imaging Reporting and Data System. It was developed by the American College of Radiology and is now in its fifth edition, published in 2013. The fifth edition is the current standard and introduced significant revisions to the ultrasound and MRI lexicons, bringing them into alignment with the mammography lexicon that had been the original foundation of the system.
BI-RADS is not a scoring system in the same sense as ACR TI-RADS, which assigns numerical points to individual features. Instead, BI-RADS is a lexicon — a standardized vocabulary — combined with an assessment framework. The radiologist selects appropriate descriptors from the lexicon to characterize a finding, then assigns an overall assessment category that reflects the integrated clinical interpretation.
This distinction matters. In TI-RADS, the algorithm generates the category from the feature inputs. In BI-RADS, the radiologist generates the category, informed by the descriptors but not mechanically derived from them. The lexicon structures the thinking; the radiologist synthesizes the conclusion.
The assessment categories span the full range of clinical scenarios:
BI-RADS 0 — Incomplete: Additional imaging or prior comparison is needed before a final assessment can be made. Common in screening mammography when a finding requires diagnostic workup or when priors are unavailable.
BI-RADS 1 — Negative: No finding to report. Routine screening interval applies.
BI-RADS 2 — Benign: A finding is present but is definitively benign. Simple cysts, intramammary lymph nodes, stable fat-containing lesions, and post-surgical changes fall here. Routine screening continues.
BI-RADS 3 — Probably Benign: Malignancy risk is greater than zero but no more than 2%. Short-interval follow-up — typically 6 months — is recommended to establish stability rather than tissue sampling.
BI-RADS 4A, 4B, 4C — Suspicious: Tissue sampling is warranted. The three subcategories reflect increasing malignancy probability: 4A carries greater than 2% to 10% risk, 4B carries greater than 10% to 50%, and 4C carries greater than 50% to less than 95%. All three subcategories share the same management recommendation (biopsy), but the subcategory grade informs how aggressively to pursue tissue diagnosis if initial sampling is non-diagnostic.
BI-RADS 5 — Highly Suggestive of Malignancy: Malignancy probability is 95% or higher. Tissue sampling is mandatory; appropriate surgical planning should proceed concurrently.
BI-RADS 6 — Known Biopsy-Proven Malignancy: The finding has already been histologically confirmed as malignant. This category is used during neoadjuvant treatment monitoring or pre-surgical imaging.
Mammography: Understanding the Lexicon
Mammography was the foundation on which BI-RADS was originally built, and the mammographic lexicon remains the most developed of the three modality sections.
Breast Composition
Before characterizing individual findings, the radiologist must assess overall breast composition. The four categories — fatty (A), scattered fibroglandular density (B), heterogeneously dense (C), and extremely dense (D) — do not directly influence the assessment category of a specific finding, but they are clinically critical for two reasons. First, dense breast tissue reduces mammographic sensitivity, which has implications for supplemental screening decisions. Second, heterogeneously and extremely dense breasts carry an independent, modest elevation in cancer risk beyond the masking effect.
Mass Descriptors
A mass is a three-dimensional space-occupying lesion with convex outward borders, visible on two projections. The mammographic lexicon describes masses along three axes: shape, margin, and density.
Shape ranges from oval (elliptical, with up to three gentle undulations) through round to irregular (neither oval nor round). Oval and round masses with circumscribed margins are the classic appearance of fibroadenomas and are generally assigned BI-RADS 3 when first detected in appropriate clinical contexts. Irregular shape raises the probability of malignancy substantially.
Margin descriptors carry significant prognostic weight. A circumscribed margin — sharply defined with an abrupt transition to surrounding tissue — is reassuring. Obscured margins reflect superimposition by adjacent tissue and may mask an underlying irregular margin. Microlobulated margins suggest multiple small surface undulations at a scale smaller than the overall shape lobulation. Indistinct margins indicate no clear demarcation from the surrounding parenchyma. Spiculated margins — with lines radiating outward from the mass — carry the highest suspicion, representing the classic appearance of invasive carcinoma eliciting a desmoplastic stromal response.
Density refers to the mass's X-ray attenuation relative to the surrounding fibroglandular tissue. High density raises concern. Fat-containing masses — with density below surrounding tissue — include hamartomas, galactoceles, lipomas, and oil cysts, all benign.
Calcifications
Calcification characterization in mammography is among the most nuanced tasks in breast imaging, and the BI-RADS lexicon reflects that complexity with a detailed morphological taxonomy.
Typically benign calcifications include skin calcifications (often with a lucent center, visible in tangential views), vascular calcifications (tram-track pattern along vessel walls), coarse or popcorn calcifications (fibroadenoma), large rod-like calcifications (plasma cell mastitis or secretory disease), round and punctate calcifications when diffuse, rim calcifications (oil cysts), dystrophic calcifications (post-traumatic or post-radiation), milk of calcium (tea-cup appearance on lateral projection), and suture calcifications.
Calcifications of intermediate concern include amorphous calcifications (small and hazy, without discernible shape) and coarse heterogeneous calcifications (irregular, clustered, but coarser than fine pleomorphic). Both warrant consideration of biopsy, particularly when segmentally distributed.
Suspicious morphologies include fine pleomorphic calcifications (varying shapes and sizes, less than 0.5 mm) and fine linear or branching calcifications (casting type, suggesting ductal filling). The latter pattern has a particularly high positive predictive value for ductal carcinoma in situ.
Distribution matters as much as morphology. Diffuse or regional distributions of typically benign morphologies can generally be left alone. Grouped (clustered) calcifications of intermediate or suspicious morphology warrant biopsy. Linear and segmental distributions are the most concerning, as they suggest ductal involvement and increase the probability that calcifications of any morphological grade represent malignancy.
Architectural Distortion and Asymmetries
Architectural distortion — a spiculated pattern without a central mass — may represent the earliest mammographic sign of invasive lobular carcinoma or radial scar. It requires tissue sampling when not associated with prior surgery.
Asymmetries are described along a spectrum of clinical significance. An asymmetry visible on only one view does not fulfill the criteria for a three-dimensional mass. Global asymmetry is a large volume of fibroglandular tissue occupying a quadrant or more, with no corresponding architectural distortion or associated calcifications. Focal asymmetry is seen on two views but lacks convex outward borders or architectural distortion. Developing asymmetry — new or growing compared to a prior study — is the most worrisome variant and warrants tissue sampling in most cases.
Ultrasound: Adding Echogenicity and Orientation
Breast ultrasound extends the BI-RADS lexicon into the real-time domain, adding descriptors that exploit the unique physical properties of sound — particularly tissue echogenicity and the direction of wavefront propagation through the lesion.
Mass Descriptors
The ultrasound mass lexicon shares the shape categories of mammography (oval, round, irregular) but adds orientation as a distinct descriptor. Orientation refers to the long axis of the mass relative to the skin surface. A parallel (horizontal) orientation — wider than tall — is reassuring, as it suggests the lesion is growing along the tissue planes. A not-parallel (vertical or taller-than-wide) orientation means the lesion is growing across tissue planes, which correlates with more aggressive biology.
Margin descriptors in ultrasound mirror the mammographic categories conceptually, but the subcategories of non-circumscribed margins are expanded: indistinct, angular, microlobulated, and spiculated. Angular margins — with acute angles at the interface between the lesion and surrounding tissue — may reflect infiltrating tumor extending along fat lobule septa.
Echo pattern describes the lesion's internal echogenicity relative to subcutaneous fat. The spectrum runs from anechoic (simple cyst, no internal echoes) through hyperechoic, complex cystic and solid, hypoechoic, isoechoic, and heterogeneous. Hypoechoic is the most common echo pattern for breast carcinoma, though it is far from specific. Anechoic masses with circumscribed margins, posterior acoustic enhancement, and no internal features define the simple cyst, which is BI-RADS 2 by ultrasound regardless of size.
Posterior acoustic features — enhancement, shadowing, or combined patterns — supplement the mass characterization. Enhancement (increased echogenicity behind the lesion) is classically associated with cysts and some fibroadenomas. Shadowing may indicate dense fibrous stroma or calcification. These features contribute to but do not independently determine the assessment category.
Special Cases
The ultrasound lexicon includes a special cases category for findings with distinctive appearances that allow confident benign characterization without full descriptor application. Simple cysts, intramammary lymph nodes, vascular abnormalities, post-surgical collections, and fat necrosis are assigned BI-RADS 2 when their characteristic features are unambiguous. Clustered microcysts and complicated cysts — which have some internal echoes but no discrete solid component — are BI-RADS 3 and merit short-interval follow-up.
MRI: The Most Complex Lexicon
Breast MRI is the most sensitive imaging modality for breast cancer detection. Its specificity, however, is limited by the enhancement of benign processes — fibroadenomas, fibrocystic changes, hormonal effects — that can mimic malignancy. The MRI lexicon is correspondingly the most complex of the three BI-RADS sections, incorporating temporal enhancement kinetics that have no analog in the other modalities.
Focus
A focus is a punctate area of enhancement smaller than 5 mm that is too small to characterize morphologically. Foci are commonly seen in background parenchymal enhancement and, when isolated and stable, are generally BI-RADS 3 or lower. A new focus or one with suspicious kinetics may warrant upgrade.
Mass
MRI mass descriptors echo the mammographic and ultrasound lexicons for shape (oval, round, irregular) and margin (circumscribed, irregular, spiculated), but add a dimension-specific to MRI: internal enhancement pattern. The six patterns — homogeneous, heterogeneous, rim enhancement, dark internal septations, enhancing internal septations, and central enhancement — reflect the underlying tumor microarchitecture.
Rim enhancement is among the most suspicious patterns, associated with central necrosis typical of high-grade invasive carcinoma. Dark internal septations, in contrast, are a reassuring feature of fibroadenomas: thin low-signal bands separating enhancing lobules. Enhancing internal septations — where those bands enhance — are suspicious. Homogeneous enhancement can be seen in both benign and malignant lesions and must be interpreted in conjunction with morphology and kinetics.
T2-weighted signal characteristics provide additional tissue characterization without contrast. Simple cysts are markedly T2 hyperintense. Fibroadenomas tend to show intermediate to high T2 signal. Many invasive carcinomas are T2 hypointense relative to glandular tissue, reflecting their dense fibrous stroma — an important distinguishing feature.
Non-Mass Enhancement
Non-mass enhancement (NME) refers to enhancement that is neither a mass nor a focus — it occupies a region of the breast without the three-dimensional convex morphology that defines a mass. NME is described by its distribution and internal enhancement pattern.
Distribution ranges from focal (occupying a small, round or oval area) through linear, segmental, regional, multiple regions, and diffuse. Segmental distribution — a triangle or cone of enhancement pointing toward the nipple, conforming to a ductal segment — is particularly concerning for ductal carcinoma in situ. Linear distribution, especially when branching, similarly suggests ductal involvement.
Internal enhancement patterns for NME include homogeneous, heterogeneous, clumped, and clustered ring. Clumped enhancement — cobblestone or grape-like clusters of individual enhancing foci — and clustered ring enhancement — rings of enhancement around multiple small regions — are the most suspicious patterns and carry high positive predictive values for malignancy.
Kinetic Curve Analysis
The kinetic curve describes how enhancement evolves over time. It is assessed in two phases: initial and delayed.
Initial phase enhancement is characterized as slow (less than 50% signal increase from baseline), medium (50% to 100%), or fast (greater than 100%). Faster initial enhancement is associated with higher vascular permeability and correlates with malignancy, though the relationship is not absolute.
The delayed phase — assessed from approximately 2 minutes after injection onward — produces three curve types. Type I (persistent): signal continues to increase throughout the examination. This is the most common pattern of benign lesions, particularly fibroadenomas. Type II (plateau): signal stabilizes at its peak and neither significantly increases nor decreases. Type III (washout): signal decreases after its peak. Washout kinetics represent the most suspicious pattern, reflecting rapid contrast clearance through arteriovenous shunts typical of high-grade carcinoma.
No kinetic pattern is pathognomonic. Up to 30% of invasive carcinomas demonstrate persistent enhancement kinetics. Conversely, some benign lesions show washout. Kinetic analysis supplements morphological assessment; neither should be interpreted in isolation.
Background Parenchymal Enhancement
Before characterizing individual findings, MRI reports should document background parenchymal enhancement (BPE) — the degree of normal glandular tissue enhancement. BPE is graded as minimal (less than 25%), mild (25% to 50%), moderate (50% to 75%), or marked (greater than 75%).
BPE has direct implications for diagnostic interpretation. Marked BPE can obscure small enhancing foci, reduce the contrast between lesions and background, and increase the rate of false-positive findings. It also influences the assessment of lesion kinetics, as background enhancement may mask washout patterns or create apparent washout in benign findings. Noting BPE grade contextualizes the sensitivity and specificity of the examination for the interpreting clinician and any subsequent reviewers.
Using the Aperivue BI-RADS Calculator
To support structured breast imaging reporting, we built a free tool at aperivue.com/rads/birads that covers all three BI-RADS modalities: mammography, ultrasound, and MRI.
The tool is designed around the clinical workflow, not the other way around. You select the modality, enter clinical context, and then characterize each finding using the appropriate lexicon descriptors. The calculator supports multiple simultaneous findings — a common scenario in breast imaging — with a sidebar that lets you switch between findings and track their individual BI-RADS categories.
For MRI cases, the tool includes a kinetic curve selector with inline clinical guidance: washout kinetics display a contextual note about their suspicious significance, while persistent kinetics note their association with benign etiology. The BPE selector documents background enhancement before you characterize individual findings.
The radiologist assigns the BI-RADS assessment category directly — the tool does not generate a category algorithmically from feature inputs. This design reflects how BI-RADS is actually meant to function. The lexicon organizes your thinking; your clinical judgment synthesizes the conclusion.
Once findings are characterized and categories assigned, the tool generates a PACS-ready structured report that can be copied with a single click. The impression section is auto-generated from the individual finding assessments but is fully editable to accommodate clinical nuance. The full report text includes modality header, clinical indication, comparison, findings section, and impression — formatted for direct incorporation into a radiology workflow.
Common Pitfalls in BI-RADS Application
Even experienced radiologists encounter situations where BI-RADS application is not straightforward. Several patterns of error recur commonly enough to deserve explicit mention.
Assigning BI-RADS 3 without a defined follow-up plan. The probably benign category requires a commitment to the follow-up interval. Assigning BI-RADS 3 and then failing to recommend a specific time interval — or not tracking whether the patient returns — negates the clinical rationale of the category. Short-interval follow-up is the management.
Mixing descriptors across modalities. The lexicons for mammography, ultrasound, and MRI are not interchangeable. "Posterior acoustic enhancement" is an ultrasound descriptor; it has no meaning in a mammographic report. Rim enhancement is an MRI descriptor; it should not appear in an ultrasound report in that precise terminology. Using modality-specific language correctly is part of producing a professionally credible report.
Defaulting to BI-RADS 4 for uncertainty. The suspicious category is not a catch-all for findings that are difficult to characterize. BI-RADS 4 implies a decision to recommend biopsy. If the uncertainty is about whether a finding is real, that is a BI-RADS 0 situation — additional evaluation is needed. If the finding is definitely real and not definitively benign but carries low probability of malignancy, that is BI-RADS 3 territory. BI-RADS 4A, 4B, and 4C represent genuinely suspicious findings with tissue sampling as the appropriate next step.
Ignoring the overall assessment in favor of individual descriptors. BI-RADS categories represent the radiologist's integrated assessment — not a mechanical summation of descriptor grades. A mass with a slightly irregular margin but otherwise entirely benign features (oval shape, equal density, no associated findings in a young woman with a palpable finding present for years) may be more appropriately assessed as BI-RADS 3 than BI-RADS 4, even though "irregular margin" is listed as a suspicious feature. Clinical context modifies the integrated assessment.
Limitations of BI-RADS
BI-RADS is an indispensable framework, but it has recognized limitations that responsible users should understand.
Observer variability. Even with standardized descriptors, inter-observer agreement on individual features is imperfect. Studies measuring agreement on mass margin descriptors, NME distribution categories, and kinetic curve type consistently report moderate kappa values. This reflects the inherent subjectivity of image interpretation, not a flaw in the lexicon design. BI-RADS reduces variability; it does not eliminate it.
Modality-specific MRI limitations. MRI kinetics are influenced by the timing of image acquisition relative to contrast injection, the field strength, and the patient's phase in the menstrual cycle — factors that can shift a benign persistent kinetic pattern toward apparent plateau or washout. The MRI lexicon is applied to images that are already the product of technical choices made during acquisition.
Clinical context is not encoded. BI-RADS categories are defined without reference to patient age, personal or family history, BRCA status, or prior breast surgery — all of which influence the appropriate management threshold. A BI-RADS 3 finding in a BRCA1 carrier may warrant biopsy rather than follow-up, while a BI-RADS 4A finding in a pregnant patient might appropriately be deferred. The radiologist must modulate the category-derived management recommendation based on clinical context that the lexicon does not capture.
Evolution of technology. The fifth edition was published in 2013. Since then, digital breast tomosynthesis, contrast-enhanced mammography, abbreviated MRI protocols, and AI-assisted detection have entered clinical practice. The BI-RADS lexicon predates these technologies, and guidance on their incorporation is still evolving. A sixth edition is anticipated.
Disclaimer
This article and the Aperivue BI-RADS calculator are intended for educational purposes and as a reference for healthcare professionals familiar with breast imaging. The content reflects the ACR BI-RADS 5th Edition (2013) lexicon and does not constitute medical advice, diagnosis, or treatment recommendations for individual patients. Clinical decisions about breast imaging findings and patient management must be made by qualified physicians based on the complete clinical context, including history, physical examination, and all relevant imaging. The calculator is a documentation support tool, not a diagnostic device, and has not been cleared or approved by the FDA, KFDA/MFDS, or any regulatory authority.