BI-RADS
The ACR system for breast findings: categories 0 to 6, associated management and the common language across mammography, US and breast MRI.
See categories →RESOURCES · CLASSIFICATIONS
The systems that
organize the decision
The classifications of routine radiology, rebuilt as a decision corpus: structured, cited and current, served over an API in RadCommons and applied by GUIDE inside the report.
IN DETAIL
The everyday systems
LI-RADS
Risk stratification for HCC in at-risk patients: categories LR-1 to LR-5, LR-M and LR-TIV, with major and ancillary features.
See categories →PI-RADS
Likelihood of clinically significant cancer on multiparametric prostate MRI, score 1 to 5 by zone.
See categories →TI-RADS (ACR)
Scoring of thyroid nodules by composition, echogenicity, shape, margin and echogenic foci, with size-based management criteria.
See categories →Lung-RADS
Management of nodules in low-dose CT lung cancer screening: categories, probability and follow-up.
See categories →O-RADS
Malignancy risk of adnexal lesions on US and MRI, from O-RADS 1 to 5, with category-guided management.
CAD-RADS
Grading of coronary artery disease on CT angiography: stenosis, modifiers and management implications.
Bosniak
Classification of complex renal cysts, from category I to IV, with probability of malignancy and management.
See categories →RECIST 1.1
Imaging criteria for tumor response, with target lesion selection, measurement and CR, PR, SD and PD categories.
See criteria →Fleischner
Follow-up recommendations for incidental pulmonary nodules, by size, morphology and patient risk.
EXPLORE A SYSTEM
Educational summary. The complete, current and cited criterion lives in RadCommons; in the report, GUIDE always proposes the category for physician review.
The problem RadCommons solves
Every classification lives in PDFs, articles and society websites, in versions that change. In daily practice, that turns into criteria from memory, old screenshots and disagreement between colleagues. For AI, it turns into something worse: a model that cites criteria that do not exist.
RadCommons rebuilds each system as structured data: criteria, categories, management and the citation of the current source, with versioning. Radiologists verify; the API serves. GUIDE uses this corpus to propose the category inside the report, with the criterion one click away.
In the product and in science
- IN THE REPORTGUIDE proposes the category for physician review, with the supporting criterion
- IN THE CORPUSStructured, cited and versioned systems in RadCommons
- IN MEASUREMENTLaiBench's GUIDE axis evaluates the correct application of the guideline
- ACCESSradcommons.laudos.ai, with API documentation
FAQ
Direct questions
Is the list above everything there is?
No, it is the showcase of the everyday systems. The full catalog, with the detail of each system and version, lives in RadCommons, the public corpus of Laudos.AI.
What happens when a guideline is updated?
The system gets a new version in the corpus, with the source cited. The log of each report records which version supported the suggestion at the time.
Can I use RadCommons outside Laudos.AI?
Yes, it is served over an API precisely to ground third-party agents and tools. The access documentation is on the subdomain.
Using classifications with traceability
Version and context matter as much as the category
A classification is not just a list of numbers. It depends on the correct system, the current version, the modality and the findings that support the category. That is why the corpus keeps criteria and sources tied to the version, and GUIDE presents the suggestion as a review point, not as autonomous output.
When the guideline changes, old cases should not be silently reinterpreted under the new rule. The log preserves the version used at the time of the report, while the current version guides new cases. This separation is essential for retrospective review and clinical governance.
What to check in each system
Before accepting a category, check whether the descriptors were applied to the right context, whether the system covers that exam and whether the recommendation matches the version adopted by the institution. The tool brings the criterion closer to the report; the final judgment still belongs to whoever saw the image and signs.
NEXT STEP
Want the full corpus over an API?
Live demo with your exam types and your institution's template, right in the product.
Questions and answers
How many items does this classifications section have?
There are 43 pages published in this section, and each one opens directly, with no sign-up and no paywall. All sections
What is this classifications section for?
To bring together the classification systems used in the report, with the category, criterion and management each one provides for. Who publishes
Is the content reviewed by a physician?
Yes. Editorial responsibility lies with a radiologist registered with CRM-SP, and each page shows the date the content last changed. Who signs
Does this replace clinical assessment?
No. The material supports the decision of whoever writes the report. The finding, the clinical correlation and the management remain with whoever signs the exam. Clinical governance