The latest CDISC Library release deepens oncology coverage and pushes the 360i vision of AI-driven, metadata-first clinical research another step forward.

CDISC has released the latest package of Biomedical Concepts (BCs) and SDTM Dataset Specializations, and this one is worth a closer look. Package 18 is now live in the CDISC Library, bringing new concepts, new dataset specializations, and a round of refinements to existing content driven by community feedback. With this release, the Library now holds 1,475 Biomedical Concepts and 1,475 SDTM Dataset Specializations — a one-to-one implementation map that keeps growing as CDISC advances its 360i Initiative.

For programmers, data managers, and standards teams, releases like this are more than a changelog. Each new BC is a reusable, machine-readable building block that removes guesswork from data collection and submission mapping. Below is a practical summary of what shipped and where it fits in the bigger picture.

A quick refresher: what Biomedical Concepts actually are

Biomedical Concepts fill the gaps that the Foundational Standards alone leave open. They add the semantics, variable relationships, and detailed metadata needed to generate CRFs or Define-XML without reinventing the wheel on every study. CDISC uses a two-layered approach:

• A conceptual layer — a standards-agnostic, unambiguous semantic definition largely anchored in NCI Thesaurus (NCIt) concepts.
• An implementation layer — SDTM Dataset Specializations that provide value-level metadata to drive automation. The model is flexible enough to accommodate other standards such as HL7 FHIR.

The payoff is consistency and automation: preconfigured value-level metadata reduces variability in implementations, promotes reuse, and provides the building blocks for Define-XML. That is exactly the direction the 360i Initiative — CDISC's multi-year program to enable standards-driven automation from study design through results — is pushing the whole industry.

What's new in Package 18

The headline of this release is breadth in two areas: clinical classifications and subject-level characteristics, plus a substantial expansion of oncology and breast cancer content.

New clinical classification and characteristic concepts

• PASI (Fredriksson version). New BCs for the Psoriasis Area and Severity Index, Fredriksson version clinical classification — a standardized, machine-readable representation of a widely used dermatology severity measure.
• Subject Characteristics. New BCs aligned with the SC Codetable Mapping File, giving teams consistent, preconfigured metadata for subject-level descriptors.
• Allergen-induced IgE Mast 7-Class. New BCs supporting allergy testing results across the seven-class IgE grading scheme.

A deeper push into oncology and breast cancer

Oncology gets the most attention in this package, reflecting CDISC's ongoing work to make Therapeutic Area content available as reusable, machine-readable metadata. New and expanded content includes:

• Surgical, therapeutic, and diagnostic procedures for breast cancer.
• AJCC Cancer Staging Manual, 7th Edition staging concepts.
• Progesterone and estrogen receptor measurements — key biomarkers in breast cancer characterization.
• Surgical margin status.
• Disease recurrence.

These concepts arrive with SDTM Dataset Specializations linked to new BCs across the CM, IS, MH, MI, PR, RS, SC, and TU domains — spanning concomitant medications, immunogenicity, medical history, microscopic findings, procedures, disease response, subject characteristics, and tumor/lesion identification. For oncology programmers, that domain coverage is the practical win: the mapping from clinical concept to submission-ready structure comes preassembled.

Alongside the new content, Package 18 folds in updates and enhancements to existing BCs and SDTM specializations based on ongoing community feedback — the iterative refinement that keeps the library trustworthy for production use.

Why this release matters

Every added BC is one fewer thing each study team has to define from scratch. The value-level metadata behind these concepts is what lets tooling generate CRFs, populate Define-XML, and validate mappings with far less manual interpretation. Concretely, this release helps:

• Standards teams reduce implementation variability and improve adherence through preconfigured metadata.
• EDC/CRF designers cut configuration time and open avenues for automated EDC builds.
• Data managers spend less time interpreting the protocol, especially when BCs are linked to a Schedule of Activities.
• Statistical programmers get end-to-end data-flow transparency and ready building blocks for Define-XML.

The reach into breast cancer staging, receptor status, and recurrence is especially timely. These are exactly the data points oncology submissions hinge on, and having them standardized at the value level reduces the risk of one-off, study-specific interpretations that complicate review.

How to get it

The new and updated content is available through the usual three channels:

• Excel exports from the Exports tab of the CDISC Biomedical Concepts website.
• The CDISC Library APIs, which deliver BCs as JSON output for programmatic use.
• The BC Browser, for searching, filtering, and reviewing the latest BCs and dataset specializations directly.

If your team maintains its own standards metadata repository, this is a good moment to diff against the latest export and pull in the oncology and subject-characteristics additions before your next study build.

The takeaway

Package 18 is an incremental release with an outsized practical impact for oncology teams, and it reinforces the trajectory CDISC set with 360i: a Library that is less a reference document and more an automation engine. At 1,475 concepts and counting, the Biomedical Concepts effort is steadily turning clinical semantics into infrastructure — the kind you build tools on top of, not around.

Source: CDISC Community announcement (July 15, 2026) and cdisc.org/cdisc-biomedical-concepts. Concept counts and release contents per the official CDISC release notice.