NCI Thesaurus: Structure, API & Research Use

The NCI Thesaurus (NCIt) is a reference terminology and biomedical ontology that provides standardized definitions, codes, and hierarchical relationships for cancer research and clinical oncology. Produced by the Enterprise Vocabulary Services (EVS) within the National Cancer Institute (NCI), the thesaurus establishes a shared semantic foundation for oncology clinical trials, translational laboratory studies, basic cancer biology, and administrative reporting across the National Institutes of Health (NIH). In modern biomedical data systems, investigators working within oncology rely on NCIt codes to eliminate terminological ambiguity across electronic health records, genomic registries, and federally funded clinical protocols.

What it is

The NCI Thesaurus is a controlled biomedical vocabulary structured as a formal ontology using description logic. Each concept within NCIt is assigned a stable, semantically meaningless alphanumeric identifier (for example, C9305 for Malignant Neoplasm or C3262 for Neoplasm), preventing identifier instability when scientific classifications evolve. Unlike general biomedical vocabularies, NCIt is specifically curated to capture the depth of neoplastic disease classification, cancer genetics, cellular pathophysiology, pharmacologic agents, and clinical trial design.

NCIt functions in close alignment with two adjacent federal biomedical resources: the NCI Metathesaurus (NCIMeta) and the Cancer Data Standards Registry and Repository (caDSR). While the NCI Metathesaurus integrates multi-source vocabularies mapped to Concept Unique Identifiers (CUIs) published on a biannual cycle, NCIt acts as an independent core vocabulary with monthly and weekly distribution tracks. Within caDSR, NCIt concepts provide the standardized semantic definitions and permissible value sets for Common Data Elements (CDEs) deployed in clinical trial case report forms, directly linking trial operations with broader NCI funding infrastructure.

Who runs it and what it covers

The NCI Thesaurus is developed and maintained by the Enterprise Vocabulary Services (EVS) team within the Center for Biomedical Informatics and Information Technology (CBIIT) at the National Cancer Institute, located in Bethesda, Maryland. The platform is distributed openly under the Creative Commons Attribution 4.0 International license (CC BY 4.0), allowing academic institutions, commercial vendors, and international research consortia to incorporate NCIt terminology into local databases and software pipelines without licensing fees.

The vocabulary is organized into 19 top-level root categories that span basic biological mechanisms to applied clinical administration:

Abnormal Cell (C12913): Cellular abnormalities, cytogenetic findings, and atypical morphology.
Activity (C43431): Clinical trial actions, diagnostic procedures, research techniques, and therapeutic interventions.
Anatomic Structure, System, or Substance (C12219): Macroscopic and microscopic anatomy, organ systems, and bodily fluids.
Biochemical Pathway (C20633): Signal transduction, metabolic cascades, and intracellular regulatory networks.
Biological Process (C17828): Physiologic, cellular, and molecular dynamics.
Chemotherapy Regimen or Agent Combination (C12218): Multi-agent therapeutic protocols, dosages, and combination therapies.
Conceptual Entity (C20181): Classification schemas, data descriptors, and biostatistical concepts.
Diagnostic or Prognostic Factor (C20047): Biomarkers, staging parameters, and clinical prognosticators.
Disease, Disorder or Finding (C7057): Detailed classifications of solid tumors, hematologic malignancies, and non-neoplastic conditions.
Drug, Food, Chemical or Biomedical Material (C1908): Small-molecule agents, immunotherapies, radiopharmaceuticals, and biological materials.
Experimental Organism Anatomical Concept (C22188): Anatomical definitions tailored to preclinical animal models.
Experimental Organism Diagnosis (C22187): Diagnostic categories for veterinary and laboratory animal oncology.
Gene (C16612): Human and model organism genetic loci.
Gene Product (C26548): Proteins, transcripts, and macromolecular complexes.
Manufactured Object (C97325): Medical devices, laboratory instruments, and surgical equipment.
Molecular Abnormality (C3910): Gene fusions, point mutations, copy number variations, and structural rearrangements.
Organism (C14250): Taxonomic classifications from viruses to mammals.
Property or Attribute (C20189): Observable qualities, physical quantities, and relational attributes.
Retired Concept (C28428): Historical concepts retained with permanent identifiers to preserve longitudinal data integrity.

NCIt publishes regular updates: development cycles produce weekly builds (such as version 26.11a), while stable baseline packages are issued monthly (such as version 26.10d). Distribution files include flat tab-delimited text archives (Thesaurus.txt), Web Ontology Language 2 files (Thesaurus.owl), and inferred ontology packages (ThesaurusInferred.owl) prepared for ingestion into the Unified Medical Language System (UMLS).

How to search or use it

Researchers and data managers interact with NCIt through web-based exploration, direct REST API consumption, or batch file downloads. When designing oncology studies, establishing compliant metadata requires standardizing definitions through the following practical steps:

Identify Required Terms: Determine the clinical parameters, biomarkers, or therapeutic agents that require standardized ontology tagging in your research database or case report form.
Query the EVS REST API: Use the public NCI EVS REST API (endpoint https://api-evsrest.nci.nih.gov/api/v1/concept/ncit/search) with the term parameter to programmatically identify corresponding concept candidates.
Verify Concept Metadata: Retrieve the full concept payload using /api/v1/concept/ncit/{code}?include=full to inspect preferred names, synonyms, formal definitions, and parent-child hierarchical links.
Check Cross-Terminology Mappings: Review the synonyms and definitions arrays to confirm mappings to standard regulatory subsets, including CDISC (Clinical Data Interchange Standards Consortium), FDA terminology, and caDSR value sets.
Incorporate Alphanumeric Codes: Store the permanent NCIt concept code (e.g., C9305) in your institutional data dictionary alongside the human-readable preferred label to ensure downstream interoperability.
Automate Version Management: Query /api/v1/metadata/terminologies on a monthly cadence to track terminology version releases and evaluate deprecated or retired concepts in your data architecture.
Download Bulk Ontology Distributions: For high-throughput bioinformatics applications or offline knowledge graphs, retrieve Thesaurus.FLAT.zip or Thesaurus.OWL.zip from the official EVS FTP server to parse native class hierarchies.

Standardizing variables according to established vocabularies simplifies downstream protocol integration, particularly when coordinating multi-center studies governed by oncology clinical trial endpoints and standardized reporting requirements.

Strengths and limits

The primary strength of the NCI Thesaurus lies in its dedicated oncology focus and rigorous institutional maintenance. Because EVS terminologists actively curate cancer biology, immunotherapy, and molecular pathology concepts, NCIt provides deeper granularity for neoplastic sub-types and multi-agent chemotherapy regimens than broad biomedical lexicons. Furthermore, its open CC BY 4.0 licensing and direct coupling with FDA and CDISC regulatory standards make it an optimal choice when constructing an institutional metadata schema alongside broader guidelines on how to choose a metadata schema.

However, NCIt also presents operational boundaries that data managers must navigate. Because NCIt is heavily centered on oncology and related biological systems, general medical concepts (such as non-oncologic chronic illnesses, specialized surgical procedures outside oncology, or primary care interventions) may lack the granular sub-classification found in comprehensive systems like SNOMED CT. Additionally, while weekly and monthly releases ensure rapid integration of emerging cancer terminology, database managers must establish automated version-control routines to manage concept retirements and hierarchy modifications across long-term clinical trial registries.

Alternatives

Oncology researchers often evaluate NCIt alongside other controlled terminologies and biomedical knowledge repositories:

Terminology / System	Primary Scope	Governing Body	Key Oncology Use Case
NCI Thesaurus (NCIt)	Oncology, cancer genetics, therapeutics, clinical trials	NCI CBIIT (NIH)	Gold standard for cancer trials, caDSR common data elements, CDISC subsets
SNOMED CT	Comprehensive clinical healthcare terminology	SNOMED International	Broad electronic health record (EHR) documentation and clinical charting
MeSH (Medical Subject Headings)	Biomedical literature indexing and cataloging	National Library of Medicine (NLM)	Bibliographic search, MEDLINE/PubMed indexing, and literature retrieval
UMLS Metathesaurus	Multi-terminology integration and cross-mapping	National Library of Medicine (NLM)	Cross-vocabulary interoperability and concept cross-walking across ontologies
ICD-O-3	Oncologic topography and tumor histology coding	World Health Organization (WHO)	Cancer registry reporting, epidemiological tracking, and tumor staging

FAQ

What is the difference between NCIt and NCI Metathesaurus?

The NCI Thesaurus (NCIt) is a standalone, curated ontology developed directly by NCI EVS, featuring monthly releases and 19 distinct root hierarchies. In contrast, the NCI Metathesaurus (NCIMeta) is a composite database modeled on the UMLS Metathesaurus that aggregates terminology from dozens of external vocabularies into Concept Unique Identifiers (CUIs), published on a biannual schedule.

How does NCIt integrate with caDSR?

The Cancer Data Standards Registry and Repository (caDSR) stores Common Data Elements (CDEs) used across NCI-sponsored clinical trials. NCIt serves as the semantic backbone for caDSR, providing the authoritative concept definitions, permissible values, and standardized codes for these clinical elements.

Is the NCI Thesaurus free to use in commercial software?

Yes. The NCI Thesaurus is released under the Creative Commons Attribution 4.0 International license (CC BY 4.0). Users may freely share and adapt the material for both academic and commercial applications, provided appropriate credit is given to the NCI Enterprise Vocabulary Services.

What formats are available for downloading NCIt?

EVS publishes NCIt in multiple standard formats through its public FTP server: a tab-delimited flat file (Thesaurus.txt), Web Ontology Language 2 format (Thesaurus.owl), and an inferred OWL file (ThesaurusInferred.owl) optimized for UMLS integration.

Sources

National Cancer Institute Enterprise Vocabulary Services. (2026). NCI Thesaurus ReadMe and Release Documentation (Version 26.10d). U.S. National Institutes of Health. https://evs.nci.nih.gov/ftp1/NCI_Thesaurus/ReadMe.txt

National Cancer Institute Enterprise Vocabulary Services. (2026). NCI Thesaurus Terms of Use (CC BY 4.0). U.S. National Institutes of Health. https://evs.nci.nih.gov/ftp1/NCI_Thesaurus/ThesaurusTermsofUse.htm

National Cancer Institute Enterprise Vocabulary Services. (2026). NCI EVS REST API Specification (Version 2.5.0.RELEASE). U.S. National Institutes of Health. https://api-evsrest.nci.nih.gov/v3/api-docs

National Cancer Institute Center for Biomedical Informatics and Information Technology. (2026). EVS REST API Terminology Metadata and Root Hierarchies. U.S. National Institutes of Health. https://api-evsrest.nci.nih.gov/api/v1/metadata/terminologies