eduKateSG · Why Science?
Give cancer-research concepts shared computable meaning—without turning a terminology code into diagnosis, evidence or treatment advice
Choose the closest route, then return to the full index whenever you need the wider evidence chain.
Science learning becomes powerful when students can tell a measured pattern from a model, a statistical decision from a mechanism, and a useful lead from a finished conclusion. NCI Thesaurus, or NCIt, is the National Cancer Institute’s reference terminology and core biomedical ontology. It provides persistent concept codes, preferred terms, synonyms, definitions and inter-concept relationships for cancer research and broader biomedical work. The official download page checked for this article identifies 26.09d as the current release, completed on 28 September 2026, and offers asserted and inferred OWL plus tab-delimited distributions. NCIt can help systems code and exchange meaning consistently, but a concept code is not a patient diagnosis, a clinical recommendation or proof that two local data fields are operationally equivalent. This guide supports Primary Science, PSLE Science, Secondary Science, O-Level Science and STEM education as a longform evidence-reading route; it is not medical advice, a laboratory protocol, an admission promise or a career guarantee.
Related eduKate reading: ICD-O-4 and comparable cancer registry data; UMLS Metathesaurus and connected biomedical meaning; HGNC and stable human gene symbols; careers by subject capability; Science Learning Hub; Education Hub; STEM Education; Careers by Subject Capability.
Primary and official sources checked on 13 October 2026: NCI Enterprise Vocabulary Services overview, checked 13 October 2026; Official NCI Thesaurus downloads showing version 26.09d, checked 13 October 2026; NCI Vocabulary for Cancer Research overview, checked 13 October 2026; NCI data standards explanation, checked 13 October 2026; 2026 Singapore–Cambridge O-Level Biology syllabus; MOE G2/G3 Lower Secondary Science teaching and learning syllabus; 2026 MOE G2 Computing syllabus. Terminologies, ontologies, interfaces and recommendations can change, so preserve dated documentation, exact source data and software versions.
A deeper reading route
Cancer research crosses pathology, genomics, medicines, anatomy, procedures, trials and public-health data. Each field brings abbreviations and local naming habits. NCIt provides a shared concept layer so a dataset can attach a persistent code to a meaning while still displaying reader-friendly terms and source-specific synonyms.
A concept code is deliberately semantically opaque: it identifies the concept without trying to pack its meaning into the characters. Preferred terms and synonyms can evolve while the code preserves continuity. That separation matters because a renamed concept should not force every database key to change, yet revisions still need release-aware review.
NCIt is more than an alphabetical thesaurus. Its is-a hierarchy and defined roles support an ontology in which software can traverse broader and narrower meanings or classify concepts. Asserted content records editor statements, while inferred distributions include relationships produced through classification. Analysts must say which distribution they used.
Definitions are valuable evidence for human interpretation, but they do not automatically specify a database field, eligibility criterion or diagnostic rule. A trial system may use an NCIt concept within a controlled value set that adds operational boundaries. Reuse therefore needs both the concept and the implementation context.
The official download page checked on 13 October 2026 lists version 26.09d, whose editing was completed on 28 September 2026. NCIt releases are versioned, and concept additions, retirements, merges or relationship changes can affect queries. A reproducible result preserves the exact release and any subset or mapping files.
Tagged synonyms can preserve terms from partners and specialist sources. They improve search and interoperability, but synonymy in a terminology does not mean that every source uses a term identically in every workflow. Mapping should compare definitions, scope, granularity and intended use—not string similarity alone.
Subsets and value sets let programmes select a governed portion of the larger ontology. They are practical for forms and exchange standards, yet membership can change. A coded field therefore needs the subset identity, version and permissible-value rules as well as individual concept codes.
For students, NCIt is a hopeful example of scientific teamwork. A community can agree on identifiers and definitions while still preserving uncertainty and change. The code answers “which concept?”, the source evidence answers “why is this claim supported?”, and clinical judgement answers “what should happen for a person?”
Mappings from NCIt to ICD-O, ChEBI, HGNC or other standards can connect datasets, but mappings are relationships with provenance. They may be exact, broader, narrower or context-dependent. Retaining both source codes and a reviewed mapping prevents a convenient crosswalk from erasing meaningful distinctions.
A trustworthy implementation treats terminology as infrastructure with tests. It validates codes against the pinned release, rejects retired concepts for new capture where appropriate, preserves historical values, and records any migration. The result is not merely cleaner data; it is a more honest path from source wording to published analysis.
Use the persistent NCIt code as the join key and store the preferred term for display. Searching by words alone can miss synonyms or combine distinct concepts.
Asserted and inferred OWL files serve different reasoning needs. A report should identify which one produced its hierarchy because classification may add relationships not stated directly by editors.
Retired or merged concepts remain important in historical data. Keep the captured code and release, then add a current mapping with status and rationale rather than overwriting the past.
A value set is a governed selection, not the whole ontology. Store its identifier, version, inclusion rules and expansion date beside exported permissible values.
Do not treat a hierarchical parent as a clinical generalisation without checking the relation and use case. Ontological subsumption and bedside decision rules are different layers.
Free-text mapping needs uncertainty and review. A term matcher can propose candidates, while domain experts decide whether the source meaning, negation and context support a code.
Terminology services should cache releases carefully and expose failures. Returning an old answer without its version can create silent inconsistency across systems.
NCIt content can support research exchange, but protected health information lives in surrounding records. Apply minimisation, access control and disclosure review independently.
| Stage | Question | Evidence | Responsible output |
|---|---|---|---|
| Name | Which concept is intended? | Persistent code, preferred term and synonyms | A resolvable biomedical meaning |
| Structure | How does it relate to other concepts? | Is-a hierarchy, roles and classifications | A computable ontology context |
| Govern | Which content is allowed here? | Release, subset, value set and local rules | A controlled implementation |
| Interpret | What claim is justified? | Source evidence, workflow and expert review | A bounded research statement |
Inside this guide
1–12 · Foundations and design
- 1. Start with the data question
- 2. Pin release 26.09d
- 3. Use concept codes as keys
- 4. Display preferred terms
- 5. Keep synonyms searchable
- 6. Read definitions in context
- 7. Traverse is-a carefully
- 8. Understand role relationships
- 9. Distinguish asserted and inferred
- 10. Treat OWL as a model
- 11. Use flat files within their limits
- 12. Version every subset
13–24 · Models and estimation
- 13. Expand value sets reproducibly
- 14. Handle retired concepts
- 15. Handle merges transparently
- 16. Map by meaning, not string
- 17. Represent mapping strength
- 18. Keep clinical evidence separate
- 19. Connect NCIt to ICD-O-4
- 20. Connect NCIt to HGNC
- 21. Connect NCIt to ChEBI
- 22. Validate free-text coding
- 23. Test terminology services
- 24. Protect patient context
25–36 · Diagnostics and interpretation
- 25. Preserve provenance
- 26. Explain ontology limits
- 27. Treat updates as governed change
- 28. Connect to Primary Science
- 29. Connect to Secondary Science
- 30. Connect to mathematics
- 31. Connect to computing
- 32. Connect to pathways
- 33. Did you know?
- 34. Write a reader-ready methods note
- 35. Build a validation ladder
- 36. Finish with a reproducible bundle
Section 1 of 36
1. Start with the data question
Decide whether NCIt will support search, coding, a value set, a data model or cross-standard mapping.
Working checkpoint — NCI Thesaurus: Write the intended use and owner first. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 2 of 36
2. Pin release 26.09d
The official page checked on 13 October 2026 names 26.09d as the current release and dates completion to 28 September 2026.
Working checkpoint — NCIt concept code: Store the release with every export. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 3 of 36
3. Use concept codes as keys
NCIt codes are persistent identifiers whose characters do not encode a clinical interpretation.
Working checkpoint — cancer terminology: Keep the code separate from display text. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 4 of 36
4. Display preferred terms
A preferred term offers consistent reader-facing wording for a concept.
Working checkpoint — biomedical ontology: Refresh labels only through controlled release updates. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 5 of 36
5. Keep synonyms searchable
Synonyms improve discovery across abbreviations and partner vocabularies.
Working checkpoint — preferred term: Retain synonym source and term type. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 6 of 36
6. Read definitions in context
Textual definitions help distinguish concepts but may not constitute an operational rule for a form or study.
Working checkpoint — ontology release: Pair the definition with the implementation specification. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 7 of 36
7. Traverse is-a carefully
Hierarchical parents and children support broader or narrower retrieval.
Working checkpoint — NCI Thesaurus: Check whether the relation and granularity fit the analysis. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 8 of 36
8. Understand role relationships
OWL role relations can express structured connections beyond a simple tree.
Working checkpoint — NCIt concept code: Use an ontology-aware tool and preserve the distribution type. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 9 of 36
9. Distinguish asserted and inferred
Asserted files contain editor-stated relations; inferred files include results of classification.
Working checkpoint — cancer terminology: Name which graph powered each query. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 10 of 36
10. Treat OWL as a model
The OWL distribution supports ontology tooling and reasoning, while flat files serve simpler lookup needs.
Working checkpoint — biomedical ontology: Choose format by use, not convenience alone. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 11 of 36
11. Use flat files within their limits
The tab-delimited distribution includes codes, hierarchy, synonyms and NCI definitions but excludes some richer properties.
Working checkpoint — preferred term: Do not infer absent roles from a flat export. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 12 of 36
12. Version every subset
Partner terminologies and subsets can be distributed within NCIt and updated on their own schedules.
Working checkpoint — ontology release: Store subset and NCIt versions together. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 13 of 36
13. Expand value sets reproducibly
A value-set expansion is a time-stamped result of rules against a terminology release.
Working checkpoint — NCI Thesaurus: Archive the expansion inputs and member list. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 14 of 36
14. Handle retired concepts
Historical records can contain concepts no longer active for new coding.
Working checkpoint — NCIt concept code: Preserve capture and attach current status. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 15 of 36
15. Handle merges transparently
A merge may redirect several earlier codes to one current concept.
Working checkpoint — cancer terminology: Keep every original code and mapping receipt. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 16 of 36
16. Map by meaning, not string
Matching labels can hide differences in scope, and different labels can be true synonyms.
Working checkpoint — biomedical ontology: Compare code, definition, hierarchy and use case. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 17 of 36
17. Represent mapping strength
Crosswalks may be exact, broader, narrower or related rather than equivalent.
Working checkpoint — preferred term: Store relation type and reviewer. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 18 of 36
18. Keep clinical evidence separate
NCIt identifies concepts used in research and clinical information; it does not evaluate a patient or recommend care.
Working checkpoint — ontology release: Link evidence and decision rules explicitly. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 19 of 36
19. Connect NCIt to ICD-O-4
ICD-O-4 classifies tumour topography and morphology, while NCIt covers a broader ontology of cancer and biomedical concepts.
Working checkpoint — NCI Thesaurus: Preserve both codes when each has a role. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 20 of 36
20. Connect NCIt to HGNC
HGNC provides authoritative human gene nomenclature; NCIt can reference gene and cancer concepts in a broader research model.
Working checkpoint — NCIt concept code: Do not replace HGNC identity with a display synonym. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 21 of 36
21. Connect NCIt to ChEBI
ChEBI specialises in chemical entities of biological interest, while NCIt includes drug and substance concepts for cancer research.
Working checkpoint — cancer terminology: Use curated mappings and retain source IDs. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 22 of 36
22. Validate free-text coding
Automated coding can miss negation, hypothetical language and local abbreviations.
Working checkpoint — biomedical ontology: Review low-confidence and high-impact assignments. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 23 of 36
23. Test terminology services
APIs should return exact codes, releases and status while handling timeouts without stale ambiguity.
Working checkpoint — preferred term: Use fixtures for current, retired and unknown concepts. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 24 of 36
24. Protect patient context
A code may be non-identifying alone while the surrounding clinical record remains sensitive.
Working checkpoint — ontology release: Apply role-based access and minimisation. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 25 of 36
25. Preserve provenance
A reusable coding record links source phrase, selected code, release, mapper, confidence, rationale and validation.
Working checkpoint — NCI Thesaurus: Make every aggregate auditable. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 26 of 36
26. Explain ontology limits
A well-formed graph can still reflect incomplete knowledge or a scope decision.
Working checkpoint — NCIt concept code: Report gaps and requests for curation. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 27 of 36
27. Treat updates as governed change
New concepts and revised relations help the terminology follow science.
Working checkpoint — cancer terminology: Regression-test dependent forms and queries before deployment. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 28 of 36
28. Connect to Primary Science
A fictional fictional cancer-concept card set lets pupils see how careful labels make observations easier to compare without turning a label into an explanation.
Working checkpoint — biomedical ontology: Use invented examples and ask what evidence still needs to be collected. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 29 of 36
29. Connect to Secondary Science
NCI Thesaurus supports lessons about classification, evidence, models, uncertainty and responsible scientific communication.
Working checkpoint — preferred term: Trace one database statement back to its source and limit. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 30 of 36
30. Connect to mathematics
Sets, graphs, hierarchies, identifiers and versioned mappings show why definitions and boundaries must be fixed before quantities are compared.
Working checkpoint — ontology release: Work through one small boundary case by hand. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 31 of 36
31. Connect to computing
OWL reasoners, APIs, subsets, release archives and validators show how software can preserve shared biomedical meaning.
Working checkpoint — NCI Thesaurus: Break one identifier, relationship, release or mapping, then validate and repair it transparently. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Section 32 of 36
32. Connect to pathways
NCIt connects biology, medicine, oncology, health informatics, statistics and data governance and can illuminate STEM education and career pathways.
Working checkpoint — NCIt concept code: Check current official education and qualification routes separately from this conceptual guide. Read every coded concept beside definition, hierarchy, source terminology, workflow and supporting evidence.
For future study, connect biology to oncology and informatics while checking pathways separately.
Section 33 of 36
33. Did you know?
The current official download checked for this article is NCIt 26.09d, completed on 28 September 2026.
Working checkpoint — cancer terminology: Explain the fact in plain language, then name what it does not prove. Preserve retired and merged codes with migration receipts; a current label is not the complete history.
For writing, cite NCIt release, code, distribution, subset, mapping and interpretation limits.
Section 34 of 36
34. Write a reader-ready methods note
State the resource and release, source data, identifiers, mapping method, validation, privacy controls and supported claim.
Working checkpoint — biomedical ontology: Keep the note beside the machine-readable record. Separate terminology identity from diagnosis, clinical evidence, eligibility and treatment guidance.
For students, ask how a stable concept number can survive a change in preferred wording.
Section 35 of 36
35. Build a validation ladder
Check identity, release, syntax, hierarchy, provenance, scientific plausibility and known fixtures in sequence.
Working checkpoint — preferred term: Stop at the earliest unresolved layer. Treat release, distribution type, value set and mapping rule as scientific inputs: each can change results.
For parents, ask whether a cancer-data claim distinguishes a terminology code from a diagnosis.
Section 36 of 36
36. Finish with a reproducible bundle
Deliver source receipts, releases, mappings, software versions, tests, hashes and limitations without exposing governed data.
Working checkpoint — ontology release: Make every published claim traceable. Keep source phrase, NCIt code, preferred term, release, subset, mapping relation, confidence and review decision together.
For a classroom model, code invented research concepts and compare broader and narrower terms.
Contents · Previous section · Continue to the Science Learning Hub