DICOM Controlled Terminology

A Context Group is a predefined set of coded concepts that defines the allowed values for a specific clinical context. Each concept is a triplet of a Code Value, Coding Scheme Designator (e.g., SNOMED-CT, RadLex), and Code Meaning.

• Example: CID 6042 'Breast Imaging Report Findings' contains codes for 'Mass', 'Calcification', and 'Architectural Distortion'.
• Purpose: Restricts data entry to a controlled vocabulary, eliminating synonymy and ambiguity in structured reports.

A Template defines the structure and content of a document or a section of a document. It specifies which coded concepts are mandatory, optional, or conditional, and how they relate to each other in a hierarchical tree.

• Example: TID 1500 'Measurement Report' dictates how to encode a measurement's value, unit, and the anatomic location it was taken from.
• Function: Templates enforce a consistent document architecture, ensuring that a 'Systolic Blood Pressure' measurement is always recorded with the same structure and codes.

Controlled Terminology defines not just flat lists of codes but also the semantic relationships between them. A Value Set is a dynamically generated list of codes that are valid for a specific template field.

• Parent-Child Hierarchies: Codes can be organized in 'is-a' relationships (e.g., 'Invasive Ductal Carcinoma' is a child of 'Breast Carcinoma').
• Cross-Mapping: Terminology maps concepts across different coding schemes, such as linking a local hospital code to a standard LOINC code for lab results.

DICOM Controlled Terminology is not an isolated standard; it heavily leverages and profiles external, domain-specific lexicons to ensure clinical relevance.

• SNOMED CT: The primary source for anatomic locations, morphologic abnormalities, and procedures. DICOM restricts the vast SNOMED CT ontology to specific, clinically relevant subsets.
• RadLex: A lexicon from the Radiological Society of North America (RSNA) used extensively for radiology-specific findings and imaging technique descriptions.
• LOINC: Used for laboratory and some clinical observation codes to ensure interoperability with non-imaging systems.

The terminology is distributed as a machine-readable resource, typically in XML (using the DICOM Content Mapping Resource schema) or JSON format. This allows software systems to programmatically load and validate coded data.

• Validation: An application can automatically check if a code used in a Structured Report is a member of the correct Context Group.
• Rendering: A display system can look up the human-readable 'Code Meaning' for a given 'Code Value' to present information to a clinician without requiring them to memorize codes.

Beyond diagnostic reporting, Controlled Terminology standardizes the description of the imaging acquisition itself. This includes coded definitions for Contrast Agent, Patient Position, and Imaging Procedure Step.

• Example: A CT scan's protocol can be unambiguously tagged with a code for 'CT Abdomen with IV Contrast' rather than relying on a free-text protocol name.
• Benefit: This enables automated protocol-driven workflows, dose management analytics, and reliable querying of imaging archives for specific acquisition techniques.

DICOM Controlled Terminology heavily relies on SNOMED CT as its primary external reference vocabulary. DICOM Part 16 defines a subset of SNOMED CT concepts that are valid for specific DICOM contexts, ensuring that clinical findings in structured reports are mapped to globally recognized, computable codes rather than ambiguous free text. This integration allows for semantic interoperability between imaging systems and broader electronic health records.

Logical Observation Identifiers Names and Codes (LOINC) are used within DICOM to standardize the names of imaging procedures and observations. By mapping a local procedure code to a universal LOINC identifier, DICOM enables cross-institutional data aggregation and clinical research. This ensures that a 'CT Abdomen with Contrast' is semantically identical across different hospital information systems.

A Context Group is a defined set of coded concepts that are permissible for a specific DICOM attribute in a particular clinical context. For example, a Context Group for 'Contrast Agent' would list all valid SNOMED CT codes for iodinated and gadolinium-based agents. This mechanism constrains terminology to a manageable, clinically relevant subset, preventing nonsensical coding and ensuring data consistency.

The DCMR is the machine-readable encoding of DICOM Controlled Terminology, defined in Part 16. It specifies the relationships between coded concepts, their human-readable meanings, and their assignment to specific Context Groups. Software systems parse the DCMR to validate coded entries in real-time, ensuring that a structured report only contains allowed terminology for each data element.

The Coded Entry Sequence (0040,A168) is a critical DICOM attribute used in Structured Reports to encode a concept. Each item in this sequence contains a triplet: - Code Value: The machine-readable identifier (e.g., '373098007') - Coding Scheme Designator: The vocabulary authority (e.g., 'SCT' for SNOMED CT) - Code Meaning: The human-readable description (e.g., 'Neoplasm, Benign') This structure provides unambiguous, multilingual semantics.

DICOM Structured Reports are built using Templates, which define the hierarchical structure of content items (e.g., 'Finding' contains 'Lesion' contains 'Measurement'). Each node in a template is associated with a Context Group that restricts the allowed coded concepts. This template-driven approach enforces a rigorous, standardized format for clinical documentation, making reports fully computable and queryable.