Glossary

DICOM Structured Report

A DICOM Structured Report (SR) is an information object that encodes clinical observations and measurements as structured, machine-readable text with coded concepts, replacing traditional free-text radiology reports.

Get in touch Learn more

Stylish WeWork-like workspace with hot desks and document wall, professional searching through enterprise knowledge base on a mounted ultrawide display, warm industrial pendants overhead.

MACHINE-READABLE CLINICAL DOCUMENTATION

What is DICOM Structured Report?

A DICOM Structured Report (SR) is a DICOM information object that encodes clinical observations and measurements as structured, machine-readable text with coded concepts, replacing traditional free-text radiology reports with computable data.

A DICOM Structured Report (SR) is a DICOM SOP Class that encodes clinical findings, measurements, and observations as a tree of content items linked by explicit relationships, rather than as a narrative block of free text. Each content item contains a coded concept drawn from a controlled terminology like SNOMED CT or RadLex, paired with a value that can be numeric, textual, or an image reference. This transforms a diagnostic report from a document meant only for human reading into a computable data structure that can be queried, analyzed, and integrated directly into clinical decision support systems.

The SR object uses a parent-child relationship model defined by the DICOM Content Mapping Resource (DCMR) to preserve the semantic context of findings, such as linking a measurement to the specific image region from which it was derived. Unlike a Secondary Capture image of a report, an SR maintains the machine-readable semantics of every observation, enabling automated tumor response tracking, clinical trial data extraction, and structured data mining across large patient cohorts without manual abstraction.

STRUCTURED REPORTING

Key Features of DICOM Structured Reports

DICOM Structured Reports (SR) transform free-text radiology narratives into machine-readable, coded information objects, enabling automated data mining, clinical decision support, and seamless integration with electronic health records.

Coded Concept Encoding

Unlike narrative text, SR documents encode clinical observations using controlled terminologies such as SNOMED CT, LOINC, or RadLex. Each measurement, finding, or anatomical reference is stored as a triplet of Code Value, Coding Scheme Designator, and Code Meaning.

Eliminates semantic ambiguity inherent in free-text reports
Enables automated cross-study comparison and cohort identification
Facilitates direct integration with clinical decision support systems

Hierarchical Content Tree

SR documents organize clinical data into a strict parent-child relationship tree defined by the Document Relationship Macro. The root node is the TID 1000 (General Relevant Patient Information) template, which branches into sections, sub-sections, and individual measurements.

Each node has a Relationship Type (CONTAINS, HAS PROPERTIES, INFERRED FROM)
Enables deterministic parsing and navigation of complex diagnostic logic
Supports nested findings such as a lung nodule with associated measurements and morphology descriptors

Template-Driven Standardization

The DICOM Part 16 standard defines Template IDs (TIDs) that prescribe the exact structure and allowed content for specific clinical use cases. For example, TID 1500 governs measurement templates, while TID 2000 covers basic diagnostic imaging reports.

Ensures consistent reporting across different vendors and institutions
Templates define mandatory, conditional, and optional content items
Enables validation of report completeness and compliance at the point of creation

Measurement and Spatial Referencing

SR documents can embed precise quantitative measurements with explicit references to spatial regions in the source images. Using DICOM Segmentation Objects or Spatial Coordinates, a report can link a finding directly to a specific pixel location or 3D volume of interest.

Stores linear distances, areas, volumes, and density values with units
References the exact SOP Instance UID and frame number of the source image
Enables downstream quantitative analysis and radiomics feature extraction

Enhanced SR IOD

The Enhanced SR Information Object Definition extends the Basic Text SR to support multimedia evidence. It allows the inclusion of by-reference or by-value images, waveforms, and spatial coordinates directly within the report structure.

Supports key images, regions of interest, and time-based waveforms
Maintains the provenance chain linking findings to original acquisition data
Required for advanced use cases like computer-aided detection results and quantitative imaging biomarkers

Completion and Verification Flags

Every SR document carries a Completion Flag (PARTIAL or COMPLETE) and a Verification Flag (UNVERIFIED or VERIFIED). These metadata attributes signal the clinical readiness and authority of the report content to downstream consumers.

A PARTIAL report indicates a preliminary or work-in-progress document
A VERIFIED status confirms the content has been reviewed and signed by a responsible observer
Critical for managing clinical workflows and medicolegal accountability

DICOM SR

Frequently Asked Questions

Clear answers to common questions about the DICOM Structured Report standard, its encoding mechanisms, and its role in transitioning radiology from free-text dictation to machine-readable, coded clinical data.

A DICOM Structured Report (SR) is a DICOM Information Object Definition (IOD) that encodes clinical observations and measurements as structured, machine-readable text with coded concepts instead of a traditional free-text narrative. Unlike a dictated prose report stored as a Secondary Capture or encapsulated PDF, an SR separates the content tree from the presentation. Each finding is a node in a hierarchical tree, linked to a coded concept from a terminology like SNOMED CT or RadLex. This allows a computer to query for specific measurements—such as 'retrieve all reports where the left ventricular ejection fraction is less than 40%'—without natural language processing. The SR maintains the parent-child relationships between observations, preserves the context of measurements, and can directly reference the source DICOM images via SOP Instance UIDs, creating a closed-loop diagnostic record.

Enabling Efficiency, Speed & Accuracy

Intelligent Analysis, Decision & Execution

We build AI systems for teams that need search across company data, workflow automation across tools, or AI features inside products and internal software.

Talk to Us

Search across company data

Give teams answers from docs, tickets, runbooks, and product data with sources and permissions.

Useful when people spend too long searching or get different answers from different systems.

Enterprise searchRAGPermissions

Automate internal workflows

Use AI to route work, draft outputs, trigger actions, and keep approvals and logs in place.

Useful when repetitive work moves across multiple tools and teams.

AI agentsWorkflow automationGovernance

Add AI to products and internal tools

Build assistants, guided actions, or decision support into the software your team or customers already use.

Useful when AI needs to be part of the product, not a separate tool.

AI integrationDecision supportModel routing

DICOM SR ECOSYSTEM

Related Terms

DICOM Structured Reports do not exist in isolation. They are built upon a foundation of coded concepts, templates, and relationships that enable machine-readable clinical documentation.

DICOM Controlled Terminology

The coded vocabulary backbone of every Structured Report. DICOM Part 16 defines standardized codesets like SNOMED CT, LOINC, and RadLex that replace free-text descriptions with unambiguous, machine-readable concepts. A finding of 'mass' is encoded as a specific SNOMED code rather than a string, enabling deterministic querying and data mining across disparate systems.

EXPLORE

DICOM Segmentation Object

A companion SOP Class that encodes regions of interest as binary or fractional maps. In an SR context, a radiologist's measurement of a tumor volume is often derived from a DICOM Segmentation Object that precisely delineates the lesion boundary. The SR captures the quantitative measurement, while the Segmentation Object provides the spatial proof and enables volumetric rendering.

SOP Class

The fundamental unit of DICOM interoperability. A Structured Report is defined by a specific SOP Class UID that combines the SR Information Object Definition (IOD) with the Storage Service Class. Key SR SOP Classes include:

Basic Text SR (1.2.840.10008.5.1.4.1.1.88.11)
Enhanced SR (1.2.840.10008.5.1.4.1.1.88.22)
Comprehensive SR (1.2.840.10008.5.1.4.1.1.88.33)
Mammography CAD SR (1.2.840.10008.5.1.4.1.1.88.50)

Query/Retrieve Level

Defines the hierarchical scope for searching SR documents in a PACS or VNA. An SR is stored at the Series Level within a Study, meaning a C-FIND at the Study level retrieves all associated reports. Understanding this hierarchy is critical for building efficient DICOMweb QIDO-RS queries that filter for specific report types without pulling entire studies.

DICOM De-identification Profile

The standardized ruleset in DICOM Part 15 for scrubbing Protected Health Information (PHI) from SR documents. Unlike pixel data anonymization, SR de-identification must parse structured content sequences to remove or pseudonymize coded entries, observer identities, and institution names embedded deep within the content tree, while preserving clinical concept integrity.

DICOMweb STOW-RS

The modern RESTful mechanism for storing SR instances. Using an HTTP POST with multipart/related content, a reporting system can transmit a Comprehensive SR document directly to a PACS or VNA without establishing a traditional DIMSE association. This is the preferred integration pattern for web-based structured reporting platforms and AI-driven analysis tools.

EXPLORE

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.

Limited slotsGet a Free AI Consultation

How We Work

Custom AI workflows for your Business

One-fit-all AI don't work for modern businesses. At Inferensys, we aim to understand your business & custom requirements; which we use to define most efficient agentic workflows, the data, and the tools for your business.