Inferensys

Glossary

DICOMweb

A set of RESTful web services defined in DICOM Part 18 for querying, retrieving, and storing medical images using HTTP-based protocols like WADO-RS, QIDO-RS, and STOW-RS.
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.
RESTful Medical Imaging

What is DICOMweb?

DICOMweb is a set of RESTful web services defined in DICOM Part 18 that enables querying, retrieving, and storing medical images using standard HTTP protocols.

DICOMweb is the modern, HTTP-based API for medical imaging, defined in DICOM Part 18. It replaces legacy DIMSE network commands with RESTful services like WADO-RS for retrieval, QIDO-RS for querying, and STOW-RS for storage. By using standard application/dicom+json and multipart/related media types, it allows web developers to integrate imaging data without specialized DICOM networking libraries.

This standard is the cornerstone of cloud-based PACS and VNA architectures, enabling zero-footprint medical viewers. Unlike traditional Association Negotiation, DICOMweb relies on standard HTTP authentication and TLS, simplifying firewall traversal and interoperability between disparate health IT systems.

RESTful Medical Imaging

Core DICOMweb Services

DICOMweb defines a set of HTTP-based RESTful services for querying, retrieving, and storing medical images, replacing legacy DIMSE protocols with modern web APIs.

06

Capabilities & Conformance

A DICOMweb server's functionality is described by its Conformance Statement, which details supported services and media types. Discovery is often automated via a Capabilities resource.

  • Capabilities Endpoint: GET / or a specific URL returns a JSON document listing all supported services (QIDO-RS, WADO-RS, STOW-RS).
  • Media Type Negotiation: Clients use standard HTTP Accept headers to request application/dicom+json or application/dicom+xml.
  • Authentication: Typically handled by standard HTTP mechanisms like OAuth 2.0 Bearer tokens.
NETWORK PROTOCOL COMPARISON

DICOMweb vs. DIMSE: A Protocol Comparison

A technical comparison of the RESTful DICOMweb services defined in DICOM Part 18 against the legacy DIMSE protocol for medical imaging network operations.

FeatureDICOMwebDIMSENotes

Architectural Style

RESTful (HTTP/HTTPS)

Stateful TCP/IP Association

DICOMweb uses standard web protocols; DIMSE requires a persistent negotiated connection

Transport Protocol

HTTP/1.1 or HTTP/2

TCP/IP (Port 104 or 2762)

DICOMweb operates over standard web ports (80/443); DIMSE uses dedicated DICOM ports

Data Encoding

JSON, XML, multipart/related

Binary DICOM (Little/Big Endian)

DICOMweb supports human-readable metadata; DIMSE uses raw binary encoding per negotiated Transfer Syntax

Authentication Mechanism

OAuth 2.0, JWT, API Keys, Mutual TLS

DICOM Association Negotiation only

DICOMweb integrates with enterprise IAM; DIMSE relies on IP-based access control and TLS

Firewall Traversal

DICOMweb uses standard HTTPS and is firewall-friendly; DIMSE often requires VPN or dedicated network configuration

Query Service

QIDO-RS (GET with query params)

C-FIND (DIMSE-C)

QIDO-RS returns JSON/XML; C-FIND returns binary DICOM datasets

Retrieve Service

WADO-RS (GET by UID)

C-MOVE / C-GET (DIMSE-C)

WADO-RS is a direct pull; C-MOVE requires a third-party destination AE Title

Store Service

STOW-RS (POST multipart/related)

C-STORE (DIMSE-C)

STOW-RS uses a single HTTP POST; C-STORE sends one instance per command

Stateless Operation

DICOMweb requests are independent; DIMSE requires an active association for the duration of operations

Load Balancing Support

DICOMweb works with standard HTTP load balancers; DIMSE requires sticky sessions or DICOM-aware proxies

Caching Support

DICOMweb responses can leverage HTTP caching headers and CDNs; DIMSE has no native caching mechanism

Metadata Retrieval

WADO-RS metadata endpoint

C-FIND or C-GET

DICOMweb returns JSON metadata without pixel data; DIMSE always returns full DICOM datasets

Frame-Level Access

WADO-RS supports byte-range requests for individual frames; DIMSE retrieves entire SOP Instances

Bulk Data Handling

multipart/related with chunked transfer

PDV fragmentation within association

DICOMweb uses standard HTTP chunking; DIMSE fragments data into Presentation Data Values

Error Handling

HTTP status codes (4xx, 5xx)

DIMSE status codes in command sets

DICOMweb uses familiar web error semantics; DIMSE uses DICOM-specific status codes

Standard Specification

DICOM Part 18 (PS3.18)

DICOM Part 7 (PS3.7)

DICOMweb is a modern supplement; DIMSE is the original network specification

Adoption in New Deployments

Increasing (cloud-native PACS)

Declining (legacy systems)

DICOMweb is preferred for new architectures; DIMSE remains for backward compatibility

Suitable for Cloud/Hybrid

DICOMweb is designed for distributed cloud environments; DIMSE is optimized for local area networks

DICOMWEB INTEROPERABILITY

Frequently Asked Questions

Clear, technical answers to the most common questions about implementing and troubleshooting DICOMweb RESTful services for modern medical imaging workflows.

DICOMweb is a set of RESTful web services defined in DICOM Part 18 that enables querying, retrieving, and storing medical images using standard HTTP protocols. Unlike the legacy DIMSE (DICOM Message Service Element), which relies on a persistent, stateful TCP/IP association negotiated on port 104, DICOMweb uses stateless HTTP/HTTPS requests. This fundamental architectural shift means DICOMweb leverages standard web infrastructure—load balancers, reverse proxies, and firewalls—without requiring specialized DICOM network configurations. The primary services include WADO-RS (Web Access to DICOM Objects), QIDO-RS (Query based on ID for DICOM Objects), STOW-RS (Store Over the Web), and WADO-URI (a legacy URI-based retrieval method). DICOMweb returns data in consumer-friendly media types like application/dicom+json and application/dicom+xml, making it significantly easier for web developers and mobile application engineers to integrate medical imaging into modern healthcare platforms without deep DICOM protocol expertise.

Prasad Kumkar

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.