DICOMweb
Category: Healthcare Standards
Node id: dicomweb
Standards: HIPAA, GDPR, SOC2
Unified DICOMweb node: query studies (QIDO-RS), retrieve images/metadata (WADO-RS), and store instances (STOW-RS)
What it does
- QIDO-RS: search studies, series, instances by patient, date, modality, UIDs
- WADO-RS: retrieve metadata, rendered images, or bulk DICOM data
- STOW-RS: store single or batch DICOM instances
- Bearer token or HTTP Basic authentication
- Per-node CA certificate for PACS behind a private CA
- Pagination support for queries
- Streaming bulkdata retrieval (one instance in RAM at a time)
Use cases
- Search a PACS for imaging studies by patient or date range
- Retrieve DICOM metadata for clinical data pipelines
- Download rendered images for AI model inference
- Store DICOM images in a cloud PACS
- Migrate imaging studies between DICOMweb archives
Runtime
- Performance tier:
variable
Input ports
| Name | Type | Required | Description |
|---|---|---|---|
patient_id | text | No | Patient ID to search for (query) |
patient_name | text | No | Patient name to search for (query) |
study_date | text | No | Study date filter YYYYMMDD (query) |
modality | text | No | Modality filter e.g. CT, MR, US (query) |
accession_number | text | No | Accession number (query) |
study_instance_uid | text | No | Study Instance UID (query/retrieve) |
series_instance_uid | text | No | Series Instance UID (query/retrieve) |
sop_instance_uid | text | No | SOP Instance UID (retrieve) |
dicom_data | any | No | DICOM instance data — bytes or string (store) |
dicom_instances | json | No | Array of DICOM instances for batch upload (store) |
Output ports
| Name | Type | Required | Description |
|---|---|---|---|
result | json | No | Operation result — shape varies by operation |
metadata | json | No | Response metadata (retrieve only) |
success | boolean | No | Whether the operation succeeded |
Settings
Basic
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| DICOMweb URL | dicomweb_url | string | Yes | Base URL of the DICOMweb server | ||
| Operation | operation | enum | Yes | query, retrieve, store | DICOMweb operation to perform |
Auth
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Auth Token | auth_token | secret | No | Bearer token for DICOMweb server authentication (not with username) | ||
| Username | username | string | No | HTTP Basic username for the DICOMweb server (not with auth_token) | ||
| Password | password | secret | No | HTTP Basic password (with username) | ||
| CA Certificate | ca_cert | text | No | CA bundle (PEM content or file path) the server certificate must chain to, for a PACS behind a private/hospital CA. Adds trust; verification stays on |
Query
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Query Level | query_level | enum | No | studies | studies, series, instances | QIDO-RS query level (query) |
| Limit | limit | number | No | 100 | min 1 | Maximum results to return (query) |
| Offset | offset | number | No | 0 | min 0 | Results to skip for pagination (query) |
| Include Fields | include_fields | string | No | Comma-separated DICOM tag names to include (query) |
Retrieve
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Retrieve Type | retrieve_type | enum | No | metadata | metadata, rendered, bulkdata | WADO-RS retrieval mode (retrieve) |
| Retrieve Timeout (s) | retrieve_timeout_seconds | number | No | 300 | min 1 | HTTP timeout for a WADO-RS bulkdata study pull. Raise for large/slow PACS studies to avoid 'context deadline exceeded'. |
| Media Type | media_type | string | No | image/png | Requested media type for rendered output | |
| Quality | quality | string | No | 90 | Rendering quality 1-100 (retrieve rendered) | |
| Viewport | viewport | string | No | Viewport dimensions e.g. 512,512 (retrieve rendered) |
Store
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Study Instance UID | study_instance_uid | string | No | Target Study UID for study-level STOW-RS (store) |
Errors
| Code | Description |
|---|---|
DICOMWEB_URL_REQUIRED | dicomweb_url configuration is missing |
OPERATION_REQUIRED | operation must be query, retrieve, or store |
STUDY_UID_REQUIRED | study_instance_uid is required for retrieve |
DICOM_DATA_REQUIRED | dicom_data input is required for store |
SERVER_ERROR | DICOMweb server returned an HTTP 4xx/5xx error |
Runtime notes
- query: result has results[], result_count, query_level, http_status
- retrieve (metadata): result has dicom_metadata[], instance_count
- retrieve (rendered): result has image_data, content_size
- retrieve (bulkdata): result has instances[], instance_count, total_bytes, temp_dir
- retrieve (bulkdata, multi-study): when the input carries a LIST of studies (upstream query result.studies/result.results array, a bare array of study objects, or multiple study_instance_uid values), ALL studies are retrieved and flattened into result.instances[] (each enriched with study_instance_uid; series_instance_uid is set only when the study was fetched series by series, absent on the whole-study fallback; sop_instance_uid is never set — treat both as optional). Also: result.study_count, result.studies_ok, result.errors[] (per-study {study_instance_uid, error} for tolerated failures). Single-study input keeps the original shape.
- store: result has status, stored_instances[], instance_count, response, http_status
See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.
Quelque chose n'est pas clair ? Envoyez vos commentaires sur la documentation par courriel.