Digital Imaging and Communications in Medicine
DICOM (Digital Imaging and Communications in Medicine) is the international standard for medical imaging. It defines formats for storing medical images and protocols for exchanging them between systems, enabling interoperability across vendors and institutions.
DICOM organizes medical images in a hierarchical structure:
DICOM uses globally unique identifiers to ensure each object can be unambiguously referenced:
| UID Type | Scope | Example Use |
|---|---|---|
| Study Instance UID | Unique per study | Links all series in one exam |
| Series Instance UID | Unique per series | Groups related images |
| SOP Instance UID | Unique per object | Identifies individual image |
| SOP Class UID | Defines object type | CT Image Storage, MR Image Storage |
DICOM files contain data elements (attributes) identified by tags. Each tag consists of a group and element number in hexadecimal format.
| Tag | Name | Description |
|---|---|---|
(0010,0010) |
Patient Name | Patient's full name |
(0010,0020) |
Patient ID | Medical record number |
(0008,0020) |
Study Date | Date study performed |
(0008,0060) |
Modality | CT, MR, US, CR, etc. |
(0008,1030) |
Study Description | Description of examination |
(0028,0010) |
Rows | Image height in pixels |
(0028,0011) |
Columns | Image width in pixels |
(7FE0,0010) |
Pixel Data | Actual image pixel values |
DICOM defines network protocols for exchanging images and worklists between systems:
| Service | Operation | Use Case |
|---|---|---|
| C-STORE | Store image | Send images to PACS |
| C-FIND | Query database | Search for studies |
| C-MOVE | Retrieve images | Pull images from PACS |
| C-GET | Retrieve directly | Alternative to C-MOVE |
| C-ECHO | Verify connectivity | Test DICOM connection |
Before exchanging data, DICOM systems establish an "association" where they negotiate supported services (SOP Classes), transfer syntaxes (encoding), and roles. This handshake ensures both parties can communicate effectively.
DICOMweb provides RESTful web services for DICOM operations, enabling modern web-based applications:
| Service | HTTP Method | Purpose |
|---|---|---|
| WADO-RS | GET | Retrieve studies/series/instances |
| STOW-RS | POST | Store DICOM objects |
| QIDO-RS | GET | Query for studies/series |
| UPS-RS | Various | Unified worklist service |
// Example QIDO-RS query for studies
GET /dicomweb/studies?PatientName=Smith*&StudyDate=20240101-20240131
// Example WADO-RS retrieve
GET /dicomweb/studies/1.2.3.../series/1.2.4.../instances/1.2.5...DICOM supports multiple transfer syntaxes for encoding pixel data:
| Transfer Syntax | Compression | Use Case |
|---|---|---|
| Implicit VR Little Endian | None | Default, widely supported |
| Explicit VR Little Endian | None | Clear VR specification |
| JPEG Lossless | Lossless ~3:1 | Diagnostic quality archival |
| JPEG 2000 Lossless | Lossless ~3:1 | Modern lossless standard |
| JPEG 2000 Lossy | Lossy 10:1+ | Review, non-diagnostic |
| JPEG-LS | Lossless/near-lossless | Efficient compression |
Lossy compression should be used with caution in medical imaging. While it reduces storage requirements, it may affect diagnostic accuracy. Regulatory guidelines and institutional policies typically specify acceptable compression ratios for different clinical uses.
DICOM SR enables structured, machine-readable clinical reports with coded content:
Integrating the Healthcare Enterprise (IHE) defines implementation profiles that specify how DICOM and other standards work together in clinical workflows:
| Profile | Purpose | Key Standards |
|---|---|---|
| SWF | Scheduled Workflow | HL7, DICOM Worklist |
| XDS-I | Cross-enterprise image sharing | DICOM, HL7, XDS |
| PIR | Patient Information Reconciliation | HL7, DICOM |
| IOCM | Imaging Object Change Management | DICOM |
Every DICOM-compliant product must publish a Conformance Statement documenting its supported SOP Classes, transfer syntaxes, and network roles. Reviewing conformance statements is essential when integrating imaging systems to ensure compatibility.
Medical image compression sits at the intersection of diagnostic quality preservation and storage and transmission efficiency. The simulator's COMPRESSION ENUM exposes eight standardised compression strategies, each with its corresponding DICOM Transfer Syntax UID and threshold parameters.
| ENUM | DICOM Transfer Syntax UID | Lossy / Lossless | Primary Use |
|---|---|---|---|
| JPEG_BASELINE | 1.2.840.10008.1.2.4.50 | Lossy | 8-bit visible imagery |
| JPEG_LOSSLESS | 1.2.840.10008.1.2.4.70 | Lossless | Legacy diagnostic imaging |
| JPEG_2000_LOSSLESS | 1.2.840.10008.1.2.4.90 | Lossless | CT and MR diagnostic standard |
| JPEG_2000_LOSSY | 1.2.840.10008.1.2.4.91 | Lossy | Long-term archive |
| JPEG_LS | 1.2.840.10008.1.2.4.80 / .81 | Lossless / near-lossless | X-ray, CR, DR |
| HEVC | 1.2.840.10008.1.2.4.107–109 | Lossy / lossless | Video and 4D imaging |
| RLE | 1.2.840.10008.1.2.5 | Lossless | Binary or low-frequency images |
| DEFLATE | 1.2.840.10008.1.2.1.99 | Lossless | Non-pixel metadata payloads |
Each compression ENUM is selectable from the simulator's Panel 1 (Algorithm) compression dropdown. When the user selects an ENUM the simulator estimates the expected compression ratio, PSNR (Peak Signal-to-Noise Ratio), SSIM (Structural Similarity Index Measure), and encoding/decoding throughput in megabytes per second. JPEG 2000 Lossless is the most widely adopted scheme for CT and MR diagnostic imaging, typically achieving 2:1 to 3:1 compression while guaranteeing pixel-level fidelity.
JPEG-LS (ISO/IEC 14495), built on the LOCO-I (Low Complexity Lossless Compression for Images) algorithm, has acquired particular significance in medical imaging because of its combination of low computational complexity and strong compression ratios. For high-resolution, high-bit-depth modalities such as X-ray, CR, and DR, JPEG-LS encodes approximately two to three times faster than JPEG 2000 Lossless while delivering five to ten per cent better compression ratios.
| Modality | Recommended | Alternative | Rationale |
|---|---|---|---|
| CT diagnostic | JPEG_2000_LOSSLESS | JPEG_LS | Absolute preservation of Hounsfield Units |
| MR diagnostic | JPEG_2000_LOSSLESS | JPEG_LS | Preservation across multiple sequences |
| X-ray | JPEG_LS | JPEG_2000_LOSSLESS | High bit depth and fast decoding |
| Ultrasound (still) | JPEG_2000_LOSSY (q90+) | JPEG_BASELINE (q95) | Real-time imaging characteristics |
| Ultrasound (video) | HEVC (CRF 18) | JPEG_BASELINE sequence | Bandwidth efficiency |
| Pathology WSI | JPEG_2000_LOSSLESS | HEVC | Microscopy gigapixel data |
| Nuclear medicine | RLE | JPEG_2000_LOSSLESS | Low-frequency SPECT/PET imagery |
| Mammography | JPEG_2000_LOSSLESS | JPEG_LS | Mandatory preservation of microcalcifications |
The acceptability of lossy compression in clinical imaging must be quantitatively validated. DICOM WG-04 (the Compression Working Group) issues per-modality recommendations: CT permits 8:1 to 10:1 compression while preserving diagnostic accuracy above 99 per cent (per Royal College of Radiologists studies); MR tolerates 6:1 to 8:1 with no measurable clinical difference (per Canadian Association of Radiologists guidance); mammography is recommended to remain lossless to preserve microcalcifications; ultrasound permits 8:1 with Doppler data retained losslessly; and nuclear medicine permits 4:1 subject to verification that quantitative SUV measurements are unaffected.
PSNR ≥ 40 dB and SSIM ≥ 0.98 are widely cited acceptance thresholds for diagnostic-grade imagery. The simulator automatically computes both metrics whenever an ENUM is selected, allowing readers to relate textbook acceptance criteria to live numerical outputs.
"Lossy compression in medical imaging shall be permitted only within clinically validated compression ratio limits. For modalities where fine structures and low-contrast lesions are diagnostically critical — mammography and high-resolution CT in particular — lossless compression shall remain the principle." — DICOM WG-04, Compression Guidelines, 2024.
HEVC (High Efficiency Video Coding, ISO/IEC 23008) achieves approximately 50 per cent better compression efficiency than its H.264/AVC predecessor. In medical imaging it has gained traction for 4D CT (a CT volume varying over time), real-time ultrasound cine loops, and angiographic cine series. The DICOM standard adopted HEVC as a formal Transfer Syntax in 2017 (UIDs 1.2.840.10008.1.2.4.107 through .109), supporting the Main, Main 10, and Lossless profiles.
In practical deployments HEVC delivers substantial storage and transmission savings, simultaneously enabling real-time remote consultation and large-scale archival. Several leading Korean teaching hospitals operate HEVC-based 4D CT archives that report storage cost reductions of approximately 60 per cent relative to legacy multi-frame encoding while preserving clinical quality.
The Korean edition (한국어판) extends this chapter with section §3.A on Korean compression regulatory infrastructure, including Article 22 of the Medical Service Act (preservation of medical records), Article 23 of the Personal Information Protection Act (sensitive health information), the Ministry of Health and Welfare Standard for Medical Image Data Preservation, and the Korean Society of Medical Informatics (KOSMI) Compression Recommendation. The K-PACS Standard, jointly published by MFDS and KOSMI in 2024, is also discussed in detail.
Compression policy in a hospital is rarely a single decision; it is a governance regime that integrates clinical, regulatory, and operational considerations. The radiology department typically chairs the policy committee, with representation from medical physics, PACS administration, information security, and information governance. The committee establishes per-modality compression rules, monitors compliance via PACS audit logs, and reviews the policy annually in light of evolving evidence and revised regulatory guidance.
A robust governance regime includes a written compression policy referencing primary sources, automated PACS rules that enforce the policy on inbound and outbound images, periodic audits of representative cases for diagnostic accuracy, and a documented exception-handling process for cases such as research, legal evidence, or non-standard external referrals.
WIA-Official/wia-standards-public/tree/main/medical-imaging — open standard initiative providing source code for simulator, spec, API, and ebook assets cited throughout this volume; serves as the canonical verification record for all primary-source citations made by the WIA standard committee in this chapter. Canonical ENUM tokens used in this volume include CT, MRI, PET, SPECT, MAMMOGRAPHY, ULTRASOUND, XRAY, DICOM_3_0, NIFTI, NRRD, JPEG_BASELINE, JPEG_2000_LOSSLESS, HEVC, DICOMWEB, QIDO_RS, STOW_RS, WADO_RS, PACS, VNA, RIS, HIS, FDA_510K, CE_MDR, MFDS_CLASS_2, HIPAA, GDPR, DICOM_CONFORMANCE, RUL, RML, RLL, LUL, LLL, MEDIASTINUM, HL7_FHIR_IMAGING, IHE_RAD, NEMA_MITA, DICOM_SR, RADLEX, U_NET, V_NET, RESNET, VIT, TRANSFORMER, CNN, MONAI.