A DICOM image is a medical image file that follows a universal standard called Digital Imaging and Communications in Medicine. Unlike a regular photo saved as a JPEG or PNG, a DICOM file bundles the image itself with detailed clinical information: who the patient is, what scanner produced the image, how the scan was configured, and exactly how the image should be displayed. This standard is what allows hospitals to share scans between different machines, software systems, and facilities without losing critical data.
Why the DICOM Standard Exists
Before DICOM, every manufacturer of medical imaging equipment used its own proprietary format. A CT scanner from one company couldn’t easily send images to a workstation made by another. The National Electrical Manufacturers Association (NEMA) created the DICOM standard to solve this interoperability problem. It defines a common language so that scanners, servers, workstations, printers, and network hardware from different manufacturers can all work together within a hospital’s imaging system.
That central system is called a PACS, or Picture Archiving and Communication System. When you get a CT scan, the scanner saves the images in DICOM format, then transmits them over the hospital network to the PACS, where they’re stored and made available to radiologists and other clinicians. Because everything speaks DICOM, a radiologist can pull up your scan on any compatible workstation in the hospital, or even at a different facility.
What’s Inside a DICOM File
A DICOM file has two main parts: a header containing metadata and the pixel data that makes up the actual image. Every file begins with a 128-byte preamble followed by the four-character prefix “DICM,” which identifies it as a DICOM file. After that comes a structured series of data elements, each identified by a standardized tag.
The metadata is what makes DICOM fundamentally different from an ordinary image format. A single file can contain dozens or even hundreds of tagged fields, organized into categories:
- Patient information: name, ID number, birth date, and sex
- Study details: the date and time of the exam, the referring physician, and a description of the procedure
- Equipment settings: the type of imaging modality (CT, MRI, ultrasound, etc.), voltage settings, exposure time, and other technical parameters the scanner used
- Image geometry: slice thickness, pixel spacing, and the orientation of the image in three-dimensional space
This embedded information is essential for diagnosis. A radiologist needs to know the exact slice thickness of a CT scan to measure a tumor accurately. A surgeon reviewing an MRI needs to understand the spatial orientation of each image slice. Without this context, the pixel data alone would be far less useful.
Which Types of Medical Imaging Use DICOM
Nearly every type of medical imaging produces DICOM files. The list goes well beyond the scans most people think of:
- Cross-sectional imaging: CT scans, MRI, and PET scans (including combined PET/CT and PET/MR)
- Projection imaging: standard X-rays, mammography, and bone density scans
- Ultrasound: general diagnostic ultrasound and specialized ophthalmic ultrasound
- Nuclear medicine: both planar nuclear medicine and SPECT imaging
- Specialty imaging: dental radiography (both intra-oral and extra-oral), ophthalmic photography, optical coherence tomography, endoscopy, and even digital pathology slides
A single CT scan of your chest might generate hundreds of individual DICOM files, one for each image slice. An MRI of your knee could produce multiple series with different contrast settings, each series containing its own set of DICOM files. This is why medical imaging generates enormous amounts of data compared to regular photography.
How DICOM Handles Image Quality
Medical images need to preserve every detail a clinician might use for diagnosis, so DICOM supports lossless compression methods that reduce file size without discarding any image data. One built-in option is Run Length Encoding (RLE), a byte-oriented lossless compression scheme. The standard also supports JPEG 2000 and other compression formats through its framework for encapsulated pixel data.
Lossy compression, which sacrifices some image detail for smaller files, is available within DICOM but used cautiously. In clinical settings, lossless storage is generally preferred because even subtle details in pixel data can matter for diagnosis. Lossy compression is more common for teaching files, teleradiology in low-bandwidth situations, or cases where regulatory guidelines permit it.
How to View DICOM Files
You can’t open a DICOM file with a standard photo viewer. The format requires specialized software that can interpret the metadata, handle the unique pixel encoding, and display images with the correct brightness and contrast settings stored in the file.
In hospitals, radiologists use professional workstations with high-resolution medical-grade monitors. These systems can display multiple image series simultaneously, allow 3D reconstruction, and provide measurement tools. But if you’ve been given a CD of your own scans or downloaded images from a patient portal, you’ll need a DICOM viewer on your personal computer.
Several free and commercial viewers are available. RadiAnt, Horos (for Mac), and 3D Slicer are popular options. System requirements are modest for basic viewing: a modern processor, a few gigabytes of RAM, and at least a 1920 x 1080 screen resolution for comfortable reading. If you want to do 3D volume rendering, you’ll benefit from a dedicated graphics card and 8 GB or more of RAM. Most viewers let you scroll through image slices, adjust window and level settings (brightness and contrast), zoom in, and take basic measurements.
Patient Privacy and De-Identification
Because DICOM files contain detailed patient information embedded directly in the header, they carry significant privacy implications. Every file potentially includes your name, birth date, medical record number, and other identifiers. This is useful within the clinical system, where the image needs to stay linked to the right patient, but it becomes a concern when images are shared for research, education, or second opinions outside the treating institution.
De-identification is the process of stripping these personal identifiers from DICOM files. Under HIPAA, there are two accepted approaches. The Safe Harbor method requires removing 18 specific categories of identifying information: names, dates (except year), geographic details smaller than a state, phone numbers, email addresses, Social Security numbers, medical record numbers, and several others. The Expert Determination method involves a qualified statistician certifying that the remaining data carries very small risk of re-identifying anyone.
Specialized software tools can batch-process DICOM files to remove or replace identifying tags while preserving the image data and clinically relevant metadata. This is standard practice in medical research, where imaging datasets may include thousands of scans that need to be shared across institutions without exposing patient identities.
DICOM vs. Regular Image Formats
If you’ve ever wondered why your doctor can’t just email you a JPEG of your MRI, the answer comes down to what DICOM preserves that other formats don’t. A JPEG or PNG stores pixel values and basic display information. A DICOM file stores all of that plus the clinical context needed to interpret the image correctly.
DICOM images also typically use higher bit depths than consumer image formats. While a standard JPEG stores 8 bits per pixel (256 shades of gray), medical images commonly use 12 or 16 bits per pixel, capturing 4,096 or 65,536 shades. This wider range matters because subtle density differences in tissue can be diagnostically significant. Radiologists adjust the “window” and “level” to view different tissue ranges within that full spectrum, something a standard JPEG simply cannot support.
The DICOM standard is maintained by the Medical Imaging and Technology Alliance, a division of NEMA, and is updated several times per year to accommodate new imaging technologies, workflow requirements, and data types. It remains the universal backbone of medical imaging worldwide.