What Is X-Ray Radiodensity Absorption? (Hounsfield Scale)
X-ray radiodensity describes how much an object reduces an X-ray beam; CT reports that effect as Hounsfield units (HU), scaled against water. Water is conventionally 0 HU and air −1,000 HU. HU help compare image values, but they are not fixed material-density labels: scan settings, reconstruction, contrast, and artifacts can change readings.
When time is short, a single number can seem like the fastest way to identify a material or compare scans. But an HU value only makes sense when you know how the CT image was acquired and how its stored pixel values were transformed. I use the sequence below to check those details before interpreting a measurement.
Planning: Define the Scan and Measurement Question
Start by stating what you want the HU value to tell you. The same number may have different meaning across scan settings, image types, or clinical tasks. Record the scanner, protocol, contrast status, and reconstruction method before comparing measurements or applying a threshold.
Radiodensity is a common informal term for how strongly a material appears to reduce X-rays. In CT, the relevant physical quantity is linear attenuation, written as μ. It describes the reduction in X-ray intensity through a material per unit distance. Attenuation depends on both the material and the X-ray spectrum.
The Hounsfield scale converts attenuation into a value relative to water:
HU = 1,000 × (μmaterial − μwater) / μwater
Here, μmaterial is the material’s linear attenuation coefficient, and μwater is water’s coefficient under the same scan spectrum. If a material attenuates the beam like water, the result is 0 HU. If it attenuates more, its HU is positive; if less, its HU is negative.
The scale provides a useful comparison, not a universal identity chart. A measured value can vary with tube voltage (kVp), beam filtration, reconstruction, contrast material, and image artifacts. Before using HU to answer a question, define the task: are you comparing two areas in one scan, checking a change between scans, or trying to characterize a material?
Access and Clearances: Confirm Data, Metadata, and ROI Placement
Use the original CT data when possible, and confirm how the image values were stored and transformed. A region of interest (ROI) must sample the intended area without including edges, neighboring materials, or artifacts. Record the image source and measurement method so another reader can review them.
A DICOM image may store pixel values that need conversion before they represent CT numbers. Check the image’s Rescale Slope (0028,1053) and Rescale Intercept (0028,1052) attributes. The conversion is:
Output value = stored pixel value × Rescale Slope + Rescale Intercept
For CT images, this output is commonly expressed in HU. DICOM also allows a Modality LUT, which maps stored values through a lookup table instead of, or as defined by, a linear rescale. Check the applicable transformation and any stated units rather than assuming that raw stored pixel values are HU. Some enhanced objects can also carry transformations at the frame level.
The ROI is the selected image area used to calculate a value such as a mean. Place it well inside the target when practical. Avoid partial-volume effects, where one voxel contains more than one material, and avoid boundaries, streaks, or bright areas that may distort the result.
Before measuring, note:
- Whether the image is original CT data or a derived image.
- The acquisition protocol, kVp, contrast status, and reconstruction kernel.
- The ROI size, location, and whether you report a mean, median, or range.
Execution: Convert Attenuation to Hounsfield Units
HU are scaled attenuation values, not direct measurements of physical density. Apply the DICOM transformation first, then interpret the result in its scan context. The reference points are water at 0 HU and air at −1,000 HU by convention, but real measurements can differ slightly.
Suppose a material has an attenuation coefficient 1.2 times that of water under a particular spectrum. Its value is 1,000 × (1.2 − 1), or 200 HU. If its coefficient is 0.8 times water’s, the result is −200 HU. These are mathematical examples, not reference values for a specific tissue or substance.
| Comparison point | HU meaning | Practical interpretation |
|---|---|---|
| Air reference | −1,000 HU by convention | A reference point; an image measurement may differ due to calibration or artifacts. |
| Water reference | 0 HU by convention | The comparison baseline for the scale. |
| Attenuation above water | Positive HU | Indicates greater attenuation than water under that scan spectrum. |
| Attenuation below water | Negative HU | Indicates less attenuation than water under that scan spectrum. |
| Bone or metal | Often high values | Exact values vary; dense objects can also cause artifacts nearby. |
Worked scenario 1: comparing two areas in one scan. If two ROIs are measured in the same image, with the same reconstruction and no major artifact, their HU difference can be informative. I would still check the ROI placement and whether contrast or mixed materials affect either value.
Worked scenario 2: comparing separate scans. A change in HU between scans does not automatically mean the material changed. Different kVp, reconstruction kernels, contrast phases, or scanners can shift values. Treat cross-scan comparisons as protocol-dependent, not as direct measurements on one fixed scale.
Worked scenario 3: near a dense object. A bright streak beside metal may produce misleading local values. The measured HU there may reflect the artifact as well as the nearby material. Do not classify tissue from one threshold reading in that region.
Verification: Check Calibration, Artifacts, and Interpretation
A reliable HU reading depends on image quality, calibration, metadata, and careful interpretation. Check scanner quality assurance under the applicable procedure, review likely artifacts, and report acquisition conditions when quantitative comparisons matter. An isolated value should not be treated as proof of tissue identity.
A CT scanner’s water and air references are conventions, not guarantees that every voxel will measure exactly 0 or −1,000 HU. Scanner performance should be checked through the manufacturer’s or facility’s applicable quality-assurance procedure. Do not substitute an informal measurement from a patient image for that procedure.
Beam hardening occurs because lower-energy X-rays are absorbed more readily as the beam passes through an object, changing the spectrum that reaches the detector. Metal-related artifacts can create streaks, dark bands, or distorted values. Both can alter local HU, especially close to dense objects. A value that appears unusual may reflect the scan rather than the material itself.
Use this review sequence:
- Confirm that the image is CT and identify the acquisition protocol and contrast status.
- Apply or verify the DICOM Modality LUT or Rescale Slope and Intercept; confirm the output units.
- Check the ROI location, size, and image quality. Avoid boundaries and visible artifacts when possible.
- Compare values only with suitable reference data for the same scanner, protocol, and task.
- Report the measurement with relevant conditions if another person needs to reproduce or compare it.
There is no universal HU range that identifies “bone,” “fat,” or another tissue class in every scan. Thresholds must be validated for the scanner, protocol, contrast status, reconstruction, and clinical question. For medical interpretation, a qualified clinician should assess the image in context.
Conclusion and FAQ
The Hounsfield scale makes CT attenuation easier to compare by expressing it relative to water. To use HU responsibly, verify the DICOM value transformation, understand the scan conditions, and check the ROI for artifacts. Treat the result as protocol-dependent evidence, not a standalone label for material identity.
Frequently asked questions
What does HU stand for in CT?
HU means Hounsfield unit. It is a CT image value that expresses X-ray attenuation relative to water.
What is the Hounsfield formula?
HU = 1,000 × (μmaterial − μwater) / μwater, where μ is the linear attenuation coefficient under the scan’s X-ray spectrum.
Why is water 0 HU?
Water is the reference material in the scale, so its attenuation is defined as 0 HU by convention.
Why is air −1,000 HU?
Air is conventionally assigned −1,000 HU relative to water. An actual image measurement may vary due to scan and calibration factors.
Are HU the same as physical density?
No. HU reflect X-ray attenuation, which depends on material composition and scan conditions. They are not a spectrum-independent measurement of physical density.
How do I convert stored DICOM pixels to HU?
Use the applicable Modality LUT or calculate stored pixel value × Rescale Slope + Rescale Intercept. Confirm the relevant attributes and units first.
Can I compare HU values from different scanners?
Only with care. Differences in scanner, kVp, reconstruction, calibration, and contrast can affect values, so comparisons need suitable protocol context.
Does one HU value identify a tissue?
No. Tissue classes do not have universal HU ranges. Interpretation depends on the scan protocol, artifacts, and the clinical task.
Why can metal distort nearby HU readings?
Metal can cause streaks and other CT artifacts that change local pixel values. A reading near metal may not represent the actual tissue attenuation.
What is an ROI in CT?
An ROI, or region of interest, is a selected image area used to measure values such as mean HU. Placement affects the result.
(This article was written by one of our staff writers, Garrett Vance. Visit our Meet the Team page.)
