4.1 Quantitative Computed Tomography (QCT & pQCT)

Read the units before interpreting a CT result

Quantitative computed tomography (QCT) uses a calibrated CT acquisition to estimate mineral density in a three-dimensional region of interest (ROI). A volumetric measurement is expressed in milligrams of mineral equivalent per cubic centimeter (mg/cm³). Dual-energy X-ray absorptiometry (DXA) measures areal BMD in grams per square centimeter (g/cm²). A raw CT number is a separate quantity: it is expressed in Hounsfield units (HU).

For linear attenuation coefficient $\mu$ and the reference coefficient for water $\mu_w$, the CT scale is:

Water is assigned 0 HU and air is approximately −1,000 HU. A bone ROI reading of 100 HU cannot be read as 100 mg/cm³. The measured attenuation depends on acquisition settings such as tube voltage and on what tissue is in the ROI. A validated calibration method relates attenuation to a mineral-equivalent concentration.

Compare area and volume with a calculation

Imagine two uniform blocks with mineral-equivalent density 1 g/cm³. The first is a 1 cm cube, with mass 1 g and projected area 1 cm². The second is a 2 cm cube, with mass 8 g and projected area 4 cm². Their areal densities are 1 and 2 g/cm². Their volumetric densities are both 1 g/cm³. The example isolates the effect of beam-axis thickness on a projection.

Volume normalization removes that particular geometric dependency. Clinical estimates still depend on segmentation and calibration. A voxel near a thin cortex can contain both bone and soft tissue or marrow. Its mean attenuation may lower the apparent cortical density. Growth, anatomical differences and reference data matter when interpreting the result.

Choose the acquisition and region

The International Society for Clinical Densitometry (ISCD) specifies L1–L3 for single-slice spine QCT and L1–L2 for three-dimensional spine QCT. For hip QCT, the scan extends from the femoral head to the proximal shaft. Follow the validated system protocol and record the levels actually analyzed. For trabecular spine QCT, place the measurement in the vertebral body interior, avoiding the peripheral cortex, posterior elements and basivertebral region as appropriate for the software. The smaller internal region avoids much of the projected overlap from osteophytes, posterior elements and vascular calcification encountered on DXA.

Inspect the image before using a number. A fracture, prior surgery, local lesion or metal artifact can compromise the planned ROI. Marrow fat can lower the density estimate from single-energy CT, and contrast and reconstruction settings can alter attenuation. Check whether the analysis is validated for the actual scan. A dual-energy implementation can account for some composition effects when calibrated for that purpose.

Choose a peripheral CT site for the compartment

Peripheral QCT (pQCT) uses a dedicated extremity scanner, commonly at the radius or tibia. A scout view and a measured limb length allow a repeatable offset from a distal landmark. For example, a published radial protocol uses a distal 4% site for trabecular and total volumetric density and a 66% shaft site for cortical density, thickness and cross-sectional area. The labels refer to locations in that protocol. They are not a universal site prescription.

Some software reports a polar strength–strain index, a surrogate incorporating the distribution of mineral and geometry across the cross section. Two bones with similar average density can have different indices because of where the material lies. The index cannot provide an observed breaking force in newtons. Check the scanner, site, motion and segmentation before comparing repeated scans.

Dedicated extremity pQCT usually involves lower effective dose than dedicated axial QCT. Dose depends on the system and protocol. When central DXA and QCT both answer the same clinical question, ISCD prefers DXA to limit exposure.

ISCD Adult Official Positions · ACR QCT Practice Parameter · IAEA CT quality assurance · Primary radial pQCT study