X-ray Computed Tomography

X-ray computed tomography (CT) is an imaging technique that uses X-rays measured from multiple directions and computational reconstruction to reveal an object’s internal structure. Unlike conventional radiography, which superimposes structures along the beam direction, CT produces cross-sectional images that can be combined into three-dimensional volumes. It is widely used in radiology and also in materials characterization and industrial inspection. The resulting images primarily represent spatial differences in X-ray attenuation rather than ordinary optical appearance. (fda.gov)

Physical principles

As an X-ray beam passes through matter, some photons are absorbed or redirected by scattering. The transmitted intensity depends on the material’s composition, thickness, and the photons’ energy. CT measures these changes along many paths through an object, then estimates the distribution of attenuation within it. Materials that appear superimposed in a radiograph can therefore be separated spatially in reconstructed sections. (fda.gov)

For an ideal monochromatic beam, the Beer–Lambert law gives

where $I_0$ is incident intensity, $I$ is transmitted intensity, and $\mu$ is the local linear attenuation coefficient. Taking the negative logarithm of $I/I_0$ yields a line integral of attenuation. Recovering the spatial distribution from these measurements is an inverse problem. Real scanners require corrections because their beams contain a spectrum of energies and their measurements include scatter and detector imperfections. (ncbi.nlm.nih.gov)

Acquisition and reconstruction

A medical CT scanner typically contains an X-ray tube and opposing detector array mounted in a rotating gantry, together with a motorized patient table. Filters shape the beam, while collimation restricts its extent. During rotation, detectors record numerous transmission profiles, which are digitized and processed by a computer. (nibib.nih.gov)

In axial acquisition, the table remains stationary during each acquisition and moves between positions. In helical, or spiral, acquisition, continuous table movement accompanies gantry rotation, producing a helical sampling trajectory relative to the patient. Multiple detector rows permit simultaneous sampling across a wider longitudinal region and support volumetric imaging. (www-pub.iaea.org)

A classic reconstruction method is filtered back-projection. Each projection is filtered and then distributed backward along its measurement paths; combining the contributions estimates the attenuation image. Its mathematical basis involves the Fourier transform. Iterative methods instead repeatedly compare predicted projections with measured data and update the image. These approaches can incorporate measurement statistics, scanner geometry, and regularization to control noise and instability. Reconstruction choices affect both image appearance and the visibility of fine structures. (ncbi.nlm.nih.gov)

Image values and limitations

Medical CT commonly expresses reconstructed attenuation using Hounsfield units (HU). On this scale, water is assigned 0 HU and air approximately −1000 HU. Images are displayed using a selected window width and level: these settings map a chosen interval of CT values to screen brightness, allowing different tissues to be examined without changing the underlying measurements. (www-pub.iaea.org)

A reconstructed volume consists of voxels, or three-dimensional image elements. Their dimensions describe sampling, but do not alone establish spatial resolution. Small structures can be averaged with surrounding material through partial-volume effects. Image quality also depends on noise, acquisition geometry, and reconstruction settings. (www-pub.iaea.org)

Artifacts are features introduced by measurement or reconstruction rather than by the object itself. Motion can blur or distort anatomy. Beam hardening occurs when lower-energy photons are preferentially removed from a polychromatic beam, potentially producing shading or streaks. Dense materials, especially metals, can cause severe inconsistencies and streak artifacts. Such effects limit the accuracy of both visual interpretation and quantitative measurements. (www-pub.iaea.org)

Medical applications and radiation exposure

CT supports the diagnosis of injuries and diseases involving the head, lungs, abdomen, bones, and blood vessels. Applications include locating hemorrhage, characterizing complex fractures, evaluating tumors, and detecting pulmonary embolism. It also guides some interventional procedures and helps monitor treatment response. (nibib.nih.gov)

Contrast agents increase attenuation differences between structures. Intravenous iodine-containing agents can improve visualization of blood vessels and organ enhancement; orally administered agents may delineate the gastrointestinal tract. Contrast administration is distinct from image acquisition and is not required for every CT examination. Possible adverse reactions constitute a separate concern from radiation exposure. (nibib.nih.gov)

X-rays are ionizing radiation and can cause biological damage, including damage to DNA. CT generally exposes patients to more radiation than a conventional single radiograph because it acquires many projections. Dose varies substantially with the examination, patient size, equipment, and operating settings. Children are more radiation-sensitive than adults. Radiation protection therefore addresses both the justification of examinations and the optimization of exposure for the required imaging task. (fda.gov)

Development and nonmedical uses

Godfrey Hounsfield developed a practical CT apparatus at EMI, while Allan Cormack contributed theoretical work on reconstructing internal attenuation distributions. The first patient examination with the EMI brain scanner took place on October 1, 1971. Cormack and Hounsfield jointly received the 1979 Nobel Prize in Physiology or Medicine for developing computer-assisted tomography. (nobelprize.org)

Industrial CT examines internal dimensions and defects without cutting objects apart. X-ray microtomography characterizes pores, powder particles, polymer foams, and composite structures. Such measurements connect internal geometry with material properties and manufacturing quality. (nist.gov)

Further developments include energy-sensitive imaging and photon-counting CT, whose detectors register individual X-ray photons rather than only an aggregate energy signal. Research also applies deep learning to image reconstruction and noise reduction, alongside improvements in sources, detectors, and imaging geometry. (content.govdelivery.com)