How CT Scans Work: Building a 3D Image From X-Rays

A conventional X-ray has a well-known limitation: it flattens a three-dimensional body into a single two-dimensional image, so structures at different depths overlap on top of each other, the same way a photograph of a stack of transparent sheets would blur all of them into one confusing picture. Computed tomography, or CT, solves this by taking many X-ray readings from different angles and using mathematics, rather than a clever camera trick, to reconstruct what's actually inside, one thin slice at a time.

The hardware: a rotating source and detector

Inside a CT scanner, an X-ray source and an opposing ring of detectors are mounted on a rotating gantry that spins continuously around the patient, who moves slowly through the center on a motorized table. As the gantry rotates, the X-ray source fires continuously or in rapid pulses, and the detector array on the opposite side measures how much radiation passed through the body along each angle, recording how much was absorbed or scattered by the tissue in its path. A single rotation captures hundreds of individual angled measurements of the same cross-section.

Why one flat X-ray can't do this, but many angled ones can

Each individual X-ray reading from a CT scanner is, by itself, just as flat and overlapping as a conventional X-ray. The difference is that a single reading at one angle isn't the final image; it's raw data. Reading the same cross-section from dozens or hundreds of different angles around a full rotation gives the reconstruction algorithm enough independent information to mathematically solve for what density of tissue must exist at every point within that slice to produce all of those different angled readings simultaneously. This mathematical inversion problem, known as tomographic reconstruction, is what actually produces the recognizable cross-sectional CT image, and it depends entirely on having many angles of data rather than any single exceptionally clear X-ray.

From raw data to a viewable slice

The reconstruction algorithm, running on dedicated computing hardware built into the scanner, processes the full set of angled readings for a given slice and calculates the X-ray absorption value at each small volume element, called a voxel, within that cross-section. Denser tissue, like bone, absorbs more X-rays and appears bright; less dense tissue, like air-filled lung, absorbs less and appears dark; soft tissue falls somewhere in between, at various shades of gray depending on its specific density. Radiologists commonly use a standardized density scale, measured in Hounsfield units, to compare tissue density consistently across different scanners and different patients.

Modern scanners repeat this whole process for many adjacent slices as the patient moves through the gantry, then stack the resulting slices into a full three-dimensional dataset that software can display as individual cross-sections or re-slice along any plane, and even render as a rotatable three-dimensional model of bone or vasculature.

Contrast dye: making specific structures stand out

Some structures, particularly blood vessels and certain soft-tissue abnormalities, have similar natural X-ray density to the tissue surrounding them, making them hard to distinguish on a plain CT scan. For these cases, a contrast agent, usually an iodine-based compound, is injected intravenously or sometimes swallowed before the scan. The contrast agent is deliberately chosen because it absorbs X-rays strongly, so blood vessels or tissues that have taken up the contrast appear noticeably brighter than surrounding tissue, letting radiologists trace blood flow or highlight the boundary of a tumor that would otherwise blend into its surroundings on a plain scan.

Why radiation dose is a genuine tradeoff

Because a CT scan requires many individual X-ray exposures from different angles rather than one, it delivers a meaningfully higher total radiation dose to the patient than a single conventional X-ray of the same body region, sometimes on the order of dozens to a few hundred times more, depending on the specific type of scan and body part. This isn't an oversight in scanner design; it's an inherent tradeoff for the additional diagnostic detail cross-sectional imaging provides. For this reason, ordering physicians and radiology guidelines from bodies like the U.S. Food and Drug Administration emphasize using CT only when its added diagnostic value justifies the radiation exposure, using the lowest dose that still produces a diagnostically useful image, and considering alternative imaging methods that use no ionizing radiation at all, such as ultrasound or MRI, when they can answer the same clinical question.

How CT differs from an MRI

CT and MRI are often confused because both produce cross-sectional images of the body's interior, but they use entirely different physics. CT relies on X-rays, ionizing radiation absorbed differently by different tissue densities, and is generally faster and better at imaging bone and detecting acute bleeding. MRI uses strong magnetic fields and radio waves to manipulate and detect the behavior of hydrogen atoms in the body's water and fat, involves no ionizing radiation at all, and generally provides better contrast between different types of soft tissue, at the cost of longer scan times and higher equipment cost. Neither technology has made the other obsolete; each is favored for different clinical situations depending on which physical property, tissue density or hydrogen distribution, is more diagnostically useful for a given question.