A tomogram is a cross-sectional image of an object, produced by a technique called tomography. Rather than flattening everything in its path into a single overlapping picture the way a standard X-ray does, tomography reconstructs individual slices through the subject, whether that subject is a human brain, a titanium engine part, or the Earth’s mantle. The word itself comes from the Greek tomos (slice) and graphein (to write). The principle is the same across wildly different technologies: send some form of energy through or around an object from many angles, record how it is absorbed or scattered, and use math to reassemble a map of the interior.
Why Slices Matter
If you have ever looked at a conventional chest X-ray, you have seen the fundamental problem tomography solves. A plain radiograph fires X-rays straight through you and captures what comes out the other side on a flat detector. The result is a shadow in which your ribs, lungs, heart, and spine are all stacked on top of each other. A tumor hiding behind a rib or a small bleed near the spine can vanish into that layered mess. A CT image, by contrast, shows dramatically improved contrast because it isolates one thin plane at a time, though it sacrifices some spatial resolution to get there.1Physics Education. X-ray computed tomography That trade-off is almost always worth it in practice, because being able to see a structure free of overlying tissue matters more than squeezing out the sharpest possible edges.
How a CT Scanner Builds Each Slice
In a medical X-ray CT scanner, you lie on a motorized table that slides through a doughnut-shaped housing called a gantry. Inside that gantry, an X-ray tube and an arc of detectors spin around you. The tube fires a thin, fan-shaped beam of X-rays through your body at one angle, the detectors on the opposite side measure how much of the beam made it through, and then the whole assembly rotates a fraction of a degree and fires again. One full rotation can happen in less than a second on modern machines, collecting hundreds of these angular snapshots.
Each snapshot is essentially a line-by-line record of how much the X-rays were weakened, or attenuated, as they passed through different tissues. Dense bone absorbs a lot, air-filled lung absorbs almost nothing, and soft organs fall somewhere in between. A computer takes all those angular measurements and runs them through a reconstruction algorithm to calculate what combination of tissue densities at each tiny point within the slice would explain the pattern of attenuation measured from every angle. The result is a grid of numbers, each one representing the density at a specific location. Those numbers are converted into shades of gray, and you get your tomogram: a clean cross-section as if someone had physically sliced through the body and photographed the cut surface.
Each pixel’s density value is expressed on the Hounsfield Unit scale, named after the engineer who built the first clinical CT scanner. Water is defined as 0 HU, air as roughly −1000 HU, and dense cortical bone can reach well above +1000 HU. In principle, every tissue has a characteristic HU range, though the relationship between HU and actual physical density is not perfectly one-to-one. In soft tissue, assuming standard compositions, accuracy within about one percent is achievable, but in bone the uncertainty can be larger because mineral content varies.2PLOS ONE. On the molecular relationship between Hounsfield Unit (HU), mass density, and electron density in computed tomography (CT) Radiologists exploit this density information constantly, adjusting the display window and level to highlight different tissues from the same raw data. Viewing the same scan with “bone window” settings versus “soft tissue” settings reveals entirely different pathology, and using multiple window settings during interpretation improves how well lesions are detected and characterized.3PubMed. Liver and bone window settings for soft-copy interpretation of chest and abdominal CT
The Inventors Behind the Scanner
The theoretical groundwork for CT came from Allan Cormack, a South African physicist. While working at Groote Schuur Hospital in the mid-1950s, Cormack noticed that the methods used for X-ray dosimetry were crude, and he realized that precise dose planning required knowing how X-ray attenuation was distributed through inhomogeneous tissue. He framed the problem mathematically: if you know how much a beam is weakened along every possible straight line through an object, can you work backward to figure out the density at every point inside? He found no existing solution, so he derived one himself, first for circularly symmetric objects and then for arbitrary shapes. By the early 1960s he had built phantoms from wood and aluminum, taken transmission measurements, and produced calculated attenuation profiles that matched independent measurements. It was arguably the first experimental demonstration of computer-assisted tomography.4PubMed Central. How CT happened: the early development of medical computed tomography
Independently, the English engineer Godfrey Hounsfield developed a practical clinical scanner while working for the Electrical and Musical Industry (EMI) company. In collaboration with radiologists James Ambrose and Louis Kreel, he introduced his machine at Atkinson Morley’s Hospital in Wimbledon in 1971.5PubMed Central. Godfrey Newbold Hounsfield (1919-2004): The man who revolutionized neuroimaging Cormack and Hounsfield shared the 1979 Nobel Prize in Physiology or Medicine, despite having worked completely independently and never having communicated before the prize was announced.
Tomography Beyond X-Rays
The slice-and-reconstruct logic at the heart of CT is not limited to X-rays. Any energy source that can penetrate a material and be measured on the other side can, in principle, be turned into a tomographic imaging system. Several major medical and scientific technologies exploit exactly this idea.
Magnetic resonance imaging builds tomograms without any ionizing radiation at all. Instead of X-rays, MRI uses strong magnetic fields and radiofrequency pulses to manipulate the behavior of hydrogen atoms in your body. By applying magnetic gradients in three steps, the scanner encodes the position of hydrogen atoms in each slice, and the resulting signals are reconstructed into images.6Progress in Cardiovascular Diseases. Basic principles of magnetic resonance imaging Because MRI is sensitive to water content, fat, and the chemical environment of tissue rather than density alone, it excels at distinguishing between different types of soft tissue in ways CT cannot easily match.
Optical coherence tomography, or OCT, uses light to create cross-sectional images of tissue at a microscopic scale. The principle underlying OCT is interferometry: a beam of broad-spectrum light is split into two paths, one directed at the tissue and one at a reference mirror. The reflected beams recombine at a detector, producing an interference pattern that encodes information about how deep the reflections came from and what the tissue looks like at each depth.7PubMed Central. Optical coherence tomography retinal imaging: narrative review of technological advancements and clinical applications OCT is now a routine tool in ophthalmology, where it produces exquisitely detailed cross-sections of the retina.
Ultrasound tomography applies the same angular-reconstruction strategy using sound waves instead of X-rays or light. In breast imaging, for example, quantitative transmission ultrasound reconstructs three-dimensional maps of sound speed and attenuation through the tissue.8Scientific Reports. Objective breast tissue image classification using Quantitative Transmission ultrasound tomography Because it uses no ionizing radiation, it is attractive for screening settings where repeated imaging is desirable. Early clinical systems have demonstrated the ability to reconstruct speed-of-sound distributions by acquiring ultrasound data from multiple angles around the breast in a single cross-sectional plane.9PubMed. Ultrasound computed tomography in breast imaging: first clinical results of a custom-made scanner
Tomography at the Molecular Scale
At the opposite end of the size spectrum from whole-body scanners, cryo-electron tomography (cryo-ET) images individual cells and protein complexes. Samples are flash-frozen to preserve their natural state, then imaged in an electron microscope from a series of tilt angles. The resulting projections are computationally reconstructed into three-dimensional volumes, just as a CT scanner reconstructs body slices from X-ray projections. Cryo-ET can visualize cellular structures under close-to-life conditions at molecular resolution.10PubMed Central. Cryo-electron tomography of cells: connecting structure and function Advances in imaging technology and image processing now enable sub-nanometer resolution, making it possible to study macromolecular organization and dynamic biological processes within intact cells.11Current Opinion in Structural Biology. In situ cryo-electron microscopy and tomography of cellular and organismal samples
Seeing Inside the Earth and Other Large Structures
Tomography scales up just as well as it scales down. Seismic tomography uses earthquakes as the energy source and the whole Earth as the object. When an earthquake occurs, seismic waves travel through the planet’s interior, and their speed depends on the temperature, composition, and density of the rock they pass through. By recording the arrival times of seismic waves at stations spread around the globe, geophysicists reconstruct three-dimensional maps of the Earth’s mantle and core. This approach has revealed fundamental processes like mantle convection and the geometry of subducting tectonic plates.12PubMed Central. Mantle dynamics and seismic tomography Recent applications have used ambient seismic noise, not just earthquake signals, to map upper-mantle shear-wave speed beneath Africa, revealing segmented low-velocity zones beneath the East African Rift System.13Geochemistry, Geophysics, Geosystems. Upper Mantle Earth Structure in Africa From Full‐Wave Ambient Noise Tomography
Muon tomography uses cosmic-ray particles to peer inside structures too thick or too precious for conventional X-rays. Muons are naturally produced when cosmic rays hit the upper atmosphere, and they rain down on the Earth’s surface constantly. Dense material absorbs more muons than empty space does, so placing muon detectors around a structure and counting how many muons arrive from each direction reveals internal voids. This technique made headlines when it was used to characterize a previously unknown corridor-shaped void in the Great Pyramid of Khufu. The corridor was estimated to be roughly 2 meters wide, about 2.2 meters tall, and 9 meters long, sitting just behind the stone chevrons on the pyramid’s north face.14Nature Communications. Precise characterization of a corridor-shaped structure in Khufu’s Pyramid by observation of cosmic-ray muons
Industrial CT scanning, sometimes called micro-CT, applies the same X-ray tomography principles used in hospitals to non-destructive testing of manufactured parts. A cost-effective micro-CT system has been shown to detect porosity and cracks in solid titanium parts up to 13 mm thick, with a voxel size of 118 micrometers and porosity estimation errors as low as 0.3 percent.15PubMed Central. Cost-effective micro-CT system for non-destructive testing of titanium 3D printed medical components For 3D-printed medical implants, where a single hidden void could lead to mechanical failure inside a patient, this kind of inspection is invaluable.
Artifacts and Common Limitations
Tomographic images are mathematical reconstructions, and the math can go wrong in predictable ways. The most common problem in X-ray CT is beam hardening. X-ray beams from clinical tubes are not a single energy; they contain a spectrum of photon energies. Low-energy photons are absorbed preferentially as the beam passes through dense material, leaving behind a “harder” beam of higher-energy photons. The reconstruction algorithm assumes the beam behaves consistently, so this energy shift produces errors in the final image. Dense objects like metal implants cause severe streaking and dark bands called metal artifacts, which can obscure the very anatomy the scan was ordered to evaluate.16PubMed Central. Computed tomographic beam-hardening artefacts: mathematical characterization and analysis
Motion artifacts are another persistent challenge. If you breathe, swallow, or shift during a scan, the anatomy moves between projections and the reconstruction blurs. Modern scanners use fast rotation speeds and cardiac or respiratory gating to minimize this, but it remains a real concern for imaging the heart or the abdomen of uncooperative patients, including young children.
Radiation Dose and the ALARA Principle
Because medical CT uses X-rays, every scan delivers a dose of ionizing radiation. The guiding philosophy in radiology is ALARA: dose as low as reasonably achievable. The key word is “reasonably.” The goal is not simply to minimize radiation; it is to produce images good enough for diagnosis while using the least dose necessary. In practice, dose optimization is more a problem of image quality than of radiation measurement: you need to know what image quality is “good enough” before you can figure out how low the dose can go.17PubMed. Optimizing CT radiation dose based on patient size and image quality: the size-specific dose estimate method
Patient size plays a direct role. Body mass index, waist circumference, and abdominal fat are all strongly correlated with the radiation dose needed to maintain image quality, which is why modern protocols try to tailor the X-ray output to each patient’s build rather than using a fixed one-size-fits-all setting.18PubMed Central. Optimizing CT Abdomen-Pelvis Scan Radiation Dose: Examining the Role of Body Metrics For cranial CT specifically, low-dose protocols have demonstrated meaningful reductions in absorbed dose to sensitive structures: thyroid doses dropped by more than 64 percent and eye-lens doses by over 50 percent compared to standard protocols, without compromising diagnostic quality.19Journal of Radiation Research and Applied Sciences. On cranial CT imaging: The role of low-dose protocols in enhancing patient safety and reducing radiation exposure
Deep Learning and Noise Reduction
Lowering the radiation dose is straightforward in principle: just turn down the X-ray tube current. The catch is that fewer photons means noisier images, and noisy images can hide pathology. This is where deep learning has made a tangible impact. Neural networks trained on pairs of low-dose and full-dose images learn to suppress noise while preserving diagnostically relevant detail.
In coronary CT angiography, applying a deep-learning denoising algorithm to images acquired with iterative reconstruction significantly reduced image noise and improved both signal-to-noise and contrast-to-noise ratios, with radiologists rating the processed images as subjectively better.20PubMed Central. Incremental Image Noise Reduction in Coronary CT Angiography Using a Deep Learning-Based Technique with Iterative Reconstruction A recent prospective study of liver CT took this further: a real-time deep-learning noise reduction algorithm allowed a roughly 73 percent reduction in radiation dose while producing images with noise levels and contrast-to-noise ratios comparable to the standard full-dose protocol, and diagnostic acceptability was rated at 100 percent for the denoised images.21PubMed Central. Ultra-low-dose hepatic computed tomography with a novel real-time deep learning-based noise reduction algorithm These are not theoretical improvements. They represent a practical path to significantly lower patient radiation exposure without asking radiologists to accept worse images.
Photon-Counting Detectors
The detectors in most CT scanners in use today are energy-integrating: they absorb all incoming X-ray photons and produce a signal proportional to the total energy deposited. They cannot tell you anything about the energy of individual photons. Photon-counting detectors, which have recently entered clinical use, count each photon individually and sort it by energy. This seemingly small engineering change has cascading benefits.
Because individual detector pixels can be made much smaller than those in conventional detectors, photon-counting CT achieves higher spatial resolution on the order of 0.25 mm, enabling visualization of small vessels, arterial walls, distal airways, and bone microstructure that conventional CT could not reliably resolve.22PubMed Central. Clinical Applications of Photon-counting CT: A Review of Pioneer Studies and a Glimpse into the Future The energy-sorting capability also means the scanner can distinguish materials that look identical on conventional CT but absorb X-rays differently at different energies, opening the door to better tissue characterization and reduced artifacts.23PubMed Central. Technical Basics and Clinical Benefits of Photon-Counting CT On top of that, the technology enables improved contrast-to-noise ratio and potential dose reduction along with elimination of electronic noise, which has plagued conventional detectors.24European Journal of Radiology. Initial clinical results of photon-counting computed tomography (PCCT): A review of the literature
Photon-counting CT is still expensive and limited to a handful of scanner models, but early clinical studies have been encouraging enough that it is widely expected to become the standard detector technology over the coming decade. For patients, the practical promise is straightforward: sharper images, better material identification, and lower radiation doses, all at once.
Solar Tomography and Imaging the Corona
One of the more unexpected applications of tomographic thinking is in solar physics. The Sun’s corona, the hot outer atmosphere visible during eclipses, is three-dimensional but we observe it from essentially one vantage point. Solar rotational tomography exploits the Sun’s own rotation: as it turns, different sides face the Earth, effectively providing new viewing angles over about two weeks. By collecting extreme ultraviolet images of the corona over time and treating each rotational position as a new projection angle, physicists can reconstruct the three-dimensional distribution of coronal density and temperature. The technique, known as differential emission measure tomography, produces global 3D maps of the corona’s electron density and temperature, helping researchers understand how energy is distributed and transported through the Sun’s atmosphere.