An X-ray image is a shadow cast by the body onto a detector, because denser tissue absorbs more of the beam and leaves a darker mark. A CT slice is not a photograph but a reconstruction computed from hundreds of X-ray projections taken around the patient. MRI uses no ionising radiation at all: it reads the behaviour of hydrogen nuclei in a strong magnetic field, using radio waves instead of X-rays.
What is an X-ray image and why is it like a shadow?
An X-ray tube fires a beam through the patient toward a detector on the far side. Every tissue in the path removes, or attenuates, some of that beam. Bone, being dense and rich in calcium, removes a great deal; air in the lung removes almost none. What reaches the detector is therefore a map of how much beam survived, and the image is a projection of that map, summed along the whole path of the beam.
That is why the shadow analogy holds. Hold your hand in front of a lamp and the wall shows a flat silhouette in which a finger and a wrist bone overlap without any clue as to which is nearer. An X-ray image, shadow of the chest works the same way: the ribs, the heart and the spine are superimposed on one another, and the radiologist reads depth from experience, from the shape of the edges, and from a second view taken at an angle. Nothing in a single plain film tells you the depth of a structure directly.
The shadow comparison also explains the familiar grey scale. Bright white means the beam was almost entirely blocked, as with metal or dense bone. Black means it passed through, as with air. The soft greys in between are soft tissue, fat and fluid, each absorbing a different fraction of the beam. A radiograph is therefore a transmission picture, not a reflection picture, and the contrast in it comes from differences in absorption rather than from differences in colour or surface.
How does a CT scanner create a computed slice?
A CT scanner takes the shadow idea and rotates it. The X-ray tube and the detector array sit opposite each other on a ring, and they spin around the patient while the table moves slowly through the opening. In one rotation the system collects hundreds or thousands of individual transmission measurements, each one a shadow from a different direction.
Those measurements are the raw data, and they are not yet an image. A reconstruction algorithm, historically filtered back projection and now often an iterative method, solves for the distribution of attenuation values inside the scanned volume that would produce exactly this set of shadows. The result is a grid of small volume elements, each assigned a number on the Hounsfield scale, where water sits at zero, air near minus one thousand, and dense bone well above four hundred.
Only then does the computer cut the volume into slices and display them. The slice is computed, in the literal sense: it is a mathematical solution, not a view through a window. This is why a CT can show a small nodule deep in the lung that no plain film would separate from the ribs behind it, and why windowing settings let the reader stretch the grey scale to emphasise bone, lung or soft tissue from the same dataset. The trade-off is dose: because many projections are needed, a CT examination delivers more ionising radiation than a single radiograph, which is why justification and optimisation matter so much in practice.
Why does an MRI scan not use ionising radiation?
MRI produces contrast from a completely different physical process. The patient enters a strong static magnetic field, typically one and a half or three tesla in clinical machines, and the hydrogen nuclei in water and fat, which behave like tiny magnetic moments, align with that field. A pulse of radio-frequency energy then tips them out of alignment, and as they relax back they emit a faint radio signal that coils around the patient detect.
The key point is the energy of that radio-frequency pulse. It is far too low to knock electrons out of atoms, which is the definition of ionising radiation. X-rays and gamma rays carry enough energy per photon to break chemical bonds and damage DNA directly or indirectly; radio waves in the megahertz range used in MRI do not. That is the whole reason MRI is described as non-ionising, and it is why it is preferred for repeated follow-up, for younger patients, and for imaging during pregnancy when the clinical question cannot wait. A genotype has the same property as an image: it answers one question well and others not at all, as the SOD1 test sets out.
Non-ionising does not mean risk-free. The strong magnetic field can pull ferromagnetic objects, so implants and foreign bodies must be screened before anyone enters the room. The radio-frequency pulses deposit some energy as heat, measured as the specific absorption rate, and the scanner software limits it. Acoustic noise from rapidly switched gradient coils is loud enough to require hearing protection. These are real safety considerations, but they are mechanical, thermal and acoustic rather than radiobiological.
How the three methods differ in what they show
Plain radiography gives a fast, cheap, low-dose overview of bone and air-filled structures, and it remains the first examination for many chest and limb problems. Its weakness is superimposition: everything along the beam path lands on the same pixels.
CT removes that weakness by computing the third dimension, at the cost of a higher radiation dose and a longer, more complex examination. It is the workhorse for trauma, for staging, and for any question that needs fine anatomical detail across a large volume in a short time.
MRI offers the best soft-tissue contrast of the three and adds no ionising dose, but it is slower, noisier, more expensive, and unsuitable for some patients with certain implants or severe claustrophobia. It is the method of choice for the brain, the spinal cord, joints, and many pelvic and musculoskeletal questions.
Ultrasound sits alongside these as a fourth approach, using reflected sound waves rather than transmitted X-rays or emitted radio signals, and it is likewise free of ionising radiation. Each method answers a different kind of question, and the referral decision usually turns on what the clinician needs to see rather than on which machine is available.
What this means if you are the patient
If you have been referred for imaging, the practical details vary by method. A radiograph takes minutes and usually needs no preparation. A CT may involve a contrast agent given by mouth or into a vein, a short period of breath-holding, and a few minutes on the table. An MRI appointment can run from twenty minutes to an hour, requires you to remove metal, and asks you to lie very still while the scanner makes a loud knocking sound.
In the United Kingdom, the framework around all of this is set by the Ionising Radiation (Medical Exposure) Regulations 2017, usually abbreviated IR(ME)R 2017, which place duties on the referrer, the practitioner and the operator. The principles behind them are justification, meaning the examination must do more good than harm for this particular patient, and optimisation, meaning the dose must be as low as reasonably practicable while still answering the question. Radiologists, radiographers and sonographers each hold defined responsibilities under that framework.
Knowing how the picture is made helps you ask better questions at the appointment: whether a contrast agent is needed, whether a non-ionising alternative would answer the same question, and when and how the report will reach the clinician who referred you. The image itself is only the middle of a process that begins with a written request and ends with a written report.