Medical Imaging Through History

Anatomical atlases, wax models, photographs, and microscope slides prepared medicine to treat mediated sight as evidence, but they are not all diagnostic imaging in the narrower clinical sense used here. Dissection showed opened dead bodies; microscopy usually showed removed and prepared specimens. X-rays altered the problem by making a shadow record of a living body's interior, while later systems calculated sections, mapped returning echoes, or followed radioactive tracers.

Before 1895, practitioners inferred internal disease from a patient's account, external signs, palpation, percussion, auscultation, exploratory procedures, surgery, and post-mortem anatomy. These methods were not simply superseded. Imaging entered clinical decisions unevenly and was most useful when its result could answer a defined question. Joel Howell's study of Pennsylvania Hospital records from 1900 to 1925 shows why later celebration should not be confused with immediate routine use. (Howell, 2016)

The apparent directness of an image can hide this history. Exposure, positioning, contrast material, reconstruction algorithms, display settings, and selection all affect what becomes visible. A radiograph, sonogram, or MR image is therefore neither a neutral window nor a mere illustration: it is evidence made for a purpose and interpreted with other clinical information.

Röntgen was investigating electrical discharges in low-pressure tubes, a field already shaped by other physicists, when he detected penetrating rays in November 1895. His ten-page preliminary communication, Eine neue Art von Strahlen ("A New Kind of Rays"), described fluorescence, differing penetration by materials, and photographic shadow pictures. Created to report physical experiments to the Würzburg Physical-Medical Society, it is a primary source for what Röntgen claimed at the end of 1895—not proof of every later story about clinical adoption. (Röntgen, 1895, digitized by Wellcome Collection)

News and reproductions circulated rapidly in early 1896. Surgeons and hospital physicians first found the new technique especially useful for fractures and radiopaque foreign bodies. The older word skiagraph—literally a shadow picture—captured the limits well: bones and metal appeared clearly, while overlapping soft tissues were difficult to distinguish. Photographers, electricians, physicists, surgeons, and assistants assembled tubes, coils, plates, and chemicals; the specialty of radiology did not arrive fully formed with the rays. (Howell, 2016)

Thomas Edison's laboratory developed a practical fluoroscope by May 1896: a fluorescent screen gave a moving shadow image without waiting for a photographic plate to be developed. This speed came at a cost. Operators worked close to poorly shielded tubes, exposures could be long, and the delayed relationship between exposure and burns, chronic injury, or cancer was initially difficult to establish. Edison's assistant Clarence Dally died after severe radiation injury, one of the most visible warnings that the new sight could harm those who produced it. (Smithsonian National Museum of American History)

1. November–December 1895, Würzburg: Röntgen investigates and reports the unknown rays and their shadow-producing effects.
2. Early 1896: demonstrations, translations, and apparatus spread through physics laboratories, photographic workshops, surgical services, and hospitals.
3. May 1896: Edison's laboratory has developed a medical fluoroscope, exchanging a permanent plate for real-time viewing.
4. By the early 1900s: injuries among operators are well reported, but standards for measurement, shielding, and occupational protection remain incomplete.

Radiography chiefly mapped attenuation. Nuclear medicine instead placed a radioactive tracer in the body and detected where it travelled or accumulated. Work with radioisotopes in physiology and thyroid medicine during the first half of the twentieth century prepared the field; Hal Anger's gamma camera, developed at the University of California's Radiation Laboratory in the 1950s, made the spatial distribution of emitted gamma rays easier to record. These images represented function and uptake rather than a simple anatomical shadow. (Williams, 2008)

Diagnostic ultrasound grew from acoustics, wartime sonar, industrial flaw detection, and several experimental medical programmes. In Glasgow during the 1950s, obstetrician Ian Donald, engineer Tom Brown, and physician John MacVicar adapted industrial equipment to distinguish abdominal masses. Their 1958 Lancet paper made the collaboration visible, but it was a milestone in a longer and international sequence, not the act of a lone inventor. (Campbell, 2013)

Ultrasound used returning sound echoes rather than ionizing radiation and could display movement, making it valuable in obstetrics and cardiology. Its obstetric spread also changed pregnancy as a social experience. Clinicians learned to read uncertain grey patterns while fetal images circulated beyond the clinic as keepsakes and political symbols. Historians therefore distinguish what a scan can support clinically from the emotional or moral certainty viewers may attach to it. (Erikson, 2007)

Physicist Allan Cormack published reconstruction work in 1963 and 1964, initially arising from a radiation-dosimetry problem. Engineer Godfrey Hounsfield, working independently at EMI in England, combined detectors, an X-ray source, and computer reconstruction in a clinical machine. Neuroradiologist James Ambrose, hospital staff, EMI engineers, tissue specimens, public funding, and access to mainframe computing were all part of the translation. The familiar phrase "Hounsfield invented the CT scanner" is therefore useful shorthand, not a complete history. (Schulz, Stein, and Pelc, 2021)

On 1 October 1971, Ambrose supervised the first patient brain scan at Atkinson Morley Hospital near London. Reconstruction occurred on an off-site computer, so the result was not available for two days. Early radiologists worried that CT's spatial detail was inferior to film and that the machine was expensive; its advantage was instead the removal of superimposition and the discrimination of small differences among soft tissues. Public demonstrations in 1972 accelerated orders and hospital adoption. (Schulz, Stein, and Pelc, 2021)

In 1979 the Nobel Prize recognized Cormack and Hounsfield jointly "for the development of computer assisted tomography." The award records an important judgment about contribution, not an exhaustive list of the clinicians, programmers, engineers, government funders, and industrial teams who made CT workable and widely available. (Nobel Prize, 1979)

MRI did not descend from X-rays. It developed from nuclear magnetic resonance (NMR), a physical phenomenon used after the 1940s to study the chemical structure of materials. In 1971 Raymond Damadian reported different relaxation times in normal and cancerous rat tissues. In 1973 Paul Lauterbur showed how magnetic-field gradients could locate signals and form two-dimensional images; Peter Mansfield and colleagues developed gradient methods, mathematical analysis, and faster acquisition. Human imaging followed through several teams in the later 1970s. (Sanghvi et al., 2024)

By the early 1980s, "NMR imaging" was increasingly called magnetic resonance imaging. Clinical systems demanded strong magnets, radiofrequency coils, computing, pulse sequences, methods to reduce motion, trained operators, and buildings able to house the apparatus. MRI offered multiple forms of soft-tissue contrast without ionizing radiation, but long scans, motion, cost, and contraindications limited what it could do. Its rise did not make CT obsolete; the modalities answered different questions.

Priority remains more distributed than retrospective accounts or prizes imply. Damadian, Lauterbur, Mansfield, Richard Ernst, John Mallard's Aberdeen group, industrial designers, and many others made distinct contributions. Naming one "inventor of MRI" collapses experimental tissue measurement, spatial encoding, image reconstruction, whole-body hardware, and clinical adoption into one achievement. (Hennig, 2023)

CT was digital at acquisition, yet for years CT and MRI images were commonly printed onto film and stored in physical jackets. As the number of images increased, hospitals built picture archiving and communication systems (PACS) to store, retrieve, display, and transmit them. This changed the image from a local object in a film library into a record that could move among departments and, later, across institutions.

Proprietary formats initially made machines from different manufacturers difficult to connect. The American College of Radiology and National Electrical Manufacturers Association formed a joint committee in 1983; it published ACR–NEMA Standard 300 in 1985. A substantial 1993 revision became Digital Imaging and Communications in Medicine (DICOM), designed for networked exchange and workflow as well as file structure. (DICOM Standards Committee, history of the standard)

Digitization improved comparison and circulation, but also created new dependencies: patient identifiers, metadata, compatible software, displays, storage, cybersecurity, and reliable networks became part of diagnostic practice. The visible image continued to rest on substantial invisible labour.