Imaging techniques and what they show
Medical imaging reveals internal structures and disease, supports treatment decisions and can provide three-dimensional models for planning and communication.
Imaging in cancer care
Imaging helps identify primary tumours, find possible metastases in organs such as the liver, bones and brain, and assess the course of disease during treatment. These assessments help determine whether treatment has achieved its goal or needs adjustment when there is no response.
How the techniques differ
Imaging techniques include methods that use X-rays and methods that do not. Their characteristics include the following.
| Technique | How it produces information | Characteristics and uses |
|---|---|---|
| Ultrasound | A transducer with a piezoelectric crystal emits sound waves into the body and converts the resulting information into images. | Uses no X-rays; can be supplemented with contrast media; depends heavily on the physician's training in the technique. |
| Plain X-ray imaging | Small doses of X-rays penetrate the body to form images of internal structures. | Sensitivity varies with the application; mammography can detect breast cancer. |
| Computed tomography (CT) | Produces high-resolution cross-sectional images that can be reconstructed in different ways. | Rapid imaging can cover the body within seconds and help identify tumours, but exposes the patient to radiation. |
| Magnetic resonance imaging (MRI) | Uses a strong magnet and signals from hydrogen atoms in the body. | Provides good soft-tissue contrast for imaging tumours and following their course; also used in orthopaedics, neurology and paediatrics. |
| Nuclear medicine | Radioactive substances can be injected and accumulate in particular diseased regions. | Can support diagnosis and some treatments, such as thyroid treatment; can be combined with other imaging, as in PET/CT. |
Three-dimensional reconstruction
Three-dimensional visualisation can be built from two-dimensional CT or MRI images rather than requiring a new imaging modality. The Plato's Cave system, described in 2011, used standard-of-care DICOM images from existing approved CT and MRI technologies. It could construct a colour three-dimensional representation in about five seconds from images typically taken a quarter to half a millimetre apart.
The model could be viewed at different depths, levels of detail and angles, with peripheral structures removed to isolate an organ. Game controllers and other interfaces allowed clinicians to travel through and rotate the images. Volumetric views supported detailed pretreatment plans that could be reviewed with patients and the treatment team. In 2011, the system was used primarily for surgical planning and complemented diagnostic imaging and radiologists' interpretation; images could also be deployed to a physician's desktop or mobile device.
Such views can reveal incidental findings whose clinical relevance is not immediately clear. In one demonstration, zooming out revealed a previously undetected aneurysm in a leg, while other views showed oesophageal varices, colonic polyps and plaque within a blood vessel. Collaborative discussion among physicians can help assess these findings.
Dental imaging and treatment planning
Radiology is used daily in dentistry for diagnosis. Cone beam CT provides three-dimensional information for a digital diagnostic and treatment workflow, rather than only the detail in one two-dimensional slice. By 2024, many models were available worldwide, although handling the information remained time-consuming in clinical practice.
Segmentation separates anatomical structures within images, making their relationships visible for planning. Dental examples presented in 2024 included viewing the relationship between a horizontally positioned wisdom tooth's root and the mandibular nerve before extraction. Injury to that nerve can permanently affect sensation in the lower jaw and lip. Another example planned the transplantation of a premolar to replace a front tooth lost in a child's cycling accident. Planning also supported repositioning a tooth growing in the wrong direction while avoiding adjacent nerves and blood vessels. Jaw surgery can cause lasting nerve injury or life-threatening arterial bleeding, which the planning aims to avoid.
Orbital reconstruction illustrates the importance of anatomical relationships: differences of a few millimetres can leave lasting facial asymmetry after a blowout fracture. Segmented models can also support sinus surgery and implant placement. For endodontic planning, internal structures can be on the scale of 100–200 micrometres.
Multimodal registration combines information from different scanners, including radiographic and soft-tissue scans, in one model of a virtual patient. Teeth can be removed from the view to inspect the relationship between bone and soft tissues, and an implant and crown planned while maintaining a safety zone from the mandibular nerve. The model also supports discussion with the patient about the proposed rehabilitation and their wishes, and communication among the surgeon, the clinician placing the crown and the dental laboratory.