Modern radiation therapy is defined by precision, but that precision was built incrementally. Many of its advances originated outside oncology—in nuclear physics, radar, computing and medical imaging—before being adapted to the problem of treating cancer while protecting normal tissue.
Cobalt-60 machines, introduced in the 1950s, emitted gamma rays of approximately 1.25 million electron volts. Compared with earlier X-ray equipment, they offered better penetration and a useful skin-sparing effect. A substantial dose could reach tumours many centimetres beneath the surface, making treatment of deeper disease more practical.
Cobalt-60 also had limitations. Its energy was still inadequate for some very deep targets, and the radioactive source decayed continuously. Sources required periodic replacement, creating technical, logistical and environmental challenges.
The medical linear accelerator addressed many of these problems. In 1953, Hammersmith Hospital in London became one of the first centres to use a linac clinically, treating a patient with retinoblastoma. Linacs accelerated electrons through a linear tube using high-frequency electromagnetic waves. Electrons could treat superficial disease directly, or generate high-energy X-rays when directed onto a metal target.
Linacs provided higher energies, improved penetration and avoided a continuously radioactive source. By the 1970s, machines could produce several photon and electron energies, giving clinicians far greater flexibility in matching the beam to the depth and geometry of disease.
Radiation delivery improved, but the next revolution came from seeing anatomy in three dimensions. CT planning allowed oncologists to localise tumours and normal organs throughout the patient’s volume. Three-dimensional conformal radiation therapy then used computer-designed beam shapes to surround the tumour more closely and reduce unnecessary dose.
Intensity-modulated radiation therapy extended this principle by varying the intensity within each field. Optimisation algorithms could sculpt complex dose distributions around sensitive structures while maintaining target coverage. Image-guided radiation therapy added imaging at the treatment machine, allowing clinicians to account for setup variation, breathing, organ motion and anatomical change.
Stereotactic radiosurgery and stereotactic body radiation therapy pushed accuracy further, delivering very high doses in one or a few sessions with sub-millimetre precision. Techniques first developed for intracranial targets expanded to tumours in the lung, liver, spine and other sites, sometimes achieving outcomes comparable with surgery in carefully selected patients.
Proton therapy uses a different physical interaction. Protons deposit most of their energy near a selected depth—the Bragg peak—with relatively little exit dose. This can reduce exposure to healthy tissue, especially in children and when tumours lie close to critical structures.
Artificial intelligence is now entering contouring, planning and adaptive workflows. Algorithms can automate the delineation of tumours and organs, optimise dose and help revise plans as anatomy changes. The opportunity is substantial, although the profession must preserve the clinical judgment needed to supervise these systems and recognise their errors.
MR-linacs combine a linear accelerator with MRI. Real-time soft-tissue visualisation during treatment can support dynamic targeting and online adaptation beyond what is possible with conventional cone-beam CT. The treatment machine is becoming not merely a radiation source, but an integrated platform for imaging, decision-making and delivery.
The arc from Cobalt-60 to MR-guided adaptive therapy is more than a sequence of machines. It is a history of progressively tighter feedback between what we can see, what we can calculate and what we can safely deliver. That convergence has improved tumour control, reduced toxicity and changed what is possible for patients.