A Short History of Positron Imaging, From a Cosmic-Ray Track to PET-CT

Positron emission tomography has a tidy name and a messy origin. Physicists found the particle, doctors realised that its signature could locate tumours, engineers built detector rings, chemists made the right molecule, and a later generation joined PET to CT. Each step had several contributors, and histories sometimes disagree about who was first. This page gives a short timeline limited to names, years and claims that appear in the sources listed at the end, and says plainly where credit or dating depends on how you define the milestone. The physics behind the story is explained in the physics of positron emission and annihilation.

The particle: 1928 to 1936

A 2023 historical review by Uta Eberlein places the theoretical groundwork in 1928, when Paul Dirac presented a relativistic wave equation that predicted an antiparticle of the electron. In 1932 Carl Anderson, studying cosmic rays, noticed that a small fraction of his photographed tracks (15 out of 13,000, according to Eberlein) curved the opposite way to electrons, which pointed to a particle of electron-like mass and positive charge. Eberlein reports that Patrick Blackett and Giuseppe Occhialini confirmed the finding in 1933. The Nobel Prize in Physics 1936 was divided equally between Victor Hess, for his discovery of cosmic radiation, and Anderson, "for his discovery of the positron"; the Nobel site calls it the first antiparticle proven by experiment.

Early medical use: the 1950s

It took nearly two decades, in Eberlein's phrasing, before positron-emitting atoms were used to localise brain tumours in the early 1950s. She cites F. R. Wrenn, M. L. Good and P. Handler in 1951 for using positron-emitting radioisotopes this way, W. H. Sweet in 1951 for work on diagnosis and treatment of brain tumours, and G. L. Brownell and W. H. Sweet in 1953 for describing coincidence detection to localise brain tumours with copper-64. The logic was the one later used in every PET scanner: detect the paired annihilation photons together to say where a tracer sits.

Tomography and the first PET scanners: 1968 to 1978

Eberlein credits David Kuhl and Roy Edwards with introducing emission and transmission tomography in 1968. In a 2020 perspective in the Journal of Nuclear Medicine, Townsend and Jones add that David Chesler, working in Gordon Brownell's group at Massachusetts General Hospital, produced the first tomographic reconstructions using annihilation coincidence detection with a planar positron camera, and applied filtered backprojection to medical imaging for the first time in 1971. They then describe the 1975 proposal by Michael Phelps, Edward Hoffman, Nizar Mullani and Michel Ter-Pogossian at Washington University in St. Louis of the first positron emission transaxial tomograph, called PETT, an array of detectors around a single imaging plane. The 24-detector prototype was tested on phantoms and a dog. In 1976 the team built a larger single-plane instrument able to accommodate a human, and in 1977 a team including Ter-Pogossian showed that radionuclide concentrations in human tissue could be measured quantitatively.

Which of these counts as "the first PET scanner" depends on the definition. Eberlein dates the first PET scanner using filtered back projection to 1975 and names Ter-Pogossian, Phelps and Hoffman. Townsend and Jones show that annihilation coincidence tomography had been demonstrated with a planar camera in 1971. Both statements can be right, because one is about a ring of individual detectors for transaxial images and the other about earlier camera-based reconstruction.

Commercial scanners followed. Townsend and Jones record that Phelps and Hoffman, by then at UCLA, worked with EG&G Ortec on the ECAT II, and that several centres bought it to measure physiological quantities such as blood flow and glucose consumption. A hexagonal brain scanner, the NeuroECAT, came in 1981, and the ECAT III of 1984 moved to a circular detector arrangement that, they say, became the design standard for future scanners. Around 1984 a block detector module allowed up to 64 scintillators to be read by four photodetectors.

FDG: 1976 to 1986

A scanner needs a tracer worth imaging. Fluorodeoxyglucose, or FDG, labelled with fluorine-18, is the best-known one, introduced in PET radiotracers explained. According to Joanna Fowler's 2020 perspective, the molecule was modelled on the carbon-14 2-deoxyglucose method for measuring brain glucose use in animals. The first synthesis route used fluorine gas made at Brookhaven National Laboratory's cyclotron, and enough product was made to be transported from Long Island to the University of Pennsylvania for the first human studies in 1976. Eberlein, by contrast, dates the development of FDG to 1978 and credits Ido and colleagues. The two years need not conflict: 1976 is when the tracer was first given to people, and 1978 is the date Eberlein gives for its development. We have not seen a source that settles which date a reader should treat as the birth of FDG, so both are given.

The early route was inefficient. Fowler writes that at most 50 percent of the fluorine-18 was incorporated and that it gave a mixture of isomers. A 1986 paper by Hamacher and colleagues used a different chemistry, with a compound called Kryptofix, to reach about 95 percent incorporation and roughly 60 percent overall yield. Fowler calls that radiosynthesis transformative because it let FDG be distributed from central production sites around the world, a point that links to the cyclotron supply chain described on the radiotracers page.

PET-CT and later: 1991 to the present

The next big step came from merging two machines. In a 2008 history, Townsend writes that the PET/CT concept was born in 1991 in work with Ron Nutt and colleagues, that the first prototype became operational in 1998 and was evaluated at the University of Pittsburgh, and that the first commercial PET/CT to be announced was the GE Discovery LS in early 2001. By 2006, he says, PET-only scanners could no longer be obtained. Townsend and Jones list PET/MR as introduced in 2010, and they describe a total-body design called EXPLORER with 564,480 detectors that covers 198 cm along the body. The trade-offs of the hybrid designs are explained in PET-CT and PET-MRI.

Timeline at a glance

Year	Milestone	Source
1928	Dirac's equation predicts the electron's antiparticle	Eberlein
1932	Anderson identifies the positron in cosmic-ray tracks	Eberlein; Nobel Prize
1936	Nobel Prize in Physics, shared by Hess and Anderson	Nobel Prize
1951 and 1953	Positron emitters used to localise brain tumours	Eberlein
1971	Chesler applies filtered backprojection to coincidence data	Townsend and Jones
1975	PETT proposed by Phelps, Hoffman, Mullani and Ter-Pogossian	Townsend and Jones
1976	Human-size PETT; first human FDG studies	Townsend and Jones; Fowler
1984	ECAT III, circular detector design	Townsend and Jones
1986	High-yield FDG synthesis (Hamacher and colleagues)	Fowler
1998 and 2001	PET/CT prototype operational; first commercial PET/CT announced	Townsend (2008)
2010	PET/MR introduced	Townsend and Jones

This list is selective. Many chemists, physicists and clinicians are missing from it, and a lack of mention is not a judgement on their work. Readers who want to go further will find starting points on the further reading page, and teachers can use the timeline alongside the page for students and teachers.

The short version

The positron was identified in 1932, put to medical use for brain tumours in the early 1950s, turned into tomographic scanners in the 1970s, paired with a practical glucose tracer in the late 1970s and 1980s, and joined to CT in 1998 to 2001. Credit for each step is shared, and several dates depend on the definition used. For how any of this affects a scan you are having, your imaging team has the final word.