PET imaging: radioactivity to see inside the body
Medical imaging is a fantastic tool available for doctors to help find a diagnosis or to help treat patients. Almost everyone knows about X-ray Computed Tomography (CT) or Magnetic Resonance Imaging (MRI). However there is another technique, widely used but yet less famous, Positron Emission Tomography (PET). That’s a shame because PET imaging is an essential tool on its own and also works well when coupled with structural imaging methods such as CT and MRI. It can measure body functions such as blood flow, oxygen use, sugar metabolism and the functioning of organs and tissues. It is also interesting to note that it is a non-invasive technique, meaning that there is no surgery involved, even though we have a ionising radiation exposure during the acquisition.
Radioactivity? Isn’t that dangerous?
There is a huge misunderstanding about radioactivity. Most people picture only A-bombs or nuclear plants but it is entirely different. Radioactivity  is just a natural, physical process where an atomic nucleus loses energy by emitting a particle. Nature likes stable things and radioactivity is here to make a stable element from an unstable one. For every element, there are different versions of it – we call them isotopes. The isotopes used in PET imaging are Fluorine, Carbon, Nitrogen or Oxygen. Why those ones? Well because they are already present in our body so it will be more likely to accept them. The most used radiotracer is Fluorideoxyglose (FDG). It is a glucose molecule coupled with an isotope of Fluorine called Fluorine-18, which is absorbed by the brain, heart muscles and tumorous tissues.
What does a PET scanner look like?
Because a picture is worth a thousand words, here comes the PET:
But what is it made of? We see the bed where patients lie and the structure containing what is the most interesting part: the PET camera. The basic element of the camera is a small detection block containing 8 rows and 8 columns of crystal scintillators  associated with 4 photomultipliers . The utility of each element is going to be explained later. We assemble several of these blocks into a ring of 72 blocks. 4 of those rings compose the PET camera to take measurements from various angles.
Where do we use it?
Many important fields of medicine, such as Neurology, Cardiology and Oncology, use PET imaging for diagnosis and treatment monitoring .
In Neurology, PET allows us to examine how the brain functions and to help detect disorders or diseases like epilepsy or Alzheimer.
In Cardiology, it is used to evaluate the risk of spreading of diseases in patients considered for transplantation or revascularisation, to detect ischemic disease or to check the viability of the cardiac muscle (myocardium).
It is for Oncology that we conduct a large proportion of procedures making use of PET because it is able to identify malignant tumors.
How does PET work?
We have our radiotracer travelling inside the body. The amount of uptake varies depending on the kind of tissue or organ, their health, the radiotracer used among other variables. The isotope emits, at a random moment, a positively-charged particle called a positron (𝛽+). This particle only travels a few millimetres through the body and disappears when encountering the negative equivalent of a positron, an electron (𝛽–). In addition to this annihilation of both particles, two gamma photons (𝛾) are emitted and projected in exactly opposite directions. See following figure.
It is thanks to these gamma photons that we are able to detect what is happening. Indeed, once they reach the crystal on the PET camera, it will create light and the photomultiplier will detect it.
This whole process happens millions of times and we store all this information about the location of the gamma photons in a datafile.
How does all that produce an image?
In mathematics reconstructing images is called an inverse problem and it is a real challenge. To visualise such a problem, we can use an analogy. Let’s take an object, like a cup. You have your cup, on a table, with a lamp facing it. This cup will produce a shadow on the table. The cup is our patient. The lamp is our radiotracer. The shadow is our data. Our problem is to find the exact shape of the cup, knowing the shadow and the direction of the light.
From there, we have two choices:
First we have a method called Filtered BackProjection (FBP). It consists in applying the Filtered Backprojection Formula or Radon Inversion Formula , reversing the path from the data to the image. But such a problem is ill-posed, meaning, unfortunately, that more than one solution to the problem exists.
The other possibility is to use statistics. The idea is to approach step-by-step the image we want. We start with a random image on which we apply a projection similar to the one of the gamma photons. We now have the same kind of data detected by the PET camera and we can compare them. If they are close enough, we can stop. Otherwise, we correct the image to make it more accurate and we repeat the process. We call such a method Expectation Maximisation (EM) .
Both methods have their upsides and downsides but in PET imaging we apply the second one more frequently nowadays.
Finally, we have our image which is going to help doctors to find the most suitable treatments for patients. Yay!
 L’Annunziata, Michael F. (2007). Radioactivity: Introduction and History. Amsterdam, Netherlands: Elsevier Science. p. 2. ISBN 9780080548883
 Duclos, Steven J. (1998). “Scintillator Phosphors for Medical Imaging”. Interface. The Electrochemical Society. 7 (2): 34–39. ISSN 1944-8783.
 Iams, H.; Salzberg, B. (1935). “The Secondary Emission Phototube”. Proceedings of the IRE. 23: 55. doi:10.1109/JRPROC.1935.227243.
 Strauss, L.G.; Dimitrakopoulo-Strauss, A (2017). Positron Emission Tomography (PET) Applications, Encyclopedia of Spectroscopy and Spectrometry (Third Edition), 2017, Pages 718-722, doi:10.1016/B978-0-12-803224-4.00046-7
 Radon, J.; Parks, P.C. (translation) (1986), “On the determination of functions from their integral values along certain manifolds”, IEEE Transactions on Medical Imaging, 5 (4): 170–176, doi:10.1109/TMI.1986.4307775, PMID 18244009.
 Shepp, L.; Vardi, Y. (1982). “Maximum likelihood reconstruction for emission tomography”. IEEE Transactions Medical Imaging. 1 (2): 113–122, doi.org/10.1109/TMI.1982.4307558
Radioactivity : Nuclear Imaging, http://www.radioactivity.eu.com/site/pages/Medical_Apparatus.htm
PET/CT – Position Emission Tomography/ Computed Tomography, https://www.radiologyinfo.org/en/info.cfm?pg=pet
Bailey, D.L; D.W. Townsend; P.E. Valk; M.N. Maisey (2005). Positron-Emission Tomography: Basic Sciences. Secaucus, NJ: Springer-Verlag. ISBN 1-85233-798-2, doi:10.1007/b136169