Time-of-flight (TOF) PET

In PET, a positron annihilates with an electron to produce two 511 keV photons travelling in almost exactly opposite directions; the scanner detects both photons in coincidence and knows the decay occurred somewhere along the line connecting the two detectors (the line of response), but conventional PET has no information about where along that line. Time-of-flight (TOF) PET measures the small difference in the two photons' arrival times with high-precision timing electronics and fast scintillator crystals, which localises the annihilation point to a smaller segment of the line of response: the positional uncertainty is Δx = c x Δt / 2, where c is the speed of light and Δt is the scanner's coincidence timing resolution.

Modern clinical TOF-PET scanners achieve coincidence timing resolutions in the range of roughly 200-400 picoseconds, corresponding to a spatial localisation of roughly 3-6 cm along the line of response; this does not by itself improve the intrinsic spatial resolution of the image (still around 4-5 mm, set by detector and reconstruction factors), but it substantially reduces statistical noise during image reconstruction because each detected event only needs to be redistributed across a much shorter segment of possible locations rather than the full line of response. The resulting gain in image signal-to-noise ratio is proportional to the square root of the object size (patient girth) divided by the timing resolution, so the benefit of TOF is greatest in larger patients.

Clinically, TOF capability improves lesion detectability, allows shorter scan times or lower injected activity for the same image quality, and is now standard on modern PET/CT and total-body PET systems used for oncological imaging, including PSMA, SSTR and FDG PET.