Matthias Würlpresents two essential steps to implement offline PET monitoring of proton dosedelivery at a clinical facility, namely the setting up of an accurate MonteCarlo model of the clinical beamline and the experimental validation ofpositron emitter production cross-sections. In the first part, the field sizedependence of the dose output is described for scanned proton beams. Both theMonte Carlo and an analytical computational beam model were able to accuratelypredict target dose, while the latter tends to overestimate dose in normaltissue. In the second part, the author presents PET measurements of differentphantom materials, which were activated by the proton beam. The resultsindicate that for an irradiation with a high number of protons for the sake ofgood statistics, dead time losses of the PET scanner may become important andlead to an underestimation of positron-emitter production yields.