Caren van Roekel

45 Radioembolization (92). Regarding the first problem: the full-width at half maximum (FWHM) of 90 Y PET/CT is around 5 mm, and scan times of about 20 minutes are enough for dosimetry. The background noise is worse in organs were the 90 Y activity concentration is lower (84). Due to the higher intrinsic resolution of 90 Y PET/CT compared to 90 Y SPECT/CT, intrahepatic distribution is more accurately visualized by 90 Y PET/CT imaging. In the next generation PET/CT systems, the photomultiplier tubes are replaced with solid-state digital photon counting detectors. Wright et al have shown that detection of internal pair production is feasible with this digital photon counting (dPET) technology. dPET images provided increased contrast and allow for more precise localization of microsphere distribution (93). There is no guideline as to when patients should undergo 90 Y PET/CT imaging. Due to the limited sensitivity, it is advised to perform these scans the day of treatment or one day later. 6.1.3. 90 Y PET/MRI The next development in 90 Y imaging is PET/MRI. PET/MRI provides a better soft tissue contrast than PET/CT. The first clinical study of 90 Y radioembolization with post-treatment PET/MRI imaging was conducted by Fowler et al. Twenty-four patients with tumors of diverse origin were treated with radioembolization with either glass or resin microspheres. Images were qualitatively assessed for microsphere distribution and the absorbed doses per lesion were calculated, as well as dose volume histograms, to measure the dose distribution within a tumor. The duration of PET/MR imaging ranged from 42-60 minutes. They found that the average dose per lesion, as well as the minimum dose to 70% of the lesion, were significant predictors of response (94). Simultaneous PET/MRI offers the opportunity to directly image liver motion due to respiration during the PET acquisition and correct for it during the PET reconstruction. Eldib et al. have performed phantom experiments with PET/ MRI. They found that motion resulted in a large loss of contrast recovery, but this was successfully corrected by MR-based data correction (95). 2