Caren van Roekel

44 Chapter 2 The shape and magnitude of these effects depend on photon energy, tissue composition, collimator and detector characteristics, the distance between the source and the collimator, and the energy window settings (86). Elschot et al. have developed a new method for 90 Y SPECT reconstruction. This Monte Carlo-based reconstruction algorithm compensates for scatter and attenuation effects and improves the quantitative accuracy of bremsstrahlung SPECT images (86). Because of the wide range of photon energy, the energy window for bremsstrahlung SPECT/CT should be wide. However, because of the high focal uptake in the liver after radioembolization, a single-window approach is sufficient (87). Reported ranges in the literature are 105-195 keV and 50-250 keV (86-88). A high-energy collimator is proven to be better than a medium energy collimator by Elschot et al (14). The specifications of the SPECT/CT are defined by Elschot et al.: a radius of rotation of camera of 260 mm, 120 projections/360°, 40-minute acquisition time, 256x256 matrix size and 1.6x1.6 mm 2 pixel size (88, 89). A largely neglected issue in clinical SPECT is respiratory motion. Bastiaannet et al. have shown that respiratory motion has a very large effect on dosimetry on the spatial scale of individual tumors, with an average decrease in activity recovery and tumor to non-tumor ratio from 90% to 66%. Thus, respiratory motion leads to an underestimation of tumor dose. This can partially be solved by using retrospective gating schemes (90). 6.1.2. 90 Y PET/CT Although annihilation photon pairs are produced about 700 times less often than bremsstrahlung photons, 90 Y PET/CT was found to be superior over bremsstrahlung SPECT/CT for the assessment of the microsphere distribution after radioembolization (88, 91). One of the first post-infusion clinical 90 Y PET/ CT scans was performed in 2009. Since then, many studies have described results from both patient and phantom imaging. There are two major issues in 90 Y PET/CT imaging: the low true-coincidence rate and the high singles rate due to bremsstrahlung x-rays. The first issue may result in noisy images, long scan times and background noise from scintillator-decay. The singles-count rate from bremsstrahlung is much higher than the positron emission rate and may cause saturation of the detector, leading to a limited quantitative accuracy (84). However, this latter issue has not been found in recent investigations