OBJECTIVE : To evaluate the infusion dynamics and residual yttrium-90 activity during and after resin microsphere radioembolization with different injection techniques and initial activities. To assess the distribution of residual activity in the administration systems to allow optimization of the procedure and the equipment. METHODS : In a setup similar to that in standard clinical practice, radioembolization procedures were performed ex vivo. The influence of different injection techniques was assessed by comparing pulsatile and continuous injections. The influence of the absolute amount of activity to the residual activity was assessed by comparing pulsatile 0.5-GBq- with 1.0-GBq-procedures. Continuous dose rate measurements were performed. Activity distribution was determined by positron-emission tomography (PET)/CT. RESULTS : Fifteen procedures were performed: 5 pulsatile 0.5-GBq-, 5 continuous 0.5-GBq-, and 5 pulsatile 1.0-GBq-procedures. Mean residual activity was 4.0% ± 1.7% (range 1.2%-6.6%), without statistically significant differences between injection techniques (P = .841) or between prescribed activities (P = .222). Dose-rate measurements revealed an exponential decrease of the activities in the vials with high variability. Activity fell rapidly to 32% ± 7.9% (range 23%-55%) after injection of 4 of 20 mL 5% dextrose solution. Residual activity accumulations were identified at the 3-way stopcock (100% of procedures), in the C-line (80%), at the microcatheter connector (20%), and in the A-line (6.7%), but not in the vials. CONCLUSIONS : Residual activity in a commercial administration system for resin microsphere radioembolization is variable and does not systematically depend on initial yttrium-90 activity or on injection technique. Predilection sites for residual activity were identified, which should receive special attention when performing resin transarterial radioembolization procedures, and for further administration system developments.