Background <p>Transarterial radioembolization (TARE) treatment efficacy critically depends on the microsphere biodistribution in the liver. Current treatment planning protocols rely on a pre-treatment scout procedure to evaluate this biodistribution for a selected catheter configuration. However, replication of the catheter configuration during treatment is challenging and often results in suboptimal tumor targeting. Here we present and experimentally validate a concept to improve tumor targeting using dynamic contrast-enhanced ultrasound (DCE-US) imaging to predict the microsphere biodistribution in real-time, using ultrasound contrast microbubbles as microsphere precursors.</p> Methods <p>DCE-US-guided TARE was performed in three ex vivo perfused porcine livers using a customized normothermic machine-perfusion platform. In each liver, one lobe was selected for targeting, and a 9L4 ultrasound transducer was fixed to image this location. A microcatheter was inserted in the hepatic artery under fluoroscopic guidance. Several catheter positions were analyzed in real-time using ultrasound contrast microbubbles and time-intensity curves (TICs). The catheter position was optimized based on peak TIC intensity, which was retained for subsequent injection of non-irradiated ¹⁶⁵Ho-microspheres. The resulting microsphere biodistribution was evaluated by R2^* mapping in MRI.</p> Results <p>Here we show that DCE-US can be used to identify feeding arteries, enabling optimization of the catheter placement to maximize peak intensity within the target lobe. MRI confirms the relative microsphere biodistribution as predicted by DCE-US in two quantifiable cases (Case 2: 99.2% [MRI] vs. 98% [DCE-US]; Case 3: 70.2% [MRI] vs. 80.9% [DCE-US]).</p> Conclusions <p>DCE-US enables accurate prediction of microsphere biodistribution, providing real-time feedback for immediate intra-operative optimization of the catheter position during TARE.</p>