Caren van Roekel

273 Discussion happened in these studies. The low incidence of serious radioembolization- related adverse effects (hepatic failure, portal vein thrombosis, radiation hepatitis; occurring in only four patients (0.8%)) contributes to this assumption (35). A different treatment planning approach was chosen in the DOSISPHERE study. In this study, 56 patients with hepatocellular carcinoma were randomly assigned to treatment with radioembolization using standard dosimetry or to treatment with radioembolization using personalized dosimetry with the goal to deliver at least 205 Gy to the tumor. For safety reasons, in the personalized dosimetry group, the maximum tolerable dose on the healthy liver tissue was set at 120 Gy in case of lobar treatment. With segmental treatment, the parenchymal-absorbed dose could be higher than 120 Gy, provided that there was a hepatic reserve >30%. Thus, some flexibility in activity prescription was possible, allowing for high enough tumor-absorbed doses. Dosimetry was based on the distribution of 99m Tc-MAA and response was evaluated after three months. In the standard dosimetry arm, the response rate was 36%, versus 71% in the personalized dosimetry arm. A significant dose-response relationship was established and mean tumor-absorbed doses were significantly higher in the personalized dosimetry group (324 Gy versus 221 Gy) (85). Although this study was done in patients with hepatocellular carcinoma, it is a very important study for mCRC patients as well, as it is the first study to show that a personalized treatment planning leads to higher response rates. Although several studies showed a dose-response relation in mCRC patients treated with radioembolization (86-88), a prospective study to validate the thresholds was not performed yet. For 90 Y-resin, the tumor-absorbed dose thresholds needed for response vary from 40-60 Gy (86-90) (Table 1). For 90 Y-glass, dose-response studies in mCRC patients are still awaited but in HCC patients, the tumor-absorbed dose needed for response is set around 200 Gy (85), although higher thresholds (up to 500 Gy) were reported in earlier studies (91). The safety and efficacy thresholds are different mainly because of differences in specific activity. In mCRC patients treated with 166 Ho-radioembolization, an evident dose- response relationship was found, both at a patient- and at a tumor-level. An 10