Caren van Roekel

198 Chapter 7 DISCUSSION Building on the establishment of a dose-response relationship in patients treated with 166 Ho-radioembolization by Bastiaannet et al. (11), this study explored the dose-response relationship in a homogenous population of patients with CRC only. Furthermore, dose-toxicity relationships were studied. Our results suggest a positive association between a higher parenchymal- absorbed dose and increase in CTCAE grade toxicity, both for clinical and laboratory toxicity. Furthermore, our data unveils, both at a lesion- and at a patient-level, a significant dose-response relationship. Also, a mean tumor- absorbed dose >90 Gy – the minimal mean tumor-absorbed dose in the group of patients with CR/PR - was associated with a significantly longer survival. In this study, treatment with radioembolization was well tolerated. The most frequent clinical adverse events were CTCAE grade 1-2 abdominal pain, nausea and fatigue. These adverse events are well-known side effects of treatment with radioembolization (19). One patient of our study died of hepatic failure. This safety profile is compliant with the results of the MORE study, which showed that treatment with 90 Y-resin radioembolization is safe in a patient population highly comparable to ours, namely CRC patients who received several lines of prior chemotherapy (20). A study on the safety of 90 Y-glass radioembolization in CRC patients showed similar results, with the most frequent side effects being fatigue, abdominal pain and nausea (21). The incidence of grade ≥3 laboratory toxicity is also comparable between the three types of microspheres (20, 21). Regarding efficacy, the metabolic response rate (CR/PR) at a tumor level was 36%, comparable to previous dose-response data on resin microspheres in a similar patient cohort, treated in the same hospital (22). In other studies, higher tumor-response rates up to 75% were found, with dose thresholds of 46 and 60 Gy (23, 24). However, it is difficult to compare these studies with our study, as those patients were less heavily pretreated or received concomitant systemic therapy (23, 24). Also, their thresholds cannot be compared to our data, as there are major differences between microsphere types in specific activity, size, number of particles and half-life. There was only minimal agreement in response between the PERCIST and the RECIST assessments. In 15 patients of our study, RECIST underestimated response according to PERCIST. This finding is