Caren van Roekel

218 Chapter 8 Qualitative analysis Redistribution was rated as successful in 26/37 (70%) cases, dubious in 5/37 (14%) cases, and no redistribution was found in 6/37 (16%) cases. Inter-rater agreement was considered high (κ = 0.82). Quantitative analyses Five patients were excluded from the quantitative analyses. One patient had an additional parasitized artery that could not be coil-embolized, one had a superselective injection of microspheres, in which the healthy liver VOI could not be determined, one had corrupted post-treatment imaging files, and two were treated sequentially and had a large volume increase in one liver lobe making accurate image registration impossible. The median ratio of the dependent to non-dependent segment activity concentration was 0.88 (range 0.26 – 2.05). This means that the activity concentration in the coiled segments amounted to 88% of the activity concentration to the rest of the treated volume. For tumors the median ratio was 0.80 (range 0.19 – 1.62). Success-rates for redistribution based on activity concentration (using cut-off ratios of 0.9, 0.8, and 0.7) were 29%, 43% and 57%, respectively (Table 2 and figure 4a). Influencing factors Redistribution of segment IV arteries showed the highest rate of successful redistribution to the tumors (31%, 46% and 69%, using cut-off ratios of 0.9, 0.8, and 0.7 respectively) and parasitized arteries the lowest (0%, 33% and 33%). There was no notable difference in success rates between the microspheres used ( 90 Y glass, 90 Y resin, or 166 Ho microspheres). Comparison between the two largest tumor categories, colorectal carcinoma (CRC) and neuroendocrine tumor (NET), showed markedly lower success rates in NET patients (27%, 55%, 72% vs 0%, 0%, 0%). In parasitized arteries, 0/4 were deemed successful on the visual assessment, 1/4 (25%) was dubious and 3/4 (75%) were unsuccessful. In the quantitative analysis the success-rate was 0%, 33%, 33%, using cut-off ratios of 0.9, 0.8, and 0.7 respectively.