Y90 radioembolization in advanced liver cancer of rarely treated tumor entities: prediction of treatment response and survival by clinical and imaging features.

Background Radioembolization is classically used in patients with hepatocellular carcinoma and secondary hepatic malignancies like colorectal cancer and neuroendocrine tumors. Up to now, data for other hepatic metastasized tumor entities are rare, and it is not clear which clinical and imaging features could predict the outcome.Methods Data of 60 patients exceptionally treated with radioembolization through the decision of an interdisciplinary board with tumor entities usually not treated by radioembolization were retrospectively analyzed. Hepatic features on cross-sectional preprocedural imaging, for example, like hypervascularization, hepatic tumor burden, maximum diameter, and clinical data, were also analyzed during follow-up. Predictive factors were calculated.Results Univariate Cox regression showed a significant impact of different features on overall survival, for example, primary malignancy histology in the case of sarcoma, with 25 months median vs. 6 months median overall survival (P = 0.015). Multivariate Cox regression revealed the two key factors for overall survival: arterial hypervascularization more than rim enhancement (15 vs. 6 months, P = 0.011, hazard ratio: 0.339) and hepatic tumor burden below 10% (18 vs. 5 months, P < 0.001, hazard ratio: 1.04).Conclusion Preprocedural hypervascularization of liver malignancies and hepatic tumor burden are overall survival predicting factors in tumor entities rarely treated by radioembolization. Consideration of these factors can help to carry out suitable patient selection in the interdisciplinary board for radioembolization.