Anatomic versus Metabolic Tumor Response Assessment after Radioembolization Treatment.

PURPOSE: To assess applicability of metabolic tumor response assessment on 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) after radioembolization (RE) in patients with colorectal liver metastases (CRLM) by comparison with one-dimensional size-based response assessment on MR imaging. MATERIALS AND METHODS: This prospective cohort study comprised 38 patients with CRLM undergoing RE. MR imaging and 18F-FDG PET/CT imaging were performed at baseline, 1 month (n = 38), and 3 months (n = 21). Longest tumor diameter (LTD) reduction on MR imaging at these time points was compared with reduction in total lesion glycolysis (TLG) on 18F-FDG PET/CT. Hepatic response was compared between RECIST and total liver TLG and correlated with overall survival (OS). RESULTS: TLG and LTD were positively correlated in 106 analyzed metastases (38 patients) at 1 month and 58 metastases (22 patients) at 3 months. Agreement was poor, with LTD underestimating TLG response. A significant association with prolonged OS was found in total liver TLG at 1 month (HR 0.64, P < .01) and 3 months (HR 0.43, P < .01). For LTD, a significant association with OS was found at 3 months (HR 0.10, P < .01). Important differences in liver response classification were found, with total liver TLG identifying more patients and situations where there appeared to be treatment benefit compared with RECIST. CONCLUSIONS: TLG response assessment on 18F-FDG PET/CT appears to be more sensitive and accurate, especially at early follow-up, than size-based response assessment on MR imaging in patients with CRLM treated by RE. Semiautomated liver response assessment with total liver TLG is objective, reproducible, rapid, and prognostic.