PURPOSE: To quantify tumor perfusion after transcatheter arterial embolization (TAE) with functional computed tomography (CT) and to validate the reproducibility of quantification measurements. MATERIALS AND METHODS: This study was conducted in accordance with an institutional animal care and use committee-approved protocol. Sixteen rats with liver tumors underwent TAE with 1 mg (group 1) or 3 mg (group 2) of polyvinyl alcohol particles. In each group, four rats underwent functional CT immediately after TAE (day 0) and four others underwent functional CT 2 days after TAE (day 2). Another four rats served as control rats. Blood flow (BF), blood volume (BV), mean transit time (IVITT), and permeability-surface area product were measured by using a functional CT software program. For evaluation of reproducibility, six additional rats with mammary tumors underwent functional CT twice, with examinations 2 hours apart. The mixed-effect model was used to assess the TAE treatment effect, and the Pearson correlation test was used to determine measurement reproducibility. RESULTS: With the exception of BF in group 1 on day 2 (P =.41), BF and BV values in both groups on both days were significantly lower than BF and BV values in the control rats (with P values ranging from .018 to <.001). BF was significantly lower in group 2 than in group 1 on days 0 and 2 (P =.043 and P =.02, respectively). BV was significantly lower on day 2 than on day 0 in group 2 (P =.016). IVITT was generally inversely related to BF. MTTs in group 2 on days 0 and 2 were significantly longer than those in the control rats (P <.001 and P =.03, respectively), and MTT was shorter on day 2 than on day 0 in group 2 (P =.02). Permeability-surface area product changes were similar to BF changes. There were no significant differences (P values ranged from .2 to .5) between perfusion parameters in the reproducibility study. CONCLUSION: The results of this study validate the use of functional CT in the quantification of tumor perfusion after TAE and the reproducibility of such quantification measurements. (c) RSNA, 2005.