Objectives: To prospectively evaluate the feasibility of perfusion CT as a follow-up modality after transcatheter arterial chemoembolization (TACE) and to compare these findings with those of histopathology as the reference standard in a VX2 tumor rabbit model.
Materials and methods: VX2 carcinoma tumors were implanted into the liver of 20 rabbits 3 weeks prior to TACE. Perfusion CT was performed prior to TACE and 1- and 4-week after TACE. After obtaining perfusion index maps on perfusion CT, 2 radiologists measured the parametric perfusion indices of blood flow (BF), blood volume (BV), mean transit time (MTT), permeability of the capillary vessel surface (PS), and hepatic arterial fraction (HAF) of primary tumors on pre-TACE perfusion CT, chemoembolized primary tumors on 1-week perfusion CT, and recurred tumors on 4-week perfusion CT. The normal liver parenchyma indices were also recorded. In addition, the radiologists investigated the presence of a recurred tumor adjacent to the chemoembolized area on perfusion index maps of 4-week CT images. The areas of higher hepatic blood flow (HBF), hepatic blood volume (HBV), PS, and HAF, and lower MTT on 4-week perfusion CT than the normal liver parenchyma and the identical area on 1-week perfusion CT were considered as recurred tumors. Histopathology revealed the presence of a recurred tumor, and mean vessel density (MVD) was determined by immunochemical staining for CD31. CT perfusion indices were compared by use of the t test. Comparisons were made for the primary tumor versus normal liver parenchyma on pre-TACE CT, the primary tumor on pre-TACE CT versus the chemoembolized tumor on 1-week CT, the recurred tumor on 4-week CT versus the identical area on 1-week CT, and the primary tumor on pre-TACE CT versus the recurred tumor on 4-week CT. For the detection of recurred tumors, the sensitivity and specificity for 4-week perfusion CT were calculated. Correlation analysis between the recurred tumor perfusion indices and the MVD of the corresponding tumor region was performed. Among 20 rabbits, 6 were excluded from the analysis, and results were based on 14 rabbits.
Results: Recurred tumors were histologically proven in 8 of 14 rabbits (57.1%). The BF, BV, PS, and HAF indices of primary tumors were significantly higher, whereas the MTT was significantly lower than that of the normal liver parenchyma on pre-TACE perfusion CT and that of chemoembolized areas on 1-week perfusion CT (P < 0.05). In addition, recurred tumors also showed significantly higher BF, BV, PS, and HAF, and lower MTT indices than the identical areas on 1-week perfusion CT (P < 0.05). The perfusion indices of recurred tumors were not significantly different from the indices of primary tumors (P > 0.05). Both sensitivity and specificity were 100% for 4-week perfusion CT. There were significant positive correlations between BF (r = 0.947), BV (r = 0.758), PS (r = 0.759), HAF (r = 0.955), and MVD in recurred tumors, and a significant inverse correlation between MTT (r = -0.782) and MVD was observed (P < 0.05).
Conclusions: We believe that perfusion CT is a feasible alternative modality for the successful early response assessment and early detection of a marginally recurred tumor after TACE. However, perfusion CT has limitations for the prediction of tumor recurrence after TACE.