Semiquantitative prediction of early response of conventional transcatheter arterial chemoembolization for hepatocellular carcinoma using postprocedural plain cone-beam computed tomography.
Hepatol Res
; 47(3): E113-E119, 2017 Mar.
Article
em En
| MEDLINE
| ID: mdl-27138382
ABSTRACT
AIM:
To investigate whether plain cone-beam computed tomography (CT) immediately after conventional transcatheter arterial chemoembolization (c-TACE) can help to predict tumor response semiquantitatively in patients with hepatocellular carcinoma (HCC).METHODS:
Analysis was carried out retrospectively on 262 targeted HCCs in 169 patients treated with c-TACE. Dynamic CT was performed at baseline and 1-4 months after c-TACE. Receiver-operating characteristic curve analysis was undertaken to evaluate whether voxel values of cone-beam CT could predict a complete response and to identify the cut-off value. Final tumor response assessment and early prediction using the retention pattern of iodized oil, the cut-off value of the density, and the combination of the cut-off density value and retention pattern of iodized oil in HCCs on postprocedural cone-beam CT were compared.RESULTS:
Complete response was obtained in 72.9% of lesions. According to the pattern of iodized oil uptake, the sensitivity, specificity, and accuracy for predicting complete response were 85.9%, 70.4%, and 81.7%, respectively by excellent uptake on cone-beam CT. The area under the curve was 0.86 with the optimal cut-off at a voxel value of 200.13. According to not only the density but also the homogeneity of iodized oil retention, the sensitivity, specificity, and accuracy values for predicting complete response were 86.4%, 95.8%, and 88.9%, respectively. The predictive accuracy was significantly better than that of the pattern of iodized oil retention only (P = 0.019).CONCLUSION:
The combination of density and visual estimate of homogeneity is superior to either alone in predicting tumor response of c-TACE in HCC patients.
Texto completo:
1
Base de dados:
MEDLINE
Idioma:
En
Ano de publicação:
2017
Tipo de documento:
Article