Usefulness of virtual parenchymal perfusion software visualizing embolized areas to determine optimal catheter position in superselective conventional transarterial chemoembolization for hepatocellular carcinoma.

Abstract

To determine the optimal catheter position during superselective conventional transarterial chemoembolization (cTACE) for hepatocellular carcinoma (HCC) using virtual parenchymal perfusion software. Patients who had newly developed HCC nodules ≤6cm and five or fewer lesions were eligible. The virtual catheter tip was placed on a tumor-feeder identified by TACE guidance software using cone-beam computed tomography during hepatic arteriography to minimize the virtual embolized area (VEA), including the tumor with a safety margin. Conventional transarterial chemoembolization was then carried out at the same position. The VEA and real embolized area where iodized oil was retained on cone-beam computed tomography after cTACE were compared using the dice similarity coefficient, linear regression analysis, and mean surface distance. Technical success of cTACE and therapeutic effects by the modified Response Evaluation Criteria in Solid Tumors were also evaluated. Ninety-one tumors in 56 patients were embolized. The mean dice similarity coefficient values in 80 VEAs and real embolized areas were 0.78±0.01. Both volumes were well correlated (r=0.957, p<0.001) with a mean surface distance of 2.78±2.11mm. Eighty-four (92.3%) tumors were embolized with a safety margin. Regarding the early response of 82 tumors, complete response was achieved in 72 (87.8%), partial response in six (7.3%), and stable disease in four (4.9%). Regarding responses of 81 tumors during the follow-up (mean, 20±4.9months), complete response was maintained in 62 (76.5%), whereas 19 (23.5%), including six that were incompletely embolized, locally progressed. Virtual parenchymal perfusion software can determine the optimal catheter position in superselective cTACE.