Performance of novel virtual parenchymal perfusion software visualizing embolized areas of transcatheter arterial chemoembolization for hepatocellular carcinoma

AIM: To evaluate the performance of novel virtual parenchymal perfusion (VPP) software in conventional transcatheter arterial chemoembolization (cTACE) for hepatocellular carcinoma. METHODS: VPP was retrospectively applied to 43 hepatocellular carcinomas treated with cTACE. The virtual embolized area (VEA) was estimated after positioning a virtual injection point on images of non-selective cone-beam computed tomography during hepatic arteriography, at the same position in superselective cTACE. The real embolized area (REA) was defined as the area where iodized oil was retained on 1-week computed tomography after cTACE. Three dimensions across the tumor (maximum [a] and minimum [b] in the axial and craniocaudal [c]) directions, and the volume of the VEA and REA were compared using linear regression analysis. It was also evaluated whether an adequate safety margin ≥5 mm could be predicted by VPP. RESULTS: The mean lengths of a, b, and c of the VEA and REA were 54.6 ± 15.9 mm (range 24.9-91.0 mm) and 55.0 ± 15.7 mm (range 23.9-92.8 mm; r = 0.9448, P < 0.001), 35.4 ± 11.7 mm (range 16.1-66.0 mm) and 35.4 ± 13.2 mm (range 12.2-69.2 mm; r = 0.9369, P < 0.001), and 42.1 ± 11.6 mm (range 25.0-75.7 mm) and 42.9 ± 11.7 mm (range 25.7-78.7 mm; r = 0.9092, P < 0.001), respectively. The mean volume of the VEA and REA was 71.8 ± 44.8 mL and 75.5 ± 46.9 mL, respectively (r = 0.9913, P < 0.0001). VPP showed no safety margins in seven tumors, including all three actually embolized without safety margins. CONCLUSION: The VEA estimated using VPP showed a good correlation with the REA of cTACE.