Purpose: To report our personal experience with helical CT evaluation of hepatocellular carcinoma treated with various percutaneous interventional procedures. We assessed both nodular response and the spectrum of changes within normal parenchyma.
Material and methods: December 1996 to September 1998 we examined with helical CT 41 patients (73 nodules in all) with hepatocellular carcinoma treated with percutaneous ablation therapies: conventional ethanol injection in 18 subjects (31 nodules), one-shot ethanol injection in 3 (8 nodules), radiofrequency thermal ablation in 16 (25 nodules), and combined chemoembolization and ethanol injection in 4 (9 nodules). CT performed was 4-27 days after the last session, acquiring biphasic volumetric images in 14 patients and triphasic volumetric images in 27. A second treatment with subsequent CT study was performed for 28 lesions; 15 underwent 3 serial studies and 6 underwent 4 studies.
Results: Compared with pretreatment findings, the diameter was unchanged in 62% of the nodules and increased in 38%. Morphology was unchanged in 63% of the lesions while in 37% a mild deformation toward the needle path or a more regular and round shape was evident. Borders were unchanged in 37% of the cases and modified in 63%, appearing well-defined in 73% and ill-defined in 27%. The necrotic portion had a low attenuation with a nodule-to-parenchyma gradient more evident on delayed than on venous and finally arterial acquisitions; 8% of the lesions were not recognizable on unenhanced scans. Residual viable tissue was identified in 44% of the nodules and quantified as 100% in 1% of all lesions, > 75% in 3%, > 50% in 4%, > 25% in 12%, < 25% in 23%. It was located centrally in 6% of the cases, peripherally in 12%, and eccentrically in 81%, and the shape was crescent in 66% of the cases, oval in 19%, and different in 16%. During the arterial phase the residual tumor appeared hyperdense in 97% of the nodules and isodense in 3%, while during the portal phase it was hyperdense in 22%, isodense in 28% and hypodense in 50%, and during the delayed phase hypodense in 100%. For what concerns the uninvolved parenchyma, lobar atrophy was present in 4% of the nodules, segmental atrophy in 8%, and subsegmental in 16%; caval thrombosis was found in 1 patient, portal thrombosis in 4. Arterioportal fistulas were identified in 5% of the nodules and areas of transient inhomogeneous attenuation in several cases. Subsegmental biliary dilation was detected in 4% of the lesions.
Conclusion: Multiple-phase helical CT allows optimal depiction of primitive liver nodules treated with percutaneous interventional procedures and has a central role in the assessment of tumor response. Accurate is also the evaluation of changes or complications involving the surrounding parenchyma. Regarding the ablation effect, two typical features can be identified: after total necrosis nodular volume is unchanged or increased, borders are well-defined, and density is low and homogeneous (especially in delayed phases), while after partial necrosis the diameter is not modified, margins are ill-defined, attenuation is less decreased, and peripheral tumor maintains high density in arterial phase and low density in delayed phases. As for the treatment procedure, we found that thermal ablation with radiofrequency and single-session ethanol injection cause more evident alterations within both nodule and hepatic parenchyma, while after multi-session ethanol injection changes are less dramatic and consequently more difficult to assess.