Preablation three-dimensional ultrasonography can predict therapeutic effect and local tumor progression after radiofrequency ablation for hepatocellular carcinoma.

Purpose: To assess the agreement between ablative margin (AM) predicted by preablation three-dimensional ultrasonography (3D-US) and AM measured on postablation computed tomography (CT)/magnetic resonance (MR) images.

Methods: Sixty patients with 73 hepatocellular carcinoma nodules were enrolled. 3D-US data were collected immediately after puncture by the electrode before ablation. The maximum distance from the electrode to the edge of the tumor in the plane perpendicular to the electrode (C-plane) was defined as "a" and the diameter of the ablation zone as "b". We classified predicted AM into "0.5b - a" ≥0 mm as AM(+) or <0 mm as AM(-), and "0.5b - a" ≥3 mm or <3 mm.

Results: Forty-eight nodules (66 %) were visualized in the C-plane. There was an agreement between the predicted and measured AMs for 39 (81 %) of the 48 nodules. Local tumor progression was observed in 3 (7%) of 43 nodules with predicted AM(+) and in 2 (40 %) of 5 nodules with predicted AM(-) but was not observed in any of 21 nodules with predicted AM ≥ 3 mm. The local tumor progression rate was significantly lower for nodules with predicted AM(+) compared with predicted AM(-)(p = 0.03), and for nodules with predicted AM ≥ 3 mm compared with predicted AM < 3 mm (p = 0.04). Local progression was detected in 2 (4.7 %) of 42 nodules with a sufficient AM (≥0 mm) on postablation CT/MR images and in 5 (83.3 %) of 6 nodules with an insufficient AM (<0 mm); the difference in progression rate was significant (p = 0.0008).

Conclusion: 3D-US allows prediction of the AM before radiofrequency ablation.