Three-dimensional registration of images obtained before and after radiofrequency ablation of hepatocellular carcinoma to assess treatment adequacy.

Abstract

The aim of our study was to evaluate the value of 3D registration images reconstructed by fusion of pre- and posttreatment CT or MRI for the assessment of ablative margins after percutaneous radiofrequency ablation (RFA) of hepatocellular carcinoma (HCC). From January 2007 to May 2011, we performed RFA in 84 patients to treat 139 HCC nodules, the margins of which had been assessed by comparing pre- and postablation images side by side. The same nodules were retrospectively assessed again with 3D registration images after classification into four margin grades. We analyzed the cumulative local recurrence rate for each grade and reviewed the origin of recurrence. Three-dimensional registration images predicted local recurrences more accurately than did the conventional side-by-side method (area under the curve, 0.678 and 0.536, respectively; p = 0.0144). The cumulative rates of local recurrence were significantly different among the margin grades assessed with 3D registration images (p = 0.0088). Three-dimensional registration images detected that the major origins of recurrence (n = 22) were residuals (n = 13) and sites of no margin (n = 6), especially proximate to blood vessels more than 3 mm in diameter. Three-dimensional registration of pre- and postablation CT or MRI more accurately assesses the ablative margin than the conventional method. It can predict a proclivity for local recurrence after RFA according to margin grade. It also indicated that residuals and sites of no margin proximate to blood vessels that are more than 3 mm in diameter are high-risk locations for local recurrence after ablation.