Clinical evaluation of spatial accuracy of a fusion imaging technique combining previously acquired computed tomography and real-time ultrasound for imaging of liver metastases

Purpose: This study was designed to evaluate the spatial accuracy of matching volumetric computed tomography (CT) data of hepatic metastases with real-time ultrasound (US) using a fusion imaging system (VNav) according to different clinical settings. Methods: Twenty-four patients with one hepatic tumor identified on enhanced CT and US were prospectively enrolled. A set of three landmarks markers was chosen on CT and US for image registration. US and CT images were then superimposed using the fusion imaging display mode. The difference in spatial location between the tumor visible on the CT and the US on the overlay images (reviewer #1, comment #2) was measured in the lateral, anterior-posterior, and vertical axis. The maximum difference (Dmax) was evaluated for different predictive factors. CT performed 1-30 days before registration versus immediately before. Use of general anesthesia for CT and US versus no anesthesia. Anatomic landmarks versus landmarks that include at least one nonanatomic structure, such as a cyst or a calcification Results: Overall, Dmax was 11.53 ± 8.38 mm. Dmax was 6.55 ± 7.31 mm with CT performed immediately before VNav versus 17.4 ± 5.18 with CT performed 1-30 days before (p < 0.0001). Dmax was 7.05 ± 6.95 under general anesthesia and 16.81 ± 6.77 without anesthesia (p < 0.0015). Landmarks including at least one nonanatomic structure increase Dmax of 5.2 mm (p < 0.0001). The lowest Dmax (1.9 ± 1.4 mm) was obtained when CT and VNav were performed under general anesthesia, one immediately after the other. Conclusions: VNav is accurate when adequate clinical setup is carefully selected. Only under these conditions (reviewer #2), liver tumors not identified on US can be accurately targeted for biopsy or radiofrequency ablation using fusion imaging.