Effect of tumor size and tumor-to-liver contrast of hypovascular liver tumors on the diagnostic performance of hepatic CT imaging.
OBJECTIVE: To assess the effect of tumor size and tumor-to-liver contrast of simulated hypovascular liver tumors on the diagnostic accuracy of hepatic computed tomography (CT). MATERIALS AND METHODS: This retrospective study was approved by the institutional review board, and informed consent was waived. A total of 153 simulated hypovascular liver tumors were embedded in 70 hepatic CT data sets that were acquired during the portal venous phase. The simulated tumors had 3 different diameters (6, 10, and 14 mm) and 3 different tumor-to-liver contrast values (20, 35, and 50 HU). There were also 30 hepatic CT data sets without liver tumors. Three radiologists independently performed tumor detection on the randomized 100 hepatic CT data sets. RESULTS: The lowest sensitivity was obtained for the 6-mm tumors with a tumor-to-liver contrast of 20 HU (4.1%), and the highest sensitivity was obtained for the 10- and 14-mm tumors with a tumor-to-liver contrast of 50 HU (100%). Increasing the contrast from 20 to 35 to 50 HU in the 6-mm tumors yielded a significant increase in sensitivity (4.1%, 48.8%, and 92.4%, respectively; P < 0.0001). The sensitivity for the 10- and 14-mm tumors also increased significantly as the tumor-to-liver contrast value increased from 20 to 35 HU (P < 0.01). However, no significant increase in sensitivity was seen for the 10- and 14-mm tumors as the tumor-to-liver contrast values increased from 35 to 50 HU (P = 0.733 and P = 1.0, respectively). CONCLUSIONS: Increasing the tumor-to-liver contrast from 20 to 35 HU results in a significant increase in the detection of hypovascular liver tumors ranging from 6 to 14 mm in diameter. Optimization of the tumor-to-liver contrast is necessary for improved detection of hypovascular liver tumors.
Schindera, ST; Hareter, LF; Raible, S; Torrente, JC; Rusch, O; Rosskopf, AB; Marin, D; Vock, P; Szucs-Farkas, Z
Volume / Issue
Start / End Page
Pubmed Central ID
Electronic International Standard Serial Number (EISSN)
Digital Object Identifier (DOI)