Liver CT for vascular mapping during radioembolisation workup: comparison of an early and late arterial phase protocol

OBJECTIVES: To compare right gastric (RGA) and segment 4 artery (A4) origin detection rates during radioembolisation workup between early and late arterial phase liver CT protocols. METHODS: 100 consecutive patients who underwent liver CT between May 2012–January 2015 with early or late arterial phase protocol (n = 50 each, 10- vs. 20-s post-threshold delay) were included. RGA/A4 origin detection rates, assessed by two raters, and contrast-to-noise ratio (CNR) of the hepatic artery relative to the portal vein were compared between the protocols. RESULTS: The first–second rater scored the RGA origin as visible in 58–65 % (specific proportion of agreement 82 %, κ = 0.62); A4 origin in 96–89 % (94 %, κ = 0.54). Thirty-six percent of RGA origins not detectable by DSA were identified on CT. Origin detection rates were not significantly different for early/late arterial phases. Mean CNR was higher in the early arterial phase protocol (1.7 vs. 1.2, p < 0.001). CONCLUSION: A 10-s delay arterial phase CT protocol does not significantly improve detection of small intra- and extrahepatic branches. RGA origin detection requires further optimization, whereas A4/MHA origin detection is adequate, with good inter-rater reproducibility. CT remains important for preprocedural planning, because it may reveal arterial anatomy not discernible on DSA. KEY POINTS: • An early arterial phase does not significantly improve RGA and A4/MHA origin detection. • RGA origin detection (58–65 %) on CT is still suboptimal. • 36 % of RGA origins undetectable on DSA can be identified on CT. • A4/MHA origin detection (89–96 %) on CT is excellent. • Inter-rater reproducibility is good for RGA and A4/MHA origin detection on CT.