Hepatic Arterial Embolization Using Cone Beam CT with Tumor Feeding Vessel Detection Software: Impact on Hepatocellular Carcinoma Response

F. H. Cornelis, A. Borgheresi, E. N. Petre, E. Santos, S. B. Solomon, K. Brown

Dedicated tumor feeding vessel detection software (TFVDS) using cone beam CT has shown a higher sensitivity to detect tumor feeding vessels during hepatic arterial embolization (HAE) of hepatocellular carcinoma (HCC) compared to 2D imaging. Our primary hypothesis was that HCC tumors treated with HAE guided by a TFVDS would show more complete response (CR) than when treated with 2D imaging alone. Secondary analysis of the impact on X-ray exposure was performed. Nineteen males and 8 females (median age: 69 year, 46-85) with 44 tumors (median size: 38 mm, 6-100) treated with selective HAE between January 2013 and December 2014 were included. Exclusion criteria were: extra-hepatic supply, >4 tumors, tumor size >10 cm, and adjunctive local therapy. Baseline patient and procedure characteristics were reviewed. Differences in CR per modified Response Evaluation Criteria in Solid Tumors were assessed by univariate and multivariate analyses for tumor size, number, location, particles size, and use of TFVDS. Median imaging follow-up was 20.1 months (2-33). Use of TFVDS (13 patients, 19 tumors) was the only factor predictive of CR (OR = 3.85 [CI95%: 1.09, 13.67], p = 0.04) on univariate analysis but not on multivariate analysis (OR = 3.26 [0.87, 12.23], p = 0.08). A higher rate of CR was observed for HAE using TFVDS guidance versus 2D imaging alone (68.4%, 13-19, vs. 36%, 9-25, p = 0.03). Median dose area product was lower when TFVDS was used (149.7 Gy.cm(2), 38-365, vs. 227.8 Gy.cm(2), 85.3-468.6, p = 0.05). HCC embolized with TFVDS may result in improved local tumor response without increasing the dose exposure.