Watershed Hepatocellular Carcinomas: The Risk of Incomplete Response following Transhepatic Arterial Chemoembolization.

ABSTRACT

PURPOSE: Hepatocellular carcinomas (HCCs) bridging two or more Couinaud-Bismuth segments of the liver ("watershed tumors") can recruit multiple segmental arteries. The primary hypothesis of this study was that fewer watershed tumors show complete response (CR) after chemoembolization, with shorter time to local recurrence. Secondary analysis on the impact on transplantation eligibility in the presence of progressive disease was also performed. MATERIALS AND METHODS: A total of 155 transplantation-eligible patients whose HCC met Milan criteria (watershed, n = 83; nonwatershed, n = 72) and was treated with chemoembolization were included. Cone-beam computed tomography (CT) was used for guidance and for confirmation of circumferential uptake. Local response to chemoembolization per modified Response Evaluation Criteria In Solid Tumors and local disease-free survival (DFS) for the index tumor were calculated. Differences were assessed by univariate and multivariate analyses. RESULTS: CR after a single of chemoembolization was observed in 55.4% of watershed tumors and in 72.2% of nonwatershed tumors (P = .045). Estimated DFS intervals were 151 days (95% confidence interval [CI], 93-245 d) and 336 days (95% CI, 231-747 d; P = .040) in the watershed and nonwatershed groups, respectively. Worse DFS was observed with a Model for End-Stage Liver Disease score > 20 (P = .0001), higher Child-Pugh-Turcotte score (P = .049), and watershed location (P = .040). Waiting list drop-off rates were statistically similar between groups. CONCLUSIONS: Hepatocellular carcinomas located in the watershed region of the liver have a poorer response to chemoembolization than those located elsewhere. These tumors are associated with worse DFS and require additional treatments to maintain transplantation eligibility per Milan criteria. Cone-beam CT can identify crossover supply and confirm complete geographic drug uptake, possibly reducing (but not eliminating) the risk of incomplete response.