Predictors of microscopic portal vein invasion by hepatocellular carcinoma: Measurement of portal perfusion defect area ratio

Abstract

Microscopic portal vein invasion (PVI) by cancer cells is a poor prognostic factor after hepatic resection for hepatocellular carcinoma (HCC). The aim of this study is to predict PVI preoperatively in patients with HCC. We studied 46 hepatectomized patients who had HCC without any portal venous invasion detected during preoperative radiographic evaluation. We defined the portal perfusion defect area ratio (PPDAR) as the following: the quotient of the maximal portal perfusion defect area, on computed tomography during arterio-portography (CTAP) is divided by the maximal tumor area on magnetic resonance imaging (MRI) or CT. The median PPDAR was 1.3 (mean 1.4 +/- 1.1; ranged from 0.7 to 5.8). The incidence of PVI was 4.5% in patients with a PPDAR <1.3, 35.7% in those with a PPDAR of 1.3-1.6, 70% in those with a PPDAR > or = 1.6 (P = 0.0005). When analyzing the preoperative value of different cut-off points for the PPDAR, the lowest P-value by Fisher's exact test was achieved when the PPDAR threshold was 1.6 (P = 0.0012). The sensitivity was 58%, and specificity was 91% with this cut-off value. On univariate analyses, factors that significantly correlated with PVI were PPDAR (P = 0.0012), serum levels of des-gamma-carboxy prothrombin (P = 0.033), and tumor size (P = 0.0126). On multivariate analysis, PPDAR was the only significant independent predictor of PVI. Our study shows that PPDAR is a new concept, which is useful in predicting PVI and that a value > or = 1.6 is predictive of PVI.