Accuracy of preoperative prediction of microinvasion of portal vein in hepatocellular carcinoma using superparamagnetic iron oxide-enhanced magnetic resonance imaging and computed tomography during hepatic angiography

Our aim was to diagnose microinvasion of the portal vein in hepatocellular carcinoma from preoperative radiological findings and to construct a scoring system. Forty-seven patients (38 men and 9 women; median age, 66.8 years) who underwent hepatic resections for hepatocellular carcinoma were selected retrospectively. Microscopically, 22 had portal vein invasion (PVI) and 25 had no PVI. All patients were examined preoperatively with superparamagnetic iron oxide-enhanced magnetic resonance imaging and computed tomography during hepatic angiography (CTHA). Perilesional enhancement on T1-weighted imaging, tumorous arterioportal (AP) shunt, and corona enhancement (contrast enhancement of the adjacent liver appearing in the late phase of CTHA) were assessed. Relative risk for PVI in terms of clinical and tumor characteristics was also assessed. The relative contribution to PVI was determined by the coefficient of a stepwise logistic regression. Each variable was given a score relative to the coefficient. On univariate analysis, distortion of corona, tumorous AP shunt, and tumor size indicated a higher prevalence of PVI. The PVI predictive score was calculated as: total score = (maximum size in cm) + (T1 ring; + = 1, - = 0) + (tumorous AP shunt; + = 3, - = 0) + (distortion of corona; + = 10, - = 0). The PVI (+) group score was four times that of the PVI (-) group (16 vs 4). At a cutoff score of 10, the sensitivity, specificity, and accuracy were 82%, 84%, and 86%. Distortion of corona, tumorous AP shunt, and tumor size are good predictors of the risk of PVI. This scoring system is simple and worth using clinically.