Helical CT screening for hepatocellular carcinoma in patients with cirrhosis: Frequency and causes of false-positive interpretation.

OBJECTIVE. The purpose of our study was to determine the specificity of helical CT for depiction of hepatocellular carcinoma in a population of patients with cirrhosis. SUBJECTS AND METHODS. Single-detector helical CT screening was undertaken in 1329 patients with cirrhosis who were referred for transplantation. The patients underwent one or more helical CT examinations over 30 months and were followed up for an additional 19 months or until transplantation. We predominantly used unenhanced and biphasic contrast-enhanced techniques with infusions of 2.5-5.0 mL/sec. Four hundred thirty patients underwent transplantation within this period. Liver specimens were sectioned at 1-cm intervals, with direct comparison of imaging and pathologic findings and histologic confirmations of all lesions. Prospective preoperative helical CT reports were used for the primary data analysis. A retrospective unblinded review was undertaken to determine characteristics of false-positive lesions diagnosed as hepatocellular carcinoma. RESULTS. Thirty-five patients (8%) had false-positive diagnoses for hepatocellular carcinoma based on helical CT. Twenty of these patients (5%) showed hypoattenuating lesions seen during one of the three helical CT examination phases. Fifteen patients (3%) had hyper-attenuating lesions seen during the arterial phase. Among the 15 hyperattenuating lesions, CT revealed the causes to be transient benign hepatic enhancement (n = 3), hemangiomas (n = 2), fibrosis (n = 2), peliosis (n = 1), volume averaging (n = 1), low-grade dysplastic nodule (n = 1), or undetermined (n = 5). Of the 20 hypoattenuating lesions, the causes were shown to be fibrosis (n = 8), focal fat (n = 4), infarcted regenerative nodules (n = 2), regenerative nodules (n = 1), fluid trapped at the dome of the liver (n = 1), hemangioma (n = 1), or undetermined (n = 3). Follow-up helical CT in 13 (72%) of 18 patients allowed a change in the diagnosis of hepatocellular carcinoma to a finding of no cancer present. CONCLUSION. Helical CT screening for hepatocellular carcinoma in patients with cirrhosis has a substantial false-positive detection rate. Although most of lesions were hypoattenuating, a few hyperenhancing arterial phase lesions were proven not to be hepatocellular carcinoma. An awareness of imaging characteristics and follow-up imaging can help radiologists avoid a mistaken diagnosis in many patients.