Detection and characterization of hepatic tumors: value of combined helical CT hepatic arteriography and CT during arterial portography.

Abstract

The purpose of our study was to evaluate the diagnostic accuracy of combining helical CT hepatic arteriography (CTA) with helical CT during arterial portography (CTAP) for detecting and characterizing hepatic tumors. From January 1994 to September 1995, combined helical CTA and CTAP were performed in 52 patients (38 men and 14 women, 37-83 years old [mean, 64.1 years old]) with suspected primary or metastatic hepatic tumors before they were evaluated for hepatic resection. Retrospectively, two radiologists reviewed the CTAP images alone and subsequently matched CTA and CTAP images. The radiologists recorded a confidence level on a five-point scale for their evaluation of visible perfusion abnormalities as true lesions, and diagnostic accuracy was assessed using receiver-operating-characteristic analysis. Using the data sets obtained at the interpretation sessions by the radiologists, we determined sensitivities for CTAP alone and for combined CTA and CTAP in 28 malignant hepatic tumors that were confirmed in 19 patients who underwent definitive surgery. CTA revealed one hepatocellular carcinoma in a patient with severe cirrhosis and six adenomatous hyperplasias in another cirrhotic patient that CTAP failed to reveal. In comparison with CTAP alone, the combination of CTA and CTAP allowed significantly more accurate detection of lesions (area under receiver-operating-characteristic curve: 0.889 for CTAP alone versus 0.954 for CTA and CTAP combined, p < .01). The combination of CTA and CTAP also significantly raised the accuracy for characterizing hepatic tumors. Sensitivities for malignant hepatic tumors revealed in the 19 patients who underwent surgery were identical: 93% (26 of 28 tumors) with CTAP alone and with CTA and CTAP combined. Although combining CTA and CTAP does not significantly increase sensitivity for detecting hepatic tumors, the combined techniques significantly raise specificity in detecting hepatic tumors and accuracy in characterizing hepatic tumors.