Hypervascular liver metastases: do unenhanced and hepatic arterial phase CT images affect tumor detection?

Abstract

To evaluate the relative roles of unenhanced and hepatic arterial phase (HAP) computed tomographic (CT) imaging in the detection of hypervascular liver metastases. Eighty-four patients with biopsy-proved liver metastases from hypervascular primary tumors other than hepatocellular carcinoma underwent unenhanced and HAP and portal venous phase (PVP) helical CT studies. Three blinded radiologists evaluated each series of images separately for the number, size, and enhancement characteristics of lesions. Sixty-nine patients had follow-up imaging proof of tumor burden. The three readers detected 381-402 lesions on the PVP images and 397-416 lesions on the unenhanced images. Unenhanced images allowed detection of 72%-80% of the lesions seen on PVP images. They detected 94-137 additional lesions on unenhanced but not PVP images. On the HAP images, 375-395 lesions were identified. HAP images allowed detection of 81%-90% of the lesions seen on PVP images. Forty-five to 78 additional lesions were detected on HAP but not on PVP images. In the 69-patient subset, maximal detection of tumor foci occurred in 94% of patients with unenhanced plus PVP images and in 78% with HAP plus PVP images. Unenhanced plus PVP images allowed detection of 96% of the 322 tumors in the subset population. Unenhanced plus PVP CT images allow detection of statistically significantly more hypervascular liver metastases than do HAP plus PVP images or imaging only in the PVP.