Hepatic malignancies: Usefulness of acquisition of multiple arterial and portal venous phase images at dynamic gadolinium-enhanced MR imaging

PURPOSE: To determine whether liver tumor detection is increased by acquiring multiple dynamic arterial phase and portal venous phase magnetic resonance (MR) images. MATERIALS AND METHODS: Dynamic MR imaging was performed in 205 patients at 1.5 T with use of a fast spoiled gradient-echo technique (repetition time, 9-12 msec; echo time, 2.1-3.0 msec; flip angle, 30 degrees). During intravenous bolus injection of gadopentetate dimeglumine or gadoteridol (0.1 mmol/kg), nine images were acquired at each of 10-12 locations over approximately 120 seconds. The number of tumors detected on arterial phase and portal venous phase images and unenhanced T1- and T2-weighted spin-echo (SE) images was evaluated separately. Data obtained in 75 patients with proved malignancies were analyzed. RESULTS: At imaging, 220 malignant tumor nodules were depicted. At prospective review of all images, 110 hepatocellular carcinomas (HCCs) were detected: 82 (75%) on unenhanced T1-weighted SE images, 83 (75%) on unenhanced T2-weighted SE images, 92 (84%) on arterial phase images, and 76 (69%) on portal venous phase images. At prospective review, eight HCCs were detected on only arterial phase images, one on only portal venous phase images, nine on both arterial and portal venous phase images, and 11 on only unenhanced SE images. The 18 additional HCCs detected prospectively on only dynamic images increased HCC detection by 21% over prospective detection on only SE images. Detection of non-HCC tumors (including metastases) did not increase with dynamic images. CONCLUSION Acquisition of multiple dynamic arterial phase and portal venous phase images increased detection of HCC but not metastases.