[Liver metastasis from carcinoid: diagnostic imaging]

Our study was aimed at describing the diagnostic imaging patterns of carcinoid liver metastases. Six patients with liver metastases secondary to carcinoid tumor were examined. The metastases were histologically proved in each patient. All patients were examined with ultrasonography (US), Computed Tomography (CT), Magnetic Resonance Imaging (MRI) and Digital Subtraction Angiography (DSA). All patients were treated with transcatheter embolization of liver metastases. Diagnostic imaging methods showed ten to many dozen metastases in each patient. Tumor size ranged 0.5 to 14 cm. US showed hypoechoic, hyperechoic and isoechoic carcinoid liver metastases with a hypoechoic halo. Large lesions had anechoic central areas due to colliquative necrosis. Hypoechoic patterns were the most frequent ones. Precontrast CT showed hypodense metastases; very small lesions were isodense relative to surrounding liver. CT during contrast agent injection showed that the metastases were hyperdense in the arterial phase; contrast enhancement was poorer in the portal phase. Large lesions showed a hypodense central area due to necrosis which remained hypodense in the late phase. The metastases were hypointense on T1-weighted and hyperintense on T2-weighted MR images. Gradient echo dynamic imaging with Gd-DTPA showed high-signal metastases in the arterial phase and lower signal intensity in the portal phase. DSA, an essential exam before embolization, showed tortuous and elongated intra- and extrahepatic arteries and tumor neovascularization with no malignant abnormalities. In the capillary phase, tumor uptake and inhomogeneous hypervascular patterns were shown. Portal veins were only displaced and compressed, but never infiltrated by the metastases. All the techniques we used contributed to assess liver involvement by carcinoid metastases. DSA must be used only before treatment; both CT and MRI showed the hypervascular patterns of the metastases, but no technique could predict their nature.