Differentiation of Malignant from Benign Portal Vein Thrombi on CT Images Using Thrombus Density

Abstract
Background: Portal vein thrombosis described as the presence of a clot in the portal vein lumen or a permanent obliteration of the portal vein as a result of prior thrombosis with replacement by numerous tortuous venous channels (termed cavernoma). Malignant portal vein thrombus, so named for its neoplastic origin, is a common complication of HCC, and, in some cases, it may be even the initial sign of an undetected HCC. Detection of malignant PVT in a patient with liver cirrhosis heavily affects the therapeutic strategy.
Aim of Study: The purpose of this study was to investigate the role of CT thrombus density (measured in Hounsfield Units) in distinguishing between neoplastic and bland portal vein thrombosis on arterial and portal venous phases.
Material and Methods: In this study, 30 patients underwent contrast-enhanced CT of abdomen & pelvis were included for characterization of portal vein thrombosis. Assessment of portal vein thrombosis was performed by measuring of CT attenuation values of the thrombi in Hounsfield Units (HU).
ROC (Receiver Operating Characteristic) curves were used to identify accuracy and optimal cutoff values.
Results: Of the 30 CT studies, 14 neoplastic thrombi and 13 bland thrombi were identified on the images. CT thrombus density (measured in Housinfield Unit) to differentiate neo-plastic from bland thrombus. The AUCs was 0.98 in arterial phase and 0.98 in portovenous phase for thrombus density.
The optimal cut off in arterial phase is 47 and in porto-venous phase is 50.
Conclusion: CT attenuation values allow reliable differ-entiation between neoplastic and bland thrombi on arterial and portal venous phase CT examination.