Imaging affects outcome in portal vein thrombosis

October 5, 2005

Imaging plays an increasingly important role in the early detection of portal vein thrombosis (PVT). It also helps determine the condition's underlying cause, which has implications for further treatment, according to U.K. researchers.

Imaging plays an increasingly important role in the early detection of portal vein thrombosis (PVT). It also helps determine the condition's underlying cause, which has implications for further treatment, according to U.K. researchers.

"Differentiation of tumor from bland thrombus can be difficult but is important. The most common cause for portal vein tumor invasion is hepatocellular carcinoma," said Dr. Euan Armstrong, specialist registrar in the department of radiology at Derriford Hospital in Plymouth.

Armstrong and his colleagues presented an educational poster at the European Society of Gastrointestinal and Abdominal Radiology meeting in May.

Medical management of PVT consists of systemic anticoagulation. Surgical thrombectomy is ineffective because clots typically extend into small portomesenteric venous radicles, which are inaccessible to standard balloon embolectomy, the authors wrote. Early initiation of anticoagulation can minimize serious complications such as peritonitis due to bowel necrosis. Chronic sequelae of PVT, including splenorenal shunts, may be seen.

Tumor thrombus in the portal vein can be detected in 30% to 62% of patients with hepatocellular carcinoma (HCC), and detection of tumor strongly affects prognosis. The median survival of patients with portal vein tumor thrombus (PVTT) is 2.7 months, compared with over two years for those without PVTT. Involvement of the portal vein in HCC is a relative contraindication to surgical resection and transplant. Bland thrombosis is also a negative predictor of postradiation survival in patients with HCC.

"Ultrasound is the modality of choice for initial investigation of PVT. The features of PVT on B-mode ultrasound include increased diameter of the portal vein, visualization of the thrombus, collateralization, and cavernous transformation. Sensitivity is improved by color Doppler and the use of contrast agents," Armstrong wrote.

Doppler sonography can differentiate bland from tumor thrombus, but Doppler parameters must be fully optimized. Pulsatile flow in portal vein thrombi occurring in patients with cirrhosis is a 62%-sensitive and 95%-specific sign for the diagnosis of malignant portal vein thrombus, according to the Plymouth team. Continuous or absent flow can be detected in benign and malignant portal vein thrombus and is not useful in differentiating between the two.

The limitations of ultrasound are its operator dependence and reduced sensitivity with increased patient body habitus, gas, and ascites. Low flow can also be mistaken for thrombosis. Portal-vein visualization is difficult in cirrhotic patients, due to sound attenuation, and may be improved with contrast agents. Overall detection of PVT by ultrasound has a reported sensitivity of 89% and specificity of 92%. Endoscopic ultrasound may be a more sensitive test than transabdominal ultrasound, but it is invasive.

Overall CT detection of bland PVT has a reported sensitivity of 90% and specificity of 99%. CT may also detect the underlying cause of PVT. CT characterization of malignant PVT has a reported sensitivity of 86% and specificity of 100%.

"Contrast-enhanced CT has gained an increasingly important role in patients with acute mesenteric venous thrombosis, and is considered the diagnostic test of choice at many centers," the authors wrote. "Multidetector CT scanners offer the advantages of significantly shorter acquisition times, retrospective thin or thick-section reconstruction from the same raw data, improved 3D rendering with decreased helical artifact, and increased z-axis coverage without compromising image quality."

Timing parameters can be optimized to allow imaging of both arterial and portal venous phases. CT can evaluate vascular structures, the bowel wall, and adjacent mesentery. Findings of PVT include persistent, well-defined intraluminal filling defects with central low attenuation, which may be surrounded by well-defined, rim-enhancing venous walls. Features suggesting malignancy include identification of a main PVT diameter greater than or equal to 23 mm, PVT enhancement, and neovascularity.

"A pitfall of portal vein imaging by CT is the appearance of pseudothrombus. This appearance occurs during the hepatic arterial phase in the main portal vein lumen and is due to mixed flow from the enhanced splenic vein return and the nonenhanced superior mesenteric vein return," the authors noted. "However, a homogeneously enhanced portal vein is seen during the late portal venous phae."

Conventional imaging of PVT by MRI employs fast spin-echo T1- and T2-weighted sequences. MR angiography (gradient-echo) using time-of-flight and phase contrast sequences is slow to acquire and is therefore more susceptible to motion artifact, according to the researchers. Gadolinium-enhanced 3D MRA can provide good-quality imaging in a single breath-hold. Images are acquired in a coronal plane as a thick maximum intensity projection with reduced flow and saturation artifacts. Overall detection of PVT by 3D contrast-enhanced MRA has a sensitivity of 98% and specificity of 99%.

MRI appears to be the most sensitive and specific test, but it is limited by time, availability, and cost, and some patients cannot tolerate MRI, the researchers concluded.