Surgical portosystemic shunting, the formation of a vascular connection between the portal and systemic venous circulation, has been used as a treatment to reduce portal venous pressure. Although the use of portosystemic shunt surgery in the management of portal hypertension has declined during the past decade in favour of alternative therapies, and subsequently surgeons and radiologists became less familiar with the procedure, it remains a well-established treatment. Knowledge of different types of surgical portosystemic shunts, their pathophysiology and complications will help radiologists improve communication with surgeons and enhance their understanding of the diagnostic and therapeutic role of radiology in the assessment and management of these shunts. Optimal assessment of the shunt is essential to determine its patency and allow timely intervention. Both non-invasive and invasive imaging modalities complement each other in the evaluation of surgical portosystemic shunts. Interventional radiology plays an important role in the management of complications, such as shunt thrombosis and stenosis. This article describes the various types of surgical portosystemic shunts, explains the anatomy and pathophysiology of these shunts, illustrates the pearls and pitfalls of different imaging modalities in the assessment of these shunts and demonstrates the role of radiologists in the interventional management of complications.

To evaluate the efficacy of CT-maximum intensity projection (CT-MIP) in the detection of gastric varices and their inflowing and outflowing vessels in patients with gastric varices scheduled to undergo balloon-occluded retrograde transvenous obliteration (B-RTO).

Sixteen patients with endoscopically confirmed gastric varices were included in this study. All patients were evaluated with CT-MIP using three-dimensional reconstructions, before and after B-RTO.

CT-MIP clearly depicted gastric varices in 16 patients (100%), the left gastric vein in 6 (32.5%), the posterior gastric vein in 12 (75.0%), the short gastric veins in 13 (81.3%), gastrorenal shunts in 16 (100%), the hemiazygos vein (HAZV) in 4 (25.0%), the pericardiophrenic vein (PCPV) in 9 (56.3%), and the left inferior phrenic vein in 9 patients (56.3%). Although flow direction itself cannot be determined from CT-MIP, this modality provided clear images of the inflowing and the outflowing vessels. Moreover, in one patient, short gastric veins were not seen on conventional angiographic portography images of the spleen, but were clearly revealed on CT-MIP.

We suggest that CT-MIP should be considered as a routine method for detecting and diagnosing collateral veins in patients with gastric varices scheduled for B-RTO. Furthermore, CT-MIP is more useful than endoscopy in verifying the early therapeutic effects of B-RTO.

The authors developed ABI alloy, which mechanically resembles stainless steel 316. The main elements of ABI alloy are palladium and silver. Magnetic resonance (MR) images and radiographs of ABI alloy and stainless steel 316 stent models and of nitinol, tantalum, and Elgiloy stents were compared. ABI alloy showed the least MR imaging artifacts and was more radiopaque than stainless steel 316. ABI alloy has the potential to replace stainless steel 316 for construction of balloon-expandable MR imaging-compatible stents.

To investigate the safety and imaging artifacts with different coronary arterial stents and magnetic resonance (MR) imaging sequences.

The heating, artifacts, and ferromagnetism with different stents were studied with a 1.5-T MR tomograph with ultrafast gradients by using turbo spin-echo, turbo gradient-echo, and echo-planar imaging sequences. Nineteen stents, which were 8-25 mm in length and 3.0-4.5 mm in diameter, were evaluated. Stent deviation induced by the magnetic field and during MR imaging, migration, and heating caused by the radio-frequency pulses were examined. The size of imaging artifacts was measured with all the stents under standardized conditions and with six stents after their implantation into the coronary arteries of freshly explanted pig hearts.

All except two types of stents showed minimal ferromagnetism. No device migration or heating was induced. Turbo spin-echo images had minimal artifacts; larger artifacts were seen on the turbo gradient-echo and echo-planar images. With ultrafast gradients, the artifacts on the echo-planar images were substantially reduced.

The studied coronary stents were not influenced by heating or motion during 1.5-T MR imaging. Artifact size differed according to the type and size of the stent and the MR imaging sequence used. Thus, patients with these stents can be safely examined.

To evaluate MR imaging features of commercially available stents before, during, and after in vitro deployment as a step toward MR-guided stent deployment.

Fourteen stents were deployed in a phantom under MR monitoring at 1.5 T by using a gradient-echo sequence. Device visibility was rated on a four-point scale (excellent, fair, poor, not visible).

The Memotherm stent and the rolling membrane (RM) Wallstent showed excellent stent visibility and at least fair scores for artifact-induced narrowing of the stent lumen. Three stents (Palmaz, AVE, Easy Wallstent) showed excellent visibility of the stent but no visible lumen. Five stents (Strecker, Accuflex, Hemobahn, Passager, Sinus) displayed fair visibility. The delivery catheters of four stent systems (Smart, Vascucoil, Symphony, ZA) displayed severe black hole artifacts.

The imaging features of several stent systems might be suitable for MR-guided intervention. The Memotherm and the Wallstent RM combine good visibility of the stent and the lumen.

The construction of a mesoinnominate shunt between the superior mesenteric vein and the left innominate vein is a recent surgical procedure for the treatment of Budd-Chiari syndrome with vena caval obstruction. The purpose of this study was to determine the role of the different imaging modalities for the follow-up of mesoinnominate shunts. Doppler US (n = 32), and MR imaging examinations (n = 32) were prospectively performed in 10 patients with mesoinnominate shunts. Shunt patency or thrombosis was confirmed with angiography (n = 32) and transhepatic portography with pressure measurement (n = 6). For each follow-up evaluation, all examinations were performed within 4 days. Shunt patency was correctly demonstrated in 28/28 cases with Doppler US, CT and MR imaging. Shunt thrombosis was correctly demonstrated in 4/4 cases with Doppler US, CT and MR imaging. Since Doppler US, CT and MR imaging have the same accuracy for diagnosing mesoinnominate shunt patency and detecting thrombosis, our study suggests that redundant screening methods can be avoided. Doppler US is accurate enough and should be the preferred technique in the evaluation of mesoinnominate shunts.

We describe two children who underwent magnetic resonance imaging (MRI) as part of the evaluation of isolated splenomegaly after Doppler ultrasound was unable to identify the portal vein. The diagnosis of cavernous transformation of the portal vein and the normalcy of the biliary tract and hepatic architecture were confirmed by MRI. Because of the risks associated with computed tomography (CT) and angiographic evaluation of such patients, the use of MRI, a noninvasive modality, in the evaluation of pediatric patients requires further consideration.

We report the case of a boy with Budd-Chiari syndrome. At 3 years of age a meso-atrial shunt was inserted for thrombosis of the hepatic veins with associated inferior vena cava occlusion by an enlarged caudate lobe. Seven uneventful years later, the patient was reassessed because of concern about insufficient length of the synthetic shunt. Ultrasound with color Doppler imaging (CDI) and magnetic resonance imaging (MRI) with magnetic resonance angiography (MRA) clearly depicted a patent graft, without kinking, stenosis or other sign of graft dysfunction. Conventional selective angiography confirmed the patency of the shunt. MRI with MRA and CDI are less invasive and less expensive than conventional angiography and might replace this technique for assessment of meso-atrial shunt patency.

Percutaneous transluminal angioplasty (PTA) was performed with the Inoue balloon catheter in nine patients with membranous obstruction of the inferior vena cava (MOVC) between November 1988 and April 1991. There were five men and four women, aged 30.8 +/- 8.5 years. Two patients had had previous surgical treatment. Three patients had complete and six had incomplete MOVC. The caval pressure below the MOVC was 24.0 +/- 5.6 and 11.8 +/- 5.3 mm Hg (p less than 0.0001) before and after PTA, respectively. The caval diameter at the site of MOVC increased from 1.5 +/- 1.7 to 20.3 +/- 2.6 mm (p less than 0.0001), and the maximal caval diameter below the MOVC decreased from 28.7 +/- 12.9 to 19.8 +/- 9.9 mm (p = 0.006), before and after PTA, respectively. One patient died of massive pulmonary embolism following successful PTA. All the other eight patients remained asymptomatic during an 18.5 +/- 11.5 months (range 3.5 to 32) follow-up period. Two-dimensional ultrasonograms showed no recurrence of MOVC. We conclude that PTA with the Inoue balloon catheter is an effective and safe alternative to surgical treatment of MOVC.

An investigation was carried out on 50 cadavers, in which the projection onto the anterior abdominal wall of the following vascular points was examined: the portal bifurcation, the direction of the course of the right and left branches of the portal vein and the terminal course of the hepatic veins near their entry into the inferior vena cava (IVC). The results are related to a transverse axis passing through the apex of the xiphoid process and the median plane in the supine position. The average position of the portal bifurcation is projected onto a point between a vertical line passing through the midpoint of the right hemithoracic width and a horizontal line passing through a point on the midclavicular line (MCL) corresponding to 57% of the height of the liver measured upwards from its inferior margin. The axis of the prehepatic course of the portal vein makes an angle of about 50 degrees, open downwards, with a vertical line drawn through the apex of the internal angle of the portal bifurcation. A line parallel to the course of the right and left branches of the portal vein is projected on to a surface line cranial to the right costochondral margin, which runs upwards at an angle of approximately 20 degrees towards the apex of the xiphoid process. The termination of the three great hepatic veins is projected at about the level of the xiphisternal joint, one sternal width to the right of the midline. Close to the IVC, the right hepatic vein runs upwards and medially at an angle of between 20 degrees and 30 degrees with the transverse plane.(ABSTRACT TRUNCATED AT 250 WORDS)

Sequential 2D inflow NMR (SDI) venography was employed to define venous anatomy and pathology in five healthy subjects and five patients, mean age 48 +/- 7 years, with documented deep vein thrombosis. SDI images were graded independently by two observers on a scale of 1,2 nondiagnostic and 3,4 diagnostic categories. All iliac and femoropopliteal SDI venograms were diagnostic. In contrast, only one out of nine calf SDI venograms was diagnostic. SDI is a promising new method to assess the proximal peripheral venous system. Improved imaging strategy is needed to produce flow images of the calf veins.

We performed in vitro tests of ferromagnetism of platinum microcoils of two manufacturers. In addition, we performed in vitro magnetic resonance imaging (MRI) on these two devices to determine their levels of magnetic susceptibility artifact production. We found that both devices were nonferromagnetic and produced a very low level of artifact during MRI. In vivo MRI of a dog and 2 patients having previous placement of platinum microcoils was performed without coil migration or significant artifact production. Platinum microcoils are highly MR-compatible.

Three patients, two of whom had histories of episodic hepatic encephalopathy, were shown by magnetic resonance imaging (MRI) to have occult spontaneous portosystemic shunts. The multiplanar imaging capabilities of MRI provided an optimal, noninvasive method of visualization of collateral anatomy in each case.

The results of ultrasound imaging combined with Doppler studies of blood flow are presented in two patients with surgical portacaval shunts performed for portal hypertension. Duplex scanning proved a quick and noninvasive method of confirming the patency of the anastomoses. The demonstration of appropriate Doppler flow signals across the shunt is the most convincing ultrasound evidence of patency. Portacaval shunts are more amenable to study by ultrasound than more peripheral shunts because the liver can be used as an acoustic window.

Magnetic forces exerted on surgical clips and the magnetic resonance imaging distortion they create in phantoms and rabbits at magnetic field strengths of 1.5 Tesla were investigated. Results are reported for both ligation and aneurysm clips manufactured from three types of stainless steel as well as titanium, tantalum and niobium metals. Paramagnetism and eddy currents were measured in a customized moving Gouy balance. Direct measurements of other magnetic forces were carried out in a 1.5T MRI system. The titanium and tantalum clips showed the least interaction with the magnetic field, both in terms of forces exerted and the observed image distortion with the larger clips generating the larger interactions. The strongest field distortions and attractive forces occurred with 17-7PH stainless steel clips. These interactions were ferromagnetic in origin and of sufficient strength to present significant risk to patients having this type of clip present during an MRI scan.