Transjugular intrahepatic portosystemic shunt creation using a three-dimensional fluoroscopy guidance system in patients with the Budd-Chiari syndrome

Abstract

When performed for Budd-Chiari syndrome (BCS), transjugular intrahepatic portosystemic shunt (TIPS) creation can be technically difficult due to hepatic congestion and asymmetric hypertrophy. We present three female patients with decompensated BCS in whom TIPS were created using a three-dimensional fluoroscopy guidance system. On a dedicated workstation using three-dimensional volumes of computed tomography imaging, a virtual needle path was created by the operator extending from the needle entry point (hepatic vein stump or inferior vena cava) to the target portal vein. Subsequently, the virtual needle path was overlaid on the fluoroscopy image for guidance of portal venous cannulation. This technology can be used for TIPS procedures in patients with BCS and other complex TIPS cases, as it may help delimit the trajectory of the needle pass and optimally result in more efficient procedures with decreased radiation dose.

The transjugular intrahepatic portosystemic shunt (TIPS) procedure—which involves creation of a shunt through the hepatic parenchyma connecting the hepatic vein to the portal vein—is one of the most complex procedures performed by the interventional radiologist. The procedure has an increasing role in the treatment of Budd-Chiari syndrome (BCS) and is generally more technically difficult in BCS patients than in those with cirrhosis. The most challenging step of TIPS creation is accessing the portal vein. Many techniques and technologies have been developed to improve localization and access to the portal vein (1). Here, we discuss the use of a three-dimensional (3D) fluoroscopy guidance system to achieve cannulation of the portal vein in BCS patients. With multiple vendors offering cone-beam computed tomography (CT) as part of flat-panel angiography systems, it has become possible to create volumetric reformat imaging in the interventional suite (2). Recently developed 3D fluoroscopy guidance systems use the cone-beam CT volumetric data to create virtual needle paths, allowing the operator to track the needle in real time. The guidance systems are typically applied to percutaneous procedures (2). To the authors' knowledge, use of this technology for TIPS creation has not been reported to date. We present three cases of female patients with decompensated BCS in which TIPS were created with the application of a 3D fluoroscopy guidance system (XperCT and XperGuide; Philips Healthcare, Andover, MA).

CASE SERIES

In each of the three cases, the TIPS procedure was performed under general anesthesia. Table 1 summarizes patient and procedure characteristics. Prior to each procedure, contrast-enhanced CT or magnetic resonance (MR) axial images (Figure 1) were transferred to a commercially available dedicated workstation (XtraVision, Philips Healthcare, Andover, MA). The patient was prepped and draped in standard fashion. Subsequently, a noncontrast cone-beam CT dataset was acquired during breath hold via a flat-panel detector C-arm system (Allura Xper FD20 and XperCT, Philips Healthcare, Andover, MA). During a 180° to 240° rotation of the C-arm, 312 projections were acquired. The rotation time was approximately 10 seconds for the low-dose abdomen protocol. Contrast resolution of the cone-beam CT was approximately 5 HU at a slice thickness of 10 mm. The highest spatial resolution was 0.4 mm, depending on field of view and resolution matrix (3).

Table 1.

Pertinent demographic and procedure characteristics in each of the three women

	Patient 1	Patient 2	Patient 3*
Age (years)	26	34	26
MELD score	39	34	18
Anesthesia	General	General	General
Needle passes to cannulate portal vein	2	2	7
TIPS entry point	Transcaval	Hepatic vein stump	Transcaval
Fluoroscopy time (min)	12	16	52
TIPS stent length (cm)	8	8	7
Post-TIPS portal venous pressure (mm Hg)	22	20	16
Post-TIPS right atrial pressure (mm Hg)	16	17	12
Post-TIPS portosystemic gradient (mm Hg)	6	3	4

MELD indicates Model for End-Stage Liver Disease; TIPS, transjugular intrahepatic portosystemic shunt.

^*See Case Series section for procedure details.

Figure 1.

Coronal contrast-enhanced CT reformat image from Case 3 loaded onto the 3D fluoroscopy guidance system workstation. There is marked hypertrophy of the caudate lobe of the liver, heterogeneous peripheral liver enhancement, and large-volume ascites, which are features characteristic of Budd-Chiari syndrome.

Using specialized software on the dedicated workstation, the preprocedure contrast-enhanced CT or MR dataset was coregistered with the cone-beam CT soft-tissue dataset. 3D reconstructions were created (Figure 2). Within the 3D dataset, a virtual needle path was drawn by the operator from the puncture site (hepatic vein stump or intrahepatic inferior vena cava [IVC]) to the target portal vein (Figure 3).

Figure 2.

3D reconstruction image of coregistered contrast-enhanced CT and intraprocedural low-dose cone-beam CT images depicted on the 3D fluoroscopy guidance system workstation.

Figure 3.

Coregistered 3D reconstruction images demonstrate chosen (a) entry and (b) target points (arrows) in the intrahepatic inferior vena cava and portal vein, respectively. The virtual needle path is displayed between the points.

Following creation of the virtual needle path, the optimal imaging angulation was programmed instantly with automated coupling with the C-arm. The 3D dataset and virtual needle path were coregistered with the real-time fluoroscopy image, and the needle path was projected onto the fluoroscopy image, displaying a highly accurate real-time image of needle progression from the puncture site to the target portal vein (3).

Movement of the C-arm was coupled with the fluoroscopy projection of the 3D dataset, and the virtual needle path was visible in the appropriate projection. For typical percutaneous procedures, two main C-arm geometry positions are used in needle-based interventional procedures: the entry point view and the progression view. The entry point view looks down the barrel of the virtual needle path. The progression view is perpendicular to the virtual needle path. Both views can be accessed with a single button on the tableside controls (3). For the TIPS procedure, only the progression view was used, as the entry point view could not be utilized due to C-arm position limitations. With the C-arm in the progression view, the virtual needle path was shown with centimeter markings and with two colored rings at the ends marking the puncture site and the target portal vein. The colored rings had an adjustable diameter (1 cm default), providing a safety margin (3).

Following this, the TIPS procedure was started in usual fashion, and a Colapinto needle (Cook Medical, Bloomington, IN) was positioned in the hepatic vein stump or intrahepatic IVC. In the progression view, the Colapinto needle was advanced towards the portal vein along the virtual needle path, and a 21-gauge coaxial needle was advanced through the Colapinto needle to access the portal vein (Figure 4). The TIPS procedure was then completed in standard fashion.

Figure 4.

Fluoroscopic image with progression view 3D CT image overlay following TIPS stent placement. The slight mismatch of the TIPS stent and virtual needle path may be due to respiration.

In our third case, the virtual marking for the target portal vein was anterior and superior to the expected location. Due to concern for extrahepatic puncture, four or five needle passes were attempted with a 21-gauge coaxial needle through a Colapinto needle via the intrahepatic IVC inferior and posterior to the marked location without successful cannulation of the portal vein. To further delineate the position, the portal vein was marked with a percutaneously inserted wire, which confirmed the position marked by the guidance system. The coaxial needle was then advanced towards the portal vein with an aggressive anterosuperior needle pass with successful cannulation. The TIPS procedure was then completed in standard fashion.

DISCUSSION

Since 1993, when the first TIPS procedure was performed for BCS, TIPS has been shown to effectively preserve functioning hepatocytes by mitigating hepatic congestion and maintaining portal venous flow (4). Garcia-Pagán et al demonstrated excellent long-term outcomes in 124 patients with BCS treated with TIPS creation at six European centers (5). Rössle et al reported 1- and 5-year transplant-free survival rates of 93% and 74%, respectively (6).

As stated, TIPS creation is generally more technically difficult in BCS patients than in patients with cirrhosis. Hepatic congestion and asymmetric liver hypertrophy, especially of the caudate lobe, can increase the distance to the target portal vein (5). Thrombotic occlusion can limit access to the hepatic veins, requiring shunt creation via the hepatic stump or the intrahepatic IVC. These factors increase the complexity of TIPS creation and can result in longer procedure times, increased radiation dose, and elevated risk of complications. Several techniques have been developed to simplify shunt creation in these complex TIPS cases, including 1) direct intrahepatic portosystemic shunt; 2) the gun sight technique; and 3) “percutaneous” TIPS (1). We believe the technique described herein offers an additional effective approach to this challenging procedure.

Recently developed 3D fluoroscopy guidance systems have demonstrated utility in percutaneous needle interventions, including biopsy, ablation, aspiration and/or drainage, nephrostomy tube placement, and vertebroplasty (3). In our three cases, a 3D fluoroscopy guidance system was used to improve cannulation of the portal vein.

In our first two cases, the target portal vein was accessed using only two needle passes, after which the TIPS was created without difficulty. In the third case, due to the unexpected position of the portal vein marker, there was concern for extrahepatic rupture. Accordingly, conservative needle passes were attempted, but portal vein access was not obtained. Ultimately, percutaneous wire placement confirmed the location of the target portal vein as marked by the guidance system, the portal vein was cannulated, and the procedure was successfully completed.

TIPS creation in BCS patients is extremely complex. Blind efforts to access the portal vein can increase total fluoroscopy time and the rate of complications (7). Delimiting the needle trajectory with a 3D fluoroscopy guidance system may decrease complications and improve overall procedure success and efficiency. In the first two cases, the guidance system allowed us to access the portal vein with ease. The third case could have been performed more efficiently had we advanced along the virtual needle path to the marked target site.

One potential limitation to this technique concerns motion of the liver with breathing, which can result in poor fusion of the cone-beam CT and preprocedure contrast-enhanced CT. This can be minimized by performing both CT image acquisitions in the same phase of breathing. Also, due to the limited size of the cone-beam CT C-arm, this technique may be impossible in patients of larger abdominal circumference.

This technology can be applied to TIPS procedures in patients with BCS and other complex TIPS cases, as it may help delimit the trajectory of the needle pass and optimally result in more efficient portal vein cannulation and decreased radiation dose.