Successful embolization of a large intrahepatic portosystemic shunt using the coil-in-plug technique: A case report assessed with four-dimensional flow magnetic resonance imaging

Abstract

Congenital intrahepatic portosystemic shunts are uncommon vascular anomalies that can lead to hyperammonemia and liver dysfunction. We report a 48-year-old woman with a large intrahepatic shunt presenting with elevated blood ammonia and progressive hepatic atrophy. She underwent percutaneous coil-in-plug embolization using an Amplatzer Vascular Plug II filled with coils. Four-dimensional Flow MRI performed before the procedure and at 2 days and 6 months afterward demonstrated complete shunt occlusion, restoration of normal portal flow, and increased liver volume. Her ammonia level normalized within 1 day, and the patient remained asymptomatic. This case highlights the safety and efficacy of the coil-in-plug technique for large intrahepatic shunts and underscores the utility of four-dimensional Flow MRI for treatment planning and follow-up.

Background

Congenital portosystemic shunts, resulting from aberrant development of the portal vein (PV), eg, residual embryonic vessels, pose significant clinical challenges, such as hepatic encephalopathy [1]. Historically, treatment options have included surgical ligation and endovascular embolization [2]. In endovascular therapy, the migration and potential recanalization of embolic materials are concerns. Additionally, portosystemic shunts involving large-caliber veins often require numerous coils, escalating costs. Specifically for intrahepatic portosystemic shunts, preservation of adjacent normal vessels is desirable, necessitating embolization over a short segment.

The Amplatzer Vascular Plug (AVP; Abbott, Abbott Park, IL, USA) is favored for its self-expanding properties, which allow it to anchor securely at the target site, reducing the risk of migration and enabling effective embolization in short segments [3]. The coil-in-plug (CIP) method further enhances this approach by filling the AVP’s lobe with coils, effectively using the plug to contain the embolic material [4,5]. Specifically, the preloading CIP (p-CIP) technique involves predeploying the AVP II outside the body (Fig. 1). Initially, a microcatheter is inserted through the lobe of the AVP II. Then, while maintaining this configuration, the AVP II is retracted back into its loader. It is subsequently delivered to the target site via a guiding sheath and deployed. Afterward, the tip of the microcatheter is withdrawn so that it remains within the AVP’s lobe, which is then filled with coils. This technique avoids the challenging task of navigating a catheter through the AVP’s lobe after the device is deployed in the body, making the procedure much simpler than traditional methods.

Fig. 1.

Illustration of the process for loading the microcatheter into the Amplatzer Vascular Plug II (AVP II) outside the body. (A) Initially, the AVP II is penetrated using an 18-gauge puncture needle. (B) The microcatheter (Coiling Support EX) and microguidewire (Jupiter) are inserted through the needle’s sheath. (C and D) The microcatheter, microguidewire, and AVP II are subsequently retracted into the loader.

Noninvasive imaging is also essential for evaluating treatment effectiveness. Three-dimensional (3D) cine phase-contrast MRI, commonly known as 4D Flow MRI, can acquire time-resolved 3D blood flow velocity data throughout the cardiac cycle [6,7]. This method has been reported to be especially useful for determining treatment indications and assessing therapeutic efficacy, including in the portal venous system [[8], [9], [10]].

In the present report, we describe a case in which the CIP technique was applied to occlude a large intrahepatic portosystemic shunt in a patient with severe hyperammonemia, achieving favorable outcomes. The results were documented both visually and quantitatively using 4D Flow MRI.

Case presentation

A 48-year-old female with a large intrahepatic portosystemic shunt diagnosed over a decade ago had been managed conservatively. She presented with progressive liver atrophy, multiple hyperplastic liver nodules due to abnormal vascular flow, and elevated blood ammonia levels. These developments led to her referral for shunt embolization. Her past medical history included hereditary breast-ovarian cancer syndrome, prior salpingo-oophorectomy, and treatment for left breast cancer.

Computed tomography identified a large direct shunt measuring 18 mm in diameter between the posterior branch of the PV and the right hepatic vein (HV), accompanied by dilation of the right PV and HV. In contrast, the left and anterior PV branches were narrowed due to the steal of blood flow by the shunt. Two lesions in liver segments 6 and 8 were biopsied and confirmed to be benign hyperplastic nodules. Laboratory tests revealed a high serum ammonia level of 121 μg/dL.

Preinterventional contrast-enhanced 4D Flow MRI (see Table 1 for imaging parameters) confirmed a markedly dilated shunt from the right PV to the right HV, with an elevated flow rate of 519 mL/min through the shunt (Fig. 3A). Conversely, flow in the left PV was very low, and this branch was not visualized on the 4D Flow images.

Table 1.

Four-dimensional Flow MRI parameters.

Magnetic field strength (Tesla)	3.0
Repetition time (msec)	6.5
Echo time (msec)	3.37
Flip angle (°)	8
Number of excitations	1
Field of view (mm)	262.5 × 300
Voxel size (mm)	1.56 × 1.56 × 1.5
Temporal resolution (ms)	68.76
Parallel imaging factor	3
Time frames per R-R intervals	15
Approximate imaging time (min)	14
Velocity encoding (cm/sec)	60

Fig. 3.

Noninvasive blood flow assessment with four-dimensional flow magnetic resonance imaging (4D Flow MRI). (A) Preoperative 4D Flow MRI reveals that almost all the blood flow in the main portal vein (MPV) was diverted through the intrahepatic portosystemic shunt to the right hepatic vein; the dashed circle indicates the constricted left portal vein (LPV). (B) Two days postinterventionally, 4D Flow MRI confirms the shunt was effectively embolized, and blood flow in the LPV had increased. (C) Three months postinterventionally, 4D Flow MRI shows the MPV had remodeled with a slight reduction in diameter, while the LPV had further expanded. (D) Six months postinterventionally, 4D Flow MRI showed the shunt remained occluded, maintaining the same findings as at 3 months postinterventionally. AB, anterior branch of the portal vein; LPV, left portal vein; RPV, right portal vein.

An occlusion test was then performed via the right internal jugular vein using a balloon catheter (Selecon MP Catheter II, Terumo, Tokyo, Japan). Portal pressure measured 5 mmHg at baseline and rose to 20 mmHg upon temporary shunt occlusion. Despite the potential risk of portal hypertension, we proceeded with embolization since the liver parenchyma was otherwise healthy and the expected benefits outweighed the risks.

The shunt’s outflow drained directly into an enlarged right HV and the inferior vena cava, raising concern that embolic materials could migrate to the lungs. Additionally, we aimed to preserve normal portal and hepatic venous branches. Therefore, the p-CIP approach was selected to achieve a robust occlusion over a short segment of the shunt.

Using the p-CIP technique, the AVP II was prepared with a microcatheter (Coiling Support EX, HI-LEX, Takarazuka, Japan) and microguidewire (Jupiter, Boston Scientific, Marlborough, MA, USA) inserted through it outside the body, then retracted back into its loader (Fig. 1). The right internal jugular vein was accessed, and an 8-French guiding sheath (Destination, Terumo, Tokyo, Japan) was advanced into the shunt to deploy the AVP II (Fig. 2A). The microcatheter was partially withdrawn to position its tip in the second lobe of the AVP II (Fig. 2B), and 13 micro hydrogel coils (AZUR 18 and AZUR 3D soft, Terumo, Tokyo, Japan; Fig. 2C) were delivered into this lobe through the microcatheter (Fig. 2C). An angiogram of the superior mesenteric artery confirmed successful occlusion of the shunt (Fig. 2D), after which the AVP II was detached. Following a 10-minute observation period with no evidence of device migration, the procedure was concluded. Due to concern about thrombus formation in the PV, postoperative prophylactic anticoagulation therapy with edoxaban (Lixiana; Daiichi Sankyo, Tokyo, Japan) (30 mg/d) was initiated.

Fig. 2.

Sequence of endovascular therapy. (A) An 8Fr guiding sheath (Destination) is advanced from the right internal jugular vein to the shunt, where the Amplatzer Vascular Plug II (AVP II) is deployed. (B) The microcatheter (arrow) is retracted to the second lobe of the AVP II. (C) Thirteen micro hydrogel coils (AZUR 18 and AZUR 3D soft) are loaded into the lobe. (D) Superior mesenteric artery angiography confirms successful embolization. MPV, main portal vein; RHV, right hepatic vein; SMV, superior mesenteric vein.

By the next day, the patient’s serum ammonia had dropped to normal (61 μg/dL). A contrast-enhanced 4D Flow MRI (Fig. 3B) performed 2 days after the procedure confirmed complete shunt occlusion with no migration of embolic material. A thrombus was noted extending from the right PV into its posterior branch, but the diameter and flow in the remaining portal branches (such as the left PV) had increased (left PV branch flow 187 mL/min). The patient was discharged on postprocedure day 3 without any signs of portal hypertension.

Follow-up 4D Flow MRI at three and 6 months (Fig. 3C and D) demonstrated that the shunt remained occluded and its channel had decreased in caliber. The previously noted thrombus in the right PV had also reduced in size. The main PV showed mild remodeling with a slight reduction in caliber, while the left and anterior PV branches exhibited further increases in flow and diameter over time (at 3 months: left PV branch 298 mL/min, anterior branch 155 mL/min; at 6 months: left PV branch 264 mL/min, anterior branch 211 mL/min). The volume of the left hemiliver increased, and the hyperplastic nodules became less prominent on imaging. At the 1-year follow-up, the therapeutic effect was maintained, with findings unchanged from the 6-month follow-up.

Discussion

This case involved the embolization of a large intrahepatic portosystemic shunt between the right posterior PV branch and the right HV, which necessitated a focused occlusion over a short segment. A primary concern was the risk of embolic material migrating to the pulmonary circulation. The CIP technique effectively addressed these challenges, allowing a safe and successful outcome by minimizing the risk of migration. The therapeutic results were clearly confirmed both visually and quantitatively with 4D Flow MRI.

Previous reports have demonstrated the effectiveness of the CIP method in achieving short-segment embolization of large-caliber vessels. For example, Katada et al. and Koganemaru et al. applied CIP in high-flow arterial lesions with excellent results [4,5]. Kotoku et al. [11] also showed that using the CIP technique for internal iliac artery embolization required fewer coils than coil embolization alone, suggesting potential cost savings. Furthermore, in pulmonary arteriovenous malformations, embolization with an AVP alone resulted in a 5%-8% recanalization rate, whereas adding coils within the plug (the CIP approach) reduced the recanalization rate to 0% over 13 months of follow-up [12]. Consistent with these findings, our case illustrates that CIP can safely achieve short-segment occlusion even in a large venous shunt. The use of far fewer coils compared to a coil-only strategy also implies a cost benefit for this approach.

Although the CIP method involves a few more procedural steps than deploying an AVP alone, it does not significantly complicate the intervention. The coils are securely confined within the plug, and the robust support of the guiding sheath minimizes the risk of any catheter or device kickback during deployment. Given the need for localized embolization in intrahepatic portosystemic shunts and the relative ease of AVP delivery, the CIP method is particularly well-suited for such cases.

4D Flow MRI proved to be an invaluable adjunct in this case. This imaging modality enables extraction of detailed blood flow data from any vessel in the field of view and permits unlimited retrospective analysis of hemodynamics after the scan [[6], [7], [8]]. It thereby overcomes several limitations of Doppler ultrasound, such as a restricted field of view, operator dependency, and lack of retrospective evaluability [8]. Recent studies have highlighted the utility of 4D Flow MRI in guiding and assessing endovascular treatments, especially by confirming the normalization of portal blood flow following shunt embolization [[8], [9], [10]]. In our patient, 4D Flow MRI was crucial both before and after embolization, facilitating comprehensive quantitative assessment of the shunt and portal circulation. The excellent compatibility of 4D Flow MRI with interventional planning and follow-up underscores its value for the management of complex vascular shunts.

Conclusion

The CIP technique enabled the safe and effective embolization of a large intrahepatic portosystemic shunt in this case. This method allows dense short-segment occlusion while using far fewer coils than a coil-only approach, suggesting a more cost-effective strategy for treating portal venous shunts. Moreover, 4D Flow MRI provided a comprehensive, quantitative evaluation of the hemodynamic changes, greatly aiding in treatment planning and postprocedural assessment.

Patient consent

Written informed consent was obtained from the patient for the publication of this case report.