Introduction

Interventional radiology (IR) uses fluoroscopy to guide minimally invasive therapies. There is currently expanding investigation into the use of these techniques in various areas of veterinary medicine, including but not limited to the arterial and venous systems. The morbidity associated with certain open surgical procedures in these areas, particularly in compromised patients, makes these minimally invasive approaches increasingly appealing. Moreover, the lack of treatment alternatives available for more complex, terminal or end-stage diseases when traditional therapies have failed has inspired research into potential uses for these techniques, many of which have become the standard of care in human medicine. This lecture will review some of the interventional procedures currently performed in the hepatic arterial and venous systems of veterinary patients and some of the lessons learned - or more importantly the new questions raised - using these techniques.

Hepatic Vascular Anomalies

The categorisation of liver vascular anomalies is often confusing, but the most recent classification suggests three separate categories of liver vascular disease:

 Congenital portosystemic shunts (intrahepatic portosystemic shunts (IHPSSs) and extrahepatic portosystemic shunts (EHPSSs)).

 Disorders associated with abnormal hepatic bloodflow or portal hypertension, currently termed primary hypoplasia of the portal vein (PVH).

 Disturbances in outflow.

The second category (PVH) remains the most confusing, and includes processes that may or may not result in portal hypertension. These are termed PVH with portal hypertension and PVH without portal hypertension. Examples of PVH with portal hypertension include non-cirrhotic portal hypertension (NCPH), and hepatoportal fibrosis / veno-occlusive disease. PVH without portal hypertension was previously termed microvascular dysplasia (MVD).

Intrahepatic (and Extrahepatic) Portosystemic Shunts

Single, EHPSSs are amenable to relatively uncomplicated surgical attenuation; however, surgical repair of IHPSSs are consistently more challenging. Numerous techniques have been described for IHPSS attenuation, but morbidity and mortality rates can be very high, even for the most experienced surgeons. The goal of IR techniques for IHPSSs is to reduce the unacceptably high, perioperative mortality rates associated with traditional open surgical techniques and hopefully improve the outcome for these cases. The author has performed over 100 percutaneous transvenous coil embolisations (PTCE) with a vena caval stent and thrombogenic coils placed within the shunt. Perioperative complications were mostly minor, and perioperative mortalities were comparatively low versus those reported for traditional surgery.

Procedure: Percutaneous Transvenous Coil Embolisation

All dogs are treated medically initially following diagnosis of the IHPSS for a period of weeks to months. When possible, computed tomography (CT) or magnetic resonance (MR) angiography is performed to delineate the shunt anatomy and obtain caval and shunt measurements under a separate anaesthetic episode. All PTCE procedures are performed under general anaesthesia using standard liver dysfunction protocols and often neuromuscular blockade to minimise respiratory artefact during digital subtraction angiography. Shunts are typically accessed by a percutaneous right jugular approach. Contrast venography is performed to delineate the portal vein, portosystemic shunt and caudal vena cava anatomy. Intravascular pressure measurements are obtained in the caudal vena cava and portal vein.

Shunt attenuation involves placement of a stent within the caudal vena cava, positioned so as to traverse the shunt entrance into the vena cava. Shunt PTCE is subsequently performed by passage of a catheter through the stent interstices and deposition of the coils within the shunt lumen. Coils are subsequently added with intermittent shunt pressure measurements being taken to avoid creating portal hypertension. Coils are typically added until the shunt mouth is covered with coils, or the shunt pressures have increased between 6 and 10 cmH₂O, or maximal pressures approach 20 cmH₂O. Ultimate shunt and caval pressures are recorded, the jugular sheath is exchanged for a 7-Fr multilumen catheter, and the animal is recovered from anaesthesia. Following the initial procedure, medications are gradually weaned over the following 4–8 weeks. Additional PTCE is recommended if clinical signs return.

Lessons Learned

Lessons learned or questions raised concerning diagnostic imaging include:

 CT and MR angiography are well tolerated and facilitate identification of uncommon shunt anatomy; however, certain abnormalities are underestimated using these techniques. For instance, multiple small intrahepatic shunts are often not identified on standard cross-sectional imaging. These abnormalities are better identified with traditional contrast portography. It is not clear whether dual-phase CT provides us with additional important information at this time (as compared to single-phase).

 MR angiography (using gadolinium) may permit single procedure IHPSS imaging and treatment in order to avoid excessive iodinated contrast use associated with CT angiography followed by angiography.

 Even 'typical' IHPSSs have variant hepatic vein anatomy. Does location of the hepatic vein entrance into the PSS affect results of attenuation? It is conceivable that intrahepatic vein shunting (acquired intrahepatic PSS) may occur more readily in these patients. Examples will be discussed.

Lessons learned or questions raised concerning treatment include:

 A certain small population of IHPSS patients have 'significant' portal:systemic venous pressure gradients (or resting portal pressures) before treatment. This is counter-intuitive in animals with PSS in that reduced portal pressure gradients would be anticipated, and this has prevented treatment in some cases. Are there small vascular windows or narrowings present that are not identified on cross-sectional imaging with relatively wide slices? This suspicion has been raised as pull-out pressure tracings confirm short, focal areas where pressure gradients exist. The presence of a developed portal system may suggest a narrowing of the shunt in some location making access more difficult. This may be the same for EHPSS, suggesting intermittent shunt compression in phrenic shunts for instance!!

 Which is more important in preventing the development of complications associated with portal hypertension following IHPSS treatment, total portal pressure or pressure gradients? During surgery we rarely measured CVP and using IR techniques, we always measure CVP.

 During portography, when multiple small intrahepatic shunts are identified, this is almost exclusively associated with an elevated portal pressure and/or pressure gradient. Are these congenital shunts or acquired IHPSS resulting from a congenitally narrowed IHPSS?

 Acquired intrahepatic portosystemic shunts: although originally it was believed that only EHPSSs were acquired, there is more evidence that IHPSSs can be acquired as well. Do the same criteria for shunt attenuation (no greater than ~10 cmH₂O rise in portal pressure and/or no greater than ~20 cmH₂O total portal pressure) hold for attenuation of IHPSSs? Although there is no documented difference between hepatic vein attenuation and portal vein attenuation, the vascular bed receiving the congestion is intrahepatic with the former and extrahepatic with the latter. Does this matter?

 Do not perform IHPSS shunt attenuation in the face of gastrointestinal ulceration/ haemorrhage. The authors currently perform endoscopy and biopsy of all IHPSS cases before treatment and the overwhelming majority of these dogs have some degree of inflammatory bowel disease, sometimes including gastrointestinal ulceration. Approximately 17% of patients have evidence of gastrointestinal bleeding before treatment. Elevation of portal pressures with the presence of gastrointestinal ulceration can lead to severe gastrointestinal haemorrhage and death. All animals are maintained on omeprazole therapy for life. Initially a long-term mortality rate of 30% in IHPSS PTCE animals was caused by gastrointestinal bleeding in ~50% of deaths. Lifelong antacid therapy has reduced the mortality rate to 12.5% with fewer than 4% secondary to gastrointestinal bleeds. Is lifelong omeprazole therapy safe?

Lessons learned or questions raised concerning follow-up include:

 Is return to normal bile acid concentrations necessary? Should this be the goal of therapy? The majority of patients receiving IHPSS PTCE do not have return to normal liver function and some have even been identified to have no development of portal branching, although pressure gradients continue to exist. Is hepatic vein congestion and reduced portal bloodflow drainage consistent with improved portal bloodflow? Maintenance of a portal gradient appears to coincide with improved clinical signs. Can this be adequate or should additional interventional/surgical treatment be instituted?

 Development of 'longer-term' (greater than 2–4 weeks) post-procedural ascites is most commonly associated with hypoproteinaemia and not excessive portal hypertension (although they can occur simultaneously). In our population of cases, this has most commonly been associated with a gastrointestinal bleed and subsequent hypoalbuminaemia and hypoglobulinaemia and the resultant reduction in oncotic pressure. This has not always been confirmed but aggressive medical therapy with gastrointestinal protectants has resulted in resolution of the clinical signs in most of the few patients we have identified with this complication. No therapy for portal hypertension has been necessary.

 Approximately 18% (13/73) of cases required additional coiling procedures.

Hepatic Arteriovenous Malformations

Vascular malformations have been classified as high-flow or low-flow, and as arterial, venous, lymphatic or mixed. They are often associated with tumours but can also been seen congenitally in the liver (hepatic arteriovenous malformations (HAVMs)). Once called arteriovenous fistulas (AVFs), they have been more recently termed arteriovenous malformations due to their vascular anatomy upon angiography. Interventional radiology techniques have allowed us to improve our understanding of these particular AVMs in terms of their vascular nature, response to treatments (glue embolisation) and resulting pathophysiology. The observations below have been made following surgery or glue embolisation of approximately 15 HAVMs.

Lessons Learned

Lessons learned or questions raised concerning diagnostics include:

 These animals often present for clinical signs associated with PSS and are often mistaken for IHPSS due to the ultrasonographic identification of a large intrahepatic vascular structure along with clinical signs consistent with PSS.

 While all of these animals have severe portal hypertension, about 25% of these cases will not present with ascites. This is presumably due to the multiple extrahepatic PSSs that are acquired (along with other physiological changes taking place) and have decompressed the portal system sufficiently to relieve the excessive hydrostatic pressure.

 Clinical signs are often less severe with HAVM than standard PSS presumably because the portal system is more developed in the former patients than in the latter ones.

 Other diagnostic signs that are less commonly discussed include: hepatofugal portal blood flow (always present), a reduction in aortic diameter caudal to the level of the coeliac artery, and differing systemic blood pressures obtained from the forelimbs (higher) and the hindlimbs (lower) (occasionally present). Identifiable abdominal bruit is rarely present in the author's experience.

Lessons learned or questions raised concerning treatment include:

 It has been suggested in the human literature that while AVFs can be ligated or coil embolised, multiple AVFs or AVMs should receive glue embolisation (or alcohol ablation) in order to destroy the nidus that will otherwise recruit additional vessels over time. This has not been confirmed in the veterinary population but angiograms shown may support this notion.

 Initial angiography (and cross-sectional imaging) often underestimates the extent of the disease. Following initial embolisation, additional previously unidentified contributing vessels open up demonstrating the true infiltrative nature of these vascular anomalies.

 Complete HAVM embolisation or resection appears to be required to prevent recurrence. Incomplete embolisation or resection will lead to revascularisation if the HAVM nidus remains.

 These patients can tolerate complete hepatic artery embolisation and the cyanoacrylate glue appears to be permanent in the cases that have been followed to date, although the radio-opacity of the glue (Tantalum and Lipiodol) may diminish over time.

 Although return to hepatopetal portal bloodflow (towards the liver) would be considered the goal, this has not been seen in any of the cases treated to date. Acquired EHPSSs are present in all patients and can be expected to remain in place providing the least resistance to portal bloodflow. Complete HAVM embolisation could conceivably result in stagnant portal bloodflow as the direction changes from hepatofugal to hepatopetal, and near complete stasis has been identified but to date has not required intervention.

 Vascular contributions to the HAVM are not only from the hepatic artery but have also been identified to arise from the gastroduodenal artery, left gastric artery and phrenic arteries. Performing an aortogram following embolisation is recommended.

Lessons learned or questions raised concerning follow-up include:

 As acquired EHPSSs will never close, continued lifelong medical therapy is often necessary (and indicated) in patients following treatment. Some animals may not need medical therapy, however.

 It is unclear which animals may benefit from treatment. In certain cases with no overt clinical signs (for instance the cases without failure to thrive and massive ascites), these patients may not benefit from the treatment of the anomalies. Do all animals with HAVM require treatment? It has been suggested that chronic portal hypertension can lead to portal vein blunting and reduce portal perfusion. If so, is this an argument that all these animals should be treated as soon as possible?

References

References are available upon request.