Epidemiology, pathogenesis, diagnosis, surveillance, and management of hepatocellular carcinoma associated with vascular liver disease

Abstract

Vascular liver disease (VLD) presents special challenges in the diagnosis, surveillance, and treatment of hepatocellular carcinoma (HCC). HCC arising in the setting of vascular liver disease is often thought to be due to elevated hepatic arterial blood flow, rather than progressive fibrosis from chronic inflammation as with other chronic liver conditions such as viral hepatitis, autoimmune, and metabolic liver diseases. Vascular alteration inherent in VLD often impedes HCC non‐invasive diagnosis and loco‐regional treatment that depend on vascular properties found in typical liver environment. Benign and pre‐malignant liver nodules such as focal nodular hyperplasia and hepatocellular adenoma are also more common in certain VLDs, further adding to surveillance and diagnostic challenges. In this synopsis, we aimed to review available literature on the epidemiology, surveillance, diagnosis, and management of HCC in patients with VLD and specifically Budd–Chiari syndrome, congenital porto‐systemic shunts, Fontan‐associated liver disease, hereditary hemorrhagic telangiectasia.

Abbreviations

AFP

alfa‐fetoprotein

BCS

Budd–Chiari syndrome

CT

computed tomography

CPSS

congenital porto‐systemic shunts

FALD

Fontan‐associated liver disease

FNH

focal nodular hyperplasia

HCC

hepatocellular carcinoma

HCA

hepatocellular adenoma

HBV

hepatitis B virus

HCV

hepatitis C virus

IVC

inferior vena cava

MRI

magnetic resonance imaging

HVOTO

hepatic venous outflow tract obstruction

HHT

hereditary hemorrhagic telangiectasia

VLD

vascular liver disease

1. INTRODUCTION

Vascular liver disease results from hepatic circulatory abnormality due to vascular obstructions, dilatations, or fistulization. This may be manifestations of the liver disease or lead to liver related complications. The pathophysiology is due to an imbalance between hepatic arterial and portal venous blood flow and ultimately lead to increased hepatic arterial inflow, which mediates the development of dysplasia and hepatocellular carcinoma (HCC). ¹ , ² , ³

HCC is now the fifth most common cancer in the world and the third cause of cancer‐related mortality as estimated by the World Health Organization Global cancer Observatory. ⁴ HCC has several risk factors with the main culprits being chronic viral infection and excessive alcohol intake. These induce chronic inflammation and subsequently fibrosis to cirrhosis development and genetic changes, finally leading to carcinogenesis. ⁵

HCC development in vascular liver disease is unique, lacking the usual inflammation‐fibrosis‐cirrhosis cascade. These lesions also pose multiple layers of challenge from diagnosis to treatment and follow up. The current LI RADS system and its built‐in criteria of arterial phase hyper enhancement and portal venous washout may be absent in HCC lesions that occur in the background of vascular liver disease; therefore, non‐invasive diagnosis relying on radiological features may not be applicable in this setting. Meanwhile, tissue based diagnosis of HCC in patients with vascular liver disease are also problematic due to the higher risk of bleeding with vascular liver pathology as compared to nonvascular liver diseases.

Therefore, we aimed to review the epidemiology, pathogenesis, and diagnosis of HCC associated with some of the more common vascular liver diseases such as Budd–Chiari Syndrome, congenital porto systemic shunt, hereditary hemorrhagic telangiectasia (HHT) and Fontan associated liver disease (FALD). Lastly, we will discuss optimal surveillance strategy and treatments of HCC in these clinical contexts based on best available evidence.

1.1. Proposed carcinogenesis model in vascular liver disease

The normal liver receives about 70–80% of its total blood supply from the portal vein and 20–30% from the hepatic artery. These two systems are both anatomically and functionally connected, and it is well recognized a decrease in portal blood flow is compensated by an increase in arterial flow. ¹ In vascular liver disease, there is an imbalance between these two systems that often results in increased hepatic arterial flow with associated level of circulating hepatic growth factors probably plays a major role in the development of HCC. ² , ³

Experiments in rat models shows that hyperplastic lesions develop in liver following vascular disruption such as surgical porto systemic shunts. ² Hyperplastic hepatic lesions following portacaval shunting surgery in childhood are reported in case reports. ⁶ In the largest case series in pediatric patient population, post porto‐systemic shunt surgery for extra hepatic portal vein obstruction, seven out of 45 children (15%) developed liver nodules after a median of 80 months. ⁷

1.2. Budd–Chiari syndrome

Budd–Chiari syndrome (BCS) is an accepted synonym for hepatic venous outflow tract obstruction (HVOTO). This includes all the consequences of an impediment to blood flow in the small, medium, or large hepatic veins, or in the suprahepatic portion of the inferior vena cava (IVC). ⁸ , ⁹ , ¹⁰ Recent meta‐analyses showed that the pooled annual incidence of BCS was 1 per million persons (95% CI = 0.225–3 per million) and the pooled prevalence of BCS was 11 per million (95% CI = 4–21 per million).

Benign regenerative nodules, mostly focal nodular hyperplasia (FNH), are common in BCS. ¹¹ In a large imaging series of BCS, liver nodules were observed in 28 out of 77 patients (36%). ¹² A multicenter radio ‐pathological series shows that 4 out of 27 FNHs, as initially reported, were not confirmed as FNH. In fact, 2 out of 4 of these lesions were hepatocellular adenomas (HCAs) (one inflammatory and one exon three beta‐catenin mutated). ¹³

One study with good sample size shows cumulative incidence of HCC in BCS was 4% over a mean follow‐up of 5 years (1–20 years), with 11 out of 97 patients in this series having biopsy proven HCC though a total of 43 patients had liver nodules. Biopsy of nodules was obtained when one of the following criteria was met: number > or = 3, diameter > or = 3 cm, heterogeneity, washout on portal venous phase, increase in size on surveillance, or increase in AFP level. Notably, AFP at cut off level of 15 was shown to have PPV of 100% and NPV 91%. ¹⁴ Another large published series with 348 Egyptians patients and other studies have confirmed the specificity of higher AFP for HCC in these cases. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Therefore, elevated AFP even at modest level may be useful criteria for further investigation, though there is no consensus about ideal cut off value.

Another factor associated with HCC diagnosis was male sex, shown to be an independent predictor for HCC in two large studies. ¹⁴ , ¹⁵ In addition, IVC obstruction was a major predictor for HCC development compared to pure hepatic vein involvement (81.8% vs 4.6%, p < 0.001) ¹⁴ , ¹⁸ , ¹⁹ , ²⁰ , as well as factor V Leiden (54.5% vs 17.5%, p = 0.01). ¹⁴

In regards to diagnosis, differentiating between HCC and benign regenerative nodules is challenging as both types share hyper‐enhancement at the arterial phase of vascular contrast injection. ¹⁴ While enhancement of most FNH lesions was center‐to‐periphery and persistently hyperechoic on portal venous and delayed phases, most HCCs were heterogeneous on arterial phase and hypoechoic on portal and delayed phases. ¹⁶ Nevertheless, the diagnosis between FNH‐like lesions and HCC remains difficult at imaging and hepatobiliary MR contrast agents may help but have not been extensively evaluated. ²¹ Surveillance in case of typical features of FNH and low AFP levels is conventional biannual tumor marker and imaging assessment (this is not for a female with normal background liver but only in the context of BCS or vascular malformation). However, we propose that lesion biopsy should be performed in presence of atypical imaging features or interval increase in size or rising serum AFP. An AFP cut off value of 15 is reasonable, though no consensus exists among studies. There is no existing evidence that benign regenerative nodules become malignant, but the number of studies is limited. ²²

Creation of porto‐systemic shunt surgically or radiologically also after the liver vascular blood supply. Role of TIPSS in the development of these lesions is unclear ²³ ; however, it has been reported that the presence of TIPSS may lead to accelerated time to HCC growth. In this setting, tumors tend to be larger and in the periphery of the right lobe. Therefore, it is proposed that HCC development is dependent on TIPSS placement, whether obstructing venous outflow or enabling decompression. No recommendation for enhanced surveillance is available on current literature. In addition, response of HCC to loco‐regional therapy is difficult to ascertain in the context of TIPSS in situ. ²⁴ , ²⁵ Preliminary data suggest a good tolerance and a good response to hepatic arterial chemoembolization and prolonged survival in the absence of transplantation. ¹⁷ , ²⁰ , ²⁶ The Milan criteria could be irrelevant in the selection of BCS patients with HCC for liver transplantation, as the natural history and determinants of recurrence may differ from those of cirrhosis.

1.3. Congenital porto‐systemic shunts

Congenital porto‐systemic shunts are rare developmental anomalies resulting in diversion of portal flow to the systemic circulation. The overall incidence of CPSS is estimated to be 1:30,000 births and 1:50,000 for those that persist beyond early life. ²⁷ The prevalence of intrahepatic shunts is estimated to be 0.0235% as reported from a random ultrasonography population screening of asymptomatic adults. ²⁴

These malformations occur during vascular embryogenesis. They carry a higher risk of both benign and malignant liver tumors. The features of shunts vary since they can be intra‐ or extrahepatic, single or multiple, and they may induce partial or complete portal blood deprivation. ²⁷ , ²⁸

Benign liver tumors in this case seem related to major portal flow deprivation. ²⁹ Associated elevated circulating levels of hepatic growth factors are seen in 25–50% of CPSS cases. Pediatric studies have shown that shunt closure by surgery can restore intrahepatic portal supply, followed by regression of benign liver lesions. ³⁰ In this study, 13/17 (76%) patients with porto systemic shunts had liver tumors. Benign tumors such as hepatocellular adenoma (HCA) or FNH in these cases show a different pathogenesis and natural history from conventional tumors and seem to carry a higher risk of malignant transformation.

A systematic review based on 250 pediatric case reports identified 64 cases of liver tumors at a mean age of 8 years, mostly consisting of FNH or large regenerative nodules, seven HCAs, and seven HCC tumors. ³¹ In another series of five patients undergoing liver transplantation, 10 out of (63%) nodules were FNH‐like lesions, five were beta‐catenin‐mutated HCAs, and one was an inflammatory HCA. ³²

Malignant tumors (HCC, hepatoblastoma, and sarcoma) have been reported to occur in these patients in the absence of liver dysfunction and/or cirrhosis. ³³ However, malignant transformation of pre‐existing benign liver lesions seems to be more common. Cases in which hepatoblastoma occur after transformation of pre‐existing FNA have been reported. ³³ A study with slightly older cohort of patients also reports HCC arising from pre‐existing HCA. This is supported by the evidence of specific mutations (beta‐catenin mutations resulting in activation of various transcription factors) discovered in hepatocytes possibly in association with the altered microvascular environment. ¹³ , ³² , ³⁴ , ³⁵

For these reasons, indications for treatment differ and close surveillance of those benign nodules is recommended. In most cases, AFP level is often elevated and the shunt appears to work as an independent risk factor in the absence of underlying liver disease. ³⁶ , ³⁷ Diagnosis is also challenging in this setting. While MRI is key to the characterization of hepatic nodules, CPSS‐associated liver nodules are particularly heterogeneous before and after vascular enhancement. ²⁷ FNH‐like lesions are more often atypical on imaging and may enlarge over time. ²⁹ , ³⁸ Liver biopsy in the setting of an indeterminate or increasing size liver nodule may be necessary.

Liver resection or transplantation has been performed where HCC or hepatoblastoma are confirmed. ³⁹ Such tumors may develop at any age even without prior shunt symptomatology.

1.4. Fontan‐associated liver disease

The Fontan procedure is a complex cardiac‐surgical procedure performed in patients with uni‐ventricle cyanotic congenital heart disease. Surgical separation of the pulmonary circulation from the systemic circulation leads to elevated systemic venous pressure and mainly manifests as hepatic congestion, a condition termed Fontan‐associated liver disease (FALD). Lemmer et al. first described hepatic fibrosis with FALD in 1983. ⁴⁰ Fibrosis and later cirrhosis develop because of chronic venous congestion and tissue hypoxia post Fontan. ⁴¹

Cirrhosis, generally a prerequisite for HCC, develops in persons under the age of 25 years (11 to 15 years after a Fontan procedure). The estimated incidence of HCC is 1.5 to 5.0% per year. In a case series from King's college hospital from the UK, ⁴² authors noted lack of cirrhosis on liver biopsy of all three patients. Despite yearly ultrasound, all were diagnosed with HCC beyond the Milan criteria. Authors concluded that cirrhosis may not be an absolute requirement for HCC development in patients with FALD and additional unrecognized factors beyond abnormal liver flow are probably involved in carcinogenesis.

This group presents a unique scenario. The American College of Cardiology recommends one to three yearly liver function tests and liver ultrasounds to identify FALD in adults. ⁴³ However, non‐invasive assessment of fibrosis has not been validated in these settings. Typical arterial enhancing pattern of HCC on CT may not be present due to complex liver hemodynamic post Fontan. ⁴⁴

Surveillance for HCC should start 10 years after Fontan, when significant liver damage may have set in. ⁴¹ , ⁴⁵ Baseline liver MRI may help to identify high‐risk potentially malignant lesions and should be followed with biannual ultrasounds and serum AFP tests, regardless of fibrosis stage. MRI should be used for follow‐up of patients with high‐risk lesions.

Local–regional therapy may be limited by a variety of reasons including the presence of cardiac pacemakers, extra hepatic shunts and abnormal vasculature. Finally, liver transplantation may be relatively contraindicated by cardiac function, unless dual liver and heart transplantation is an option.

1.5. Hereditary hemorrhagic telangiectasia

Hereditary hemorrhagic telangiectasia (HHT) is an autosomal dominant disorder characterized by mucocutaneous and visceral telangiectasia. ⁴⁶ The average worldwide prevalence is estimated to be 1:5000–8000 persons. ⁴⁷ Due to the founder effect, this index is higher in Jura in France, Funen in Denmark, and the Netherlands Antilles.

Liver involvement is observed in 67–84% of HHT patients. ⁴⁸ Microscopic telangiectasia and direct arteriovenous, arterioportal, and porto‐venous shunts exist. These anomalies in vascular supply lead to increased hepatic arterial flow and favors development of FNH‐like lesions. ⁴⁹ The diagnosis of FNH‐like lesions in HHT patients is based on traditional imaging. Pre‐contrast and contrast‐enhanced MRI with hepatobiliary contrast agents may differentiate FNH‐like lesions from large telangiectasia. Liver biopsy is normally contraindicated in HHT because of a higher risk of bleeding. However, liver biopsy is often necessary for diagnostic purposes, and there have been reported cases with no major complications. In a Belgian HHT case series, 6 out of 30 patients had biopsy without any complication. ⁴⁹

An ultrasound‐based screening study of 274 participants from HHT families with cross sectional imaging confirmation found FNH in 2.9%. ⁵⁰ In this study, most were women and most patients with FNH had large hepatic shunts. Hepatocellular tumors (HCA and HCC) seem to be extremely uncommon in HHT. Sempoux et al did not observe any HCA or HCC cases in their pathological study, ¹³ though rare cases of HCC associated with HHT have been reported. ⁵¹ To the best of our knowledge, only two case reports have been published since 1975, and there have been no reports of HCC in the documented absence of viral hepatitis B or C.

2. CONCLUSION

Vascular liver disorders are associated with hepatocellular tumors both benign and malignant, which arise in the setting of increased hepatic arterial flow. FNH are the most common but other benign tumors such as HCA also can occur in this clinical context. HCC is rare in HHT but more common in patients with BCS, CPSS and FALD (Table 1).

TABLE 1.

Vascular liver disease and association with HCC

Vascular liver disease	HCC association (yes/no)
Budd‐Chiari Syndrome (BCS)	Yes
Hereditary Hemorrhagic Telangiectasia (HHT)	Rare if no other liver disease
Congenital Porto‐systemic Shunts (CPSS)	Yes
Fontan‐Associated Liver Disease (FALD)	Yes
Idiopathic Noncirrhotic Intrahepatic Portal Hypertension	No
Nonobstructive Sinusoidal Dilatation and Peliosis	No

Optimal surveillance strategy overall is not well defined. Conventional imaging criteria are often absent due to abnormal hemodynamic in vascular liver diseases. MRI with hepatobiliary contrast plays a major role in establishing diagnosis though with greater difficulty than those in livers without a vascular disease. Tissue diagnosis is often associated with more bleeding risks but more often needed than with other liver diseases. Loco‐regional therapy may be helpful in HCC but may have more risk of ischemia in vascular liver disease. Liver transplantation outcomes for HCC associated with vascular liver diseases are generally favorable (Table 2).

TABLE 2.

Risk factors, diagnosis, and surveillance for patients with vascular liver diseases

Vascular liver disease	Risk factor for HCC	Diagnosis	Surveillance
Budd‐Chiari Syndrome (BCS)	Male sex Inferior vena cava obstruction Factor V Leiden mutation	MRI with contrast better CT ‐ heterogeneous arterial phase plus hypo ‐echoic on portal and delayed phases Higher AFP is specific	Biannual AFP and imaging Biopsy if increasing size or AFP > 15 ng/dL
Fontan‐Associated Liver Disease (FALD)	11–15 years post Fontan	MRI better	Baseline MRI at ten year, then biannual AFP and ultrasound
Hereditary Hemorrhagic Telangiectasia (HHT)	HCC or HCA rare without coexisting HBV or HCV	Ultrasound mostly helpful MRI for atypical or large lesions only Cautious Liver Biopsy	No specific follow up for confirmed FNH
Congenital Porto systemic Shunts (CPSS)	Presence of shunt Elevated AFP	MRI with contrast better Biopsy if increasing size or indeterminate lesion	Biannual AFP and imaging

CONFLICT OF INTEREST

All authors declare no conflict of interest.

Prasad D, Nguyen MH. Epidemiology, pathogenesis, diagnosis, surveillance, and management of hepatocellular carcinoma associated with vascular liver disease. Kaohsiung J Med Sci. 2021;37:355–360. 10.1002/kjm2.12368