Introduction
Pancreatic pseudocyst-portal vein fistulization is a rarely described phenomenon within
the literature [1]
[2]
[3]
[4]
[5]
[6]. We present two cases of patients with pancreatic pseudocysts fistulizing into the
portal venous system and seeding to the liver with subsequent hepatic pseudocyst formation.
Complete and long-term resolution of the liver pseudocysts failed with more conservative
measures, including parenteral antibiotics and percutaneous drainage, and was achieved
only after endoscopic ultrasonography (EUS)-guided pancreatic cyst-gastrostomy and
metal stent insertion. We have reviewed the current literature on diagnosis and management
of hepatic pseudocysts and pseudocyst-portal vein fistulization.
Case reports
Patient 1
A 64-year-old alcoholic man presenting with abdominal pain, weight loss and anorexia,
underwent abdominal computed tomography (CT), which revealed a 3.6 × 2 cm pancreatic
cyst, portal vein thrombosis and a multiloculated, septated and hypoattenuating liver
lesion within segment eight, measuring 6.7 × 5.6 cm. Linear EUS confirmed diffuse
calcification throughout the pancreas, but no pancreatic duct dilatation, and a 4.5-cm
pancreatic head pseudocyst. Percutaneous fine-needle aspiration (FNA) of the liver
cyst yielded serosanguinous fluid. The patient was treated with antibiotics and percutaneous
drainage of presumed liver abscess and discharged, but he presented 3 months later
with abdominal pain. At that time, abdominal CT showed enlargement of the liver cyst,
which measured 7.6 × 7.2 cm in diameter and a significant increase in size of the
pancreatic pseudocyst, now 8.2 × 4.8 cm and communicating with the pancreatic duct.
Interestingly, hyperdense material was seen extending from the pseudocyst into the
portal vein ([Fig. 1] and [Fig. 2]). Further FNA of the liver cyst demonstrated fluid rich in amylase and lipase (1950 U/L
and 3389 U/L, respectively), suggesting a pancreatic origin. Despite CT-guided percutaneous
drainage of the pancreatic pseudocyst, follow-up imaging several weeks later revealed
an increase in its size. Endoscopic retrograde pancreatography (ERP) demonstrated
an abrupt cut-off approximately 2 cm into the pancreatic duct, but with no evidence
of communication with a cystic cavity. Given the documented increase in size of the
pseudocyst, another linear EUS was performed, confirming a 5 × 5.5 cm hypoechoic lesion
adjacent to the posterior stomach wall. Murky brown fluid was aspirated and sent for
culture and an EUS-guided cyst-gastrostomy was performed with placement of two 7-Fr
plastic double-pigtail stents within the pseudocyst. Repeat abdominal CT documented
a significant decrease in size of both the pancreatic pseudocyst and the intrahepatic
fluid collection. The patient was discharged home on oral antibiotics. On follow-up
CT 2 months later, however, the pancreatic pseudocyst had grown again to 7.4 cm in
diameter. The patient subsequently underwent another EUS-guided cyst-gastrostomy with
removal of the plastic stents and insertion of a 4-cm fully-covered self-expanding
metal stent (FCSEMS). A follow-up abdominal CT 2 months later showed almost complete
resolution of both the liver and pancreatic collections. The patient has not required
any further admissions to the hospital.
Fig. 1 Axial CT image of pancreatic pseudocyst (black arrow) with extension to portal venous
system (white arrow).
Fig. 2 Coronal image of liver pseudocyst (black arrow) and hyper-dense fluid in portal vein
(white arrow) and pancreatic pseudocyst.
Patient 2
In a 54-year-old man presenting with abdominal pain and weight loss, abdominal CT
identified similar features of chronic pancreatitis and an ill-defined, multiloculated,
hypoattenuating liver mass in segments 8 and 4A, measuring 6.8 × 8.8 cm. Extensive
occlusive thrombus was also seen involving the portal, superior mesenteric and splenic
veins. Percutaneous aspiration of the liver mass suggested an abscess, prompting treatment
with intravenous (IV) antibiotics and drainage. Three months later, the patient presented
with abdominal pain and ascites. Repeat abdominal CT showed evidence of complex fluid
within the portal venous system, suspicious for a pancreatic pseudocyst origin ([Fig. 3]). ERP demonstrated communication between a dilated pancreatic duct and an irregular
tubular-shaped structure, consistent with a pseudocyst. A 7-Fr plastic pancreatic
stent was inserted. Follow-up magnetic resonance cholangiopancreatography (MRCP) confirmed
complete distal migration of the pancreatic stent and a pancreatic ductal stricture
with decompression into a pseudocyst and evidence of secondary cavernous transformation
of the portal vein extending into the liver. Percutaneous aspiration of the large
liver cyst confirmed fluid rich in amylase (> 2000 U/L). The patient remained hospitalized
on IV antibiotics and a further attempt at ERP and pancreatic duct stenting was unsuccessful.
Because the man was not felt to be a suitable surgical candidate, we proceeded with
linear EUS-guided cyst-gastrostomy and deployment of a 4-cm FCSEMS. Significant bleeding
at the time of stent deployment led to urgent angiography, however, no active bleeding
point was identified, and the patient was managed conservatively. Abdominal CT 2 weeks
later showed successful decompression of the pancreatic pseudocyst and resolving hepatic
pseudocyst ([Fig. 4]).
Fig. 3 Coronal CT image shows hypodense fluid in pancreatic pseudocyst (black arrow) communicating
with the main portal vein (white arrow).
Fig. 4 Coronal image showing cyst-gastrostomy stent (white arrow) in place with drainage
of the pancreatic pseudocyst and portal venous system.
Discussion
Pancreatic pseudocysts can present various complications, some more common such as
infection, rupture, and gastrointestinal obstruction, to the slightly rarer. Hepatic
pseudocysts are a recognized but rare phenomenon described within published case reports
[7]
[8]
[9]
[10]
[11]
[12]. In one report of a patient presenting with a large liver cystic lesion in association
with a pancreatic tail pseudocyst, the authors proposed several potential mechanisms
for development of hepatic pseudocysts, including through direct invasion or possibly
presence of heterotopic pancreatic tissue within the liver. Interestingly, they also
reported a 90 % cure rate seen with percutaneous drainage, subsequently recommended
to be considered the first-line treatment. Interestingly, our two cases did not demonstrate
successful resolution following this approach [7].
Having reviewed the literature, there are very few published reports of liver pseudocyst
development as a direct consequence of invasion of the portal vein, or “fistulization”
of a pancreatic pseudocyst [2]
[3]
[4]
[5]
[6]. In one case report outlining development of a pancreatic pseudocyst-portal vein
fistula and thrombosis, the authors postulated the mechanism of a “high-pressure system”
within the pseudocyst promoting deposition of pancreatic fluid material into the portal
vein. In their case, percutaneous liver abscess drainage and insertion of a plastic
pancreatic stent across a proximal pancreatic duct stricture led to improvement [4].
Another case report describes a patient presenting with a large pancreatic pseudocyst
adjacent to a dilated, low-density fluid-filled portal vein and liver abscess. Percutaneous
transhepatic portography confirmed the diagnosis of a pseudocyst-portal vein fistula.
At ERP, the patient had evidence of a pancreatic duct stricture and communication
with the pseudocyst, and pancreatic duct stenting was successful in achieving resolution
of both pancreatic and hepatic fluid collections [3].
One report describes several cases of pseudocyst-portal vein fistulization, the authors
postulating that portal vein thrombosis occurs in these patients as a consequence
of rather than encouraging development of the fistula. In their discussion, they suggest
a strong relationship between alcohol excess and development of pseudocyst-portal
vein fistula, particularly in association with chronic pancreatitis. CT may be limited
in correctly diagnosing pseudocyst-portal vein fistula, findings often initially suggesting
only portal vein thrombosis, although in our cases, pseudocyst-portal vein fistula
development in association with portal vein thrombosis was recognized on review of
the images by a highly-experienced radiologist [6].
In another case report, the authors refer to 18 other cases of pseudocyst-portal vein
fistulization; nine had been treated supportively, six required surgical intervention,
such as pancreaticoenterostomy, and three improved with endoscopic pancreatic duct
stenting. Mortality occurred in five of 18 patients, likely due to massive release
of pancreatic enzymes into the portal circulation and resulting systemic lipolysis.
While they describe a role for surgical cyst-gastrostomy in pseudocyst-portal vein
fistula, this had failed to prevent portal vein rupture and death in their patient.
The authors propose that erosion from mass effect of the pseudocyst and release of
digestive enzymes into the portal vein may lead to development of intravascular thrombosis.
They recommend that patients with portal vein thrombosis in the setting of a pseudocyst-portal
vein fistula be considered for pancreatic duct stenting and percutaneous pseudocyst
drainage [5].
Generally, most described cases of hepatic pseudocysts within case reports were managed
conservatively or drained surgically or percutaneously. Our cases are among the first
to demonstrate that secondary or “seeded” liver pseudocysts, refractory to percutaneous
drainage and antibiotics, resolve completely after decompression of their feeding
“primary” pancreatic pseudocyst through EUS-guided cyst-gastrostomy and metal stenting.
Failure of percutaneous drainage alone in treating hepatic pseudocysts might be explained
by persistent feeding with proteolytic material via the connection between the primary
pseudocyst and portal vein. Interestingly, both of our patients demonstrated pancreatic
ductal strictures at ERP, but only one underwent pancreatic duct stenting, later found
to have migrated, and the other was not amenable to stenting. Despite this, each patient
showed complete pancreatic and hepatic pseudocyst resolution with endoscopic cyst-gastrostomy
and metal stent placement alone.
Conclusion
In conclusion, hepatic pseudocysts may develop in association with pancreatic pseudocysts
through an “embolic” phenomenon mediated through the portal venous system. Furthermore,
failure of liver cyst resolution could be attributed to ongoing “feeding” from its
parent pancreatic pseudocyst. We propose that primarily treating hepatic pseudocysts
without targeting the pancreatic pseudocyst in this setting will result in failure
of complete resolution and therefore EUS-guided pancreatic cyst-gastrostomy and metal
stent placement should be strongly considered as first-line therapy to achieve decompression
in pancreatic pseudocyst-portal vein fistula with secondary hepatic pseudocyst, or
second-line therapy in those who do not improve with endoscopic pancreatic duct stenting.
