DOI:10.2214/AJR.04.1796
AJR 2006; 186:556-561
© American Roentgen Ray Society
Postbiopsy Arterioportal Fistula in Patients with Hepatocellular Carcinoma: Clinical Significance in Transarterial Chemoembolization
Hong Suk Park1,
Sang Hyun Lee1,
Young Il Kim1,
Jong Seok Lee2,
Min Kyung Lim3,
Joong-Won Park1,
Joo Hyuk Lee2 and
Chang-Min Kim1
1 Center for Liver Cancer, National Cancer Center, 809, Madu 1-dong, Ilsan-gu,
Goyang-si, Gyeonggi-do 411-764, South Korea.
2 Department of Radiology, National Cancer Center, Gyeonggi-do 411-764, South
Korea.
3 Division of Cancer Control and Epidemiology, National Cancer Center,
Gyeonggi-do 411-764, South Korea.
Received November 19, 2004;
accepted after revision January 27, 2005.
Address correspondence to H. S. Park
(hpark{at}dreamwiz.com).
Abstract
OBJECTIVE. The purpose of this study was to determine,
retrospectively, the frequency of postbiopsy arterioportal fistula in
hepatocellular carcinoma and its significance in transarterial
chemoembolization (TACE).
MATERIALS AND METHODS. Forty-one patients who underwent percutaneous
liver biopsy for diagnosis of hepatocellular carcinoma were referred for TACE.
The control population comprised 161 patients referred during the same period
who underwent TACE without biopsy. We determined that an arterioportal fistula
was present by opacification of the portal vein during the arterial phase of
angiography or by opacification with iodized oil during TACE. We considered
hepatocellular carcinoma to be responsive to TACE when the sum of iodized oil
retention in the tumor and a low-attenuation area on CT was greater than 50%
of tumor size. We compared the frequency of arterioportal fistula and the rate
of tumor response to TACE in both groups and also evaluated possible factors
associated with postbiopsy arterioportal fistula, such as age, sex, Child-Pugh
score, tumor size, average number of needle passes, average distance that the
needle traversed normal liver before reaching the mass, and average interval
between biopsy and TACE.
RESULTS. Twenty-three (56.1%) of 41 patients in the biopsy group and
19 (11.8%) of 161 patients in the control group had an arterioportal fistula
(p < 0.001). The rate of tumor response to TACE was 87.8% (36/41)
in the biopsy group and 87.0% (140/161) in the control group (p =
0.5932). Of the possible related factors, only tumor size correlated
negatively with the occurrence of arterioportal fistula.
CONCLUSION. Percutaneous liver biopsy in hepatocellular carcinoma
patients apparently increases the rate of arterioportal fistula but does not
seem to affect the rate of tumor response to TACE.
Keywords: angiography • biopsy • digital subtraction • embolization • interventional radiology • liver disease
Introduction
The diagnosis of hepatocellular carcinoma (HCC) is usually based on
positive findings on liver CT, an elevated serum 
-fetoprotein level,
and the presence of risk factors such as liver cirrhosis and viral markers.
However, in some cases of equivocal findings on CT and a normal
-fetoprotein level, histologic confirmation with biopsy is
indispensable
[1-4].
The complication rate of percutaneous liver biopsy is low, and the rate of
major complications requiring hospitalization is only 1-3%
[5-7].
Because most vascular complications are asymptomatic
[8], little attention has been
paid to them and they are not included in this number. Among the vascular
complications, arterioportal fistula is the most common, but the influence of
postbiopsy arterioportal fistula on transarterial chemoembolization (TACE) is
not known. Therefore, in this retrospective, case-controlled study, we
evaluated the frequency of postbiopsy arterioportal fistula, the rate of tumor
response to TACE with versus without biopsy, and possible factors associated
with postbiopsy arterioportal fistula.
Materials and Methods
From April 2002 to March 2004, we retrospectively reviewed the radiologic
studies and medical records of 72 patients who had undergone percutaneous
liver biopsy because of suspected HCC. Sixty-six patients were confirmed as
having HCC, and 60 patients were referred for TACE. All patients underwent
liver CT before biopsy, and those showing a diffuse growth pattern, tumor
invasion into the portal vein, thrombosis in the portal vein, or a preexisting
arterioportal shunt on liver CT were excluded. Patients with a history of
liver surgery or any kind of intervention related to the liver were also
excluded. Nineteen patients were excluded according to these criteria;
therefore, 41 patients were enrolled in this study.
One hundred sixty-one patients who received TACE as a first line of
treatment without biopsy during the same period were enrolled as a control
group. We applied the same exclusion criteria to them. Baseline
characteristics of the patients in the biopsy and control groups are
summarized in Table 1. The two
groups did not differ significantly in age, tumor size, or Child-Pugh
score.
For liver CT, 120 mL of nonionic contrast material, iopromide (Ultravist
300, Schering), was injected IV at a rate of 3 mL/sec with a mechanical power
injector, and arterial, portal, and equilibrium phase imaging was performed
with a helical scanner (HiSpeed Advantage, GE Healthcare) at 30, 65, and 180
sec, respectively, after the start of the injection.
The indications for biopsy were a low -fetoprotein level in 35
patients, atypical CT findings in three patients, and unknown in three
patients. Biopsy was performed with an automated gun biopsy set using an
18-gauge needle (Acecut biopsy system, TSK Laboratory) under sonographic
guidance. After local anesthesia had been administered to the abdominal wall
and liver surface, the automated gun biopsy was performed. The average number
of needle passes, the average distance that the needle traversed normal liver
before reaching the mass, and the average interval between biopsy and TACE
were recorded. We consider these and other factors, including sex, age,
Child-Pugh score, and tumor size, to possibly be related to postbiopsy
arterioportal fistula.
All patients underwent celiac, proper hepatic, and selective arteriography
using digital subtraction angiography with an Integris V5000 unit (Philips
Medical Systems). For celiac angiography, 25 mL of nonionic contrast material,
iodixanol (Visipaque 270, Nycomed), was injected at a rate of 5 mL/sec with a
5-French catheter, and for proper hepatic and selective arteriography, 9 mL
was injected at a rate of 3 mL/sec with a 2.8- or 3-French microcatheter. We
determined that an arterioportal fistula was present by focal opacification of
the portal vein adjacent to the tumor during the arterial phase of angiography
or by early opacification with iodized oil during TACE (early opacification
with iodized oil can be distinguished from late filling with iodized oil in
satisfactory TACE). TACE was performed by inserting a catheter to the level of
the segmental or subsegmental arteries or lobar branches and infusing a
solution containing 10-50 mg of doxorubicin hydrochloride (ADM, Dong-a
Pharmacy) and 2-20 mL of iodized oil (Lipiodol Ultrafluid, Guerbet) through a
5-French catheter or a 2.8- or 3-French microcatheter. The end point of
infusion was stasis of the iodized oil mixture in the feeding arteries with or
without the presence of iodized oil in the portal vein adjacent to the tumor.
In cases in which most of the iodized oil mixture drained into the portal vein
through the arterioportal fistula at the beginning of the infusion or, because
of the large size of the tumor or arterioportal fistula, stasis of the iodized
oil mixture could not be obtained after infusion of the maximum amount of
iodized oil (20 mL), embolization was performed with absorbable gelatin sponge
particles (Gelfoam, Upjohn). All details of the TACE techniques and findings,
including the amount of iodized oil and drug used, whether gelatin sponge
embolization was used, the level of embolization, and the time of portal vein
opacification, were recorded.
All patients underwent liver CT 1 month after TACE for evaluation of
response and recurrence. HCC was considered to be responsive to TACE when the
sum of iodized oil retention in the tumor and a low-attenuation area without
enhancement in any phase was greater than 50% of the tumor size
[9,
10]. Two radiologists
interpreted both the CT and the angiography images retrospectively and by
consensus, without knowledge of the biopsy history.
Results
An arterioportal fistula was found in 23 patients (56.1%) of the biopsy
group (Figs. 1A,
1B,
1C, and
1D) and in 19 patients (11.8%)
of the control group (Figs. 2A,
2B, and
2C), and whether an
arterioportal fistula was detected during angiography or during TACE did not
significantly differ between the two groups
(Table 2).
View larger version (160K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —58-year-old man who underwent biopsy 4 days before transarterial
chemoembolization. Arterial phase of celiac arteriography performed with
digital subtraction angiography technique shows filling of portal vein
branches (arrowheads) via arterioportal fistula. Because
arterioportal fistula steals blood flow from tumor-feeding vessel,
hepatocellular carcinoma is only faintly visualized (arrow).
|
|
View larger version (148K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —58-year-old man who underwent biopsy 4 days before transarterial
chemoembolization. On arteriography after superselection of tumor-feeding
vessel with microcatheter, tumor staining and portal vein
(arrowheads) are more clearly visualized. Chemoembolization was
performed at this level.
|
|
View larger version (154K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —58-year-old man who underwent biopsy 4 days before transarterial
chemoembolization. Spot radiograph obtained immediately after
chemoembolization shows retention of adjacent portal vein and tumor with
iodized oil.
|
|
View larger version (125K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1D —58-year-old man who underwent biopsy 4 days before transarterial
chemoembolization. Unenhanced CT scan obtained 1 month after transarterial
chemoembolization shows complete retention of iodized oil in tumor.
|
|
View larger version (144K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —55-year-old man who had arterioportal fistula without biopsy.
Arteriography after superselection of tumor-feeding vessel with microcatheter
shows tumor staining and portal vein (arrowheads).
|
|
View larger version (129K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C —55-year-old man who had arterioportal fistula without biopsy.
Enhanced liver CT scan obtained 1 month after transarterial chemoembolization
shows complete retention of iodized oil in tumor.
|
|
The rate of tumor response to TACE was 87.8% (36/41) in the biopsy group
and 87.0% (140/161) in the control group. The difference between these rates
was not statistically significant (p = 0.5932). No patient in either
group had a major complication requiring further management and a prolonged
hospital stay. We used gelatin sponge embolization in 19 patients of the
biopsy group and 58 patients of the control group. The rate of gelatin sponge
embolization did not significantly differ between the two groups (p =
0.28). In nine patients of the biopsy group and only one patient of the
control group, we first embolized with the gelatin sponge to occlude the
fistula tract before infusing iodized oil (Figs.
3A,
3B,
3C, and
3D).
View larger version (103K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —51-year-old man who had postbiopsy arterioportal fistula that
required gelatin sponge embolization to occlude fistula tract before iodized
oil infusion during transarterial chemoembolization. Arterial phase of CT scan
obtained before biopsy shows ill-defined heterogeneously enhancing tumor
without evidence of arterioportal shunt.
|
|
View larger version (119K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —51-year-old man who had postbiopsy arterioportal fistula that
required gelatin sponge embolization to occlude fistula tract before iodized
oil infusion during transarterial chemoembolization. In early arterial phase
of selective arteriography 7 days after biopsy, portal vein (black
arrows) is well visualized. Tumor is indistinct and outlined only by fine
peritumoral arteries (white arrows).
|
|
View larger version (152K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —51-year-old man who had postbiopsy arterioportal fistula that
required gelatin sponge embolization to occlude fistula tract before iodized
oil infusion during transarterial chemoembolization. Because nearly all
infusate passed into portal vein through arterioportal fistula, it was
embolized with gelatin sponge first. Spot radiograph obtained after
chemoembolization shows successful saturation of tumor with iodized oil and
little retention in portal vein.
|
|
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —51-year-old man who had postbiopsy arterioportal fistula that
required gelatin sponge embolization to occlude fistula tract before iodized
oil infusion during transarterial chemoembolization. One-month follow-up CT
scan shows compact uptake of iodized oil in tumor.
|
|
With regard to biopsy, the average number of needle passes was 1.9 (range,
1-5) and the average distance that the needle traversed normal liver before
reaching the mass was 3.4 cm (range, 0-8.5 cm). The average interval between
biopsy and TACE was 6.2 days (range, 2-30 days).
No correlation was found between arterioportal fistula and age, sex,
Child-Pugh score, number of needle passes, distance that the needle traversed
normal liver, or interval between biopsy and TACE, and a negative correlation
was found between arterioportal fistula and tumor size
(Table 3). The incidence in
patients with tumors smaller than 4 cm (77.8%, or 14/18) was significantly
higher (p = 0.0254) than that in patients with tumors larger than 4
cm (39.1%, or 9/23). With regard to the time between biopsy and TACE, the
incidence of postbiopsy arterioportal fistula was 57.6% (19/33) when the two
were performed within 1 week of each other and 50% (4/8) when the interval was
more than 1 week.
Discussion
The rate of postbiopsy arterioportal fistula in our study was 56.1%, which
is as high as the 54% confirmed with iodized oil CT
[11] and much higher than the
24.6% [12] and 5.4%
[13] confirmed with
conventional angiography. Various mechanisms are involved in the development
of an arterioportal fistula or arterioportal shunt
[14,
15]: transsinusoidal filling,
as in liver cirrhosis; a transvasal route when tumor invades the portal vein;
a transplexal or peribiliary route, as in portal vein occlusion; a
transtumoral route, as in hypervascular tumors such as HCC; and trauma, either
iatrogenic or accidental. We tried to exclude the possibility that patients
with an arterioportal fistula of another cause could be enrolled in either
group. However, because liver CT is less sensitive than angiography in
detecting arterioportal fistulas
[16], not all patients with a
preexisting arterioportal fistula could be excluded, and preexisting
arterioportal fistulas may have been present at a rate of 11.8%, as seen in
the control group. A similar rate of arterioportal fistulas of other causes
might have been included in the biopsy group.
We evaluated possible factors associated with postbiopsy arterioportal
fistula: age, sex, Child-Pugh score, tumor size, number of needle passes,
distance that the needle traversed normal liver, and interval between biopsy
and TACE. The only significant factor was tumor size. Arterioportal fistula
was more frequent in patients with tumors smaller than 4 cm than in those with
tumors larger than 4 cm. The plausible explanation may lie in the difficulty
of targeting a small tumor during needle biopsy. The smaller the tumor is, the
more we need to vary the direction of the needle and the greater the
likelihood that vascular injury will occur.
According to a report of Hellekant
[12], the incidence of
postbiopsy vascular complications depends on the interval between biopsy and
angiography; the incidence rate was 61% for an interval of less than 1 week
and dropped to 11% for an interval of more than 1 week, and he suggested that
postbiopsy arterioportal fistulas tend to close spontaneously. In our study,
however, the incidences—at 57.6% and 50%, respectively—showed no
such abrupt decrease.
The response rate of tumors to TACE did not differ significantly between
the biopsy and control groups. Arterioportal fistula in 14 patients was
visualized not on angiography but during iodized oil infusion; thus, most
postbiopsy arterioportal fistulas were small. Unlike other arterioportal
shunts, the shunt tract of postbiopsy arterioportal fistulas may have little
or no endothelial lining; therefore, arterioportal fistulas can be embolized
more easily. In only nine patients did we have to use gelatin sponge
embolization to occlude the shunt. Most of the patients were treated with the
usual technique: infusion of a mixture of iodized oil and doxorubicin, with or
without gelatin sponge embolization afterward. All were followed up with
angiography, at a mean interval of 2.4 months (range, 28-135 days), which
showed no recurrent or residual arterioportal fistula. In contrast, TACE
itself can be a cause of arterioportal fistula
[16]. The mechanism is not
known but may be vascular injury resulting from the chemotherapeutic agent or
microwire and hemodynamic alteration after embolization of the hepatic artery
[17].
In our study, patients were categorized into a biopsy group or a control
group, not into groups with or without an arterioportal fistula. The
interventional radiologist performing TACE may be interested in the difference
in response to treatment between tumors with and without an arterioportal
fistula. However, in clinical practice, TACE has been decided on first, and
the arterioportal fistula is detected afterward, during angiography, right
before embolization. Angiography is not performed only to detect an
arterioportal fistula, and the presence of an arterioportal fistula does not
alter the therapeutic decision. Therefore, the clinical importance lies in the
fact that treatment response does not differ even though biopsy increases the
incidence of arterioportal fistula, not in whether the presence of an
arterioportal fistula influences the treatment response.
The limitation of this study was that the effect of biopsy or postbiopsy
arterioportal fistula on TACE was evaluated through the response of the tumor
as seen on follow-up CT at 1 month, not by a long-term result such as the
patient's survival. There have been several reports of local spread of HCC
after liver biopsy
[18-20]
and hematogenous dissemination after surgical resection
[21]. On the other hand, there
is also a report of no evidence of hematogenous dissemination of HCC after
biopsy, as shown by quantitative analysis of circulating tumor DNA
[22]. However, HCC usually
spreads in the liver through the portal vein
[23]; therefore, further study
is needed on the effect of biopsy-induced arterioportal fistulas on
intrahepatic metastasis of HCC.
In conclusion, percutaneous liver biopsy in HCC patients apparently
increases the rate of arterioportal fistula but does not seem to affect the
response to TACE.
References
- Pasha T, Gabriel S, Therneau T, Dickson ER, Lindor KD.
Cost-effectiveness of ultrasound-guided biopsy.
Hepatology 1998;27
: 1220-1226[CrossRef][Medline]
- Rocken C, Meier H, Klauck S, Wolff S, Malfertheiner P, Roessner A.
Large-needle biopsy versus thin-needle biopsy in diagnostic pathology of liver
diseases. Liver 2001;21
: 391-397[CrossRef][Medline]
- Horigome H, Nomura T, Saso K, Itoh M, Joh T, Ohara H. Limitations
of imaging diagnosis for small hepatocellular carcinoma: comparison with
histological findings. J Gastroenterol Hepatol1999; 14:559
-565[CrossRef][Medline]
- Park JW. Practice guideline for hepatocellular carcinoma.
Korean J Hepatol 2004;10
: 88-98[Medline]
- Lee SJ, Lim JH, Lee WJ, Lim HK, Choo SW, Choo IW. Transient
subsegmental hepatic parenchymal enhancement on dynamic CT: a sign of
postbiopsy arterioportal shunt. J Comput Assist Tomogr1997; 21:355
-360[CrossRef][Medline]
- Bravo AA, Smeth SG, Chopra S. Liver biopsy. N Engl J
Med 2001; 344:495
-500[Free Full Text]
- Machicao VI, Lukens FJ, Lange SM, Scolapio JS. Arterioportal
fistula causing acute pancreatitis and hemobilia after liver biopsy.
J Clin Gastroenterol 2002;34
: 481-484[Medline]
- Wallace S, Medellin H, Nelson RS. Angiographic changes due to
needle biopsy of the liver. Radiology1972; 105:13
-18[Medline]
- Lee H, Kim JS, Choi IJ, Chung JW, Park JH, Kim CY. The safety and
efficacy of transcatheter arterial chemoembolization in the treatment of
patients with hepatocellular carcinoma and main portal vein obstruction.
Cancer 1997; 79:2087
-2094[CrossRef][Medline]
- Takayasu K, Arii S, Matsuo N, et al. Comparison of CT findings with
resected specimens after chemoembolization with iodized oil for hepatocellular
carcinoma. AJR 2000;175
: 699-704[Abstract/Free Full Text]
- Lim JH, Lee SJ, Lee WJ, Lim HK, Choo SW, Choo IW. Iodized oil
retention due to postbiopsy arterioportal shunt: a false positive lesion in
the investigation of hepatocellular carcinoma. Abdom
Imaging 1999; 24:165
-170[Medline]
- Hellekant C. Vascular complications following needle puncture of
the liver. Acta Radiol Diagn (Stockh)1976; 17:209
-222[Medline]
- Okuda K, Musha H, Nakajima Y, et al. Frequency of intrahepatic
arteriovenous fistula as a sequela to percutaneous needle puncture of the
liver. Gastroenterol 1978;74
: 1204-1207[Medline]
- Bookstein JJ, Cho KJ, Davis GB, Dail D. Arterioportal
communications: observation and hypotheses concerning transsinusoidal and
transvasal type. Radiology 1982;142
: 581-590[Abstract/Free Full Text]
- Choi BI, Lee KH, Han JK, Lee JM. Hepatic arterioportal shunts:
dynamic CT and MR features. Korean J Radiol2002; 3:1
-15[Medline]
- Kim TK, Choi BI, Han JK, Chung JW, Park JH, Han MC. Nontumorous
arterioportal shunt mimicking hypervascular tumor in cirrhotic liver:
two-phase spiral CT findings. Radiology1998; 208:597
-603[Abstract/Free Full Text]
- Lee K, Sung K, Lee D, Park SJ, Kim KW, Yu J. Transcatheter arterial
chemoembolization for hepatocellular carcinoma: anatomic and hemodynamic
considerations in the hepatic artery and portal vein.
RadioGraphics 2002;22
: 1077-1091[Abstract/Free Full Text]
- Takamori R, Wong LL, Dang C, Wong L. Needle tract implantation from
hepatocellular cancer: is needle biopsy of the liver always necessary?
Liver Transpl 2000;6
: 73-75[Medline]
- Chapoutot C, Perney P, Fabre D, et al. Needle-tract seeding after
ultrasound-guided puncture of hepatocellular cancer: a study of 150 patients.
Gastroenterol Clin Biol 1999;23
: 552-556[Medline]
- Schotman SN, Man RAD, Stoker J, Zondervan PE, Ijzermans JNM.
Subcutaneous seeding of hepatocellular carcinoma after percutaneous needle
biopsy. Gut 1999;45
: 626-627[Free Full Text]
- Wong IHN, Lau WY, Leung T, Yeo W, Johnson PJ. Hematogenous
dissemination of hepatocytes and tumor cells after surgical resection of
hepatocellular carcinoma: a quantitative analysis. Clin Cancer
Res 1999; 5:4021
-4027[Abstract/Free Full Text]
- Yu SCH, Lo DYM, Ip CB, Liew CT, Leung TWT, Lau WY. Does
percutaneous liver biopsy of hepatocellular carcinoma cause hematogenous
dissemination? An in vivo study with quantitative assay of circulating tumor
DNA using methylation-specific real-time polymerase chain reaction.
AJR 2004; 183:383
-385[Abstract/Free Full Text]
- Yamada R, Nakamura K. Hepatic artery chemoembolization and infusion
chemotherapy. In: Han MC, Park JH, eds. Interventional
radiology. Seoul, Korea: Ilchokak, 1999:75
-95
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
