Note: This copy is for your personal non-commercial use only.

To order presentation-ready
copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

ORIGINAL RESEARCH
Focal Hepatic Lesions: US-guided
Biopsy—Lessons from Review of

䡲 GASTROINTESTINAL IMAGING
Cytologic and Pathologic Examination
Results1
Liat Appelbaum, MD
Purpose: To retrospectively assess factors affecting the success of
Robert A. Kane, MD
ultrasonographically (US)-guided core liver biopsy of focal
Jonathan B. Kruskal, MD, PhD
lesions on the basis of experience when both cytologic and
Janet Romero, MD pathologic examination results were available.
Jacob Sosna, MD
Materials and This HIPAA-compliant retrospective study was granted an
Methods: exemption from the institutional review board. All percu-
taneous US-guided biopsies of focal liver lesions performed
at one institution from January 2000 through February
2006 for which both cytologic and pathologic examination
results were available were included. Specimen adequacy
was determined with on-site cytologic examination per-
formed with a “touch prep” technique. Of 1910 liver biop-
sies, 240 (12.6%) revealed focal lesions, and cytologic and
pathologic examination results were available for 208
(86.7%) of these 240 lesions. The number of biopsy passes
and concordance between cytologic and pathologic find-
ings were evaluated, and correlation between lesion size,
type, and location and the number of passes was assessed.
The Pearson correlation ␹2 test and the Wilcoxon test
were used.

Results: Biopsy specimens were diagnostic in 205 cases (98.6%)

and were nondiagnostic in three cases (1.4%); 85.9% of
the lesions were malignant. There was a single lesion in 89
patients (42.8%), and there were multiple lesions in 119
patients (57.2%). One biopsy pass was sufficient in
58 patients (27.9%); two passes were sufficient in 75
patients (36.1%); and three, four, five, and six passes
were sufficient in 51 (24.5%), 17 (8.2%), five (2.4%),
and two (1.0%) patients, respectively. There was no
relationship between lesion size or location and the
number of passes, according to the Pearson correlation
and ␹2 test (P ⫽ .16 and P ⫽ .22, respectively). On
average, 1.9 passes were required for metastatic le-
sions, versus 2.8 for nonmetastatic lesions (P ⬍ .001,
Wilcoxon test). Cytologic and histopathologic findings
were discordant in 25 cases (12.0%).

1
Conclusion: The size and location of liver lesions sampled for biopsy do
From the Department of Radiology, Hadassah Hebrew
not influence the number of passes needed, while meta-
University Medical Center, POB 12000, Jerusalem, Israel
91120 (L.A., J.S.); and Department of Radiology, Beth static lesions require fewer passes. Without the on-site
Israel Deaconess Medical Center, Boston, Mass (R.A.K., cytologic examination service, a predetermined number of
J.B.K., J.R., J.S.). Received January 28, 2008; revision three passes would be diagnostic in almost 90% of all
requested March 25; revision received June 10; accepted cases.
August 15; final version accepted August 27. Address
correspondence to J.S. (e-mail: jacobs@hadassah
娀 RSNA, 2008
.org.il ).

姝 RSNA, 2008

Radiology: Volume 250: Number 2—February 2009 453

GASTROINTESTINAL IMAGING: US-guided Biopsy of Focal Hepatic Lesions
M
odern high-sensitivity and high-
spatial-resolution imaging tech-
niques, including ultrasonography
(US), computed tomography (CT), and
magnetic resonance imaging, have in-
creased the detection of expected and inci-
dental focal liver lesions. Tissue diagnosis,
usually by means of needle biopsy, is often
required by the referring physician to es-
tablish a definitive diagnosis and to guide
management.
The role of image-guided percutaneous
hepatic biopsy as a safe and accurate diag-
nostic procedure for the evaluation of focal
or diffuse hepatic disease has been well es-
tablished, especially with the advent of
spring-loaded biopsy needles and improve-
ments in image quality (1). US and CT guid-
ance are the techniques of choice for liver
biopsy. In comparison with CT, US is often
more readily available. US-guided biopsies
may be easier to perform, faster, and less
expensive and do not expose the patient to
radiation (1).
Most of the studies concerning the
safety and efficacy of liver needle biopsies
have been performed in patients with dif-
fuse liver disease rather than focal liver le-
sions (2–4). Many of the core liver biopsies
performed at our institution are for focal
lesions. The number of biopsy passes is
usually determined by the radiologist per-
forming the biopsy by consulting results of
an on-site cytologic evaluation (performed
by a cytologist or cytotechnologist), on the
basis of findings from a rapid touch prepa-
ration smear of the core. This on-site cyto-
logic evaluation guides the radiologist to
continue to another biopsy pass and avoid a
nondiagnostic biopsy result on one hand
Advances in Knowledge
䡲 With US-guided liver biopsy,
fewer passes are required for
metastatic lesions compared with
benign lesions or primary liver
tumors (1.9 vs 2.8).
䡲 Lesion size and location do not
influence the number of biopsy
passes needed (P ⫽ .16 and P ⫽
.22, respectively).
䡲 Without the presence of a cytolo-
gist, a predetermined number of
three passes would be diagnostic
in almost 90% of cases.
454
and potentially reduces the rate of repeat
biopsies on the other hand. However, the
service of an on-site cytologist is not always
available. This brings up a question: Can we
predict in which biopsies the cytology ser-
vice would be essential and which could be
managed without it? To the best of our
knowledge, there are no published data ad-
dressing this question. We assumed that
on-site cytologic evaluation would be more
important in small lesions, which are more
challenging to sample, and less critical in
larger ones. We also assumed that a metas-
tasis would be easier to differentiate from a
lesion such as a hepatocellular carcinoma
originating from hepatocytes. Our purpose
was to assess factors predicting the success
of US-guided liver biopsy of focal lesions on
the basis of experience when results of both
“touch prep” smear cytologic examination
and core-sample pathologic examination
were available.
Materials and Methods
Patients
The institutional review board of Beth
Israel Deaconess Medical Center granted
us an exemption for this retrospective
study, which was compliant with the
Health Insurance Portability and Ac-
countability Act. Informed consent was
not required for this retrospective anal-
ysis. From an existing database of US-
guided procedures, we identified all pa-
tients who underwent percutaneous US-
guided biopsy of a focal liver lesion
between January 2000 and February
2006. The database had been main-
Implications for Patient Care
䡲 Three core biopsies of a focal
liver lesion are sufficient for diag-
nosis in the majority of cases
(88%), although the yield is likely
to be somewhat lower in the set-
ting of benign lesions and primary
liver cancer.
䡲 On-site cytologic examination
(touch prep) can be helpful for
limiting the number of passes re-
quired to achieve the diagnosis
and, in a limited number of cases,
may be the only diagnostic test.
Appelbaum et al
tained and prospectively updated by the
US medical staff in our institution. The
database included clinical data, as well
as technical parameters related to the
procedure performed, such as needle
size, needle type, the number of passes
performed, and whether a cytologist
was present during the biopsy.
During this period, a total of 1910
liver biopsies were performed, includ-
ing 240 (12.6%) for focal lesions. Of
these 240 biopsies, 208 (86.7%) with
cytologic and pathologic examination
results (performed in 97 women and
111 men; mean age, 62 years; age
range, 24 –94 years) were included in
the study. There was no meaningful dif-
ference in age or sex distribution be-
tween the excluded patient group and
the study group. There were no mean-
ingful differences in terms of signs,
symptoms, or disease state between pa-
tients in the group where a cytologist
was present and those in the group
where a cytologist was not present at
the time of the biopsy.
For these 208 focal lesion biopsies,
a cytologist was present in the room,
and both pathologic examination and a
touch prep cytologic examination
were performed. In the other 32 cases
for which a cytologist was not present,
the procedure was terminated when
the radiologist was satisfied that the
needle had traversed the mass and a
core specimen had been obtained. For
the purposes of this study, only the
208 biopsies that had both cytologic
and pathologic examination results
were included.
Published online before print
10.1148/radiol.2502080182
Radiology 2009; 250:453– 458
Author contributions:
Guarantors of integrity of entire study, R.A.K., J.B.K., J.S.;
study concepts/study design or data acquisition or data
analysis/interpretation, all authors; manuscript drafting or
manuscript revision for important intellectual content, all
authors; manuscript final version approval, all authors;
literature research, L.A., R.A.K., J.B.K., J.S.; clinical stud-
ies, all authors; statistical analysis, all authors; and
manuscript editing, all authors
Authors stated no financial relationship to disclose.
Radiology: Volume 250: Number 2—February 2009
GASTROINTESTINAL IMAGING: US-guided Biopsy of Focal Hepatic Lesions
US-guided Biopsy Technique
Biopsies were performed by one of four
radiologists from the US staff (including
R.A.K. and J.B.K.), each of whom had
more than 10 years of interventional US
experience, or by a board-certified abdom-
inal imaging fellow (four in every academic
year) under the supervision of a staff radiol-
ogist. All patients considered for liver bi-
opsy met preestablished laboratory crite-
ria: a platelet count of greater than 60 ⫻
109/L and an international normalized ratio
of less than 1.5 at the time of the proce-
dure. Patients with coagulopathy were
treated with either fresh-frozen plasma or
platelet transfusion prior to the biopsy.
Informed consent for the biopsy was
obtained, and conscious sedation was ini-
tialized and monitored by a registered
nurse. The liver was scanned with a 5-MHz
curved-array transducer (HDI 3000 or HDI
5000; Philips Medical Systems, Best, the
Netherlands). The patient was positioned
to facilitate access to the lesion with the
shortest and safest needle trajectory. As-
cites, if present, was graded. A large
amount of ascites was considered a relative
contraindication to the procedure, and
paracentesis was performed prior to biopsy
in two patients.
After proper skin disinfection and
administration of a local anesthetic (2%
lidocaine), an 18-gauge automated bi-
opsy gun with a 2.5-cm needle throw
length (BioPince; InterV and Medical
Device Technologies, Gainesville, Fla)
was advanced into the lesion under real-
time US guidance with the use of a guide
mounted on the transducer, and a sam-
ple was taken (Fig 1). In large tumors,
the needle was advanced to the periph-
ery to avoid central necrosis. Coaxial
needles were not used, and in all cases,
separate punctures were performed for
repeated sampling.
Cytologic and Pathologic Examination
Techniques
The cytologist was either a qualified cy-
tology technician with 2–18 years of ex-
perience or an attending cytopatholo-
gist with 5–21 years of experience. Tis-
sue cores were rolled on a glass slide
(touch prep), and two smear prepara-
tions were made. One slide was air
dried and stained with a modified
Radiology: Volume 250: Number 2—February 2009
Giemsa stain for rapid on-site inter-
pretation. The other slide was fixed in
alcohol and stained with the Papanico-
laou method. This procedural protocol
was performed for each and every bi-
opsy pass, so that the result of each
pass was checked prior to proceeding
to the next pass.
The biopsy procedure was termi-
nated when at least one solid core was
obtained from the lesion and the cytolo-
gist immediately outside the biopsy
suite considered the touch prep results
to be satisfactory, with a sufficient
amount of tissue for analysis. The radi-
ologist in charge was then told that the
specimen was adequate and stopped the
procedure. Patients were kept for 4
hours of postprocedure observation.
Tissue cores were then preserved in
formalin and sent to the Department of
Pathology for histologic examination by
a histopathologist.
Data Evaluation
A computerized database was rou-
tinely updated, including all relevant
data, such as the size, number, and
location of the lesions and patient de-
mographic details. Lesion size was
measured with US, and the largest di-
ameter was recorded. The computer-
ized database was searched by one ab-
dominal imaging fellow (J.R., with 4
years of experience) for lesion type,
number, and location in the liver. The
number of passes and the concor-
dance or discordance between cyto-
logic and pathologic results for differ-
ent lesion sizes were evaluated.
The correlation between the number of
passes and lesion size was assessed, as well
as the correlation between the number of
passes and the finding of metastatic disease
versus a primary liver neoplasm. We also
evaluated whether diagnostic accuracy in
“difficult-to-access” lesions in the right and
left lobe subdiaphragmatic areas was
lower.
Statistical Analysis
We used the Pearson correlation ␹2 test to
assess the correlation between lesion size
and number of biopsy passes and the Wil-
coxon test to correlate the type of lesion
and the number of passes. Analysis was
Appelbaum et al
performed with statistical analysis software
(SAS, version 9.1; SAS Institute, Cary,
NC). P ⬍ .05 was considered to indicate a
significant difference.
Results
Biopsy Success and Complications
The biopsy specimen was sufficient for di-
agnosis in 205 patients (98.6%) and was
insufficient in three (1.4%). Tissue necrosis
with poorly preserved cell structure was the
reason in these cases.
No major complications were ob-
served, and most patients tolerated the
procedure well. Ten patients (4.8%) expe-
rienced moderate pain, which required the
use of analgesics in several instances in
which pain radiated to the right shoulder,
but symptoms subsided within a couple of
hours in all cases, and no bleeding was
noted at postprocedural US performed in
these patients. None of the patients was
hospitalized as a result of the biopsy.
Lesion Size and Type
Definitive diagnoses were categorized as
primary or metastatic malignant tumors,
benign tumors, and benign nonneoplastic
conditions (Table 1). The majority of the
lesions biopsied (85.9%) were malignant.
Metastatic lesions originated from a variety
of primary tumors (Table 2). Eighty-nine
patients (42.8%) had a single liver lesion,
Figure 1
Figure 1: Percutaneous US-guided biopsy in
45-year-old man. Transverse gray-scale US image
shows a hypoechoic focal liver lesion, with the
needle traversing the lesion. The final diagnosis
was colorectal carcinoma metastasis.
455
GASTROINTESTINAL IMAGING: US-guided Biopsy of Focal Hepatic Lesions Appelbaum et al

and 119 patients (57.2%) had multiple le- lesions (7.3%) were 0–0.99 cm in great- diaphragmatic) in the liver dome. Ten
sions. Forty-three patients had a history of est diameter, 62 lesions (30.2%) were (13%) of the left lobe lesions were lo-
known viral hepatitis; of these patients, 28 1.00 –1.99 cm, 50 lesions (24.4%) cated in the upper subdiaphragmatic
(65.1%) had a single hepatic lesion, and 15 were 2.00 –2.99 cm, 46 lesions area of segment IVa or II. There was no
(34.9%) had multiple lesions. Fifteen (22.4%) were 3.00 – 4.99 cm, 20 le- relationship between the lesion’s ana-
sions (9.8%) were 5.00 –7.99 cm, and tomic position in the liver and the diag-
Table 1 12 lesions (5.9%) were 8.00 –13.0 cm. nostic accuracy of the specimen or the
number of passes needed (P ⫽ .22). For
Definitive Diagnosis in 205 Focal Number of Biopsy Passes the 22 lesions (10.6%) in the difficult-
Liver Lesions at US-guided Targeted One pass was performed in 58 patients to-access right and left lobe subdia-
Liver Biopsies (27.9%); two passes were performed in 75 phragmatic areas, only one tissue core
Diagnosis No. of Patients patients (36.1%); and three, four, five, and was insufficient for diagnosis (because
six passes were performed in 51 (24.5%), of tissue necrosis in the sample).
Malignant tumors (n ⫽ 176) 17 (8.2%), five (2.4%), and two (1.0%)
Hepatocellular carcinoma 38 (18.5) Discrepancy between Cytologic and
patients, respectively. There was no statis-
Cholangiocarcinoma 3 (1.5) Pathologic Results
tically significant relationship between le-
Metastasis 128 (62.4)
sion size and the number of passes (P ⫽ In 25 cases (12.0%), there was discor-
Lymphoma 7 (3.4)
.16) (Fig 2). The mean number of passes dance between the cytologic and patho-
Benign, nonneoplastic
for metastatic lesions was 1.9 ⫾ 0.9 (stan- logic results, and diagnosis was achieved
conditions (n ⫽ 29)
Cirrhosis 5 (2.4)
dard deviation), compared with 2.8 ⫾ 1.1 by using either set of results.
Focal nodular hyperplasia 5 (2.4) for all other lesions; the difference was sta- There were seven instances where re-
Hepatitis and chronic tistically significant (P ⬍ .001, Wilcoxon sults of pathologic examination were incon-
inflammation 9 (4.4) test). In 183 patients (88.0%), three or clusive and the diagnosis was based solely
Hemangioma 3 (1.5) fewer passes were diagnostic. For 83% of on the results of cytologic analysis. The
Regenerative nodule in benign lesions (24 of 29), three or fewer slides were evaluated by a cytopathologist
cirrhosis 3 (1.5) passes were diagnostic. for a formal interpretation and were indeed
Liver abscess 2 (1.0) diagnostic. In four of these seven cases, the
Focal steatosis 1 (0.5) Lesion Location core contained mainly normal liver tissue,
Normal findings 1 (0.5) A total of 133 lesions (63.9%) were lo- with insufficient tumor tissue for histologic
cated in the right lobe of the liver, and diagnosis, although the cytologist did see a
Note.—Data in parentheses are percentages.
75 lesions (36.1%) were located in the few specific malignant cells in the smear. In
left lobe. Among right lobe lesions, 12 the other three cases, the core material was
Table 2
(9.0%) were located in the superior nondiagnostic because of impaired tissue
portions of segments VII and VIII (sub- structure in necrotic or mucinous material.
Origin of 128 Metastatic Liver Lesions
at US-guided Targeted Liver Biopsies
Figure 2
Diagnosis No. of Patients

Colorectal carcinoma 22 (17.2)

Lung carcinoma 17 (13.3)
Pancreaticobiliary carcinoma 17 (13.3)
Adenocarcinoma, unknown
primary 16 (12.5)
Pancreatic cancer 16 (12.5)
Neuroendocrine cancer 9 (7.0)
Breast cancer 7 (5.5)
Lymphoma 7 (5.5)
Melanoma 6 (4.7)
Poorly differentiated cells,
unknown primary 4 (3.1)
Esophageal cancer 2 (1.6)
Non–small cell cancer,
Figure 2: Transverse gray-scale US images show 1.2-cm hypoechoic left lobe liver lesion in 52-year-old
unknown primary 2 (1.6)
woman. (a) Note the granularity of the liver texture in this patient with cirrhosis (hepatitis B virus). (b) The
Renal cell carcinoma 2 (1.6)
biopsy needle is seen in an eccentric location in the lesion. In this patient, on-site cytologic examination was
Cervical cancer 1 (0.8)
very helpful. The first-pass sample showed sufficient tissue, with no need for a second pass. The final diagno-
Note.—Data in parentheses are percentages. sis was hepatocellular carcinoma.

456 Radiology: Volume 250: Number 2—February 2009

GASTROINTESTINAL IMAGING: US-guided Biopsy of Focal Hepatic Lesions
A few recognizable viable cells seemed
enough for the cytologist, although there
was not enough viable tissue for histologic
diagnosis.
In the 18 other discordant cases, the
pathologic sample was used for diagnosis
because the cytologic sample was not diag-
nostic, even when the initial slides were
evaluated by a cytopathologist for a formal
interpretation. In nine of these cases
(50%), results of pathologic examination
were benign. Three patients had hepatitis,
and six patients had either abscess, cirrho-
sis, fibrosis, focal steatosis, inflammatory
pseudotumor, or hemangioma. In these be-
nign conditions, the cytologic sample dem-
onstrated only atypical hepatocytes. The cy-
tologist, recognizing benign cells with no
signs of malignancy, assumed the lesion to
be of benign origin but could not make a
more specific diagnosis. In the nine remain-
ing discordant cases, results of pathologic
examination were malignant. Four patients
had hepatocellular carcinoma, and five had
metastatic lesions from primary melanoma,
carcinoid tumor, pancreatic cancer, lym-
phoma, or cholangiocarcinoma. Cytologic
examination demonstrated atypical hepa-
tocytes in the four patients with well-
differentiated hepatocellular carci-
noma. In five patients with metastatic
disease, cytologic examination demon-
strated atypical glandular, lymphatic, or
spindle cells. In only one of these cases,
the sample was poorly preserved. Cells
were seen, but no further information
could be obtained.
No correlation was found between le-
sion size and discrepancy. The 25 cases
with discrepancies included lesions of all
sizes, and there was no significant differ-
ence in the distribution of lesion size com-
pared with that among 180 cases with con-
cordance between cytologic and his-
topathologic findings (P ⫽ .28).
Discussion
We found no statistical relationship be-
tween the number of needle passes and
lesion size in our study. Small hepatic
lesions are more challenging to target,
and one might expect higher miss rates
in small versus large lesions. But on the
other hand, small tumors may have a
more uniform distribution of cancerous
Radiology: Volume 250: Number 2—February 2009
tissue, without the hemorrhage, necro-
sis, or sclerotic changes that are com-
mon and often make diagnosis challeng-
ing in large lesions (5). If the presence
of a diagnostic cytologic evaluation can
be ensured for biopsy specimens of
small lesions, it could lower the number
of passes, because such an evaluation
can ascertain that the sample contains
cells. In large lesions, cytologic exami-
nation can confirm that the sample con-
tains diagnostic cells and is most likely
from a viable nonnecrotic area.
Lesion type, however, did affect the
number of passes. The average number
of passes for metastatic lesions was sig-
nificantly lower than that for all other
lesion types. Metastatic lesions usually
demonstrate characteristic cells, which
are different from both liver cells and
benign or nonspecific tissue. Thus, a di-
agnosis can often be reached more eas-
ily in these cases than in patients with
primary tumors or nonneoplastic liver
disease, where cells more closely re-
semble normal liver texture. In meta-
static lesions, a fine-needle aspiration
biopsy is often sufficient for diagnosis
for the same reason. Jhala et al (6) have
shown that a preliminary diagnosis of
malignancy at the on-site evaluation of
endoscopic US-guided fine-needle aspi-
ration biopsy results is highly accurate
and specific and can therefore replace
the final laboratory cytologic interpreta-
tion, thus saving caseload and workflow
time in the cytopathology laboratory.
In our study, the anatomic location
of the lesion within the liver was not a
factor in the success of the biopsy. Of all
biopsies in “difficult-to-access” lesions,
only one resulted in a sample that was
insufficient for diagnosis, and the rea-
son was necrotic tissue. This may be
explained by the biopsy equipment we
used—a “user-friendly” automatic bi-
opsy gun that facilitates fast biopsies,
enabling patient breathing cooperation
and real-time imaging guidance.
On the basis of our finding that the
size and location of the lesion do not
affect biopsy success, we can suggest
that the performance of three passes
would be diagnostic in almost 90% of all
cases, in the setting of use of 18-gauge
core biopsy needles. Only in the remain-
Appelbaum et al
ing 10% of cases would a repeat biopsy
be necessary. For benign lesions, this
approach would be diagnostic in 83% of
cases. For presumed metastasis, only
1.9 passes were needed, on average, in
our series, and for these lesions, the
success rate might be even higher. How-
ever, institutional constraints may not
enable provision of a cytologist on site
for every biopsy procedure. In our per-
sonal opinion, it is, of course, better to
achieve a diagnostic procedure in 98.6% of
the time when on-site cytology service is
available. In the other case, the use of a
predetermined approach would result
in a diagnosis in almost 90% of cases.
This percentage might increase even
further if smears are sent for cytologic
evaluation even when a cytologist is not
present on site.
In our study, there was high concor-
dance between the histologic and cyto-
logic results, with discordance in only
25 cases (12.0%), for which a diagnosis
was achieved with either the histologic
or the cytologic results. In seven of
these cases, the core was nondiagnos-
tic, even though the cytologic sample
demonstrated atypical cells. In these
cases, cytologic examination was the di-
agnostic test, and the biopsy procedure
was terminated because sufficient diag-
nostic material had been obtained to
render a cytologic diagnosis; the final
diagnosis was made only from this sam-
ple. In four of these seven cases, the
problem was that tissue in the core con-
tained mainly normal liver, and there
was insufficient tumor tissue for diagno-
sis. In the other three cases, the core
material was nondiagnostic because it
contained a high proportion of necrotic
tissue or mucinous material. If a prede-
termined number of passes had been
performed, these patients might well
have had to undergo a repeat biopsy
because the histologic results would
have been nondiagnostic.
In the 18 remaining discordant
cases, a diagnosis was achieved based
only on the pathologic sample. In these
cases, the procedure was terminated
because the smear cytology slide dem-
onstrated adequate numbers of abnor-
mal cells, but at histopathologic exami-
nation, the cells were nonspecific, and
457
GASTROINTESTINAL IMAGING: US-guided Biopsy of Focal Hepatic Lesions
during formal interpretation, the at-
tending cytopathologist could not draw
further conclusions from them. Half of
these lesions were eventually found to
be benign, as opposed to the 14.1% of
benign lesions overall in the study. As
expected, histologic examination or
core biopsy provided a substantial ad-
vantage in the diagnosis of benign le-
sions because of the better preservation
of tissue architecture. With a predeter-
mined number of passes, these patients
would have received a proper diagnosis.
Previous studies have shown many ad-
vantages of US-guided liver biopsy in the
diagnosis of parenchymal liver disease. The
major complication and main cause of post-
procedural death is intraperitoneal bleed-
ing. Substantial hemorrhage is described in
0.35%–0.5% of all liver biopsies, and sub-
clinical bleeding occurs in up to 23% of pa-
tients (3,4,7–9). In our study, none of the
patients developed clinically important
hemorrhage, a finding that may be attrib-
uted to our strict adherence to preestab-
lished coagulation criteria, coupled with the
experience of our operators. Because we
do not routinely image patients after liver
biopsy and do not obtain hematocrit levels
unless patients show symptoms or have
other signs that suggest bleeding, we did
not check all participants for subclinical
bleeding. Previous studies have shown
many advantages to obtaining liver tissue
core histologic samples with large-caliber
biopsy needles, as compared with obtaining
fine-needle aspiration cytologic samples.
These advantages include increased speci-
ficity and accuracy in diagnosing malig-
nancy and subtyping tumors (10,11), supe-
riority in detecting benign lesions such as
hemangiomas and granulomas (10), de-
creased biological sampling errors (12),
and provision of an adequate sample for
special staining and preservation of tissue
architecture to enable proper histologic di-
agnosis (11,12). This last advantage is of
particular importance in achieving a confi-
dent diagnosis of hepatocellular carcinoma
(13). A final diagnosis with tumor subtyping
was achieved for most of our patients from
the biopsy sample alone.
Automatic biopsy guns simplify the
technique. The actual biopsy can be
performed by one radiologist holding
the US transducer in one hand while
458
performing the biopsy with the other
hand. This is an advantage in biopsy of
small hepatic lesions, such as the 127
lesions (61.1%) in our study that were
less than 3 cm in maximal diameter
(13). The larger caliber 18-gauge needle
can also be guided more easily than fine,
highly flexible needles, and its greater
visibility on the US image permits more
precise targeting of small lesions (13).
In our experience, the automated 18-
gauge biopsy gun is very accurate and
reliable, with a diagnostic utility of more
than 98.5%. None of the three nondiag-
nostic cases at biopsy performed with
these needles was limited by sampling
error.
There were several limitations to
our study. No long-term follow-up for
complications was performed, and we
did not check for subclinical bleeding.
We used 18-gauge needles only. The use
of larger or smaller needles could have
affected the success rates because of the
different amounts of tissue obtained,
but we believe that the use of a single
needle size makes our results more ho-
mogeneous. There was an inclusion bias
in our study in that it represents a sin-
gle-center experience and CT proce-
dures were not analyzed; however, the
majority of our liver biopsies are US-
guided. Because even pathologic results
may be inaccurate, especially for pri-
mary liver tumors, we did not aim to
perform long-term follow-up on patient
outcome. Another limitation may be re-
lated to sampling error in the lesion
sampled for biopsy. In 10 cases, biopsy
disclosed either inflammatory changes
or normal liver. These diagnoses may
not represent the real underlying pro-
cess in the targeted lesions.
In conclusion, successful biopsy of met-
astatic liver lesions requires fewer passes
than required for benign lesions or primary
liver tumors. Our findings suggest that le-
sion size and location do not influence the
number of passes needed and that without
the presence of a cytologist, a predeter-
mined number of three passes would be
diagnostic in almost 90% of all US-guided
focal liver biopsies.
Acknowledgment: The authors thank Shifra
Fraifeld, MBA, our research associate, for her
Appelbaum et al
editorial assistance in the preparation of this
work.
References
1. Thanos L, Zormpala A, Papaioannou G,
Malagari K, Brountzos E, Kelekis D. Safety and
efficacy of percutaneous CT-guided liver biopsy
using an 18-gauge automated needle. Eur J In-
tern Med 2005;16:571–574.
2. Caturelli E, Giacobbe A, Facciorusso D, et al.
Percutaneous biopsy in diffuse liver disease:
increasing diagnostic yield and decreasing com-
plication rate by routine ultrasound assessment
of puncture site. Am J Gastroenterol 1996;91:
1318 –1321.
3. McGill DB, Rakela J, Zinsmeister AR, Ott BJ. A
21-year experience with major hemorrhage af-
ter percutaneous liver biopsy. Gastroenterol-
ogy 1990;99:1396 –1400.
4. Piccinino F, Sagnelli E, Pasquale G, Giusti G.
Complications following percutaneous liver
biopsy: a multicentre retrospective study on
68,276 biopsies. J Hepatol 1986;2:165–173.
5. Yu SC, Liew CT, Lau WY, Leung TW,
Metreweli C. US-guided percutaneous biopsy of
small (⬍ or ⫽ 1-cm) hepatic lesions. Radiology
2001;218:195–199.
6. Jhala NC, Eltoum IA, Eloubeidi MA, et al. Pro-
viding on-site diagnosis of malignancy on endo-
scopic-ultrasound-guided fine-needle aspirates:
should it be done? Ann Diagn Pathol 2007;11:
176 –181.
7. Gilmore IT, Burroughs A, Murray-Lyon IM,
Williams R, Jenkins D, Hopkins A. Indications,
methods, and outcomes of percutaneous liver
biopsy in England and Wales: an audit by the
British Society of Gastroenterology and the
Royal College of Physicians of London. Gut
1995;36:437– 441.
8. Knauer CM. Percutaneous biopsy of the liver as
a procedure for outpatients. Gastroenterology
1978;74:101–102.
9. Minuk GY, Sutherland LR, Wiseman DA,
MacDonald FR, Ding DL. Prospective study of
the incidence of ultrasound-detected intrahe-
patic and subcapsular hematomas in patients
randomized to 6 or 24 hours of bed rest after
percutaneous liver biopsy. Gastroenterology
1987;92:290 –293.
10. Dusenbery D, Ferris JV, Thaete FL, Carr BI.
Percutaneous ultrasound-guided needle biopsy
of hepatic mass lesions using a cytohistologic
approach: comparison of two needle types.
Am J Clin Pathol 1995;104:583–587.
11. Pagani JJ. Biopsy of focal hepatic lesions: com-
parison of 18 and 22 gauge needles. Radiology
1983;147:673– 675.
12. Martino CR, Haaga JR, Bryan PJ, LiPuma JP, El
Yousef SJ, Alfidi RJ. CT-guided liver biopsies: 8
years’ experience—work in progress. Radiology
1984;152:755–757.
13. Yu SC, Lau WY, Leung WT, Liew CT, Leung
NW, Metreweli C. Percutaneous biopsy of
small hepatic lesions using an 18 gauge auto-
mated needle. Br J Radiol 1998;71:621– 624.
Radiology: Volume 250: Number 2—February 2009