The American Journal of Surgery 184 (2002) 601– 605

Scientific paper

Preoperative evaluation of hepatic lesions for the staging of

hepatocellular and metastatic liver carcinoma using endoscopic
ultrasonography
Samir S. Awad, M.D.a,*, Shawn Fagan, M.D.a, Suhaib Abudayyeh, M.D.b,
Nagla Karim, M.D.b, David H. Berger, M.D.a, Kamran Ayub, M.D.b
a
Michael E. DeBakey Department of Surgery, Houston VAMC, Surgical Service (112), 2002 Holcombe Blvd., Houston, TX 77030, USA
b
Department of Medicine, Division of Gastroenterology, Baylor College of Medicine, Houston VAMC, Houston, TX, USA
Manuscript received July 31, 2002

Presented at the 54th Annual Meeting of the Southwestern Surgical Congress, Coronado, California, April 7–10, 2002.

Abstract
Background: Noninvasive imaging techniques, such as dynamic computed tomography (CT), magnetic resonance imaging and transab-
dominal ultrasonography are limited in their ability to detect hepatic lesions less than one cm. Intraoperative ultrasonography (IOUS) is
currently the most sensitive modality for the detection of small hepatic lesions. However, IOUS is invasive requiring laparoscopy or formal
laparotomy. We sought to evaluate the feasibility of using endoscopic ultrasonograhpy (EUS) for the detection and diagnosis of hepatic
masses in patients with hepatocellular cancer (HCCA) and metastatic lesions (ML). We hypothesized that EUS could detect small (⬍1.0
cm) hepatic lesions undetectable by CT scan and could be used for biopsy of deep-seated hepatic lesions.
Methods: Consecutive patients referred for EUS with suspected liver lesions were evaluated between July 2000 and October 2001. All
patients underwent EUS using an Olympus (EM30) radial echoendoscope. If liver lesions were confirmed and fine needle aspiration (FNA)
was deemed necessary, a linear array scope was used and an FNA performed with a 22-gauge needle. Two passes were made for each lesion.
Results: 14 patients underwent evaluation with dynamic CT scans and EUS. In all 14 patients, EUS successfully identified hepatic lesions
ranging in size from 0.3 cm to 14 cm (right lobe: n ⫽ 3, left lobe: n ⫽ 1, bilobar: n ⫽ 8). Moreover, EUS identified new or additional lesions
in 28% (4 of 14) of the patients, all less than 0.5 cm in size (HCCA: n ⫽ 2, ML: n ⫽ 2), influencing the clinical management. In 2 of 14
patients EUS identified liver lesions, previously described as suspicious by CT scan, to be hemangiomas. Nine patients underwent
EUS-guided FNA of hepatic lesions (deep seated: n ⫽ 3, superficial: n ⫽ 6). All FNA passes yielded adequate specimens (malignant: n ⫽
8, benign: n ⫽ 1).
Conclusions: Our preliminary experience suggests that EUS is a feasible preoperative staging tool for liver masses suspected to be HCCA
or metastatic lesions. EUS can detect small hepatic lesions previously undetected by dynamic CT scans. Furthermore, EUS-guided FNA can
confirm additional HCCA liver lesions or liver metastases, in deep-seated locations, upstaging patients and changing clinical management.
© 2002 Excerpta Medica Inc. All rights reserved.

Keywords: Endoscopic ultrasonography; Hepatocellular cancer; Metastatic liver carcinoma; Fine needle aspiration; Staging

Accurate preoperative staging remains an integral part of
the surgical treatment of Hepatocellular and metastatic liver
carcinoma, as the extent of liver involvement may change
clinical stage and management. Current preoperative radio-
logic imaging techniques, such as dynamic computed to-
* Corresponding author. Tel.: ⫹1-713-794-7139; fax: ⫹1-713-794-
7352.
E-mail address: sawad@bcm.tmc.edu
mography (CT), CT angioportography (CTAP), magnetic
resonance imaging (MRI), and transabdominal ultrasonog-
raphy (US) have each been limited in their ability to detect
small hepatic lesions. Intraoperative ultrasonography
(IOUS) with careful intraoperative palpation is currently the
gold standard for the detection of small hepatic lesions. This
technique has consistently been shown to be the most ac-
curate method with greater sensitivity and specificity than
any preoperative radiologic examination [1–5]. However,

0002-9610/02/$ – see front matter © 2002 Excerpta Medica Inc. All rights reserved.
PII: S 0 0 0 2 - 9 6 1 0 ( 0 2 ) 0 1 0 9 2 - 9
602 S.S. Awad et al. / The American Journal of Surgery 184 (2002) 601– 605

Fig. 2. Endoscopic ultrasonography image of right lobe liver (segments

Fig. 1. Endoscopic ultrasonography image of left lobe liver (segments 2
and 3).
IOUS is invasive requiring laparoscopy or formal laparot-
omy.
Recently, endoscopic ultrasonography (EUS) with EUS-
guided fine needle aspiration (EUS-FNA) has emerged as an
important tool in the diagnosis and staging of malignant
gastrointestinal tumors, especially pancreatic cancer [6 – 8].
Traditionally EUS has not been utilized in the evaluation of
the liver even though most of the liver segments can be
visualized with the echoendoscope through the stomach and
duodenum [9]. Because of the demonstrated accuracy of
IOUS in the detection of small lesions and because the
echoendoscope can placed in close proximity to the liver,
we sought to evaluate the feasibility of EUS for the detec-
tion and diagnosis of hepatic masses in patients with known
or suspected hepatocellular (HCCA) and metastatic liver
carcinomas (ML).
5).
and the region of interest were assessed with color flow
Doppler prior to FNA. The linear scope was also used to
examine the livers of all patients in whom the radial scope
could not identify a mass.
Results
14 patients underwent evaluation with dynamic CT scans
and EUS for known or suspected HCCA and metastatic
liver carcinoma. The CT and EUS results are shown in
Table 1. In all 14 patients, EUS successfully identified
hepatic lesions ranging in size from 0.3 cm to 14 cm (right
lobe: n ⫽ 3, left lobe: n ⫽ 5, bilobar: n ⫽ 6). Moreover,
EUS identified new or additional lesions in 28% (4 of 14) of
the patients, all less than 0.5 cm in size (HCCA: n ⫽ 2, ML:
n ⫽ 2), influencing the clinical management (Table 1). In 2
of 14 patients EUS identified liver lesions, previously de-

Methods

Consecutive patients referred for EUS with suspected or

known liver lesions were evaluated between July 2000 and
October 2001. After informed consent, all patients under-
went EUS using an Olympus (EM30) radial echoendoscope
at 7.5 MHz. All examinations were performed in a sequen-
tial manner as follows: the left lobe and hilum of the liver
were examined from the gastric body and fundus after the
other surrounding structures were identified (Fig. 1). The
right lobe of the liver was imaged from the duodenum and
antrum (Fig. 2). If liver lesions were confirmed and an FNA
was deemed necessary, a linear array scope was used and an
FNA performed with a 22-gauge needle (Fig. 3). Two Fig. 3. Endoscopic ultrasonography-guided fine needle aspiration of right
passes were made for each lesion. The vascularity of legion lobe lesion.
 S.S. Awad et al. / The American Journal of Surgery 184 (2002) 601– 605 603

Table 1
Dynamic computed tomography (CT) and endoscopic ultrasonography (EUS) results

Dynamic CT scan results Pre EUS diagnosis EUS results FNA Additional
lesions
detected

Bilobar lesions Metastatic colon cancer Bilobar solid lesions Yes

Bilobar lesions Metastatic colon cancer Bilobar solid lesions No
Bilobar lesions HCCA Bilobar solid lesions Yes
6 cm hypervascular lesion segment 5 HCCA Bilobar solid lesions, right lesion Yes Yes
confirmed. Two additional left lobe
lesions detected
2 cm hypodense right lobe lesion Hepatic lesion, possible metastatic Hemangioma right lobe No
0.9 cm hypodense right lobe lesion Hypodense liver lesion No masses No
3 cm irregular hypodense left lobe lesion Left hepatic mass No masses No
Multiple hypodense lesions right lobe Metastatic colon cancer Bilobar disease Yes Yes
liver
Bilobar hypodense liver lesions Squamous cell cancer head and neck Bilobar disease Yes
⫹ lever lesions
2 cm low attenuation right lobe Right liver lesion Right lobe lesion Yes
No liver lesions GIST stomach now with recurrence Right lobe lesion Yes Yes
Numerous enhancing lesions right lobe Pancreatic head mass Bilobar liver lesions Yes Yes
Multiple bilobar low attenuation lesions Gastric cancer Left lobe lesion Yes
1.5 cm hypodense lesion right lobe Liver lesion, Hx rectal cancer Hemangioma right lobe No

FNA ⫽ fine needle aspiration; HCCA ⫽ hepatocellular carcinoma; GIST ⫽ gastrointestinal stromal tumor.

scribed as suspicious by CT scan, to be hemangiomas. Nine the improvement in these imaging techniques, it is esti-
patients underwent EUS-guided FNA of hepatic lesions mated that that only 50% of primary or secondary liver
(deep seated: n ⫽ 3, superficial: n ⫽ 6). All FNA passes carcinomas less than 1 cm can be detected in patients
yielded adequate specimens (malignant: n ⫽ 8, benign: n ⫽ undergoing liver resection [10]. The overall sensitivity of
1; Table 2). transabdominal US for the detection of primary or second-
ary liver carcinoma ranges from 50% to 70% [11,12], drop-
Comments ping to 20% for lesions less than 1 cm [13]. Dynamic
computed tomography has been found to have a higher
Because additional occult hepatic lesions often exist in sensitivity of 40% to 80% when compared with US [14] but
patients who present with primary HCCA and secondary in a similar manner the sensitivity decreases to 48% for
metastatic liver carcinoma, accurate preoperative staging is lesions less than 1 cm [13]. Spiral CT with arterial portog-
critical in the clinical management of these patients and can raphy has emerged as a new technique, which can provide
potentially prevent unnecessary laparotomies. Current im- enhanced visualization of the liver (78% to 94% sensitivi-
aging modalities for the preoperative staging of hepatic ty). However, this imaging technique is invasive and fre-
lesions include transabdominal ultrasonography, dynamic quently overcalls perfusion defects as lesions, thus having a
contrast enhanced CT, CT portography, and MRI. Despite high false positive rate of diagnosis up to 15% [15]. Fur-
thermore, it has not been demonstrated to be superior to
MRI, which has comparable detection rates (63% to 95%
Table 2
Location and results of endoscopic ultrasonography (EUS)-guided fine sensitivity) [16]. Of these techniques, CTAP and MRI have
needle aspiration of liver lesions emerged as the most accurate preoperative imaging tech-
EUS location of lesion Number Pathology Changed
niques for hepatic lesions that are larger than 2 cm but IOUS
of management with intraoperative palpation has been shown to be superior
passes to both of these imaging modalities [1–5].
Bilobar 5 Adenocarcinoma 5/5 No The increased sensitivity and specificity of IOUS relates
Bilobar 2 Hepatocarcinoma 2/2 No to the proximity of the US probe to the liver parenchyma
Bilobar 2 Hepatocarcinoma 2/2 Yes and the use of color flow Doppler in the evaluation of
Bilobar 2 Hepatocarcinoma 2/2 Yes
hepatic lesions. In a similar manner, EUS placed in the
Bilobar 2 Poorly differentiated Yes
Left lobe 2 Benign Yes stomach and duodenum provides close proximity to many
Right lobe 2 Metastatic GIST Yes of the hepatic segments and can provide the opportunity for
Bilobar 2 Adenocarcinoma Yes FNA of additional or suspicious lesions. Preliminary infor-
Left lobe 2 Large cell lymphoma No mation regarding the ability of EUS to evaluate the liver has
GIST ⫽ gastrointestinal stromal tumor. been obtained from patients who have undergone evaluation
604 S.S. Awad et al. / The American Journal of Surgery 184 (2002) 601– 605
of primary upper gastrointestinal malignancies such as gas-
tric and pancreatic cancer [9].
The primary goal of this study was to examine the
feasibility of EUS in the detection of hepatic lesions de-
tected by our conventional preoperative screening using
helical CT. Secondary goals were to determine if additional
lesions could be identified, and to determine the diagnostic
potential through EUS-guided FNA. Our preliminary expe-
rience suggests that EUS is a feasible preoperative staging
tool for liver masses suspected to be HCCA or metastatic
lesions. In our series, EUS detected all of the lesions de-
tected by helical CT. EUS also detected additional small
hepatic lesions in 28% of the patients examined which were
previously undetected by dynamic CT scans. Moreover, in 2
of 14 patients, EUS examination with the use of color flow
Doppler identified two lesions previously identified by dy-
namic CT scan to be solid lesions as hemangiomas.
With EUS, most lesions detected can also be aspirated
with EUS-FNA for pathological diagnosis. Additionally, the
pathology of any additional detected lesions in deep-seated
locations suspected to be either HCCA or metastatic carci-
noma can be also determined and can influence the clinical
management of the patient often upstaging their TNM stage.
In our series, EUS-FNA was successfully accomplished in 9
of 9 patients and yielded adequate tissue for pathological
diagnosis. In 67% (6 of 9) of these patients, the pathological
diagnosis changed the clinical management.
Despite the feasibility of this technique demonstrated in
our series, there are some limitations of EUS in the evalu-
ation of the liver. These include areas near the dome of the
diaphragm and the inferior-posterior most portion of the
right lobe of the liver. Additionally, peripheral lesions and
those far from the probe of the echoendoscope are not well
visualized. Therefore, a prospective evaluation of the depth
of penetration into the liver is necessary. Furthermore, ul-
trasonography visualization may be improved with the use
of phase-inversion sonography after injection with contrast
[17].
Limitations of our current study include (1) the unknown
sensitivity and specificity of EUS because of small sample
size, (2) EUS was not directly compared with CTAP or MRI
but rather to helical CT which in our institution is the
currently available preoperative radiological staging tool;
therefore no conclusions can be drawn with regards to those
modalities, and (3) no direct comparison to IOUS was
performed. Despite these limitations, this study demon-
strates the feasibility of EUS and EUS-FNA for the staging
of patients with suspected hepatocellular and metastatic
carcinoma of the liver. Further study of this modality with
direct comparisons to CTAP, MRI, and IOUS is warranted.
References
[1] Clarke MP, Kane RA, Steele G, et al. Prospective comparison of
preoperative imaging and intraoperative ultrasonography in the de-
tection of liver tumors. Surgery 1989;106:849 –55.
[2] Machi J, Isomoto H, Kurohiji T, et al. Accuracy of intraoperative
ultrasonography in diagnosing liver metastasis from colorectal can-
cer: evaluation with postoperative follow-up results. World J Surg
1991;15:551–7.
[3] Soyer P, Elias D, Zeitoun G, et al. Surgical treatment of hepatic
metastases: impact of intraoperative sonography. AJR Am J Roent-
genol 1993;160:511–14.
[4] Machi J, Isomoto H, Kurohiji T, et al. Detection of unrecognized liver
metastases from colorectal cancers by routine use of operative ultra-
sonography. Dis Col Rectum 1986;29:405–9.
[5] Rafaelson SR, Kronborg O, Larsen C, et al. Intraoperative ultrasonog-
raphy in the detection of hepatic metastases from colorectal carci-
noma. Dis Colon Rectum 1995;38:355– 60.
[6] Rosch T, Lorenz R, Braig C, et al. Staging of pancreatic and amp-
ullary carcinoma by endoscopic ultrasonography: comparison with
conventional sonography, computed tomography, and angiography.
Gastroenterology 1992;102:188 –99.
[7] Rosch T, Lorenz R, Braig C, et al. Endoscopic ultrasonography in
diagnosis and staging of pancreatic and biliary tumors. Endosocopy
1992;24:304 – 8.
[8] Chang KJ, Katz KD, Durbin TE, et al. Endoscopic ultrasound-guided
fine needle aspiration. Gastrointest Endosc 1994;40:694 –9.
[9] Nguyen P, Feng JC, Chang KJ. Endoscopic ultrsound (EUS) and
EUS-guided fine needle aspiration (FNA) of liver lesions. Gastroin-
test Endos 1999;50:357– 61.
[10] Robinson PJ. Imaging liver metastases: current limitations and future
prospects. Br J Radiol 2000;73:234 – 41.
[11] Scholmerich J, Volk BA, Gerok W. Value and limitations of abdom-
inal ultrasound in tumour staging: liver metastasis and lymphoma.
Eur J Radiol 1987;7:243–5.
[12] Lundstedt C, Ekbert H, Hederstrom E, et al. Radiologic diagnosis of
liver metastases in colo-rectal carcinoma: prospective evaluation of
the accuracy of angiography, ultrasonography, computed tomography
and computed tomographic angiography. Acta Radiol 1987;28:
431– 8.
[13] Wernecke K, Rummeny E, Bongartz G, et al. Detection of hepatic
masses in patients with carcinoma: comparative sensitivities of
sonography, CT and MR imaging. AJR Am J Roentgenol 1991;157:
731–9.
[14] Knol JA, Marn CS, Francis IR, et al. Comparisons of dynamic
infusion and delayed computed tomography, intraoperative ultra-
sound, and palpation in the diagnosis of liver metastases. Am J Surg
1993;165:81–7.
[15] Karl RC, Choi J, Yeatman TJ, et al. Role of computed tomographic
arterial portography and intraoperative ultrasound in the evaluation of
patients for resectability of hepatic lesions. J Gastrointest Surg 1997;
1:152– 8.
[16] Ward B, Miller D, Frank J, et al. Prospective evaluation of hepatic
imaging studies in the detection of colorectal metastases: correlation
with surgical findings. Surgery 1989;105:180 –7.
[17] Albrecht T, Hoffmann CW, Schmitz SA, et al. Phase-inversion
sonography during the liver-specific late phase of contrast enhance-
ment: improved detection of liver metastases. AJR Am J Roentgenol
2001;176:1191– 8.
Discussion
Jeffrey Bender (Oklahoma City, OK): Endoscopic ul-
trasound promises to add further benefit to our patients. It is
for now, however, just that, a promise. I have one question
for the authors or rather I have one question to ask three
times. In the manuscript you mentioned that six patients
underwent alteration in management because of the endo-
scopic ultrasound findings but do not comment on what