Professional Documents
Culture Documents
CLINICAL STUDIES
Keywords Abstract
CA19-9 – diagnosis – hepatobiliary Background: Hepatobiliary cystadenomas and cystadenocarcinomas are rare
cystadenocarcinoma – hepatobiliary and often misdiagnosed. Aims: We report our experience with 33 cases over
cystadenoma – treatment 20 years to discuss an algorithm for these diseases. Methods: Patients present-
ing with a diagnosis of hepatobiliary cystadenomas and cystadenocarcinomas
were retrospectively reviewed from January 1991 to October 2010. Clinical
Correspondence
data were collected by examining hospital records and by follow-up ques-
Yilei Mao, MD, PhD, Department of Liver
Surgery, Peking Union Medical College
tionnaire interviews. Results: Thirty-three patients had pathologically diag-
Hospital, Chinese Academy of Medical Sciences
nosed hepatobiliary cystadenomas (19/33, 17 females and two males) or
& PUMC, Shuai-Fu-Yuan 1, Dongcheng District, cystadenocarcinomas (14/33, five females and nine males). Symptoms of
Beijing 100730, China cystadenomas at hospitalization were abdominal bloating or pain (9/19). Nine
Tel:186 10 65296042 patients had an elevated level of carbohydrate antigen (CA) 19-9. The surgical
Fax:186 10 65296043 procedures, i.e. cyst enucleation, segmentectomy, sectionectomy and hemi-
e-mail: sunyongliang1982@hotmail.com; hepatectomy, were performed with satisfactory outcomes. Symptoms of
maoy@public3.bta.net.cn cystadenocarcinomas included abdominal bloating or pain (8/14) and fever
(3/14). Seven patients had elevated CA19-9. The imaging characteristics of
Received 28 March 2011 cystadenocarcinomas were similar to those of cystadenomas. The clinical
Accepted 17 May 2011 outcomes for cystadenocarcinomas were mostly poor after either surgical or
conservative treatment. Conclusions: Clinical symptoms are unreliable for
DOI:10.1111/j.1478-3231.2011.02560.x these diagnoses and their differential diagnosis. Imaging evaluations and
CA19-9 are of value for the recognition of cystadenoma and cystadenocarci-
noma, but not for their differential diagnosis. Any recurrence of liver cyst after
surgery or other treatments should lead one to suspect one of these diseases.
Invasive examination and percutaneous fine-needle aspiration cytology are
not recommended. Complete excision or careful enucleation should be the
first treatment choice for a better prognosis.
Hepatobiliary cystadenomas and cystadenocarcinomas a leading teaching hospital in China, offer a pertinent
are rare neoplasms, with the first case having been review of the literature and discuss current treatment
reported in 1892 by Keen (1). They account for 1% of modalities. This series is one of the largest clinical reports
liver cystic lesions and approximately 5% of symptomatic in the published literature.
hepatic cysts (2–4). Because of their rarity and presenta-
tions similar to those of other cystic liver lesions, the
diagnosis is often delayed; this results in inappropriate Methods
treatment modalities, such as aspiration, causing unne- This study was approved by the Ethics Committee of the
cessary morbidity and mortality (5, 6). A second-stage Peking Union Medical College (PUMC) Hospital. The
operation is sometimes needed because of the misdiag- hospital records were searched for patients with diag-
nosis of the disease. Reports of these diseases have been noses of hepatobiliary cystadenoma, cystadenocarcino-
increasing because of advances in imaging techniques ma and other hepatic cysts according to the International
such as computed tomography (CT) and magnetic Classification of Diseases (Ninth Revision). Once identi-
resonance imaging (MRI) in recent years. Here, we fied from the hospital records, the pathological speci-
present our experience with 19 hepatobiliary cystadeno- mens from operative resection were retrieved and
mas and 14 cystadenocarcinomas over the last 20 years in reviewed by two specialists in gastrointestinal pathology
who were blinded to the original diagnoses; no final
Contributed equally. pathological diagnoses were changed after reviewing.
Patient records were examined for the details of demo- Table 1. Patient characteristics
graphical characteristics, medical history, symptoms, Cystadenoma
pre-operative examination, treatments including details subgroup Cystadenocarcinoma
of surgeries and patient outcomes. (n = 19) subgroup (n = 14)
Follow-up data were obtained by reviewing the hospi- Age 44.2 14.4 57.0 14.0
tal records and filling out a questionnaire via telephone Gender (M/F) 2/17 9/5
interviews. The follow-up period was defined from the Length of symptoms 18.1 25.6 29.4 43.8
date of the operation to that of the patient’s death or the (months)
last follow-up point. Deaths from other causes were HBV ( ) 1/18 0/14
treated as censored cases. The length of hospitalization Serum CA19-9 (U/ml) 838.4 2485.7 337.9 723.3
(LOH) was defined as the time from admission to Tumour size (cm) 13.0 8.1 8.3 2.9
discharge of the patients. Definitive surgery
Cyst enucleation 3 0
Fenestration 1 1
Surgery Segmentectomy 5 1
The procedure terminology of the liver resections per- Sectionectomy 5 2
Hemihepatectomy 5 4
formed in this series was standardized according the
Trilobectomy 0 1
Brisbane Terminology (7). All surgeries were performed Operating time (hours) 3.9 1.7 4.2 1.9
following a standardized operation method. Taking left Volume of blood 514.2 621.1 556.8 592.5
hemihepatectomy as an example, after dissecting the porta loss (ml)
hepatis, the left hepatic artery and left portal vein branches TACE 0 2
were transected. The left biliary duct was also transected Length of hospitalization 18.4 9.4 23.1 15.0
only if it could be clearly dissected in the hilar plate. The (days)
second porta hepatis of hepatic veins was then fully Follow-up time (months) 22.5 (0.4–79.5) 5.5 (0.2–37.7)
dissected and exposed. The left hepatic vein was ligated
F, female; HBV, hepatitis B virus; TACE, transcatheter arterial chemoem-
and amputated if the left hepatic vein and middle hepatic bolization; M, male.
vein were not in the same outflow; an electrocoagulator was
used to divide the liver tissue within a depth of 0.7 cm from two males. Symptoms upon admission had been present
the liver surface, and then a Cavitron ultrasonic surgical for 18.1 25.6 months (range, 3 days to 7 years). Nine
aspirator (CUSA, Soering GmbH Medizintechnik, Justus- had abdominal bloating or pain (range, 2 weeks to 7
von-Liebig-Ring, Quickborn, Germany) was applied after months); among them, one had a palpable abdominal
the coagulator to further dissect the liver tissue. All the mass for 3 weeks, one showed intermittent nausea/
branches of blood vessels and bile ducts in the cutting edge vomiting for 6 months and jaundice for 1 month, and
were carefully explored and ligated or clipped. The middle one was accompanied by aggravated jaundice for 2
hepatic vein was protected, and the small branches were weeks. Ten patients were asymptomatic. LOH was
closed with fine Prolene sutures. The cutting edges were 18.4 9.4 days (range, 7–49 days) (Table 1).
either left open or closed with fine sutures according to the Pre-operative evaluations, such as laboratory studies
actual conditions. Sometimes, a drainage tube was not including complete blood count and serum chemistries,
needed if the cutting edge was treated satisfactorily. were unremarkable in all patients; hepatitis virus markers
were negative in all but one patient, who was positive for
Results the hepatitis B virus. The tumour marker carbohydrate
From January 1991 to October 2010, physicians at our antigen (CA) 19-9 was tested in 11 patients; among
institution identified 33 patients with a diagnosis of them, nine (81.8%) had levels higher than those in the
hepatobiliary cystadenomas or cystadenocarcinomas. normal range (0–37 U/ml), with an average value of
The cases were divided into a cystadenoma subgroup 838.4 2485.7 U/ml (range, 0–10 000 U/ml). Other
and a cystadenocarcinoma subgroup. The 19 hepatobili- tumour markers, including carcinoembryonic antigen
ary cystadenoma patients were hospitalized from April (CEA), a-fetoprotein (AFP), CA153 and CA242, were
2002 to October 2010, while the 14 cystadenocarcinoma unremarkable.
patients were hospitalized from October 1991 to Septem- Different abdominal imaging studies were performed
ber 2009 (Table 1). The serum levels of CA19-9 in 36 before definitive treatment. Ultrasound in 16 and CT in
patients with simple hepatic cysts were also measured. 14 patients were the most commonly used imaging
They were all within the normal range, with an average modalities; MRI was performed in six patients. The
value of 11.4 8.7 U/ml (range, 3.2–20.9 U/ml). common manifestations on ultrasound included multi-
locular or unilocular hypoechoic cysts with thickened,
irregular walls; numerous hyperechoic intracystic straps;
Cystadenoma subgroup
and thin internal septations with high echogenicity
The 19 patients were 44.2 14.4 years of age at the time within the cystic masses. Similarly, CT findings revealed
of presentation (range, 19–67 years), with 17 females and low-density, well-defined, lobulated, multilocular, thick-
Fig. 1. (A) Convex papillate on the cyst wall of a well-defined, unilocular cystic mass on computed tomography (CT) scan; the convex papillate
can be enhanced with an intravenous intensifier. The final pathological diagnosis was cystadenoma. (B) Low-density, well-defined,
unilobulated, thick-walled cystic mass with internal septa on CT scan; the internal septa and wall could be enhanced with an intravenous
intensifier. The final pathological diagnosis was cystadenoma. (C) Exogenous, unilocular cystic mass with convex papillate on the thickened
wall on CT scan. The characteristics on CT scan were mostly similar to those of cystadenomas, but the final pathological diagnosis was
cystadenocarcinoma. (D) Multiloculated cystic mass located in segment VI with an irregular, thickened wall, convex papillate and internal septa;
coarse calcifications could be seen on the internal septa. The final pathological diagnosis was cystadenocarcinoma, as we suspected
pre-operatively.
walled cystic masses with internal septae and convex medial sectionectomy in one, left hemihepatectomy in
papillae. The internal septae and walls were enhanced four, right segmentectomy in one, right anterior sectio-
after the administration of an intravenous contrast (Figs nectomy in one, right posterior sectionectomy in one,
1A and B). On MRI examination, a fluid-containing, right hemihepatectomy in one (Fig. 2) and laparoscopic
multilocular, septated mass with a homogenous low- fenestration in one. Tumours averaged 13.0 8.1 cm
signal intensity was present on T1-weighted images, the (range, 2–32 cm). The cystic fluid was clear, faint yellow
wall and septa of which became enhanced after the or brown, or jelly-like. The cystic fluid CA19-9 levels
administration of gadolinium–DTPA. were tested in three patients; they were 83.42, 5000 and
The pre-operative work-up included liver-protective 4 10 000 U/ml respectively. The operative duration was
and nutritional support treatments. The liver-protective 3.9 1.7 h (range, 1.75–9 h). Blood loss during surgery
medications of Polyene phosphatidylcholine and/or ade- was 514.2 621.1 ml (range, 20–2500 ml). Ten patients
metionine 1,4-butanedisulphonate were used 3 days required no transfusion during the operative procedure
before surgery in the patients with Child–Pugh scores or subsequent hospital stay. Of the nine patients who
Z7; no patient in our series had cirrhosis or Child–Pugh required transfusion, the mean requirement was
scores 4 9. Nutritional supportive treatment was im- 6.1 3.8 units of packed red blood cells and
plemented 3 days before surgery at 83.7 kj/kg/day of 866.7 722.9 ml of plasma. Regretfully, only four pa-
energy in patients with nutritional risks after a risk tients underwent frozen-section biopsy intra-operatively;
assessment or those older than 65 years. three of them were proven hepatobiliary cystadenomas,
Five patients had previous surgical procedures that while the other patient was diagnosed with a simple
resulted in incomplete surgery or recurrences, including hepatic cyst. All of the final pathological diagnoses were
percutaneous aspiration in two, laparoscopic fenestra- hepatobiliary cystadenomas.
tion of the cyst in two, one attempted fenestration twice Follow-up information was available for all 19 pa-
and percutaneous aspiration four times. tients. The follow-up time ranged from 0.4 to 79.5
For definitive surgeries, cyst enucleation was per- months, with a median follow-up time of 22.5 months.
formed in three, left segmentectomy in two, left biseg- Two patients were treated by percutaneous drainage
mentectomy in two, left lateral sectionectomy in two, left because of post-operative biloma. One patient had a
Fig. 2. Procedures of the operation of a patient with right hepatic cystadenoma who underwent right hemihepatectomy. (A) Pre-operative
computed tomography (CT) scan that showed a low-density, well-defined, unilobulated, thick-walled cystic mass with internal septa; the
internal septa and wall could be enhanced with an intravenous intensifier. (B) Overall appearance of the cystadenoma after laparotomy.
(C) Dissection of the portal triad. (D) Dissection of the right hepatic vein in the second porta hepatis. (E) Hepatic cut surface after right
hemihepatectomy. (F) Macroscopic appearance of the cystadenoma after surgery; the final pathological diagnosis was cystadenoma.
laparoscopic fenestration, and the intra-operative frozen- 2 months to 10 years), and among them, three were
section biopsy revealed a simple hepatic cyst; however, accompanied by fever (20 days, 9 months and 12 months
she refused to undergo a second resection after the final respectively). Three patients were asymptomatic. LOH
pathological diagnosis of hepatobiliary cystadenoma and was 23.1 15.0 days (range, 10–69 days) (Table 1).
showed no recurrence after 14.4 months of follow-up. The pre-operative blood evaluations were unremark-
Eight months after left lobectomy, one patient died of able, as in the cystadenoma subgroup; hepatitis virus
widespread metastasis of a malignant tumour despite an markers were negative in all patients. Tumour marker
original pathological diagnosis of hepatobiliary cystade- CA19-9 levels were tested in 11 patients. Seven (63.6%)
noma combined with moderate dysplasia. The other 17 had values higher than those of the normal range, and
patients showed no recurrence or clinical symptoms the average value was 337.9 723.3 U/ml (range,
upon our last follow-up (Fig. 3). 4.8–2461 U/ml). Other tumour markers, such as AFP,
CA153 and CA242, were also unremarkable.
Abdominal imaging evaluations included ultrasound
Cystadenocarcinoma subgroup
in 10, CT in 10, positron emission tomography (PET) in
The 14 patients were 57.0 14.0 years of age at the time one, nuclear medicine uptake study in one and angio-
of presentation (range, 26–73 years), with five females graphy in two. The characteristics of the cystadenocarci-
and nine males. Symptoms upon admission had been nomas on imaging evaluations were mostly similar to
present for 29.4 43.8 months (range, 20 days to 10 those of cystadenomas (Fig. 1C). Rarely, capsular, mural
years). Eight had abdominal bloating or pain (range, or septal coarse calcifications could be seen on
Discussion Table 2. Serum CA19-9 in patients with simple hepatic cyst, carbo-
hydrate antigen and cystadenocarcinoma
Hepatobiliary cystadenomas and cystadenocarcinomas Cystadeno-
may occur at any age. Approximately 85–95% of those Cystadenoma carcinoma
affected are women (8, 9). However, our cystadenocarci- subgroup subgroup
noma subgroup did not show the same female tendency Serum (n = 11) (n = 11) P-value
(9/14 cystadenocarcinomas were in males), as did other CA19-9 838.4 2485.7 337.9 723.3 0.735
reports. The cases reported so far may be so limited that Simple hepatic 11.4 8.7
the gender tendency could not be clarified. The appear- cyst (n = 36)
ance of the initial manifestations of the lesions may take P-value 0.0001 0.001
years or the peak frequency was in the fifth decade of life
(10, 11). The ages at presentation in our cystadenocarci-
noma subgroup were older than those of our cystadeno-
ma subgroup (57.0 14.0 years vs. 44.2 14.4 years cystic lesions, but not between cystadenoma and cysta-
respectively; P o 0.05). This phenomenon further sup- denocarcinoma, or other cancers originating from the
ports the hypothesis of malignant transformation of biliary tract. Whereas the level of CA19-9 in the cystic
cystadenomas to cystadenocarcinomas. fluid may be invaluable in differential diagnoses among
Most of our cases were identified over the past 10 liver cystic lesions, it may also be significantly elevated in
years, especially in the past 5 years. This may be because the fluid of a simple hepatic cyst in our experience and in
of the advances and commonly applied tools in favour the published literature (16–18).
of detection, as well as pathological awareness of the In addition, CT is less sensitive than ultrasound in
diseases. visualizing intracystic blood clots for differential diag-
The exact aetiology of hepatobiliary cystadenomas nosis. Although the imaging characteristics of hepato-
remains unknown. Both congenital and acquired causes biliary cystadenoma and cystadenocarcinoma have been
have been proposed. It is uncertain whether biliary recognized as we described in our patients, the differen-
cystadenocarcinomas are de novo cancers or are derived tial diagnosis between them on the basis of imaging alone
from cystadenomas. They are generally thought, how- has appeared to be difficult (17, 19, 20, 21). CT or MRI
ever, to arise from pre-existing benign cystadenomas can delineate the anatomic relations within the liver that
because benign epithelium has been demonstrated in may be helpful in planning the surgical procedure (22).
over 90% of cystadenocarcinomas. Many cystadenocar- Intra-operative frozen sections may be more helpful than
cinomas contain areas of cystadenoma in the same imaging characteristics in the differentiation of cystade-
sample. nomas and cystadenocarcinomas from other cystic hepa-
The most typical clinical manifestations in the litera- tic lesions. Careful histopathological evaluation of the
ture are abdominal discomfort or pain, which were resected specimen is the only safe diagnostic modality,
present in 51.5% (17/33) of our patients. Asymptomatic while malignant degeneration or transformation of cy-
patients comprised 39.4% (13/33) in our series. Other stadenomas can only be detected after thorough section-
non-specific manifestations such as jaundice and fever ing (23, 24). Histologically, both cystadenomas and
may be associated with biliary obstruction, rupture, cystadenocarcinomas have been divided into two sub-
bacterial infection or intracystic haemorrhage (12, 13). groups according to the presence of mesenchymal (ovar-
Thus, clinical manifestations are considered unreliable ian-like) stroma between the inner epithelial lining and
for the diagnosis. We believe that any patient presenting the outer connective tissue capsule (2, 11, 23). We also
with recurrence of their liver cyst after surgery or other consider that percutaneous fine-needle aspiration cytol-
treatments should be suspected of having a cystadenoma ogy should not be generally recommended in cases of
or cystadenocarcinoma. Significantly increased levels of overt or suspected cystadenocarcinoma, particularly
serum CA19-9 can be encountered in cystadenomas and when surgery is planned. The characteristic of the fluid
cystadenocarcinomas (10, 14, 15). This was also true in content of the lesion is not meaningful to diagnosis in
our series, with nine of 11 patients with cystadenomas our experience. Clear, haemorrhagical, faint yellow,
having increased CA19-9 values (838.4 2485.7 U/ml) bilious and mixed fluid contents were observed in our
and seven of 11 patients with cystadenocarcinomas series.
having increased CA19-9 values (337.9 723.3 U/ml). Because distinguishing between cystadenoma and
Other tumour markers, such as CEA, AFP, CA153 and cystadenocarcinoma pre-operatively and determining its
CA242, were unremarkable in both subgroups. We malignant potential are difficult, we suggest that total
compared serum CA19-9 in simple cysts, cystadenomas excision is ideal, may offer the best chances of cure and
and cystadenocarcinomas subgroups in order to evaluate might be the only definitive therapy. Surgery also allows a
the role of CA19-9 in differentiating diagnosis (Table 2). proper histological examination of the cyst and a detailed
Based on the data from the comparisons, we conclude classification. Once cystadenoma or cystadenocarcinoma
that the measurement of serum CA19-9 could thus be is diagnosed regardless of what the results are from the
helpful in the differential diagnosis of simple hepatic intra-operative frozen section or/and the final pathology,
18. Emre A, Serin KR, Ozden I, et al. Intrahepatic biliary cystic 23. Devaney K, Goodman ZD, Ishak KG. Hepatobiliary cysta-
neoplasms: Surgical results of 9 patients and literature denoma and cystadenocarcinoma. A light microscopic and
review. World J Gastroenterol 2011; 21: 361–5. immunohistochemical study of 70 patients. Am J Surg
19. Hai S, Hirohashi K, Uenishi T, et al. Surgical management of Pathol 1994; 18: 1078–91.
cystic hepatic neoplasms. J Gastroenterol 2003; 38: 759–64. 24. Del Poggio P, Jamoletti C, Forloni B, et al. Malignant
20. Lewin M, Mourra N, Honigman I, et al. Assessment of MRI transformation of biliary cystadenoma: a difficult diagno-
and MRCP in diagnosis of biliary cystadenoma and cysta- sis. Dig Liver Dis 2000; 32: 733–6.
denocarcinoma. Eur Radiol 2006; 16: 407–13. 25. Yu J, Wang Y, Yu X, et al. Hepatobiliary mucinous
21. Lim JH, Jang KT, Rhim H, et al. Biliary cystic intraductal cystadenoma and cystadenocarcinoma: report of six cases
papillary mucinous tumor and cystadenoma/cystadenocarci- and review of the literature. Hepatogastroenterology 2010;
noma: differentiation by CT. Abdom Imaging 2007; 32: 644–51. 57: 451–5.
22. Matsumoto S, Miyake H, Mori H. Case report: biliary 26. Aijun LI, Mengchao WU, Weiping Z, et al. Diagnosis and
cystadenoma with mucin-secretion mimicking a simple Treatment of Liver Cystadenocarcinoma: Report of 18
hepatic cyst. Clin Radiol 1997; 52: 316–8. Cases. Chin. Ger. J Clin Oncol 2005; 4: 267–70.