Efficacy of a Surveillance Program for Early Detection
of Hepatocellular Carcinoma
Marco Zoli, M.D.
Donatella Magalotti, M.D.
Giarnpaolo Bianchi, M.O.
Cristina Gueli, M.D.
Giulio Marchesini, M.D.
Emilio Pisi, M.D.
lstituto di Clinica Medica Generale e Terapia
Medica. Uni'versity of Bologna, Bologna, Italy.
Supported by a grant from Assessorato alla
Sanita. Regione Eniilia Romagna, Italy.
Address for reprints: Marco Zoli, M.D., Clinica
Medica (I, Policlini8:o S. Orsola, Via Massarenti
9. 40138--Bologna, Italy.
Received February 5, 1996; revision received
May 9, 1996; accepted May 9, 1996.
BACKGROUND. Contrasting data have so far been reported on the utility and efficacy
of screening patients with cirrhosis for early detection of hepatocellular carcinoma
(HCC). The goal of this study was to evaluate the efficacy of a regular ultrasono-
graphic and laboratory follow-up for the early detection of small HCC, and to
identifp parameters correlated with a higher risk of developing HCC.
METHODS. One hundred and sixty-Sour consecutive patients with liver cirrhosis
living in Emilia Romagna, Italy, were enrolled in the period 1989-1991. All patients
underwent clinical, biochemical, and ultrasonographic evaluations at entry and at
3- and 6-month intervals during follow-up.
RESULTS. By April 1995, 34 patients had developed HCC. In 76% of the patients.
ultrasonography identified HCC when it was still single and small (14 cm). At
discriminant, logistic regression and univariate analyses, sex and the entry concen-
tration of alkaline phosphatase, a-fetoprotein, gamma-glutamyl transpeptidase,
and albumin were associated with a higher risk of developing HCC, whereas at
multivariate analysis (Cox's model), only sex and the entry concentration of alkaline
phosphatase, albumin, and a-fetoprotein were independently and significantly
related to the appearance of HCC.
CONCLUSIONS. A regular ultrasonogrpahic follow-up, timed at 3- to 6-month inter-
vals according to the risk of HCC development in patients with cirrhosis, allows
the detection of liver carcinoma at an early stage in a high proportion of patients,
possibly improving the prognosis of the disease. Cancer 1996; 78:977-85.
0 1996 American Cancer Society.
KEYWORDS: a-fetoprotein, hepatocellular carcinoma, cirrhosis, ultrasonography.
H ferent worldwide distribution. It is common only throughout
Sub-Saharian Africa and the Far East (incidence: 30 cases per
100,000 per year). In northern European countries, the incidence is
lower (3 cases per 100,000) in comparison with southern European
countries, in which the incidence is considered intermediate (6-
10 cases per 100,000).'-3 The different geographic distribution
seems to be related to the prevalence of factors that carry a risk
for the disease, namely hepatitis viruses (B and C), which are fre-
quently associated with HCC. In Italy, the high prevalence of hepa-
titis B and C viral infection is the likely cause for the high prevalence
of HCC." In addition, cirrhosis is associated with HCC in greater
than 80% of patients, in both high and low incidence area^.',^,^
Patients with cirrhosis, irrespective of the etiology of the disease,
are at risk for HCC, which is one of the leading causes of death in
Several prospective studies have been performed on the inci-
dence of HCC in cirrhotic patients and on the importance of observ-
epatocellular carcinoma (HCC) is a malignant disease with dif-
CI 1996 American Cancer Society
978 CANCER September 1,1996 / Volume 78 / Number 5
ing patients with chronic liver disease for the early
diagnosis of HCC. Contrasting data have so far been
reported on the utility and efficacy of a strict clinical
and ultrasonographic (US) follow-up in the diagnosis
of HCC at an early stage.”“ Some authors did not find
follow-up useful for early diagnosis, probably because
it was not regularly conducted.8.”.’2 Other authors dis-
agreed on the real utility of early HCC diagnosis in
relation to the poor results obtained by currently avail-
The goal of this study was: 1) to evaluate the effi-
cacy of a regular follow-up for the detection of HCC
at an early stage; 2) to prospectively assess the inci-
dence of HCC in a group of cirrhotic patients living in
Emilia Romagna, an area in the north of Italy; and 3)
to identify clinical, biochemical, and US parameters
carrying a higher risk of HCC development.
PATIENTS AND METHODS
Between January 1989 and December 1991, all patients
with biopsy-proven liver cirrhosis, attending the Insti-
tute of Clinica Medica Generale e Terapia Medica of
the University of Bologna, underwent US examination
of the upper abdomen. The patients who met the fol-
lowing criteria were enrolled in the present study: 1)
satisfactory or good US visualization of liver paren-
chyma; 2) no space-occupying liver lesions at US; 3)
willing to cooperate by visiting the Institute at sched-
uled intervals; and 4) living in Emilia Romagna, Italy.
Patients coming from other areas were not included
to avoid drop-outs.
One hundred sixty four cirrhotic patients (75
males and 89 females) were included. Their age at
entry ranged from 23 to 84 years (median, 61 years).
The etiology of their liver disease is reported in Ta-
The rejected log included 42 patients living out-
side Emilia Romagna, 8 patients with poor US visual-
ization of hepatic parenchyma, and 12 patients with
HCC at first US examination.
All patients selected for the study gave their in-
formed consent to participate. No ethical committee
was operating in our department. The protocol was
submitted to and approved by the Technical Commit-
tee for Public Health of Emilia Romagna, Italy.
Data Collected at Entry
At entry, all patients underwent a complete clinical
examination and the following clinical and biochemi-
cal parameters were recorded for statistical purposes:
etiology of liver disease, ascites, encephalopathy, and
the plasma concentration of albumin, prothrombin
time, bilirubin, alkaline phosphatase (AP), gamma-
Etiology of Liver Cirrhosis in the 164 Patients Enrolled in the Study
HBV t HCV
Alcohol t HBV
Alcohol t HCV
Alcohol t HBV t HCV
HCC: hepatocellular carcinoma: HBV: hepatitis B virus: HCV: hepatitis C iirus; PBC: primary biliary
Clinical, Biochemical and Ultrasound Parameters in the 164 Cirrhotic
Patients at Entry Into the Study (Mean -t SD [Range] or
Number of Patients)
Patients free of
60 t 9 123-841
3.74 5 0.52
63 Z 6 146-761
3.66 ? 0.43
59 t I0 123-841
3.76 ? 0.55
Prothrombin time (%I
Total bilirubin (mgidl)
74 t 15
1.5 ? 1.4
71 2 12
1.5 t 1
75 i- 16
1.5 2 1.5
Child-Pugh class iA1BiC)
Alkaline phosphatase (UIL)
Echo pattern (homin-hom)
273 5 140
108 t 127
10 i- 35
6 2 1.4
318 t 125
103 t i9
6 t 1.2
262 t 142‘
109 t 137
15 i- 2gb
HCC: hepatocellular carinoma; hl: male; F: female; GGT: gamma-glutnnivl transprptidase: ham: homog-
enous; n-ham: nonhomogeneous.
a tiraded according to the Child-Pugh score.
” Significantly different from patients who developed hepatocellular carcinoma iP < 0.051.
glutamyl transpeptidase (GGT), and a-fetoprotein (a-
FP). The Child-Pugh score” was also calculated (Ta-
All US examinations were performed by two expert
operators by means of high-resolution real-time
equipments (Aloka SSD-280, Esaote Hitachi AU-590)
with both a linear and a sectoriaUconvex 3.5 mega-
hertz probe. The following US parameters were regis-
Early Detection of HCC/Zoli et at.
tered: quality of liver visualization: satisfactory or
good; ascites: absent, mild, moderate or severe; and
liver echo-pat tern: homogeneous or nonhomoge-
neous without evidence of space-occupying lesions.
All patients were prospectively examined every 6
months by means of physical examination, biochemi-
cal assessmenr, and US.
At each control, the clinical, laboratory, and IJS
parameters considered at entry were registered. IJS
examinations were performed by the same operators.
The 6-month interval was reduced to a 3-month inter-
val if a-FP was greater than 5 times the upper normal
limit and if the liver echo-pattern was classified as
In the presence of space-occupying lesions, the
following US parameters were also registered: number,
size, volume, echogenicity, localization of the lesion,
and the presence/absence of portal vein thrombosis.
'I'hroughout the study period, patients received
treatments to avoid or to treat complications of their
liver disease, such as lactulose and/or diuretics.
Thirty-four patients received 0-blockers or underwent
endoscopic sclerotherapy as primary or secondary
prophylaxis of gastrointestinal bleeding.
Patients were observed until the appearance of
HCC, death, or until April 1995.
1)atiunts with uncertain liver echo patterns were
further examined by means of CT and/or hepatic angi-
ography. However, the final diagnosis of HCC was
based on IJS-guided biopsy of the lesions and Lipiodol
CT was used as the gold standard to confirm HCC
Patients with I ICC were treated with hepatic re-
section (HR), whenever it was considered possible by
the surgical team, transcatheter arterial chemoemboli-
zation ('I'ACH), or percutaneous ethanol injection
(PHI, according to several well defined protocols
based on the patient's clinical status and consensus.
They were observed until death or November 1995.
Data arc reported as mean ? standard deviation (SD)
0 1 median and ranges.
Discx-iminant analysis and logistic regression were
used to identify parameters able to predict or corre-
lated with HCC appearance.
The time-course of continuous variables between
patients who developed HCC and those who remained
tumor free in the course of the study were analyzed
by means of two-factor repeated measures analysis of
To evaluate the prognostic value of significant pa-
Univariate Analysis of Clinical, Biochemical, and Ultrasound
Parameters Relative to the Appearance of Hepatocellular Carcinoma
patients Variables Cutoff points
Prothronihin time (%I
Iota1 bilirubin (mgidl.]
Alkaline phosphatase (Uil.)
Alcohol T Lira1
= . 70
> . i O
0 . w 1 1
IiK hepa~occllular rarcincinia: CG I garnma.giuiam~i-rrans~~ep~idase,
lin1i1: honicigenou,: n-horn.
rameters in relation to time, univariate survival analy-
sis was performed by means of the Kaplan-Meier
method.I6 Plots were constructed clustering continu-
ous variables as given in Table 3.
To establish whether data obtained by univariate
analysis might have independent prognostic value, the
multivariate regression analysis proposed by Cox",'"
was performed by the forward stepwise procedure. In
this model, the entry order of any variable was deter-
mined by using the maximum log likelihood value and
the statistical significance was assessed by the Wald
The RMDP statistical package was used for all
The etiology of liver disease is reported in Table 1,
whereas Table 2 shows the clinical, biochemical, and
US parameters of the 164 cirrhotic patients at time of
entry into the study. In 41 patients (250/0), the cause
of liver cirrhosis was rnultifactorial.
CANCER September 1,1996 / Volume 78 / Number 5
Clinical Features of the 34 Patients at the Time of Hepatocellular Carcinoma Diagnosis
thrombosis Age Sex Etiology
C t Alcohol
B t C
B t C t Alcohol
B t C t Alcohol
B t C
C t Alcohol
B t C
B t C
B t C
B t Alcohol
Ir-FP: rr-fetoprotein: M: male; F: female; C: hepatitis C virus; 6: hepatitis B virus; PBC: primary biliaq cirrhosis.
At entry, the US visualization of liver parenchyma
was good in 133 patients and satisfactory in 31, due
either to meteorism and/or adipose tissue and/or asci-
During the follow-up period, which ranged from
7 to 77 months, HCC was identified in 34 patients
(Table 4). The cumulative rate of tumor free survival
among the 164 cirrhotic patients is reported in Fig-
Twenty-eight patients were lost during follow-up:
22 died of complications of liver disease (i.e., liver fail-
ure, or gastrointestinal bleeding) and 2 patients under-
went liver transplantation, whereas 4 patients died of
causes not related to cirrhosis.
Among the 34 patients who developed liver can-
cer, HCC appeared within 1 year in 8 patients, within
2 years in 11 patients, and within 3 years in 5 patients.
The overall incidence of HCC was 4.9% during the first
year, 7.7% during the second year, 4.8% during the
third year, and 7.4% during the fourth year. These val-
ues were weighted for the decrease in the number of
patients during follow-up at the times of evaluation.
The cumulative hazard of HCC increased from 3% at
1 year, to 10% at 2 years, to 15% at 3 years, and to
24% at 4 years.
The diagnosis of HCC was first obtained by means
of US in all patients. Liver carcinom was unifocal in
26 patients, bifocal in 4 patients, and trifocal in 1 pa-
tient, in 3 patients, portal vein thrombosis was already
present at first diagnosis. Among these last 8 patients,
Early Detection of HCClZoli et al.
PL.. at rbk:
500 1000 1500
85 5 . 3
FIGURE 1. Cumulative percentage (Kaplan-Meier) of patients alive and
free of hepatocellular carcinoma. HCC: hepatocellular carcinoma; Pts: pa-
in whom HCC was multifocal or complicated at first
diagnosis, the US interval period had been reduced to
3 months before cancer diagnosis in 4 patients,
whereas a regular &month interval had been main-
tained in the remaining 4 patients.
In those individuals with unifocal tumors, HCC
was localized in the right lobe in 18 patients and in
the left lobe in 8 patients. In the multifocal lesions,
HCC was in the same lobe in one patient and in both
lobes in lour patients. The unifocal tumors had, at first
(IS visualization, a maximum diameter ranging from
0.8 to 4 cm, whereas in multifocal lesions, the maxi-
mum diameter of the single nodules ranged between
0.4 and 3 cm. 1 ti the 3 patients with portal vein throm-
bosis, US showed a partial thrombosis of the left
branch of the portal vein together with a 1.6-cm space-
occupying lesion of the fourth segment of the liver in
1 patient (Patient 2), a partial thrombosis of the right
portal branch together with a 3.8-cm lesion of the sixth
segment in thc second patient (Patient 141, and a par-
tial thrombosis of the right portal branch without any
clear space-occupying lesion of the hepatic paren-
chyma in the third patient (Patient 15). In the last
two patients, Lipiodol-CI showed the presence of a
At the time of HCC diagnosis, a-FP was within the
normal range (<20 nglmI.) in 13 patients whereas in
4 patients it was greater than 5 times the upper normal
limit. I n the 3 patients with portal vein thrombosis, u-
FP was 72, 80, and 122 nglmL, respectively.
Among the clinical and the biochemical parame-
ters at entry, only AP and a-FP were significantly dif-
ferent between patients who developed HCC and pa-
tients who remained tumor free at the end of follow-
up (Table 2). When data of the last control before HCC
appearance or before the end of follow-up were ana-
lyzed, only a-FP was significantly different between
the two group (41 z 64 nglml. vs. 16 -e 28; P = 0.018).
Prognostic Variables in Cox’s Regression Model
AP: alkaline phosphaiase; u-FP: o-feloproreiii.
Discriminant analysis showed that male sex and
AI’ at entry were able to identify patients who devel-
oped HCC, whereas u-FP and sex appeared to be sig-
nificant predictors when data regarding the last con-
trol before HCC appearance or before the end of fol-
low-up were considered.
By means of logistic regression analysis, sex and
entry values of AP, a-FP, and GGT appeared to be
related to HCC development, whereas at the last con-
trol a-FP, sex, Pi‘, and AP were found to be significant
Repeated measures (ANOVA) showed that none of
the investigated parameters varied significantly during
the study period in relation to the appearance of IICC.
At univariate analysis (Table 31, only three vari-
ables were significantly associated with the probability
of developing liver carcinoma. The probability was sig-
nificantly higher in males, with albumin < 3.5 g/dL
and a-FP > 20 nglmt.. The levels of AP were slightly
significant. The etiology of liver disease was not statis-
tically significant, however analyzed (i.e., viral [B + C]
vs. not viral, P = 0.398; HCV-related vs. all other cases,
P = 0.2104).
Sex, albumin, bilirubin, PI’, AP, a-FP, and liver
echo pattern were then included in a multivariate
analysis (Cox’s model). Only sex, AP, albumin, and
a-FP appeared to be independently and significantly
related to the appearance of 11CC (Table 5).
Data regarding treatment and outcome of the 34
patients with HCC are reported in Table 6. Thirteen
patients underwent TACF., 4 underwent PBI, and 2 un-
derwent HR, whereas 14 patients were not treated or
were treated only with oral tamoxifen because of poor
clinical status or because they refused any aggressive
treatment. One patient underwent liver transplanta-
tion; while the patient was awaiting transplantation,
one cycle of TACE was performed.
At the end of the observation period, 14 patients
with HCC were still alive (41% of the total patients
who developed HCC), in spite of liver carcinoma being
diagnosed by 7 to 60 months. The cause of death was
982 CANCER September 1,1996 / Volume 78 / Number 5
Treatment and Outcome of the 34 Patients with Hepatocellular
up (mos) Treatmenl
Cause of death
TACE t LT
1.F t HCC
GB t HCC
LF t HCC
LF t HCC
TACF: transcatheter arterial chernoembolizatioo; PEI: percutaneous ethanol injection; HR: heparic
rcseclion; Tam: oral ramoxifen; LT: liver transplantation; HCC: hepatocellular carcinoma expansion;
L F liver failure: GB: eastrointestinal bleeding.
related to complications of liver cirrhosis in 11 pa-
tients, whereas in 8 patients death was due to tumor
expansion. One patient died of hernoperitoneum due
to diffusion of a second cancer originating from the
In the course of follow-up, a-FP levels increased
to more than 20 ng/mL in 6 of the 13 patients with
normal levels at diagnosis. Five of these patients died
of liver failure and/or disseminated HCC; only 1 sur-
vived 23 months. In the remaining seven patients, a-
FP remained in the normal range throughout the ob-
servation period (four patients survived, three died).
Among the 8 patients with multifocal or compli-
cated HCC at diagnosis, 4 died within 1 year, 2 died
after 3 years, and 2 were alive after 12 and 18 months,
The general goal of any screening procedure is to de-
tect tumors when they are still isolated and small in
size, and thus eligible for effective treatment, namely
surgery. The current study shows that a regular follow-
up of cirrhotic patients, based on clinical, biochemical,
and US evaluations, allows in most instances the de-
tection of HCC at an early stage. According to our
previous experience, in nonregularly screened patients
with cirrhosis, HCC may be diagnosed for the first time
when it is as large as 12 cm in diameter.
HCC is a malignant tumor that, in most cases de-
scribed worldwide (even in high incidence areas), oc-
curs in cirrhotic liver.',3," This consideration supports
the hypothesis that cirrhosis per se, independent of
etiology, is an important oncogenic factor, as con-
firmed by data on the recurrence of HCC after surgical
resection; some authors reported a rate of occurrence
between 75-100%.'0 Even in our experience, HCC in
normal liver is rare. In the period between 1989 and
1991, approximately 12,000 individuals were examined
by US in our clinic and in only one was HCC found
in the absence of liver cirrhosis, thus confirming that
in our area, the development of HCC is part of the
natural history of cirrhosis.
Several studies prospectively screened cirrhotic
patients to evaluate the role of laboratory and US fol-
low-up in detecting HCC at an early stage.""
results of these studies were extremely variable; in a
few studies HCC was only diagnosed when it was
larger than 5 cm or was already rnultifocal."'"l2 These
disappointing results were probably due to the proto-
cols used for follow-up, which were based either on
very high titres of a-FP (grater than 400 ng/mL) or on
a 1-year interval in US examination. On the basis of
poor results obtained by treatment of HCC, the same
authors concluded that the cost/benefit of a screen-
ing/follow-up of cirrhotic patients was too high to be
justified because there was no evidence that this might
increase the detection rate of operable tumors, poten-
tially reducing the HCC mortality rate."."-'4
Emilia Romagna is an area of northern Italy in
which liver cirrhosis has a multifactorial origin, both
viral and alcoholic. Within a program for tumor pre-
vention, supported by the Public Health Department
of Emilia Romagna-Italy,
of patients with liver cirrhosis. The screening was
based on simple clinical and laboratory parameters
and on an accurate US examination of the abdomen
performed by expert operators by means of high-reso-
lution real-time equipments. US was performed at 6-
we scheduled a follow-up
Early Detection of HCClZoli et al.
month intervals but the interval period was reduced
to 3 months when ru-FI' increased or when the liver
echo pattern appeared clearly non-homogenous even
in the ahsence of definite space-occupying lesions.
Iluring follow-up, 34 of the 164 cirrhotic patients
developed HCC. The incidence was similar to that re-
ported in other studies performed in Italy and in west-
ern European (:ountries;' but in our study, IICC was
detected in 76% of patients when it was single and
small. In only five patients was HCC already multifocal
at first U S diagnosis and in three patients portal vein
thrombosis was already present. Portal vein thrombo-
sis is a negative prognostic factor, as observed in Pa-
tients 14 and 15, who died of HCC expansion within
few months. These two last patients underline that US
has a low diagnostic accuracy in the early detection
of rnicroinvasive IiCC, hut these are very specific cases
in which an early diagnosis is not likely to modify the
very poor prognosis.
Our data showed that male sex, reduced levels of
albumin and fY, and increased levels of rw-FP, AP, and
GGI' were indicators of a higher probability of devel-
oping liver cancer, whereas the multivariate analysis,
by means of the Cox's model, identified male sex, to-
gether with AP. albun~in, and rr-FP, as significant inde-
pendcnl predictors of the risk of developing HCC. The
combination of these variables was therefore able to
identify a subgroup of cirrhotic patients at a higher
risk of tumor development.
A higher incidence of I-ICC in males has already
becm demonstrated in all previous studies,'-.' whereas
only a few reports recognized the prognostic role of
AP." 'fhe observation that the worsening of liver func-
tion parameters, stich as reduced PI' and albumin,
map be predictive of 1 ICC appearance might support
the hypothesis that H K develops mainly in an ad-
vanced stage of chronic liver disease. However, it can-
not be exclutled that reduced I T and albumin, seen
at the last control, may be an early sign of a developing
tumor, relatively independent of the stage of disease.
Also, the relationship between the duration of liver
cirrhosis and 1 ICC occurrence, shown by several stud-
cannot be taken as definite proof of an associa-
tion between advanced disease and cancer. In patients
with slowly progressive disease, the rate of occurence
of I ICC mighi. also be high in the presence of relatively
preserved liver function because of long-lasting dis-
ease, whereas patients with rapidly progressive disease
die of liver failure before developing HCC. This would
explain why most of our patients developing HCC: are
in Chiltl-Pug11 Class A at entry.
a-I;P at entry was mildly but statistically increased
in pationts who developed cancer, but in only four
patients did it exceed five times the upper normal lim-
its at the time of HCC diagnosis. The finding that a-
FP is selected by univariate and multivariate analyses
as a risk factor for IICC development means also that
a moderate increase in a-FP must he considered a
risk factor. Unfortunately, in individual patients, a-FI'
levels cannot be given large clinical relevance, due to
the low sensitivity of the parameter for the early diag-
nosis of HCCH I" In the course of follow-up, rr-FP lev-
els increased to more than 10 times upper normal
limits (>ZOO ng/mI.) in 11 patients with IICC before
outcome, and in none of the patients who did not
develop HCC. This confirms that in approximately 70%
of HCC patients, a-FP does not increase to values use-
ful for diagnostic purposes, possibly in relation to early
diagnosis, to factors intrinsic to tumor biology, and/
or to the mass reductive effects of specific treatments.
a-FP maintains a high specificity, but the marker ap-
pears to be of no use in the early detection of small
I 1 CC .2.'
In the current series, the viral etiology of liver cir-
rhosis was not related to a significant increase in the
risk of I ICC, whereas previous studies showed a corre-
lation between hepatitis C, virus infection and HCC."".'"
A study utilizing a greater number of patients is proba-
bly necessary to confirm this linkage. According to our
data the incidence of HCC was 26% in hepatitis C virus
positive patients versus 16% in nonhepatitis C virus
In spite of the successful detection of I1CC at a
very early stage, effective removal of the lesion was
achieved in only two patients. 'The low number of re-
sections was mainly due to high surgical risk for ad-
vanced cirrhosis, and, in some cases, to patient refusal
to undergo surgery and to tumor site. However, 17
patients could be treated by TACE or PEI. and in the
final study period, a single patient was treated by
There is currently much controversy regarding the
best approach to the treatment of IICC. The results
obtained by different authors are sometimes con-
trasting, and few comparative prospective randomized
clinical trials have been performed.2"- '!' Also, 'I'ACX,
which appeared effective in nonrandoniized clinical
trials,:'" failed to increase the survival rate significantly
when tested in a randomized clinical trial.:" The nega-
tive results were related mainly to the high incidence
of recurrence of HCC in the other segments of the
liver. This further supports the view that cirrhosis per
se is a preneoplastic lesion that is not modified by
these treatments. Other important causes of death in
these patients are complications of liver cirrhosis. Also,
in our experience in early diagnosed IICC, 11 patients
died before cancer expansion, deaths were due to liver
CANCER September 1, 1996 / Volume 78 / Number 5
failure in 7 patients and to gastrointestinal bleeding
in 4 patients.
Liver transplantation could therefore be the ideal
treatment of HCC because it removes both the cancer
and the diseased, preneoplastic cirrhotic liver, but it
was abandoned due to the high incidence of HCC re-
currence promoted by immunosuppressive ther-
apy,3L,33 which adds to the limited number of available
organs and to the high total cost (approximately
$125,000 in Italy). In the last years, it has been per-
formed again in decompensated cirrhotic patients
with small HCC, and recent studies have demon-
strated that posttransplantation tumor free survival is
good if patients are well ~elected.~""' HCC should be
single, small in size (less than 3-5 cm), and without
lymph node spread and portal vein invasion. These
results give new stimulus to efforts to detect HCC at
an early stage and the current study shows that an
accurate US examination allows for detection of a
group of cirrhotic patients who may benefit frotn liver
transplantation to radically treat both HCC and de-
compensated liver cirrhosis. In these patients, TACE
andlor PEI may be proposed while they are on the
waiting list for a transplant.
In conclusion, real-time US is a very effective way
to observe cirrhotic patients at risk of HCC. When per-
formed at scheduled intervals, on the basis of the risk
probability, it is definitely the most cost-effective tech-
nique, and allows the detection of liver carcinoma at
an early stage in a high proportion of patients, possibly
improving the prognosis of the disease.
Editorial. Hepatocellular cancer: differences between high
and low incidence regions. Lancet 1987;ii:1183-4.
Okuda K. Epidemiology and clinical aspects of hepatocellu-
lar carcinoma. J Gustroenterol Heparol 1993;8:S1-4.
Colombo M. Epidemiology of hepatocellular carcinoma. Ztnl
J Gastroenterol 1995;27:83-7.
Kew MC, Popper H. Relationship between hepatocellular
carcinoma and cirrhosis. Seiniri Liver Dis 1984;4:136-46.
Tretnolada F, Benvegnh L, Drago C, Casarini C, Cecchetto
A, Realdi G, et al. Early detection of hepatocellular carci-
noma in patients with cirrhosis by alphafetoprotein, ultra-
sound and fine-needle biopsy. Heputogastroelitero/ogy
Oka H, Kurioka N. Kim K, Kanno T, Kuroki T, Mizoguchi Y,
et al. Prospective study of early detection of hepatocellular
carcinoma in patients with cirrhosis. Hepatolog,] 1990;
Tanaka S, Kitamura T, Katsumi N, Okuda S, Yamazaki H,
Hiyama T, et al. Effectiveness of periodic checkup by ultraso-
nography for the early diagnosis of hepatocellular carci-
noma. Cancer 1990;66:2210-4.
Colombo M. De Franchis R, Del Ninno E, Sangiovanni A,
De Fazio C, Tommasini M, et al. A prospective study of the
development of hepatocellular carcinoma in Italian patients
with cirrhosis. N EnglJ Merl 1991;325:675-80.
Cottone M, Turri M, Parisi P, Orlando A, Fiorentino G, Vir-
done R, et al. Screening for hepatocellular carcinoma in pa-
tients with Child's A cirrhosis: an 8-year prospective study
by ultrasound and alphafetoprotein. J Heputol 1994;21:
Oka H, Tamori A, Kobayashi K, Yamamoto S. Prospective
study of a-fetoprotein in cirrhotic patients monitored for
development of hepatocellular carcinoma. Hepatology
Pateron D. Game N, Trichet JC, Aurousseau MH, Ma1 F,
Meicler C. et al. Prospective study of screening for hepato-
cellular carcinoma in Caucasian patients with cirrhosis. J
Hepatol 1994; 20:65-71.
Sherman M, Peltekian KM, Lee C. Screening for hepatocellu-
lar carcinoma in chronic carriers of hepatitis B virus: inci-
dence and prevalence of hepatocellular carcinoma in a
North American urban population. Hepatology 1995; 22:432-
Cottone M, Virdone R, Fusco G, Orlando A, Turri M, Caltagi-
rone M, et al. Asymptomatic hepatocellular carcinoma in
Child's A cirrhosis: a comparison of natural history and sur-
gical treatment. Gastroenterology 1989; 96: 1566-71.
Kang JY, Guan R. Early detection of hepatocellular carci-
noma. Dig Surg 1995; 1279-84.
Pugh RNH, Murray-Lyon IM, Dawson JL. Transection of the
oesophagus for bleeding oesophageal varices. Br 1 Surg
Kaplan EL, Meier P. Nonparametric estimation from incom-
plete observations. J Am Stat Assoc 1958;53:457-81.
Cox DR. Regression models and life tables. J R Stat SOC
Christensen E. Multivariate survival analysis using Cox's re-
gression models. Hepnfology 1987; 7: 1346-58.
Dixon WJ, Brown MB, Engelman L, Frane JW,
rich RI, et al. BMDP statistical software. Berkley: University
of California Press, 1983.
Livraghi T, Giorgio A, Marin G, Salmi A, deSio I, Bolondi L,
et al. Hepatocellular carcinoma and cirrhosis in 746 patients:
long-term results of percutaneous ethanol injection. Radial-
OD 1995; 197:lOl-8.
Hill MA, Jen-
Tiribelli C, Melato M, Croce LS, Giarelli L, Okuda K, Ohnishi
K. Prevalence of hepatocellular carcinoma and relation to
cirrhosis: comparison of two different cities of the world-
Trieste, Italy and Chiba, Japan. Hepatology 1989; 10:998-
Fattovich G, Giustina G, Schalm SW, Hadziyannis S, San-
chez-Tapias J, Almasio P, et al. Occurrence of hepatocellular
carcinoma and decompensation in western European pa-
tients with cirrhosis type B. Hepatology 1995;21:77-82.
Taketa K. a-Fetoprotein: reevaluation in hepatology. Hepa-
tology 1990; 12~1420-32.
Tsukuma H, Hiyama T, Tanaka S, Nakao M, Yabuuchi T,
Kitamura T, et al. Risk factors for hepatocellular carcinoma
among patients with chronic liver disease. N Engl J Med
1993; 328: 1798-801.
Takano S, Yokosuka 0, Imazeki F, Tagawa M, Omata M.
Incidence of hepatocellular carcinoma in chronic hepatitis
B and C: a prospective study of 251 patients. Hepato/ogJ,
Castells A, Bruix J, Bru C, Fuster J, Vilana R, Navasa M, et
al. Treatment of small hepatocellular carcinoma in cirrhotic
patients: a cohort study coinparing surgical resection and
percutaneous ethanol injection. Hepatology 1993; 18:1121-
Early Detection of HCC/Zoli et al. Download full-text
27. Kaneniatsu ‘I, Matsumata T. Shirabe K, Sugimachi K, Saka-
mot0 !i, Nawata H, et al. A comparative study of hepatic
resection and transcatheter arterial embolization for the
treatment of primary hepatocellular carcinoma. Cancer
28. Lai ECS, Fan ST, Lo CM, Chu KM, Liu CL, Wong J. Hepatic
resection for hepatocellular carcinoma: an audit of 343 pa-
tients. Ann Svrg 1995;221:291-8.
29. Livraghi T, Elolondi L, Buscaririi L, Cottone M, Mazziotti
A, Morabito A, et al. No treatment, resection and ethanol
injection in hepatocellular carcinoma: a retrospective analy-
sis of survival in 391 patients with cirrhosis. J Hepatol
30. Chung; JW. Park JH, Choi BI, Han MC. Hepatocellular carci-
noma and portal vein invasion: results of treatment with
transc,atheter oily chemoembolization. AJR Am J Roentgenol
31. Groupe d’Etude et de Traitement du carcinotne hapatocellu-
laire. A comparison of lipiodol chemoembolization and con-
servative treatment for unresectable hepatocellular carci-
noma. N EiigIJ Merl 1995;332:1256-61.
32. Bismuth H, Castaing D, Ericzon BG, Otte JB, Roues K, Ringe
B, et d. Hepatic transplantation in Europe: first report of the
European Liver Transplant Registry. Laricet 1987; ik674-6.
33. Koneru B, Cassavilla A, Bowman J, lwatsuki S, Starzl TE.
Liver transplantation for malignant tumors. Gastroerzrerol
Clin North Am 1988; 17:177-93.
34. Bismuth H, Chiche L, Adam R, Castaing D, Diamond T, Den-
nison A. Liver resection versus transplantation for hepato-
cellular carcinoma in cirrhotic patients. A m Surg 1993;
35. McPeake JR, O’Grady JG, Portmann B, Wight DGD, Tan KC,
Calne RY, et al. Liver transplantation for primary hepatocel-
Mar carcinoma: tumor size and number determine out-
come. J Hepatol 1993; 18226-34.
36. Rogiers X, Knoefel WT, Malago M, Sternek M, Broelsch CE.
Liver transplantation for hepatocellular carcinoma. Dig Surg