Blackwell Science, LtdOxford, UK

JGHJournal of Gastroenterology and Hepatology0815-93192002 Blackwell Science Asia Pty Ltd

174April 2002
2734
Natural history of hepatocellular carcinoma
K Okuda
10.1046/j.0815-9319.2001.02734.x
Review Article401405BEES SGML

Journal of Gastroenterology and Hepatology (2002) 17, 401–405

CLINICAL AND PUBLIC HEALTH CHALLENGES OF CANCER

Natural history of hepatocellular carcinoma including fibrolamellar

and hepato-cholangiocarcinoma variants

KUNIO OKUDA

Emeritus Professor of Medicine, Department of Medicine, Chiba University School of Medicine, Chiba, Japan

Abstract The natural history of hepatocellular carcinoma (HCC) varies greatly with the global region,
because the carcinogenic factors are not the same among countries. Besides the clinicopathological
factors such as tumor characteristics, sex, and age, background liver disease is a major determinant of
prognosis. Hepatocellular carcinoma, mainly associated with chemical carcinogens such as aflatoxin,
does not have severe background cirrhosis, and grows quickly, whereas HCC developing in association
with a virus in a cirrhotic liver generally grows more slowly, and the severity of cirrhosis is the major prog-
nostic factor. The median survival of untreated sub-Saharan African patients is less than 1 month from
diagnosis, contrasted by an average survival of 4 months in virus-induced HCC associated with cirrhosis.
Tumor characteristics, such as size, number, and growth speed, which vary considerably from case to
case, affect the prognosis. Vascular (portal) invasion portends a poor prognosis, and α-fetoprotein levels
also correlate with prognosis. Several distinct clinical types of HCC occur, namely diffuse-type HCC
caused by rapid portal spread of cancer cells, febrile-type caused by poorly differentiated sarcomatoid
cancer cells, and cholestatic HCC caused by intraductal invasion; all have a short survival. There are sev-
eral histological variant forms: combined hepato-cholangiocarcinoma behaves like HCC, with a poorer
prognosis because of more frequent lymph node metastases; fibromellar carcinoma, which is relatively
common in young Caucasian adults, has a good prognosis if diagnosed early, permitting resection; and
cholangiolocellular carcinoma, which derives from the canalicular epithelium, is indistinguishable from
HCC, with a similar prognosis.
© 2002 Blackwell Publishing Asia Pty Ltd

Key words: aflatoxin, cholangio-hepatocellular carcinoma, fibrolamellar variant, geographic variation,

hepatitis viruses, hepatocellular carcinoma, natural history.

INTRODUCTION AND VARIABILITY
OF NATURAL HISTORY
The natural history of hepatocellular carcinoma (HCC)
varies greatly with the global region, because the etio-
logic or carcinogenic factors are not the same among
countries. Although carcinogenesis is known to require
a number of genetic changes of the cell, a certain etio-
logic factor plays a major role and, in the case of HCC,
aflatoxin B1, hepatitis B virus (HBV) and hepatitis C
virus (HCV) are the most important known epi-
demiological factors. Worldwide, hemochromatosis-
associated HCC accounts for only a small fraction of
patients. In southern Africa, particularly in Mozam-
bique where the incidence of HCC is the highest in the
world, aflatoxin is thought to be the major etiologic fac-
tor.1 There, the liver of patients who develop HCC has
no ongoing chronic liver disease, and with a normally
functioning liver, the patient lives until the mass reaches
a huge size, sometimes 6 or 7 kg (Fig. 1).2 In contrast,
HCV infection is the most important carcinogenic fac-
tor in Japan, and the liver is usually highly cirrhotic
when HCC evolves. The patients more frequently die
from hepatic failure,5 and the liver weighs, on average,
less than 2 kg at autopsy.3
There are many distinct differences in pathology and
clinical features between aflatoxin-induced and hepati-
tis virus-associated HCC. In some parts of Asia, both
virus and chemical carcinogens are suspected to be
involved. Thus, the natural history of HCC differs with
the area and ethnic group. One extreme example is the
peak age of HCC patients, which is 25–34 years among
Mozambican males6 and 60–70 years among the
Japanese.7 In South Africa, patients with HCC die

Correspondence: Professor Kunio Okuda, Emeritus Professor of Medicine, Department of Medicine, Chiba University School
of Medicine, Chiba, Japan. Email: okuda@med.m.chiba-u.ac.jp
402 K Okuda

20 100

90
10
80

70
0
No.

% Surviving
1000 2000 3000 4000 5000 6000 7000 Liver weight 60
0
50
50 40

30
100
20
Figure 1 Liver weight at autopsy. Comparison between
patients in Mozambique (n = 51) and Japan (n = 232). Repro- 10
duced with permission.3,4
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Months after diagnosis

within a month or two from diagnosis and, in a study in Figure 2 Survival curve for liver cancer in Uganda. Repro-
1975, none survived more than 9 months (Fig. 2).8 By duced with permission.8
contrast, the 1-year survival rate in Japan in a 1985
study was 25.9% (Fig. 3),9 and in the USA from 1992
to 1996 it was 22.6%.10 Progress in diagnosis over 100
10 years does not explain these differences. The pathol-
ogy of HCC also varies with the region. Thus, HCC not
associated with hepatitis virus is more often poorly
% Survival

differentiated2 and grows quickly, whereas virus-

50
associated HCC is usually well differentiated before it
reaches a certain size, and grows slowly, often acquir-
ing a fibrous capsule;11 HCC of non-viral etiology does
not acquire a capsule.12 The following are the clinico-
pathological factors that influence the prognosis. 0
2 4 6 9 12 18 24 30 36 42
Months

CLINICOPATHOLOGICAL Figure 3 Survival curves for untreated hepatocellular carci-

PROGNOSTIC FACTORS
Sex and age
Virtually all epidemiological studies have shown that
males are more prone to hepatocarcinogenesis. The
prognosis is somewhat better in females,10 perhaps
because of the more favorable outcome of cirrhosis itself
in this gender.
Background liver disease and major
etiologic factors
All studies in which the prognosis of patients was cor-
related with the Child’s grades showed that Child C
patients fared poorest, suggesting that cirrhosis rather
than mass size is the determinant of outcome. The clas-
sification of Okuda et al.9 and other schemes of grading
all emphasize the clinical parameters of cirrhosis.13
Prognosis of untreated HCC also depends on the asso-
ciated etiologic factor. In patients with a non-viral back-
ground of the liver, as among Mozambicans, the growth
of HCC is so fast that, as shown in a study in Uganda,8
the median survival was half a month; the correspond-
noma in Japan. (––) Stage I (n = 33), (---) Stage II (n = 134),
(–·–) Stage III (n = 82). Reproduced with permission.9
ing figure among Japanese untreated patients was
1.6 months.
Whereas no patient survived for more than 9 months
in Uganda, some patients lived up to 3 years without
treatment in Japan.9 Although no comparison is avail-
able between hemochromatotic cirrhosis and viral cir-
rhosis for prognosis, it is expected to be about the same
because both are cirrhosis based. Between HBV- and
HCV-associated HCC, the former seems to have a
somewhat better prognosis because the background
liver disease is usually less advanced; advanced cirrhosis
is less common in HBV-associated HCC compared
with HCV-associated HCC.
Tumor characteristics
The size of tumor, number of tumors, and growth
speed all affect the prognosis. The smaller the tumor,
the better the survival (Fig. 4), and the prognosis of a
solitary HCC is better than that of multiple tumors.
The growth speed can be calculated as the doubling
 Natural history of hepatocellular carcinoma 403

Table 1 Hepatocellular carcinoma compared to hepato-cholangiocarcinoma (Japan Liver Cancer Study Group, 1996–1997)

Type No. of cases Cirrhosis AFP (≥ 200 ng/mL) HBsAg Anti-HCV 5-year survival Lymph node metastases

HCC 33 394 83% 34% 17% 76% 47% 29%

HCC + CC 219.00 75% 36% 22% 57% 33% 42%

AFP, α-fetoprotein; CC, cholangiocarcinoma; HBsAg, hepatitis B surface antigen; HCC, hepatocellular carcinoma;
HCC + CC, hepato-cholangiocarcinoma, HCV, hepatitis C virus.

100 Table 2 Comparison of fibrolamellar carcinoma and hepato-

90 cellular carcinoma from the literature25
80
Survival rate (%)

70
FLHC HCC
60 No. 136 2320
50 Resection 63/68 180/369
40 Median survival (months)
30 Resected 6–252 10–60
20 Not resected 3–36 1–6
10 Cirrhosis 3.6% 76.5%
0 AFP (+) 6.7% 82.5%
0 12 24 36 48 60 72 84 96 108120 132 144 156 168 180 192 HBsAg 5.8% 64.8%
Survival period (months)
AFP, α-fetoprotein; FLHC, fibrolamellar carcinoma;
Figure 4 Survival curves for different sizes of hepatocellular
HBsAg, hepatitis B surface antigen; HCC, hepatocellular
carcinoma after resection. (---) < 2 cm (n = 3866), (–·–) 2–
carcinoma.
5 cm (n = 10 675), (····) 5–10 cm (n = 3872), (––) 10 cm
(n = 1771). Reproduced with permission.7

time by measuring the diameter at intervals; it varies a
great deal with each HCC14 and, in general, a fast-
growing HCC has a poorer prognosis compared to a
slow-growing cancer;15 HCC may change growth-
speed some time during the course, however.
prothrombin-positive but AFP-negative patients also
tend to have more advanced HCC, with a poor
prognosis.21

Vascular invasion
Hepatocellular carcinoma has a propensity to invade the
portal vein, and portal vein invasion is one of the deter-
minants of prognosis. Patients who have developed
portal invasion at an early stage fare poorly thereafter.
Hepatocellular carcinoma also invades large hepatic
veins, but less frequently compared with the portal inva-
sion.16 If HCC invades a large hepatic vein, the growth
is usually active and quite often extends into the right
atrium, or even into the pulmonary vein through the tri-
cuspid valve.17 The prognosis of patients with hepatic
vein invasion is much poorer than for those who do not
have it.
α-Fetoprotein
Studies that analyzed the relationship between prog-
nosis and α-fetoprotein (AFP) levels all showed that
patients with a high AFP level at diagnosis tend to have
a shorter lifespan compared with those who have a low-
level AFP.18 This trend becomes more apparent if one
compares patients with positive lectin-binding AFP
(fraction L3),19 and those with negative L3.20 Des-γ-
Metastasis
The effect of extrahepatic metastasis of HCC on the
natural history of the disease is unclear, because it
usually occurs late in the clinical course, and does not
significantly shorten the lifespan. However, intra-
abdominal spread following tumor rupture portends a
rapid downhill course. Tumor rupture may be managed
by embolization of the bleeding artery transarterially
using gel-foam particles or coils, but the outcome is
usually poor.
The following are rare but characteristic clinical types
of HCC; they have distincitive types of prognosis:
• Diffuse-type HCC. Diffuse-type HCC develops follow-
ing intrahepatic portal vein spread of HCC occurring
within a short period of time. Numerous small tumors
form in the liver, and the clinical course is rapidly
downhill.22
• Febrile-type HCC. According to Berman, this form of
HCC occurs in 8% of Bantu patients,4 but it is very
uncommon elsewhere. This type of HCC mimics liver
abscess, with high fever and leukocytosis. Histologi-
cally, cancer cells are very poorly differentiated and
resemble a sarcoma23 (an alternative term is sarcoma-
404
Table 3 Results of resection for fibrolamellar carcinoma
Author Year n Resected
Berman 1980 12
Farhi 1983 10
Starzl 1986 14
Stevens 1995 10
Pinna 1997 28 + 13 Tx
Tx, transplantation.
toid HCC). The disease progresses rather rapidly and
has a poor prognosis.
• Cholestatic HCC. Patients present clinical signs sug-
gestive of obstructive jaundice, and jaundice is caused
by tumor invasion into a large bile duct, often in the
hilum. According to Nakashima and Kojiro, this type
occurs in 6% of all HCC cases.16 The symptoms may
mimic gallstone disease with pain that is intense when
hemobilia occurs. The clinical course is rapidly downhill
from the onset of cholestasis.
VARIANT HISTOLOGICAL FORMS
Combined hepato-cholangiocarcinoma
This form accounts for about 0.65% of all primary liver
tumors in Japan.7 Hepatocellular carcinoma and cho-
langiocarcinoma (CC) coexist in the same liver in three
different anatomical relationships. In the first, the mass
contains a mixture of HCC and CC cells; in the second,
HCC and CC occur at different sites of the liver as
synchronous double cancers; and the third relationship
occurs when the two types of cancer are neighboring
each other. Clinically, the combined form is very much
like HCC; cirrhosis is common, virus markers are fre-
quently positive, and AFP is frequently elevated. How-
ever, the prognosis is somewhat poorer in the combined
form because lymph node metastases occur more fre-
quently and the 5-year survival is lower (Table 1).
Fibrolamellar carcinoma
This variant form of HCC occurs in Caucasian patients
mostly 5–35 years of age. The gross anatomy is charac-
terized by expanding growth with separate fibrosis,
mimicking focal nodular hyperplasia.24 This variant
is very uncommon in the Far East. In one autopsy
series of 439 cases at Kurume University, there was
not a single case of fibrolamellar carcinoma. Clini-
cally, hypertranscobalaminemia is common, and des-γ-
prothrombin, but not AFP, is elevated. The prognosis is
generally much better compared with conventional
HCC. This is because the liver does not have cirrhosis
and the expanding nature of tumor growth permits a
high rate of surgical resection25,26 (Tables 2,3).
K Okuda
Median survival 5-year survival
– 68 months 63%
6 5/10 living 1.5–8 years
8 Recurrence 1/8 None died
7 9 months –
– – 82%
Cholangiolocellular carcinoma
This histological type was first described by Steiner and
Higginson.27 The cancer cells are derived from the canal-
icular epithelium or the cells that constitute Hering’s
canal. The clinical features are not distinguishable from
those of ordinary HCC, and the prognosis is similar.
REFERENCES
1 Van Renzburg SJ, Cook-Mozaffari P, Van Schalkwyk DJ
et al. Hepatocellular carcinoma and dietary aflatoxin in
Mozambique and Transkei. Br. J. Cancer 1985; 51: 713–
26.
2 Steiner PE. Cancer of the liver and cirrhosis in trans-
Saharan Africa and the United States of America. Cancer
1960; 13: 1085–166.
3 Nakashima T, Okuda K, Kojiro M et al. Pathology of
hepatocellular carcinoma in Japan. 232 consecutive cases
autopsied in ten years. Cancer 1983; 51: 863–77.
4 Berman C. Primary Carcinoma of the Liver. London:
Lewis, 1951.
5 Okuda K, Nakashima T. Primary carcinoma of the liver.
In: Berk JE, Haubrich WA, Kalser MH, Schaffner F, eds.
Bockus Gastroenterology, 4th edn. Philadelphia: Saunders,
1985; 3315–76.
6 Higginson J. The epidemiology of primary carcinoma of
the liver. In: Pack GT, Islami AH, eds. Tumors of the Liver.
Berlin: Springer-Verlag, 1970; 38–52.
7 Liver Cancer Study Group of Japan. 14th report on liver
cancer follow-up. 1996–97. Kyoto: Shinko Publication,
2000.
8 Primack A, Vogel CL, Kyalwazi SK et al. A staging sys-
tem for hepatocellular carcinoma: Prognostic factors in
Ugandan patients. Cancer 1975; 35: 1357–64.
9 Okuda K, Ohtsuki T, Obata H et al. Natural history of
hepatocellular carcinoma and prognosis in relation to
treatment. Study of 850 Patients. Cancer 1985; 56: 918–
28.
10 El-Serab HB, Mason AC, Key C. Trends in survival of
patients with hepatocellular carcinoma between 1977 and
1996 in the United States. Hepatology 2001; 33: 62–5.
11 Okuda K, Musha H, Nakajima Y et al. Clinicopathologi-
cal features of encapsulated hepatocellular carcinoma: a
study of 26 cases. Cancer 1977; 40: 1240–5.
12 Okuda K, Peters RL, Simson IW. Gross anatomical fea-
tures of hepatocellular carcinoma from three disparate
Natural history of hepatocellular carcinoma
geographic areas. Proposal of new classification. Cancer
1984; 54: 2165–73.
13 Calvet X, Bruix J, Bru C et al. Natural history of hepato-
cellular carcinoma in Spain. Five year’s experience in 249
cases. J. Hepatol. 1990; 10: 311–17.
14 Sheu JC, Sung JL, Chen DS et al. Growth rate of asymp-
tomatic hepatocellular carcinoma and its implications.
Gastroenterology 1985; 89: 259–66.
15 Okazaki N, Yoshino M, Yoshida T et al. Evaluation of the
prognosis for small hepatocellular carcinoma on tumor
volume doubling time. A preliminary report. Cancer 1989;
63: 2207–10.
16 Nakashima T, Kojiro M. Hepatocellular Carcinoma. An
Atlas of its Pathology. Tokyo: Springer, 1987.
17 Okuda K. Clinical aspects of hepatocellular carcinoma—
analysis of 134 cases. In: Okuda K, Peters RL, eds.
Hepatocellular Carcinoma. New York: Wiley, 1976; 387–
436.
18 Nomura F, Ohnishi K, Tanabe Y. Clinical features and
prognosis of hepatocellular carcinoma with reference to
serum alpha-fetoprotein levels. Analysis 606 Patients.
Cancer 1989; 64: 1700–7.
19 Taketa K, Izumi M, Ichikawa E. Distinct molecular spe-
cies of human α-fetoprotein due to differential affinities to
lectins. Ann. N. Y. Acad. Sci. 1983; 417: 61–8.
405
20 Yamashita F, Tanaka M, Satomura S, Tanikawa K. Prog-
nostic significance of Lens culinaris agglutinin A-reactive
α-fetoprotein in small hepatocellular carcinoma. Gastro-
enterology 1996; 111: 996–1001.
21 Okuda H, Nakanishi T, Takatsu K et al. Comparison of
clinicopathologic features of patients with hepatocellular
carcinoma seropositive for α-fetoprotein alone and those
seropositive for des-g-carboxy prothrombin alone. J.
Gastroenterol. Hepatol. 2001; 16: 1290–6.
22 Okuda K, Noguchi T, Kubo Y et al. A clinical and patho-
logical study of diffuse type hepatocellular carcinoma.
Liver 1981; 1: 280–9.
23 Okuda K, Kondo Y, Nakano M et al. Hepatocellular car-
cinoma presenting with pyrexia and leukocytosis: report
of five cases. Hepatology 1991; 13: 695–700.
24 Rolfes DB. Fibrolamellar carcinoma of the liver. In:
Okuda K, Ishak KG, eds. Neoplasms of the Liver. Tokyo:
Springer, 1987; 137–42.
25 Ruffin MT. Fibrolamellar carcinoma. Am. J. Gastroen-
terol. 1990; 85: 577–81.
26 Pinna AD, Iwatsuki S, Lee RG et al. Treatment of fibro-
lamellar hepatoma with subtotoal hepatectomy or trans-
plantation. Hepatology 1997; 26: 877–83.
27 Steiner PE, Higginson J. Cholangiolocellular carcinoma
of the liver. Cancer 1959; 12: 753–9.