Diagnosis, staging and treatment of hepatocellular carcinoma

1689

Braz J Med Biol Res 37(11) 2004

Hepatocellular carcinoma

Diagnosis, staging and treatment ofhepatocellular carcinoma

1Divisão de Gastroenterologia, and 2Serviço de Radiodiagnóstico,Departamento de Clínica Médica, Faculdade de Medicina de Ribeirão Preto,Universidade de São Paulo, Ribeirão Preto, SP, Brasil3Setor de Hepatologia, Departamento de Gastroenterologia,Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brasil

A.V.C. França1,J. Elias Junior2,B.L.G. Lima1,

A.L.C. Martinelli1

and F.J. Carrilho3

Abstract

Hepatocellular carcinomas are aggressive tumors with a high dissemi-nation power. An early diagnosis of these tumors is of great impor-tance in order to offer the possibility of curative treatment. For an earlydiagnosis, abdominal ultrasound and serum alpha-fetoprotein deter-minations at 6-month intervals are suggested for all patients withcirrhosis of the liver, since this disease is considered to be the main riskfactor for the development of the neoplasia. Helicoidal computedtomography, magnetic resonance and/or hepatic arteriography aresuggested for diagnostic confirmation and tumor staging. The need toobtain a fragment of the focal lesion for cytology and/or histology fora diagnosis of hepatocellular carcinoma depends on the inability ofimaging methods to diagnose the lesion. Several classifications arecurrently available for tumor staging in order to determine patientprognosis. All take into consideration not only the stage of the tumorbut also the degree of hepatocellular dysfunction, which is known tobe the main factor related to patient survival. Classifications, however,fail to correlate treatment with prognosis and cannot suggest the idealtreatment for each tumor stage. The Barcelona Classification (BCLC)attempts to correlate tumor stage with treatment but requires prospec-tive studies for validation. For single tumors smaller than 5 cm or upto three nodules smaller than 3 cm, surgical resection, liver transplan-tation and percutaneous treatment may offer good anti-tumoral re-sults, as well as improved patient survival. Embolization orchemoembolization are therapeutic alternatives for patients who donot benefit from curative therapies.

CorrespondenceA.V.C. França

Divisão de Gastroenterologia

Departamento de Clínica Médica

FMRP, USP

Av. Bandeirantes, 3900

14048-900 Ribeirão Preto, SP

Brasil

Fax: +55-16-633-6695

E-mail: [email protected]

Publication supported by FAPESP.

Received September 16, 2003

Accepted June 14, 2004

Key words• Hepatocellular carcinoma• Tumor diagnosis• Tumor staging• Carcinoma therapy• Liver cirrhosis• Barcelona Classification

Introduction

Hepatocellular carcinoma (HCC) is themost frequent primary solid tumor of theliver. Its aggressiveness and extensive dis-semination lead to a poor patient prognosis.HCC is estimated to account for 5% of allmalignant neoplasias (1). Its prevalence isconsidered to be high (>20 cases/100,000

inhabitants/year) in the far East and Africa,medium (5 to 20 cases/100,000 inhabitants/year) in Europe, and low (<5 cases/100,000inhabitants/year) in South America. In Bra-zil, its prevalence is considered to be low,despite geographic variations (2).

Cirrhosis of the liver, regardless of etiol-ogy, is considered to be the main risk factor forthe onset of HCC. In Africa and in South Asia,

Brazilian Journal of Medical and Biological Research (2004) 37: 1689-1705ISSN 0100-879X Review

1690


A.V.C. França et al.

hepatitis B is the main cause of hepatic dis-ease. In these regions, HCC may developamong young patients, even when they do nothave cirrhosis of the liver because infectionwith the hepatitis B virus occurs during deliv-ery or soon after birth, with the prolongedperiod of infection being the major determi-nant of the onset of HCC in these patients. Inthe Western World and in Japan, hepatitis Cvirus is the main factor related to the presenceof cirrhosis of the liver in patients with HCC.In a survey conducted by the Brazilian Societyof Hepatology in Brazil in 1995 cirrhosis ofthe liver was present in 71% of patients withHCC, with the cause of liver disease beinghepatitis B in 39% of cases, hepatitis C in 27%and alcoholism in 37%. In our patient series,hepatitis C virus (43%) and alcoholism (38%)are the major causes of liver cirrhosis in pa-tients with HCC (3).

Cirrhosis of the liver is present in 60 to100% of patients with HCC (4,5) dependingon regional variations. Among our patientswith HCC, 90% have cirrhosis of the liver (3).

With the identification of cirrhosis of theliver as the main risk factor for the develop-ment of HCC, and in view of the high aggres-siveness of this type of neoplasia when diag-nosed in advanced stages, periodic follow-up is necessary for an early diagnosis of thetumor.

HCC is a tumor that satisfies the require-ments of neoplasias for which screening isjustified. There is a well-defined risk group(cirrhosis of the liver), there are effective,noninvasive and low cost diagnostic tech-niques (ultrasound, US, and alpha-1 fetopro-tein, AFP), and curative therapeutic tech-niques are available which can increase pa-tient survival (resection, liver transplanta-tion and percutaneous treatment).

Diagnosis

An early diagnosis of HCC is requiredfor the institution of treatments consideredto be curative. On this basis, screening of

each patient with cirrhosis of the liver re-gardless of etiology is of primordial impor-tance for the detection of tumors in the initialstages of development. It has been suggestedthat monitoring should be performed by USand by measurements of serum AFP at 6-month interval. The option for a 6-monthinterval is based on the mean time for tumorduplication, which is about 6.5 months andmay range from 1 to 20 months (6,7). Shorterintervals (3 to 4 months) or the use of otherimaging methods such as helicoidal com-puted tomography have been suggested forHCC screening in patients considered to beat high risk. However, the definition of whatshould be considered a high risk varies amonginvestigators and also among the popula-tions studied. A comparison of the differentschedules (reduced intervals between ex-ams) or the use of other imaging techniquesis lacking in the literature.

However, only patients who would benefitfrom curative treatment should be submittedto screening. On this basis, in patients withcirrhosis of the liver classified as Child-PughC with no indication for liver transplantation,monitoring with US and AFP will be of nobenefit in terms of patient survival.

Ultrasound

US is considered to be the technique ofchoice for the diagnosis of focal hepaticlesions (8), permitting the detection of tu-mors of small size (1 cm) still in the earlyphase of development and thus being justi-fied for the screening of HCC in patientswith cirrhosis of the liver (8,9). The ultraso-nographic characteristics of HCC depend onnodule size. Small nodules of less than 3 cmare frequently hypoechogenic. As they in-crease in size, they start to acquire isoecho-genic characteristics with a peripheral haloor hyperechogenic or heterogeneous charac-teristics due to the neoformation of bloodvessels and intratumoral necrosis. Other pos-sible findings are lateral shadows, a poste-

1691



rior reinforcement and a perilesional halothat corresponds to the presence of a peritu-moral fibrous capsule consequent to the com-pression of the adjacent hepatic parenchyma(8). US can also be used to assess the perme-ability of vascular structures and the exist-ence of hilar adenopathies suggestive of tu-moral extension.

The sensitivity of US for the detection ofHCC is directly related to tumor size. Itsdiagnostic sensitivity for tumors smaller than1 cm is about 42% (10,11), reaching 95% fortumors of larger size (12).

The combination of Doppler with US canbe useful for the identification of portalthrombosis in patients with HCC, with 89 to92% sensitivity and 100% specificity in theidentification of tumor thrombosis (13). Thepresence of a hepatofugal pulsatile flow in-side the thrombus is suggestive of vascularinvasion (14). Hepatic arteriography for theidentification of portal thrombosis can beavoided when US-Doppler reveals perme-ability of the portal system (13). In patientswith tumors submitted to alcoholization,chemical thrombosis can be differentiatedfrom tumoral thrombosis by means of US-Doppler based on the presence of blood flowin the thrombus. The presence of a pulsatilethrombus rules out the possibility of benignthrombosis (by alcohol), confirming the pres-ence of tumoral invasion of portal vessels. Incases in which doubts persist about the causeof portal thrombosis, fine needle aspirativepuncture can be used, a highly sensitiveprocedure for the confirmation of tumoralthrombosis which also involves low risks ofcomplications (14).

In addition to providing an imaging diag-nosis, US can also be used to guide theneedle for aspirative puncture or for a biopsycarried out to obtain a tissue fragment fromthe tumor nodule (8).

Alpha-1 fetoprotein

Under physiological conditions, AFP is

synthesized by the embryonic liver, by cellsof the vitellin sac and by the fetal intestinaltract. Patients with chronic liver disease,especially those with a high degree of hepa-tocyte regeneration, can express AFP in bloodin the absence of malignant neoplasia.

Measurement of serum AFP levels is notuseful for the early detection of HCC be-cause, even though 80% of the patients withHCC have serum AFP concentrations ex-ceeding normal levels (10-20 ng/ml), pa-tients with a high liver regenerative activitymay express higher than normal AFP valueswithout having HCC (8,15,16).

AFP serum levels above 400-500 ng/mlare considered to be diagnostic of HCC incirrhotic patients with focal hepatic lesions(8). However, only 1/3 of patients with HCChave AFP levels higher than 100 ng/ml, withlevels above 400 ng/ml being quite infre-quent in the presence of small tumors (<5cm) (8).

Even though they do not reach levelsconsidered to be diagnostic (400-500 ng/ml), progressively increasing AFP concen-trations during screening are suggestive of adiagnosis of HCC. These patients should besubmitted to helicoidal computed tomogra-phy in order to rule out the diagnosis of atumor. The values and the time interval be-tween AFP measurements needed for themto be considered as “progressively increas-ing AFP levels” have not yet been estab-lished.

With the development of early detectionprograms, an increase in the number of caseswith small tumors and normal AFP levels(29%) has been observed (16). In our experi-ence, 42% of patients with HCC presentserum AFP levels within normal limits (3).

At the time of tumor diagnosis, AFP seemsto be of prognostic value. This may be due tothe fact that well-differentiated tumors ex-press less AFP and patients with normal AFPhave a lower incidence of tumoral vascularinvasion and tend to present better hepaticfunction, which is known to be one of the

1692



major prognostic factors for these patients(16-18).

Other tumor markers

Des-gamma carboxyprothrombin (DCP)is another tumor marker used for the diagno-sis of HCC. Its diagnostic efficacy has beeninvestigated by various groups, with contra-dictory results. There are reports of the supe-riority of DCP over AFP in the diagnosis ofHCC, especially in the Orient and in NorthAmerica. However, studies conducted inEurope have not shown a better performanceof DCP compared to AFP. Racial and etio-logical factors of liver disease may be re-sponsible for the discrepant results. Despitethe contradictory results, the combination ofthe two markers is suggested to increase thediagnostic efficacy. Studies on large patientseries and on diverse ethnic populations areneeded to clarify the real role of DCP in thediagnosis of HCC.

Other markers such as interleukin-2, uri-nary tumor growth factor-ß1 and MAGE-4protein receptors have also been used inresearch to aid the diagnosis of HCC buttheir clinical applicability requires scientificconfirmation.

In the presence of a diagnostic suspicionby US and/or AFP, the patient should befurther investigated by helicoidal computedtomography (CT) and/or magnetic resonance(MR) and, in selected cases, by hepatic arte-riography, in order to confirm the suspecteddiagnosis and determine the stage of thetumor.

Helicoidal computed tomography

After the suspicion of HCC by US in apatient with cirrhosis of the liver, the use ofCT is suggested. The sensitivity of CT forthe diagnosis of HCC is similar to that of US.Its diagnostic efficacy depends on technicalfactors, mainly the injection of contrast, andon factors inherent to the tumor, the most

important of which are tumor size and vas-cularization. CT should be performed by thespiral (or helicoidal) technique with intrave-nous injection of contrast and images shouldbe obtained in the basal, arterial, portal, andequilibrium phases. The main characteristicof HCC detected by CT is the early uptake ofcontrast in the arterial phase of the exam.Due to the hypovascularization of small-sized tumors, the diagnostic efficacy of CTis reduced in tumors measuring less than 2cm (19,20). A second phase after the intrave-nous contrast increases the diagnostic sensi-tivity when associated with the arterial phase,when the HCC is found to be iso- or hypo-dense (20). When images and intravenouscontrast are associated, a greater capacity todetect HCC is achieved.

The use of standard CT causes technicaldifficulties such as respiratory artifacts anddifficulty in capturing images during the ar-terial phase of the contrast dye. To obviatethese technical difficulties, the helicoidal(spiral) technique has been used which, dueto the rapid image acquisition, permits thecapture of sections of the entire liver duringa single moment of apnea. This techniqueincreases by about 15% the diagnostic sensi-tivity for the detection of hepatic tumorssmaller than 2 cm compared to standard CT.CT also reveals the presence of tumor in-volvement of lymph nodes, vascular inva-sion and extrahepatic involvement with highsensitivity, specificity and diagnostic accu-racy.

Intra-arterial injection of lipiodol fol-lowed by CT (CT-lipiodol) presents low di-agnostic sensitivity (37%) which is not higherthan that of helicoidal CT, in addition tobeing an invasive method (10). Thus, theCT-lipiodol technique has not been routinelyused for the diagnosis of HCC.

Magnetic resonance

MR has been used to obtain a bettercharacterization of hepatic lesions sugges-

1693



tive of HCC and also for their differentiationfrom benign lesions (21). HCC is better evalu-ated in potentiated sequences in T2, where itfrequently appears to be hyperintense (22,23).In potentiated sequences in T1, because ofthe larger amount of water inside them, whichincreases the relaxation time, HCC are foundto be hypointense. The hyperintensity in T1may be attributed to the presence of steato-sis, to the formation of light cells and tointratumoral copper accumulation. Intrave-nous injection of paramagnetic contrast (ga-dolinium-DTPA) is also used to better char-acterize the lesions. As observed with CT,HCC appears as a hyperdense image withcontrast uptake. The sensitivity of MR de-pends on tumor size. In tumors larger than 2cm the level of detection is about 95%. How-ever, in tumors smaller than 2 cm this level isreduced to 30% (22). The diagnostic effi-cacy of MR for the detection of HCC seemsto be similar to or even lower than that of CT.However, this imaging technique is very use-ful for the demonstration of the internal ar-chitecture of the tumor, of the tumoral mar-gins, of the presence of a peritumoral cap-sule, and of intrahepatic vascular invasion(22).

One of the major useful properties of MRis the differential diagnosis from hepatichemangioma (8). Because of the lack ofanatomical limitation, MR and CT are supe-rior to US for the diagnosis of tumors closeto the lung and of isoechogenic lesions (15).

Hepatic arteriography

The diagnostic efficacy of hepatic arteri-ography (HA) depends on tumor size and onthe extent of tumor vascularization (19).Small tumors tend to be well differentiatedand consequently present low vasculariza-tion, thus being difficult to detect by thistechnique (24). For tumors smaller than 5cm, HA has 82 to 93% diagnostic sensitivity,73% specificity and 89% diagnostic accu-racy, with these values being even more

reduced when the tumors are smaller than 2cm (24,25).

Only in selected cases is it possible to uselaparoscopy for the diagnosis of HCC, sincethis technique only analyzes the presence ofsuperficial lesions.

The diagnostic accuracy of imaging tech-niques for the detection of HCC depends onthe characteristics of the tumor and on theexperience of each study group. In our expe-rience (12,26), the ability to diagnose HCCis 84% for US, 79% for CT, 77% for MR,and 64% for HA. Thus, because of its lowcost and availability, we believe that UScontinues to be the exam of choice for anearly diagnosis of HCC, being useful for themonitoring of cirrhotic patients. However,the experience and awareness of the opera-tor is of primordial importance during theexecution of the exam.

Cytology and/or histology

Cytologic and/or histopathologic exami-nation of the suspected lesion can also beused for the diagnosis of HCC. The materialto be used for cytology is obtained by fineneedle aspirative biopsy (FNAB). This is asafe technique with minimal risks of compli-cations due to the procedure, which providesadequate material when performed by trainedpersonnel. Its diagnostic accuracy may varyfrom 60 to 90% (8,27) depending on the sizeof the lesion, on the examiner and on thediameter of the puncturing needle. The speci-ficity and positive predictive value of thistechnique are higher than 90%, reaching100% in our experience (27).

Histopathological examination is the mainmethod for a sure diagnosis of HCC. A non-tumoral liver biopsy is important to rule outor to confirm the presence of liver cirrhosis,since the presence of cirrhosis may contrain-dicate surgical treatment. The two pathologytechniques can be used for the diagnosis ofHCC and their combination can increase thediagnostic accuracy (27). To prevent com-

1694



plications, the nodule should always be punc-tured through the non-tumoral liver, whichwill serve as a “buffer” preventing bleeding.

Diagnostic criteria

In a recent publication, the EuropeanAssociation for the Study of the Liver (EASL)(28) proposed guidelines and criteria for thediagnosis of HCC in cirrhotic patients (Fig-ure 1). The presence of a nodule larger than2 cm with hypervascular characteristics de-tected by at least two imaging techniquesconfirms the presence of HCC, with no needfor cyto-histopathological confirmation.FNAB and/or a biopsy of the suspected le-sion with a cutting needle are suggested forcirrhotic patients with nodules smaller than2 cm, since in about 50% of cases the diag-nosis of HCC cannot be confirmed by imag-ing methods (1).

The diagnostic criteria (28) are cytologicaland/or histological criteria, or noninvasive cri-teria, which include 1) radiologic criterion:two coinciding imaging techniques demon-strating a focal hepatic lesion >2 cm witharterial hypervascularization, or 2) combinedcriteria: an imaging technique associated with

increased serum AFP levels, a focal lesion >2cm with arterial hypervascularization, and se-rum AFP levels >400 ng/ml.

The imaging techniques to be evaluatedare US, CT, MR, and HA. However, thetechnology for the execution of all imagingexams suggested by the EASL is not avail-able at all hospital centers. Similarly, thecurative therapies also require properlytrained specialized teams that are not avail-able at all medical services. This fact leads totwo observations: 1) screening with US andAFP should be performed only in cases inwhich the medical team can offer curativetherapies (surgical resection, liver transplan-tation and percutaneous treatment); 2) thelack of access to helicoidal CT, MR and HAdoes not exclude the possibility of diagnos-ing HCC. In this situation we believe thatFNAB and/or a biopsy of the lesions shouldbe obtained for a sure diagnosis regardlessof tumor size, as long as the patients can betreated with available techniques.

Staging

The main objective of tumor staging is todetermine the prognosis of the disease and to

Cirrhotic patients* (US/AFP 6/6 months)

Hepatic nodule Absence of nodules

>1 cm <1 cm Elevated AFP** Normal AFP

<2 cm >2 cm US 3/3 months Helicoidal CT

Cyto/Histology AFP ≥400 ng/ml

CT/MR/HA

Without HCC

US/AFP 6/6 monthsHepatocellular carcinoma***

Figure 1. Diagnosis of hepato-cellular carcinoma according tothe Barcelona-2000 Conferenceof the European Association forthe Study of the Liver (EASL)(28). *Patients who can be sub-mitted to curative treatment ifan HCC is diagnosed; **Unde-fined AFP level; ***Confirma-tion by pathology or by a nonin-vasive criterion. AFP = alpha-1fetoprotein; CT = computed to-mography; HA = hepatic arteri-ography; HCC = hepatocellularcarcinoma; MR = magnetic reso-nance; US = ultrasound.

1695



Table 2. Barcelona Clinic Liver Cancer (BCLC) Group classification of hepatocellular carcinomas (29).

Stage PST Tumor stage Okuda Portal hypertension Total bilirubin Child-Pugh

AA1 0 Single I No NormalA2 0 Single I Yes NormalA3 0 Single I Yes AlteredA4 0 3 <3 cm I-II A-B

B 0 >5 cm I-II A-BMultinodular

C 1-2 Vascular invasion I-II A-Band/or metastasis

D 3-4 Any stage III C

PST = Performance status test.

establish the best therapeutic method for thepatient. Several factors should be analyzedin cirrhotic patients with HCC. Prognosisand treatment mainly depend on the degreeof hepatic dysfunction, tumor stage and gen-eral patient condition. It should be kept inmind that most patients with HCC have as-sociated cirrhosis of the liver, this being afactor frequently as important or even moreimportant than tumor stage.

Various classifications have been adoptedto stage HCC. The first was developed in1985 by Okuda et al. (5) (Table 1) in a studyon 850 patients with HCC using data con-cerning tumor size (> or < than 50% of theliver size), serum bilirubin levels (> ou <3mg/dl), serum albumin levels (> or <3 g/dl),and the presence of ascites, classified HCCinto three stages. The mean survival rate forpatients with stages I, II and III was 11.5, 3and 0.9 months, respectively, without con-sidering the treatment variable. When thepatients submitted to surgical resection ofthe tumor were analyzed, survival reached25.6 months for patients with Okuda stage I.

Several other prognostic classificationsof HCC are currently being used:

Barcelona Clinic Liver Cancer (BCLC)Group (29) (Table 2). This classificationtakes into consideration hepatic function,portal hypertension, bilirubins, symptoms

related to the tumor, tumor morphology, pres-ence of distant metastases, or vascular inva-sion. This is the only classification that cor-relates prognostic data with therapeutic pos-sibilities.

TNM classification (30) (Table 3). Thisclassification considers the size and numberof nodules, vascular invasion, and bilobarinvolvement. Hepatic function is not consid-ered as a factor in staging, although it is animportant factor for the prognosis of thesepatients. Evaluation of this classification inpatients submitted to liver transplantationdid not show benefits in prognostic terms forpatients with HCC (30). The inclusion of thehepatic fibrosis factor in staging by TNMhas been recently suggested.

French classification (31) (Table 4). Thisclassification includes as prognostic factors

Table 1. Okuda classification of hepatocellularcarcinomas (5).

Negative Positive

Tumor size <50% of liver >50% of liverAscites Absent PresentSerum albumin >3 g/dl <3 g/dlBilirubin <3 mg/dl >3 mg/dl

Okuda I: No positive factor; Okuda II: 1 or 2positive factors; Okuda III: 3 or 4 positive factors.

1696



Table 4. French classification of hepatocellular carcinomas (31).

0 1 2 3

Karnofsky index (%) ≥80 <80Bilirubin (µmol/l) <50 ≤50Alkaline phosphatase (MNL) <2 ≥2Alpha-1 fetoprotein (µg/l) <35 ≥35Portal obstruction (US) No Yes

Karnofsky index >80% = complete patient autonomy; MNL = maximum normal limit;US = ultrasound. Group A (low risk): 0 point; group B (intermediate risk): 1-5 points;group C (high risk): ≥6 points.

intermediate risk (score 1 to 5), and 7 and 3%for high risk patients (score ≥6), respec-tively. It should be pointed out that thisclassification was based on the analysis ofpatients (47%) submitted to some type oftreatment. In addition, mean patient survivalwas only 4.3 months, suggesting selection ofpatients with an advanced stage of the disease.

Chinese University Prognostic Index (32)(Table 5). This index includes the TNMclassification associated with serum levelsof bilirubin, alkaline phosphatase, AFP, pres-ence of ascites, and absence of clinical symp-toms at the time of diagnosis. The scoreranges from -7 to 12. The mean survival is 10months for low risk patients (score ≤1), 3.7months for patients at intermediate risk (scorefrom 2 to 7), and 1.4 months for high riskpatients (score ≥8). It should be pointed outthat the patients analyzed in the cited studywere Chinese, most of them (79%) withcirrhosis of the liver due to hepatitis B virus.Of the patients evaluated, 41.6% were sub-mitted to some antitumoral treatment whichmight have interfered with data assessment.The low mean survival for these patients (19weeks) may also suggest the selection ofpatients with advanced tumors.

Cancer of the Liver Italian Program (17)(Table 6). This program includes hepaticfunction, tumor morphology, presence ofthrombosis of the portal vein, and serumAFP levels. It should also be pointed out thatpart of the patients were submitted to eitherregional (57%) or systemic (18%) treatment,a fact that may change the data of this prog-nostic system. The tumors of the patients aredivided into 7 stages (0 to 6). Mean survivalis 35, 8 and 3 months for stages 0, 2 and 4-6,respectively (18).

As is the case for diagnosis, populationdifferences also influence the prognosticmodels. In contrast to other types of tumorsin which the neoplasia is responsible formortality, in HCC, cirrhosis of the liver isthe main factor related to patient survival. Ingeneral, hepatocellular function is consid-

the Karnofsky index, serum bilirubin levels,AFP, alkaline phosphatase, and the presenceof portal thrombosis detected by US. Thescore ranges from 0 to 11. The 1- and 2-yearsurvival rate was 72 and 51% for low riskpatients (score 0), 34 and 17% for patients of

Table 3. TNM classification of hepatocellular carcinomas (30).

Classification Definition

T1 Solitary tumor ≤2 cm in the widest diameter without vascular invasion.

T2 Solitary tumor ≤2 cm in the widest diameter with vascular invasion;or multiple tumors limited to one lobe, none of them >2 cm in thewidest diameter, without vascular invasion;or solitary tumor >2 cm in the widest diameter, without vascularinvasion.

T3 Solitary tumor >2 cm in the widest diameter with vascular invasion;or multiple tumors limited to one lobe, none of them >2 cm in thewidest diameter, with vascular invasion;or multiple tumors limited to one lobe, some of them >2 cm in thewidest diameter, with or without vascular invasion.

T4 Multiple tumors in more than one lobe;or tumor(s) invading a large portal branch or one or more suprahepaticveins. Solitary tumor >2 cm in the widest diameter with vascularinvasion.

Stage of TNM classification

I T1 / N0 / M0II T2 / N0 / M0III A T3 / N0 / M0III B T1 / N1 / M0

T2 / N1 / M0T3 / N1 / M0

IV A T4 / any N / M0IV B any T / any N / M1

T = tumor; N = nodes; M = metastasis; N0 = without regional lymph nodes; N1 = withregional lymph nodes; M0 = without distant metastasis; M1 = with distant metastasis.

1697



ered to represent one of the major variablesrelated to patient survival. In addition, he-patic dysfunction determines the type andefficacy of the treatment proposed.

Another factor just as important is thepresence of neoplastic vascular invasion.Four of the classifications presented earlierconsider the presence of vascular invasion tobe a factor related to survival, indicatingtumor dissemination. However, in most cases,vascular invasion is diagnosed only duringhistopathological analysis, i.e., after the ap-plication of therapeutic procedures such asresection and liver transplantation. Unfortu-nately, currently available imaging methodsare not sufficient for the diagnosis of tu-moral thrombosis of small hepatic vessels.

Serum AFP levels are also considered tobe of prognostic value in three classifica-tions. Due to the low general survival ratesof the patients evaluated by these classifica-tions, it is possible that selection of patientswith advanced tumors occurred, since earlyHCC tend not to express high AFP levels.

One of the great benefits of tumor classi-fication, in addition to providing an estimateof survival, is the selection of patients whocan be submitted to treatment. The treat-ments currently used for HCC have shownan effect on patient survival, especially sur-gical and percutaneous therapies. However,none of the classifications available consid-ers this variable (treatment). On this basis,the new classifications should evaluate thetreatment options as a prognostic factor andcorrelate the tumor stages with the form oftreatment to be adopted. The only classifica-tion that tries to correlate tumor stage withtreatment is that of the Barcelona Group(BCLC). However, this classification stillawaits prospective validation. Regional mul-ticenter studies on large patient series areneeded to clarify the best time and types oftreatment for each patient with HCC.

The main sites of HCC metastases are theadrenal glands, the bones and the lungs.Thus, it is imperative to perform bone scin-

Table 6. Cancer of the Liver Italian Program (CLIP)classification of hepatocellular carcinomas (17).

Variable Points

Child-PughA 0B 1C 2

Tumor morphologyUninodular and extension <50% 0Multinodular and extension ≤50% 1Diffuse (massive) or extension >50% 2

Alpha-1 fetoprotein<400 0≥400 1

Thrombosis of the portal veinNo 0Yes 1

CLIP classification: 0 to 6 points.

tigraphy and chest and abdomen tomogra-phy to rule out tumor dissemination. Al-though uncommon, brain metastasis mayoccur (33). For patients who benefit fromradical or curative treatment, mainly livertransplantation, we believe that skull tomog-raphy should be performed routinely to ex-

Table 5. Chinese University Prognostic Index(CUPI) classification of hepatocellular carcinomas(32).

Variable Score

TNM classificationI and II -3III A and III B -1IV A and IV B 0

Asymptomatic at diagnosis -4Ascites 3AFP ≥500 ng/ml 2Bilirubin (µmol/l)

<34 034-51 3>51 4

Alkaline phosphatase ≥200 IU/l 3

Score ≤1: low risk; score 2-7: intermediate risk;score ≥8: high risk. Risk of death within 3 months:>70% (high risk); 30 to 70% (intermediate risk);<30% (low risk). AFP = alpha-1 fetoprotein. Forstages of the TNM classification, see legend toTable 3.

1698



clude the presence of brain metastases.

Treatment

The small number of patients with thesame tumor stage impairs the execution ofrandomized and controlled studies, with aconsequent difficulty in reaching a reliableconclusion about the best treatment for HCC.However, the consensus is that for effectivetreatment the tumor must be detected in theearly phases of development. A tumor isconsidered to be in an early stage when itssize does not exceed 2 cm. However, singletumors of less than 5 cm or up to three

nodules, none of which exceeds 3 cm, areconsidered to be candidates for curative treat-ment. With follow-up programs for cirrhoticpatients by US and AFP, the number ofpatients that can be submitted to curativetreatment has increased. Tumors diagnosedin advanced phases, with vascular invasion,multinodular, and with distant metastasescannot be treated with the objective of im-proving patient survival. However, despitethe programs of early detection of HCC,only 1/3 of patients with HCC will benefitfrom treatment considered to be curative. Inour experience, we were able to performradical treatment in 25% of patients withHCC (3).

The fibrolamellar variant of HCC is morecommon among non-cirrhotic young patients.Due to the slow evolution and low metasta-sis rates of this tumor, the treatment of choiceis surgical resection even for tumors of largevolume since the hepatic functional reserveneeded for the procedure is maintained andthe remaining liver in most cases is normal.When the tumor is not resectable, liver trans-plantation may be indicated. However, mostHCC are not of the fibrolamellar variant andoccur in the presence of cirrhosis, with con-sequent impairment of treatment. Thus, weshall comment on the treatment of HCC inpatients with cirrhosis.

Surgical treatments (surgical resectionand liver transplantation) and percutaneoustreatment (alcoholization, radiofrequency andmicrowaves) are considered to be curative orradical. A 74% 5-year survival rate can bereached by patients submitted to liver trans-plantation (Table 7).

Surgical resection

Surgical resection is considered to be theoption of choice for the treatment of patientswith HCC. However, due to the frequentoccurrence of postoperative hepatic decom-pensation, this treatment modality should beindicated only for patients with preserved

Table 7. Survival after radical treatment of hepatocellular carcinoma.

N 1 year 2 years 5 years

Surgical resectionFong et al.,1999 (35) 100 83 - 42

<5 cm - - 57 -Llovet et al.,1999 (36)

No PH and TB NL 35 91 87 74PH and TB NL 15 93 59 50PH and TB >1 27 74 35 25

Arii et al.,2000 (34)Stage 1 <2 cm 1318 - 88 71Stage 1 2-5 cm 2722 - - 58

Yamamoto et al.,2001 (37) 58 97 84 61

Liver transplantationFigueras et al.,1997 (41) 38 82 75 63França,1997 (12)/Llovet et al.,1998 (26) 58 84 74 74Jonas et al.,2001 (42) 120 90 - 71Yao et al.,2001 (44)

≤pT2 46 91 - 72

AlcoholizationLivraghi et al.,1995 (50)

<5 cmChild A 293 98 79 47Child B 149 93 63 29Child C 20 64 0 0

Arii et al.,2000 (34)Stage 1 <2 cm 767 - 81 54Stage 1 2-5 cm 587 - - 39

Yamamoto et al.,2001 (37) 39 100 82 59

RadiofrequencyBuscarini et al.,2001 (52)

≤3.5 cm 88 89 62 33

NL = normal value; PH = portal hypertension; TB = total bilirubin.

1699



hepatic function. Non-judicious patient se-lection for surgical resection may not lead toincreased survival when surgical treatmentis compared to the natural history of HCC oreven to other less invasive therapies (34-37).The Child-Pugh classification and indocya-nine green clearance have been used to as-sess the extent of liver function impairmentbefore the indication of the surgical proce-dure. The presence of portal hypertensionrepresented by the portal pressure gradient(difference between occluded and free he-patic venous pressure) ≥10 mmHg has beenused as one of the main factors predictive ofhepatic decompensation after surgical resec-tion (36). This suggests that resection shouldbe indicated for patients without portal hy-pertension. When it is not possible to meas-ure portal pressure by the angiographic route,an invasive method which is not available atmost services in Brazil, other signs of portalhypertension can be determined, such assplenomegaly, presence of esophageal va-rices upon upper gastrointestinal endoscopy,and a platelet count of less than 100,000/mm3, and used as criteria to rule out resec-tion (36).

A single tumor measuring less than 5 cmin patients with preserved hepatic functionand in sufficiently good clinical conditionsto withstand the procedure is the criterionmost often used to indicate resection. Nod-ules larger than 5 cm present a higher prob-ability of invasion of the tumor capsule, withthe presence of satellite nodules indicatinglocal tumor dissemination.

Only about 10% of patients have thesetumoral characteristics, with a consequentinfrequent indication of surgical resection(38). Tumor localization, especially in aperihilar situation, may be a criterion forcontraindication of resection regardless ofthe characteristics of the tumor. Intraopera-tive US should be routinely used both todefine the safety margins and to excludeother lesions not visualized by preoperativeimaging techniques (39).

The survival rate after resection can reach97% for the 1st year and 74% for the 2nd,depending on residual hepatic reserve andtumor staging (39). When possible, conser-vative surgery should be performed, such assegmentectomy or sub-segmentectomy whichwill preserve a functioning liver mass. Tu-mor recurrence is observed in about 12, 60and 70% of patients after 1, 3 and 5 years,respectively (36,38) and is related to thepresence of satellite nodules and tumor dif-ferentiation (40). Recurrence may be localor may consist of the appearance of meta-chronic tumors since the cirrhotic liver, es-pecially when involved by extensive inflam-matory activity, continues to be a risk factor.Tumor size (>5 cm), vascular invasion, thepresence of satellite nodules, bilobar involve-ment, and the involvement of regional lymphnodes are considered to be factors related totumor recurrence (36,38).

Liver transplantation

Liver transplantation is the treatment ofchoice in cases of HCC limited to the liverthat cannot be submitted to surgical resec-tion due to poor hepatic function or to tech-nical impossibility. Liver transplantation notonly eliminates the neoplasia, but can alsocure the base liver disease. Some authorsadopt post-resection tumor recurrence as anindication for liver transplantation. Othersadopt liver transplantation as the treatmentof choice before resection. When strict se-lection criteria are used, such as a single,small tumor (<5 cm) without satellite nod-ules, without vascular invasion, without in-vasion of regional lymph nodes, without dis-tant metastases, and without an indicationfor resection, a satisfactory survival can beobtained (12,26,41-44). In our experience(12,26), with single tumors smaller than 5cm, the possibilities of survival reach 84, 74and 74% in the 1st, 2nd and 5th years afterliver transplantation, with a rate of tumorrecurrence of only 3.5%. These survival rates

1700



are similar to those observed for liver trans-plantation in patients without neoplasias(41,45). Thus, the ideal candidate for livertransplantation is a patient with a single HCCsmaller than 5 cm or with up to 3 nodules,none of them larger than 3 cm, without signsof neoplastic invasion of the portal system orof distant metastases.

Despite its advantages, the procedure alsoinvolves some disadvantages. The lack ofdonors with a consequent increase in thetime on the waiting list, the high cost of theprocedure, the possibility of tumor recur-rence, the frequent postoperative infections,the high rates of perioperative morbidity,and the quality of postoperative life are as-pects that should be taken into account at thetime when the decision for an indication ofliver transplantation is made.

Pre-liver transplantation co-adjuvanttreatment in order to prevent tumor progres-sion until the time for the surgical procedurehas been adopted at some transplant centerswhere the time on the waiting list is morethan 6 months. The real efficacy of thesetreatments for patient prognosis is still amatter of controversy. Some groups use arte-rial embolization with or without chemo-therapy as co-adjuvant pre-liver transplanta-tion treatment and localized chemotherapypost-liver transplantation (46). The combi-nation of techniques, such as embolizationand alcoholization, has demonstrated a satis-factory antitumoral effect and may be usefulfor co-adjuvant treatment before liver trans-plantation (47).

Strategies to increase the number of do-nors are of great importance. “Domino” andsplit liver transplants are options used toincrease the number of organs for transplan-tation. Living donor liver transplantationshould also be considered for patients withHCC in groups in which time on the waitinglist is more than 7 months (39). The use of an“expanded” criterion for living donor livertransplantation in HCC has been discussed.This criterion is based on the following fea-

tures (39): single tumor <7 cm or up to 3nodules, none >5 cm or up to 5 nodules,none >3 cm or a partial response to any of thetreatments fulfilling standard criteria. How-ever, these criteria still need validation andshould not be used in routine clinical prac-tice.

Immunosuppressive agents such as cy-closporine and tacrolimus are known to bestimulators of hepatic regeneration. How-ever, their interference with tumor progres-sion is still a matter of controversy.

Percutaneous treatment

Several types of percutaneous treatmentare available for HCC, all of them aiming atdestruction of the tumor with a safety marginof non-tumoral liver. The techniques mostcommonly used are alcoholization andradiofrequency. However, substances suchas boiling saline solution and acetic acid canalso be used. Coagulation by radiofrequency,microwaves, laser therapy, and electrocau-terization are percutaneous techniques usedfor the treatment of HCC. Percutaneous etha-nol injection (PEI) is the technique for whichmost experience has been obtained, withvarious studies showing its efficacy.

Absolute alcohol causes cell dehydrationand extensive coagulative cell necrosis inaddition to leading to thrombosis of theintratumoral vessels. PEI is a procedure ofeasy execution, good tolerability and lowcost, which can be applied during repeatedsessions (48,49). Using ultrasound, the alco-holization needle is introduced until it reachesthe nodule. A 22-G needle or a needle withspecific side perforations for the procedureis used. After reaching the tumor, alwaysunder US visualization, the injection of ab-solute alcohol is started. The amount of alco-hol injected per session depends on tumorsize and tolerability on the part of the pa-tient, ranging from 1 to 10 ml, with an aver-age of 5 ml per session. Larger volumes areinadvisable because of the risk of the occur-

1701



rence of an extensive area of hepatic necro-sis. However, injection of a large volume ofalcohol in a single session without the occur-rence of marked side effects has been re-ported in the literature. The number of alco-holization sessions depends on the size andconsistency of the tumor and on the distribu-tion of alcohol through the tumor. The finalobjective of this treatment is to obtain totalHCC necrosis.

Serious complications are rare. Pain dur-ing the procedure is common and is relatedto alcohol reflux towards the hepatic capsuleor to alcohol escape through the portal vein,visualized by US during the procedure. Inour routine we do not use sedation or sys-temic analgesia before PEI. There have beenreports of hemoperitoneum, pleurisy, hemo-bilia, hepatic abscess, cholangitis, and he-patic decompensation (50). Portal thrombo-sis may also occur after PEI as a conse-quence of vascular invasion by the tumor orof chemical thrombosis caused by alcohol.Differentiation of these events can be ob-tained using US-Doppler or FNAB of thethrombus (14).

A single tumor smaller than 3 cm or up tothree nodules, none of them larger than 3 cm,without extrahepatic metastases and withonly slightly deteriorated hepatic functionalreserve (Child-Pugh A and B) and withoutsurgical indication (34,37,48) are the majorindications for PEI.

The therapeutic efficacy of PEI dependson various factors. Vilana et al. (49), in ananalysis of alcoholization of tumors measur-ing less than 5 cm, observed that tumorssmaller than 3 cm gave a better response andconcluded that the success of PEI is relatedto tumor size at the beginning of treatment,as also reported by others. PEI may have atherapeutic efficacy similar to resection oreven to liver transplantation, especially whenpatients with poor hepatic function are se-lected for surgical treatment (48).

Local recurrence is observed in 17% ofcases. The appearance of new lesions after

treatment may reach 60% at 2 years and 80%at 5 years and is related to tumor size, inad-equate necrosis, presence of pretreatmentintrahepatic metastases, or even to the ap-pearance of metachronic lesions during pa-tient follow-up (49,50).

Depending on the degree of hepatic dys-function, the survival of patients submittedto PEI may reach 85% in the first year, 60%in the third, and 30% in the fifth, rates simi-lar to those obtained with surgical resection(34,37,48,50).

Radiofrequency has also been used suc-cessfully for patients with HCC. The indica-tions are similar to those for PEI (51). Norandomized studies comparing the two per-cutaneous techniques in terms of antitumoraleffect and patient survival are available. Theadvantage of radiofrequency over PEI is thesmaller number of sessions needed to obtaintumor necrosis. However, PEI is less expen-sive and is easy to perform, requiring nohospitalization. Radiofrequency should beavoided in superficial lesions because of therisk of tumoral dissemination through thepath of the needle. The 5-year survival ratefor patients submitted to radiofrequency isabout 33% (52). The option for PEI andradiofrequency depends on the experienceof each group.

Transarterial embolization orchemoembolization

Since HCC is a tumor predominantlyirrigated by the arterial system of the liver,blockade of the blood supply to the tumor isused as treatment. The surgical route hasbeen abandoned due to its severe side effectsand has been replaced with the use of arterialobstruction by a peripheral route using inter-ventionist radiology. Tumors that cannot besubmitted to radical treatment are consid-ered for transarterial embolization (TAE)/chemoembolization (TACE). Usually, pa-tients with multiple diffuse tumors or singletumors larger than 5 cm, with barely deterio-

1702



rated hepatic function are the main candi-dates for treatment by TAE (28).

Using fluoroscopy, the tumor nourishingartery is located and selective obstructionwith gelatin particles or preformed polyvinylmicroparticles is started after advancing thecatheter as close as possible to the HCC.This procedure may be combined with theuse of chemotherapeutic agents such as doxo-rubicin, epirubicin, mitomycin, or cisplatinin combination or not with lipiodol, a sub-stance retained by neoplastic cells. Thesecombinations seem to improve the antitu-moral effect, although in most cases they donot improve patient survival (53).

In a recent study, Llovet et al. (54) dem-onstrated improved survival of patients sub-mitted to TACE compared to control pa-tients (no antitumoral treatment). Patientswith multinodular HCC and asymptomaticwith respect to the tumor, without vascularinvasion or distant metastases, with preservedhepatic function (preferentially Child-PughA) and not considered to be candidates forliver transplantation may benefit from TACE.

Thrombosis of the portal vein and hepa-tofugal portal flow are considered to be acontraindication of TAE due to the risk ofsevere hepatic insufficiency after the proce-dure. The post-embolization syndrome, char-acterized by fever, abdominal pain, nauseaand vomiting, and frequently self-limited, isobserved in most patients submitted to TAEor TACE. Fifteen to 30% of the patients maypresent severe symptoms such as gallblad-der infarction by embolization of the cysticartery, arteritis, thrombosis of the intimalayer of the hepatic artery, pulmonary edema,pancreatitis, or even hepatic insufficiencydue to poor hepatic function. For this reason,this therapeutic modality is avoided in Child-Pugh C patients (53,55). A higher frequencyof post-embolization syndrome and of gas-trointestinal toxicity tends to occur in casesin which lipiodol is used as the chemothera-peutic vehicle. Antibiotic prophylaxis shouldnot be routinely applied to patients submit-

ted to TAE (56).TAE can reduce the size of the tumor,

leading to ischemic necrosis of more than80% of the HCC, while the cells that infil-trate the tumoral capsule and the portal veincontinue to be viable. Relapses are also fre-quent. In cases of relapse or persistence ofviable cells, the procedure can be repeated orpost-TAE PEI can be performed (55). How-ever, the probability of maintenance of thislevel of therapeutic response is about 10% at2 years. Larger tumors tend to respond toTAE at higher frequency, probably due tothe hypervascularization of larger HCC (53).Both TAE and TACE should be periodicallyrepeated. The time interval has not beendefined, but the suggestion is to repeat theprocedure every 3 to 6 months.

Patients with a single tumor larger than 3cm that cannot be treated surgically maybenefit from the combination of TAE andPEI. The combination of these two therapeu-tic techniques is based on the persistence ofviable neoplastic cells after TAE, especiallyat the periphery of the lesion, where thepredominant circulation seems to be thevenous one (47). In addition, large tumorsrequire larger amounts of alcohol to be ne-crotized. Thus, when TAE is performed first,necrosis of the central part of the HCC isobtained, and PEI is later performed at theperiphery. This combination of techniquesseems to be effective in reducing tumor sizeand increasing patient survival (57,58).

TACE has been used as co-adjuvant treat-ment when the patient is on the waiting listfor liver transplantation (46). In our experi-ence, the combination of TAE and PEI hasshown a good antitumoral effect in patientswho are on the waiting list for liver trans-plantation (47).

However, the possibility of tumor dis-semination after treatment, detected by thepresence of messenger RNA for AFP (58)and by a higher incidence of pulmonarymetastases after TAE, should be consideredat the time when a decision has to be made

1703



about co-adjuvant treatment in addition toradical therapy.

Evaluation of the therapeuticresponse

The therapeutic response to percutaneousmethods (PEI and radiofrequency) and to TAEis determined by using CT in the arterial phase,at least 1 month after the procedure. Beforethis period, areas suspected of tumor viabilitymay be detected, corresponding to areas ofperitumoral edema caused by the therapeuticprocedure. The persistence of a hyperdenseimage in the arterial phase is suggestive oftumor viability. The absence of contrast up-take is indicative of tumor necrosis. In MR, theloss of signal intensity in potentiated sequencesin T2 and after the administration of paramag-netic contrast in potentiated sequences in T1 isalso suggestive of tumor necrosis. US-Dop-pler has been used for the control of treatmentwith satisfactory results. The presence of anintratumoral pulsatile flow indicates the per-sistence of viable tumor cells. However, theabsence of a Doppler signal does not rule outthe possibility of tumor viability.

A quarterly US, measurement of serumAFP and dynamic CT at 6- to 8-month inter-vals are suggested for post-treatment follow-up. The absence of an increase in the lesion,in AFP levels or in contrast uptake by CT isindicative of successful treatment. The WorldHealth Organization adopts the followingcriteria for response to treatment (59):

Complete response: complete disappear-ance of known lesions and no occurrence of

new lesions as evaluated during two obser-vations separated by an interval of at least 4weeks.

Partial response: reduction of the tumormass ≥50% as evaluated during two obser-vations separated by an interval of at least 4weeks.

Stable disease: not classified as completeresponse, partial response or progressive dis-ease

Progressive disease: an increase ≥25% inthe tumoral mass of one or more knownlesions or appearance of new lesions.

Serum AFP levels can be used as a pa-rameter of response to treatment only incases in which their levels were elevatedbefore the procedure. Elevation of their lev-els after treatment is suggestive of tumorrecurrence.

Other therapies

In HCC, both estrogen and androgen re-ceptors can be found on the membrane ofneoplastic cells, theoretically justifying hor-monal therapy for this type of neoplasia. Theuse of antiestrogens has been suggested forthe treatment of HCC. However, tamoxifen,a non-steroid antiestrogen, did not prove tobe useful in improving the quality of life ofpatients with HCC even when administeredat high doses (60).

Radiotherapy, systemic chemotherapyand the use of interferon have not shownsatisfactory results in terms of antitumoraleffect or survival of patients with HCC (14).

References

1. Bosch FX, Ribes J & Borras J (1999). Epidemiology of primary livercancer. Seminars in Liver Disease, 19: 271-285.

2. Okuda K (1997). Epidemiology. In: Livraghi T, Makuushi M &Buscarini L (Editors), Diagnosis and Treatment of HepatocellularCarcinoma. GMM, London, UK.

3. Lescano M, Carneiro M, Elias Junior J, Martinelli A & França A(2002). Experiencia inicial en la evaluación de pacientes with carci-noma hepatocelular em um Hospital terciario. Gastroenterología y

Hepatología, 25 (Suppl 2): I-9-I-43.4. Carrilho FJ (1993). Carcinoma hepatocelular e cirrose hepática. Es-

tudo caso-controle de variáveis clínicas, bioquímicas, sorológicas ehistológicas. Doctoral thesis, Departamento de Gastroenterologia,Universidade de São Paulo, São Paulo, SP, Brazil.

5. Okuda K, Ohtsuki T, Obata H, Tomimatsu M, Okazaki N, HasegawaH, Nakajima Y & Ohnishi K (1985). Natural history of hepatocellularcarcinoma and prognosis in relation to treatment. Study of 850

1704



patients. Cancer, 56: 918-928.6. Barbara L, Benzi G, Gaiani S, Fusconi F, Zironi G, Siringo S, Rigamonti

A, Barbara C, Grigioni W & Mazziotti A (1992). Natural history ofsmall hepatocellular carcinoma in cirrhosis: a multivariate analysisof prognostic factors of tumor growth rate and patient survival.Hepatology, 16: 132-137.

7. Sheu JC, Sung JL, Chen DS et al. (1985). Early detection of hepato-cellular carcinoma by real-time ultrasonography. A prospectivestudy. Cancer, 56: 660-666.

8. Ebara M, Ohto M & Kondo F (1989). Strategy for early diagnosis ofhepatocellular carcinoma (HCC). Annals of the Academy of Medi-cine, Singapore, 18: 83-89.

9. Solmi L, Primerano AMM & Gandolfi L (1996). Ultrasound follow-upof patients at risk for hepatocellular carcinoma. Results of a pro-spective study on 360 cases. American Journal of Gastroenterol-ogy, 91: 1189-1194.

10. Bizollon T, Rode A, Bancel B, Gueripel V, Ducerf C, Bauliex J &Trepo C (1998). Diagnostic value and tolerance of lipiodol-computedtomography for the detection of small hepatocellular carcinoma:correlation with pathologic examination of explanted livers. Journalof Hepatology, 28: 491-496.

11. Dodd III, Miller WJ, Baron RL, Skolnick ML & Campbell WL (1992).Detection of malignant tumors in end-stage cirrhotic livers. Efficacyof sonography as a screening technique. American Journal of Roent-genology, 159: 727-733.

12. França AVC (1997). Carcinoma hepatocelular e transplante hepá-tico. Valor dos métodos de imagem no diagnóstico e estadiamentotumoral e na sobrevida de 58 pacientes. Doctoral thesis, Departa-mento de Gastroenterologia, Universidade de São Paulo, São Paulo,SP, Brazil.

13. Tanaka K, Numata K, Okazaki H, Nakamura S, Inoue S & Takamura Y(1993). Diagnosis of portal vein thrombosis in patients with hepato-cellular carcinoma. Efficacy of color Doppler sonography comparedwith angiography. American Journal of Roentgenology, 160: 1279-1283.

14. Vilana R, Bru C, Bruix J, Castells A, Solé M & Rodes J (1993). Fine-needle aspiration biopsy of portal vein thrombus. Value in detectingmalignant thrombosis. American Journal of Roentgenology, 160:1285-1287.

15. Llovet JM & Beaugrand M (2003). Hepatocellular carcinoma: pres-ent status and future prospects. Journal of Hepatology, 38 (Suppl):I-136-I-149.

16. Nomura F, Ohnishi K & Tanabe Y (1989). Clinical features andprognosis of hepatocellular carcinoma with reference to serumalpha-fetoprotein levels. Cancer, 64: 1700-1707.

17. The Cancer of the Liver Italian Program (CLIP) Investigators (1998).A new prognostic system for hepatocellular carcinoma: A retro-spective study of 435 patients. Hepatology, 28: 751-755.

18. The Cancer of the Liver Italian Program (CLIP) Investigators (2000).Prospective validation of the CLIP score: A new prognostic systemfor patients with cirrhosis and hepatocellular carcinoma. Hepatolo-gy, 31: 840-845.

19. Ikeda K, Saitoh S, Koida I, Tsubota A, Arase Y, Chayama K &Kumada H (1994). Imaging diagnosis of small hepatocellular carci-noma. Hepatology, 20: 82-87.

20. Ohashi O, Hanafusa K & Yoshida T (1993). Small hepatocellularcarcinomas. Two-phase dynamic incremental CT in detection andevaluation. Radiology, 189: 851-855.

21. Rummeny E, Weissleder R, Stark DD, Saini S, Compton CC, BennettW, Hahn PF, Wittenberg J, Malt RA & Ferrucci JT (1989). Primaryliver tumors. Diagnosis by MR imaging. American Journal of Roent-

genology, 152: 63-72.22. Ebara M, Ohto M, Watanabe Y et al. (1986). Diagnosis of small

hepatocellular carcinoma. Correlation of MR imaging and tumorhistology studies. Radiology, 159: 371-377.

23. Shimamoto K, Sakuma S, Ishigaki T, Ishiguchi T, Itoh S & Fukatsu H(1992). Hepatocellular carcinoma. Evaluation with color Doppler USand MR imaging. Radiology, 182: 149-153.

24. Takayasu K, Shima Y, Muramatsu Y et al. (1986). Angiography ofsmall hepatocellular carcinoma. Analysis of 105 resected tumors.American Journal of Roentgenology, 147: 525-529.

25. Sumida M, Ohto M, Ebara M, Kimura K, Okuda K & Hirroka N (1986).Accuracy of angiography in the diagnosis of small hepatocellularcarcinoma. American Journal of Roentgenology, 147: 531-536.

26. Llovet JM, Bruix J, Fuster J et al. (1998). Liver transplantation forsmall hepatocellular carcinoma: the tumor-node-metastasis classifi-cation does not have prognostic power. Hepatology, 27: 1572-1577.

27. França A, Giordano H, Trevisan M, Escanhoela C, Seva-Pereira T,Zucoloto S, Martinelli A & Soares E (2003). Fine needle aspirationbiopsy improves the diagnostic accuracy of cut needle biopsy offocal liver lesions. Acta Cytologica, 47: 332-336.

28. Bruix J, Sherman M, Llovet JM, Beaugrand M, Lencioni R,Burroughs AK, Christensen E, Pagliaro L, Colombo M & Rodés J(2001). Clinical management of hepatocellular carcinoma. Conclu-sions of the Barcelona-2000 EASL conference. Journal of Hepatolo-gy, 35: 421-430.

29. Llovet JM, Bru C & Bruix J (1999). Prognosis of hepatocellularcarcinoma: the BCLC staging classification. Seminars in Liver Dis-ease, 19: 329-338.

30. Greene F, Page D, Fleming I, Fritz A, Balch C, Haller D & Morrow M(2002). AJCC Cancer Staging Handbook. 6th edn. Springer-Verlag,New York, 145-153.

31. Chevret S, Trinchet J-C, Mathieu D, Rached AA, Beaugrand M &Chastang C (1999). A new prognostic classification for predictingsurvival in patients with hepatocellular carcinoma. Journal of Hepa-tology, 31: 133-141.

32. Leung TWT, Tang AMY, Zee B, Lau WY, Lai PBS, Leung KL, Lau JT,Yu SC & Johnson PJ (2002). Construction of the Chinese UniversityPrognostic Index for hepatocellular carcinoma and comparison withthe TNM staging system, the Okuda staging system, and the can-cer of the liver Italian program staging system. Cancer, 94: 1760-1769.

33. França AVC, Martinelli ALC, Castro e Silva O & Grupo Integrado deTransplante de Fígado do Hospital das Clínicas da Faculdade deMedicina de Ribeirão Preto da Universidade de São Paulo (2004).Brain metastasis of hepatocellular carcinoma detected after livertransplantation. Arquivos de Gastroenterologia (in press).

34. Arii S, Yamaoka Y, Kutagawa S, Inoue K, Kobayashi K, Kojiro M,Makuushi M, Nakamura Y, Okita K & Yamada R (2000). Results ofsurgical and nonsurgical treatment for small-sized hepatocellularcarcinoma: a retrospective and nationwide survey in Japan. Hepa-tology, 32: 1224-1229.

35. Fong Y, Sun RL, Jarnagin W & Blumgat LH (1999). An analysis of412 cases of hepatocellular carcinoma at a Western center. Annalsof Surgery, 229: 790-799.

36. Llovet JM, Fuster J & Bruix J (1999). Intention-to-treat analysis ofsurgical treatment for early hepatocellular carcinoma: resection ver-sus transplantation. Hepatology, 30: 1434-1440.

37. Yamamoto J, Okada S, Shimada K, Okusata T, Yamasaki S, Ueno H& Kosuge T (2001). Treatment strategy for small hepatocellularcarcinoma: comparison of long-term results after percutaneousethanol injection therapy and surgical resection. Hepatology, 34:

1705



707-713.38. Fuster J, García-Valdecasas JC, Grande L et al. (1996). Hepatocellu-

lar carcinoma and cirrhosis. Results of surgical treatment in a Euro-pean series. Annals of Surgery, 233: 297-302.

39. Bruix J & Llovet JM (2002). Prognostic prediction and treatmentstrategy in hepatocellular carcinoma. Hepatology, 35: 519-524.

40. Michel J, Suc B, Fourtanier G, Durand D, Rumeau JL, Rostaing L &Lloveras JJ (1995). Recurrence of hepatocellular carcinoma in cir-rhotic patients after liver resection or transplantation. Transplanta-tion Proceedings, 27: 1798-1800.

41. Figueras J, Jaurrieta E, Valls C et al. (1997). Survival after livertransplantation in cirrhotic patients with and without hepatocellularcarcinoma: a comparative study. Hepatology, 25: 1485-1489.

42. Jonas S, Bechstein WO, Steinmüller T, Herrmann M, Radke C, BergT, Settmacher U & Neuhaus P (2001). Vascular invasion and histo-pathologic grading determine outcome after liver transplantation forhepatocellular carcinoma in cirrhosis. Hepatology, 33: 1080-1086.

43. Mazzaferro V, Regalla E, Doci R, Andreola S, Pulvirenti A, Bozzetti F,Montalto F, Ammatuna M, Morabito A & Gennari L (1996). Livertransplantation for the treatment of small hepatocellular carcinomain patients with cirrhosis. New England Journal of Medicine, 334:693-699.

44. Yao FY, Ferrel L, Bass NM, Watson JJ, Bachetti P, Venook A,Ascher NL & Roberts JP (2001). Liver transplantation for hepatocel-lular carcinoma: expansion of the tumor size limits does not ad-versely impact survival. Hepatology, 33: 1394-1403.

45. Lohmann R, Bechstein WO, Langrehr JM, Knoop M, Lobeck H,Keck H, Lemmems HP, Blumhardt G & Neuhaus P (1995). Analysisof the risk factors for recurrence of hepatocellular carcinoma afterorthotopic liver transplantation. Transplantation Proceedings, 27:1245-1246.

46. Vennok AP, Ferrel LD, Roberts JP, Emond J, Frye JW, Ring E,Ascher NL & Lake JR (1995). Liver transplantation for hepatocellularcarcinoma. Results with preoperative chemoembolization. LiverTransplantation and Surgery, 1: 242-248.

47. França AVC, Lescano MAL & Martinelli ALC (2002). Tratamientocombinado coadyuvante para el carcinoma hepatocelular previo altrasplante hepático. Gastroenterología y Hepatología, 25: 153-155.

48. Livraghi T, Bolondi L, Buscarini L, Cottone M, Mazziotti A, MorabitoA, Torzilli G & The Italian Cooperative HCC Study Group (1995). Notreatment, resection and ethanol injection in hepatocellular carcino-ma: a retrospective analysis of survival in 391 patients with cirrho-sis. Journal of Hepatology, 22: 522-526.

49. Vilana R, Bruix J, Bru C, Ayuso C, Solé M & Rodés J (1992). Tumorsize determines the efficacy of percutaneous ethanol injection for

the treatment of small hepatocellular carcinoma. Hepatology, 16:353-357.

50. Livraghi T, Giorgio A, Marin G et al. (1995). Hepatocellular carcino-ma and cirrhosis in 746 patients. Long-term results of percutaneousethanol injection. Radiology, 197: 101-108.

51. Livraghi T, Goldberg SN, Lazzaroni S, Meloni F, Solbiati L & GazelleGS (1999). Small hepatocellular carcinoma: treatment with radio-frequency ablation versus ethanol injection. Radiology, 210: 655-661.

52. Buscarini E, Di Stasi M, Vallisa D, Quaretti P & Rocca A (2001).Percutaneous radiofrequency ablation of small hepatocellular carci-noma: long-term results. European Radiology, 11: 914-921.

53. Spreafico C, Marchiano A, Regalia E, Frigerio LF, Garbagnati F,Andreola S, Millena M, Lanocita R & Mazzaferro V (1994).Chemoembolization of hepatocellular carcinoma in patients whoundergo liver transplantation. Radiology, 192: 687-690.

54. Llovet JM, Real M, Montaña X et al. (2002). Arterial embolization orchemoembolization versus symptomatic treatment in patients withunresectable hepatocellular carcinoma: a randomized controlledtrial. Lancet, 359: 1734-1739.

55. Groupe D’etude et de Traitment du Carcinome Hepatocellulaire(1995). A comparison of lipiodol chemoembolization and conserva-tive treatment for unresectable hepatocellular carcinoma. New Eng-land Journal of Medicine, 332: 1256-1261.

56. Castells A, Bruix J, Ayuso C, Bru C, Montanyà X, Boix L & Rodés J(1995). Transarterial embolization for hepatocellular carcinoma. An-tibiotic prophylaxis and clinical meaning of postembolization fever.Journal of Hepatology, 22: 410-415.

57. Koda M, Murawaki Y, Mitsuda A, Oyama K, Okamoto K, Odobe Y,Suou T & Kawasaki H (2001). Combination therapy with transcath-eter arterial chemoembolization and percutaneous ethanol injectioncompared with percutaneous ethanol injection alone for patientswith small hepatocellular carcinoma. A randomized control study.Cancer, 92: 1516-1524.

58. Boix L, Bruix J, Castells A, Vianney A, Llovet JM, Rivera F & RodésJ (1996). Circulating mRNA for alpha-fetoprotein in patients withhepatocellular carcinoma. Evidence of tumor dissemination aftertransarterial embolization. Hepatology, 24: 349A (Abstract).

59. Miller AB, Hoosgstraten B, Staquet M & Winkler A (1981). Report-ing results of cancer treatment. Cancer, 47: 207-214.

60. Chow PK, Tao BC, Tan CK, Machin D, Win KM, Johnson PJ & SooKC (2002). High-dose tamoxifen in the treatment of inoperablehepatocellular carcinoma: a multicenter randomized controlled trial.Hepatology, 36: 1221-1226.

Documents

Diagnosis, staging and treatment of hepatocellular carcinoma