12 Clinico-Pathological Features of Hepatocellular Carcinoma

12 Clinico-Pathological Features of Hepatocellular Carcinoma

Massimo Colombo and Guido Ronchi

M. Colombo, MD; G. Ronchi , MD

Department of Gastroenterology and Endocrinology, IRCCS Maggiore Hospital, University of Milan, Via Pace 9, 20122 Milano, Italy

CONTENTS

12.1 Introduction 169

12.2 The Pathological Classifi cation 169 12.3 Early Detected Tumors 170 12.4 Special Types of Tumors 171 12.5 Diagnosis 171

12.6 Staging 172

12.7 Natural History of the Tumor 173 12.8 Conclusions 174

References 175

12.1

Introduction

Hepatocellular carcinoma (HCC) is a major health problem worldwide due to its high incidence (ap- proximately 600,000 new cases in 2000), and severe natural history. Indeed, the incidence and mortal- ity rates associated with this disease signifi cantly overlap worldwide (Parkin et al. 2001). The identi- fi cation of chronic liver disease as the relevant risk factor for this tumor has made surveillance aimed at early detection of HCC possible and surveillance is now universally recognized to be the practical approach for improving the treatment of HCC pa- tients (Bruix et al. 2001). The few cases (<5%) of HCCs that do not develop with a background of chronic liver disease present late and usually have poor prognosis (Bralet et al. 2000). The under- standing of both the natural history and staging of HCC is hampered by the epidemiologic and clinical variability of the tumor. This, in turn, is infl uenced by the concurrence of multiple co-morbidity fac- tors in the same patient as well as by the presence of multiple distinct cell lines in the liver that may develop into liver cell cancer (Sell 2002).

12.2

The Pathological Classifi cation

HCC is classifi ed as nodular, massive or diffuse. The nodular type occurs as a nodule sharply delineated from the surrounding liver. The massive type occupies a large area and infi ltrates the neighboring hepatic tis- sue with satellite nodules. The diffuse type is charac- terized by the diffuse involvement of the liver (Kojiro 1997). All three forms of HCC occur with a background of chronic liver disease or of an otherwise normal liver.

The growth pattern of HCC may be infi ltrative, expand- ing, multinodular and mixed type. Based on histology, the WHO proposed a classifi cation of HCC into tra- becular, acinar, compact and scirrhous (Gibson and Sobin 1978). In the trabecular type, tumor cells are arranged in cords of variable cell thickness separated by sinusoids, with minimal or no fi brosis (Fig. 12.1).

The acinar (pseudoglandular) type is characterized by cells arranged in gland-like structures, fi lled with cellu- lar debris, exudates and macrophages (Figs. 12.2, 12.3).

The compact type shows tumor cells that are packed in a solid mass with inconspicuous sinusoids (Fig. 12.4).

In the scirrhous type, signifi cant fi brous tissue sepa- rates cords of tumor cells. Each histological type is

Fig. 12.1. Trabecular hepatocellular carcinoma. Trabecular

pattern of well-differentiated neoplastic hepatocytes arranged

in plates which are between three and four cells in thickness

(×30)

further classifi ed according to different grades of cell differentiation. Well differentiated HCC is a trabecular tumor with two- to three-cell thick cords. The anaplas- tic tumor usually shows a solid growth pattern, with pleomorphic and giant syncytial cells (Fig. 12.5).

12.3

Early Detected Tumors

Surveillance of patients with cirrhosis has led to an in- creasing number of cancers detected early in the form of small nodules that fi rst appear as well-differentiated tumors and proliferate along with gradual dediffer- entiation (Kojiro 1998). A sizable number of tumors arising in cirrhotic livers seem to occur in a multi- centric distribution and a certain proportion of them may arise from dysplastic nodules (International Working Party 1995). HCCs ranging from 1–2 cm in size may present with a fi brous capsule and/or fi - brous septa in contrast to other indistinct nodular small cancers that have indistinct margins despite such tumors being clearly detected as hypoechoic or hyperechoic focal lesions on ultrasound (US) exami- nation. The latter have been considered carcinoma in situ of the liver due to the absence of invasion into the portal vein branches and intrahepatic metastases (Kojiro 2002). Minute HCCs of the indistinct nodu- lar type are diffi cult to differentiate from high grade dysplastic nodules. The majority of small (less than 1.5 cm) HCCs of the indistinct nodular type are not detected as hypervascular tumors by contrast imaging, whereas distinct nodular type tumors almost invari- ably show hypervascular features during the arterial phase of contrast imaging (Kojiro 2002). A combina- tion of the lack of fi brotic capsule and reduced num- ber of unpaired arteries per square millimeter in less than 1.5 cm tumors accounts for many false negative diagnoses of HCC with contrast imaging. Since well- differentiated tumors in the early stages proliferate along with the occurrence of gradual dedifferentiation (Kojiro 1998), more histological grades are seen in tu- mors greater than 1 cm in size. A “nodule-in-nodule”

Fig. 12.3. Acinar clear-cell hepatocellular carcinoma. Clear-cell hepatocytes are arranged in glands with lumen fi lled by bili- ary plugs (×120)

Fig. 12.4. Solid hepatocellular carcinoma. Nodular growth of neoplastic hepatocytes with prominent nucleoli. The cells grow in solid sheets with few vascular channels (×120)

Fig. 12.5. Undifferentiated hepatocellular carcinoma.

Pleomorphic liver cells growing in solid pattern with evidence of many syncytial giant cells (×75)

Fig. 12.2. Acinar hepatocellular carcinoma. Prevalence of

acinar pattern of well-differentiated hepatocytes arranged in

acini with dilated lumen (×30)

appearance has been documented in less differenti- ated tumors expanding with a clear boundary within a well-differentiated tumor.

Biopsy examination of 1–2 cm nodules in patients with cirrhosis often implies differential diagnosis between a well-differentiated HCC and a large regen- erative nodule. Histological changes seen in a biop- sied nodule are best evaluated in comparison with those of control extranodular tissue from the same liver (Table 12.1) (Borzio et al. 1994). Moderately differentiated HCC has a typical trabecular pattern, whereas poorly differentiated HCC may show a tra- becular, solid or sarcomatous like pattern.

12.4

Special Types of Tumors

Special types of HCC have distinct histologic pat- terns and natural history: fi brolamellar carcinoma, clear cell HCC and pedunculated HCC. Fibrolamellar carcinoma is composed of large eosinophilic cells arranged in thin or thick trabeculae that are sur- rounded by fi brous bounds with lamellar stranding

makes fi brolamellar carcinoma more often suitable for resection than the usual HCC (Craig 1997).

12.5 Diagnosis

Both cytohistologic criteria and non-invasive crite- ria allow diagnosis of HCC (Table 12.2). In patients with chronic liver disease, large tumors are easily diagnosed by combining clinical and radiological procedures. Arterial hypervascularization by tripha- sic spiral CT or MR of a liver mass identifi ed by abdominal US, is diagnostic for HCC (Bruix et al.

2001). Diagnosis is further confi rmed by greater than 400 ng/ml serum levels of AFP. During the arterial phase of spiral CT highly vascularized HCCs appear against a background of relatively unenhanced liver that is primarily contrasted during the late portal vein phase. In patients not fulfi lling these diagnostic criteria, the diagnosis of HCC is made by echo-guided aspiration cytology or microhistology (Bruix et al.

2001). The histological diagnosis of HCC is needed in the absence of contraindications and when a defi nite diagnosis may infl uence the choice of treatment. The diagnosis of small HCCs (1–2 cm in diameter) identi- fi ed by chance or during surveillance, may be diffi - cult. In principle, a lesion seen as either a hypoechoic or hyperechoic nodule in the liver of a patient with chronic liver disease should be presumed to be a pre- neoplastic lesion, like a macroregenerative nodule, or an HCC, and should be investigated accordingly.

For tumors of 1–2 cm in diameter the risk of false negative diagnoses with contrast imaging technique could be as high as 50% due to immature arterial vascularization of the nodule (Kojiro 2002). The highest diagnostic accuracy (85%) for these nodules was provided by the combined use of fi ne needle as- piration cytology plus intranodular and extranodu- lar fi ne needle microhistology (Table 12.1) (Borzio et al. 1994). Complications in patients subjected to

Fig. 12.6. Fibrolamellar hepatocellular carcinoma. Cords of neoplastic liver cells are separated by lamellar fi brous strands (×75)

Table 12.1. Diagnostic accuracy of fi ne-needle aspiration (A) and biopsy (B) of 36 ≤2-cm nodules developing in patients with compensated cirrhosis

Liver sampling Accuracy

Intra- + extranodular A+B 85%

Intranodular A+B 78%

Intra- + extranodular B 67%

Intranodular B 54%

Intranodular A 31%

(Fig. 12.6). The tumor occurs primarily in non-cir-

rhotic livers of young adults with equal frequency

in males and females. In the USA the incidence of

this tumor is approximately 1% of all HCCs. Serum

α-fetoprotein (AFP) is elevated in a minority of the

patients. In half the patients, an abdominal X-ray

reveals minute calcifi cations within the tumor that

are uncommon in HCC. The fact that the tumor is

sharply demarcated and arises in non-cirrhotic livers

fi ne-needle biopsy are hemoperitoneum (<0.5%) and tumor seeding along the needle track (3%–5%) (Takamori et al. 2000; Kim et al. 2000). Recently, 3D image reconstruction techniques by MR was proven to be superior to spiral CT in the diagnosis of HCC nodules of 1–2 cm in size (84% vs 47% detection rates). As expected both imaging techniques failed to recognize tumors of less than 1 cm in diameter (32%

vs 10% detection rates) as a consequence of poor arterialization of small nodules (Table 12.3) (Burrel

et al. 2003). Cases not resolved by imaging or liver biopsy should be followed up with imaging tech- niques performed at 3-month intervals (enhanced follow-up), until diagnosis is obtained (Fig. 12.7). The differential diagnosis of HCC or fi brolamellar carci- noma includes focal nodular hyperplasia, metastatic carcinomas, neuroendocrine carcinoma and cholan- giocarcinoma (Craig 1997).

12.6 Staging

Staging is a crucial variable in treatment outcome since many therapeutic failures have resulted from incorrect patient selection. Tumor dedifferentiation and vascular invasion by tumor cells have constantly emerged as independent predictors of shortened survival in patients undergoing hepatic resection or transplantation for HCC. Although, tumor size and number appear to be clinical surrogates predicting tumor dedifferentiation and vascular invasion, tu- mor-related criteria like the tumor, node and me- tastases (TNM) classifi cation do not accurately pre- dict patient survival. The latter is better predicted by criteria combining tumor characteristics, func- tional status and liver function. Triphasic spiral CT and dynamic MR are currently used for assessing the number, size and vascular invasiveness of the tumor.

In the Barcelona Clinic Liver Cancer staging classi- fi cation the functional status of the patient and the liver status are measured by the Performance Status and Child-Pugh score system, respectively (Llovet et al. 1999). The Barcelona classifi cation comprises four stages that select the best candidates for the best therapies available, i.e. from early tumor stage (Stage

Fig. 12.7. Surveillance of patients with compensated cirrhosis as suggested by the Conference of the European Association for the Study of the Liver

Table 12.2. Diagnostic criteria for hepatocellular carcinoma as proposed during the Conference of the European Association for the Study of the Liver in Barcelona

• Cytohistologic criteria

• Non-invasive criteria (cirrhotic patients) 1. Radiological criteria:

Two coincident imaging techniquesa

Focal lesion >2 cm with arterial hypervascularization 2. Combined criteria:

One imaging technique associated to AFP

Focal lesion >2 cm with arterial hypervascularization AFP levels >400 ng/ml

a

Four techniques considered: US, spiral CT, MR and angiography.

Table 12.3. Superiority of magnetic resonance (MR) angiog- raphy with respect to helical computerized tomography (CT) for detection of early hepatocellular carcinoma prior to liver transplantation

Tumor size Number Detected Detected by CT pValue (mm) of tumors by MR

<10 22 7 (22%) 2 (10%) n.s.

11–20 19 16 (84%) 9 (47%) p=0.04

>20 6 6 (100%) 6 (100%) n.s.

A) that includes asymptomatic patients with small tumors suitable for radical therapies to late tumor stage (Stage D) that includes patients with untreat- able disease (Table 12.4).

The Cancer of the Liver Italian Program (CLIP) system allocates points for four variables that af- fect prognosis including Child-Pugh stage, tumor morphologic features (single, multiple or massive tumor), serum AFP level and portal vein thrombosis (Table 12.5) (CLIP 1998). Although this scoring sys- tem has been partially validated (CLIP 2000) and is easy to use, the CLIP score has suboptimal sensitivity for tumor invasiveness, since patients with a score of 0 may have from 0% to 50% of their liver replaced by HCC. Since the score is defi nitively skewed toward more severely affected patients whose disease is not amenable to curative treatment, too many patients with a CLIP score of 0 will not meet currently ac- cepted criteria for surgery or locoregional ablation of the tumor that have been proven to be effi cacious in patients in whom there is one tumor node of less than 5 cm in size (Bruix et al. 2001). In recent years, other staging systems have been proposed including the Chinese University Prognostic Index (Leung et al.

2002), the modifi ed TNM (Henderson et al. 2003), a French score system (Chevret et al. 1999) and a German score system (Rabe et al. 2003). Tumor ag- gressiveness appears to be predicted also by the tissue expression of genetically modifi ed estrogen recep- tors (VER): VER identifi ed tumors with shorter dou- bling time and was a negative predictor of survival compared to patients with a HCC expressing wild ER (Villa et al. 2000). In a comparative study, patient classifi cation based on ER was a better predictor of survival in patients with inoperable HCC compared to CLIP, Barcelona and French staging scores (Villa et al. 2003). However, one important limitation of the ER score system is its invasive approach to obtain a sample of liver tissue. Since staging scores developed thus far refl ect differences in demographic features

of the patients seen locally, expertise and treatment algorithms adopted in different centers, one wonders whether it is worth attempting to reach consensus on a single model for staging HCC. From a clinical point of view, it appears mandatory that prognostication of liver cancer should always incorporate treatment-de- pendent variables.

12.7

Natural History of the Tumor

The tumor size when HCC is fi rst detected does not predict the course of the disease in all cases. In fact, the median time of doubling volume for a small HCC may range from 1 to 20 months (Barbara et al. 1992;

Ebara et al. 1986; Okazaki et al. 1989). The tumor is a clinically indolent disease during the early phases of growth, whereas in the advanced stages it often presents with painful hepatomegaly and/or jaundice.

In the majority of patients with compensated cirrho- sis undergoing surveillance, HCC is fi rst detected as a single node (Table 12.6). The multinodular pattern of the tumor appears to be more common in pa- tients with multiple etiological factors than in those with a single etiologic factor (Benvegnù et al. 2001;

Fasani et al. 1999). Primary and secondary HCCs may be differentiated by matching radiological and histopathological fi ndings on explanted or resected livers only. Distinction between these two conditions bears important clinical implications, since second primary tumors appear to be less aggressive than metastatic tumors and recur less frequently after ab- lation than the former tumors (Kumada et al. 1997).

The growth pattern of HCC varies greatly from one tumor to another and may have clinical implications, since it infl uences the choice and outcome of treat- ments. Slowly expanding tumors (Franco et al. 1990) are more commonly seen in Caucasian and Asian pa-

Table 12.4. The Barcelona Clinic Liver Cancer Staging Clas- sifi cation of patients with hepatocellular carcinoma Staging Performance Tumor stage Child-

status Pugh

(A) Early 0 Single <5 cm

3 nodes <3 cm A & B (B) Intermediate 0 Large/multinodular A & B (C) Advanced 1–2 Vascular invasion A & B

extrahepatic spread

(D) End-stage 3–4 Any of the above C

Table 12.5. Cancer of the Liver Italian Program (CLIP) staging classifi cation of hepatocellular carcinoma

Score Tumor Child- AFP Vascular

morphology Pugh invasion

0 Uninodular

<50% of the liver A <400 mg/dl No 1 Multinodular

>50% of the liver B >400 mg/dl Yes

2 Massive C - -

tients than in South African patients who have more fast growing, replacing type tumors (Anthony 1973).

Further complicating the assessment of tumor course is that some HCC nodes have constant rates of growth during follow-up, while others either have a declining growth rate in the late phases of follow-up or, after an initial phase of resting, increase in volume exponen- tially (Ebara et al. 1986; Okazaki et al. 1989). This great diversity of the tumor growth patterns makes the predictive power of the size of the tumor at diag- nosis not absolute and explains why prognostication in HCC patients can be more reliably obtained by combining tumor size with liver function.

One controversial issue that bears important clinical implication is the presence of microscopic vessel invasion by the tumor, that is considered di- rect evidence of intrahepatic metastasis. Although macroscopic venous invasion seen with CT or MR scan is a well-established prognostic indicator and is one of the variables in the pathologic staging of HCC, the clinical signifi cance of microscopic ve- nous invasion in patients with operable HCC, re- mains unclear. Patients with microscopic venous invasion have higher serum levels of AFP, a larger tumor size and more nodules lacking a fi brous cap- sule (Tsai et al. 2000). Interestingly, up to 40% of less than 2-cm explanted tumors show microscopic venous invasion, a feature that overestimates the actual risk of tumor recurrence in patients with less than 5-cm tumors undergoing liver transplantation (Mazzaferro et al. 1996). Interestingly, circulating tumor cells have been demonstrated in the blood of 23 of 44 patients with HCC using a cytomorpho- logical approach but only one patient ultimately developed extrahepatic metastases during a 3-year follow-up period (Vona et al. 2004). This clearly suggests that the presence of tumor cells within tu-

mor vessels is not synonymous of tumor metastasis in all cases, thus attenuating the predictive value of this pathological feature.

12.8 Conclusions

The understanding of the natural history of HCC has led to signifi cant improvements in the management of patients with this tumor and made surveillance programs aimed at early diagnosis of HCC, possible.

Surveillance programs with abdominal US and se- rum AFP have been of strategic importance since HCC treatment could only be improved in tumors diagnosed early. Chronic carriers of hepatitis B and patients with cirrhosis are the ideal target popula- tion for prospective surveillance (Bruix et al. 2001).

Abdominal US is the best tool for surveillance with its greater predictive value for HCC than serum AFP (54% vs 32%) (Colombo et al. 1991; Oka et al. 1994). Since the average volume doubling time of HCC arising in patients with cirrhosis is 6 months, a 6-month interval between surveillance is considered cost-effective by many (Bruix et al. 2001). Although surveillance can identify tumors at an early stage and increase the chances of successful treatment, it is still not clear whether it reduces liver-related mortal- ity in parallel. Ageing of the patient population and deterioration of liver function during surveillance, occurrence of multinodular tumors and limited ac- cess to liver transplantation may hamper surveillance program effectiveness. Two non-randomized studies in the Far East (Chen et al. 2002; Yuen et al. 2000) and one in Italy (Bolondi et al. 2001) provided confl ict- ing results in terms of survival when patients with

Table 12.6. Prevalence of single, small nodes of hepatocellular carcinoma (HCC) detected during surveillance programs with abdominal ultrasound (US) of patients with compensated cirrhosis

Study Patients with US periodicity HCC × Single HCC cirrhosis (months) year <5 cm

Oka et al. 140 3 6.5% 82%

1990

Colombo 447 12 3.2% 54%

et al. 1991

Cottone 147 6 4.4% 83%

et al. 1994

Bolondi 313 6 4.1% 80%

et al. 2001

an HCC detected during surveillance were compared to patients with an incidentally diagnosed tumor.

However, those studies did not adequately assess pa- tient survival in relation to the signifi cant improve- ments in management of HCC that have occurred in recent years. The reanalysis of a cohort of 417 HCC-free patients with compensated cirrhosis who had been under prospective surveillance for 148 months, showed a fall in liver-related mortality rates in HCC patients identifi ed between 1997 and 2001.

Mortality rates fell from 45% in the fi rst 5-year pe- riod (1986–1991) to 37% in the second (1991–1996) and 10% in the third (1997–2001; fi rst vs second not signifi cant, fi rst vs third p=0.0009, second vs third p=0.018) in parallel with a reduction in yearly mor- tality of treated patients (34%, 28% and 5%: fi rst vs second not signifi cant, second vs third p=0.036; fi rst vs third p=0.0024) (Sangiovanni et al. 2004). During the last 5-year surveillance period, there was a shift of more patients from surgery towards the less ag- gressive locoregional ablative techniques, favored by the application of stringent criteria for patient selection to hepatic resection and the limited avail- ability of donated organs for treating HCC with liver transplantation (Bruix et al. 1996; Mazzaferro et al. 1996). Also, fewer patients with a single small tu- mor were left untreated or missed radical treatment compared to previous periods (46% vs 38% vs 26%), and fewer patients treated with hepatic resection or locoregional ablative therapies died of causes unre- lated to cancer (35%, 25%, 0%). The gain in survival of cirrhotic patients developing an HCC during the last 5 years was likely to be the consequence of im- proved management of the tumor and complications of cirrhosis.

References

Anthony PP (1973) Primary carcinoma of the liver. A study of 282 cases in Ugandan Africans. J Pathol 110:37–48 Barbara L, Benzi G, Gaiani S, et al (1992) Natural history of

small untreated hepatocellular carcinoma in cirrhosis: a multivariate analysis of prognostic factors of tumor growth rate and patient survival. Hepatology 16:132–137

Benvegnù L, Noventa F, Bernardinello E, et al (2001) Evidence for an association between the aetiology of cirrhosis and pattern of hepatocellular carcinoma development. Gut 48:110–115

Bolondi L, Sofi a S, Siringo S, et al (2001) Surveillance pro- gramme of cirrhotic patients for early diagnosis and treatment of hepatocellular carcinoma: a cost effectiveness analysis. Gut 48:251–259

Borzio M, Borzio F, Macchi R, et al (1994) The evaluation

of fi ne-needle procedures for the diagnosis of focal liver lesions in cirrhosis. J Hepatol 20:117–121

Bralet MP, Regimbeau JM, Pineau P, et al (2000) Hepatocellu- lar carcinoma occurring in nonfi brotic liver: epidemiologic and histopathologic analysis of 80 French cases. Hepatol- ogy 32:200–204

Bruix J, Castells A, Bosch J, et al (1996) Surgical resection of hepatocellular carcinoma in cirrhotic patients: prognostic value of preoperative portal pressure. Gastroenterology 111:1018–1022

Bruix J, Sherman M, Llovet JM, et al (2001) Clinical manage- ment of hepatocellular carcinoma conclusions of the Bar- celona-2000 EASL Conference, Barcelona September 15-17, 2000. J Hepatol 35:421–430

Burrel M, Llovet JM, Ayuso C, et al (2003) MRI angiography is superior to helical CT for detection of HCC prior to liver transplantation: an explant correlation. Hepatology 38:1034–1042

Chen TH, Chen CJ, Yen MF, et al (2002) Ultrasound screening and risk factors for death from hepatocellular carcinoma in a high risk group in Taiwan. Int J Cancer 98:257–261 Chevret S, Trinchet JC, Mathieu D, et al (1999) A new prog-

nostic classifi cation for predicting survival in patients with hepatocellular carcinoma. Groupe d’Etude et de Traitement du Carcinome Hepatocellulaire. J Hepatol 31:133–141

CLIP, the Cancer of the Liver Italian Program Investigators (1998) A new prognostic system for hepatocellular carci- noma: A retrospective study of 435 patients. Hepatology 28:751–755

CLIP, the Cancer of the Liver Italian Program (CLIP) Inves- tigators (2000) Prospective validation of the CLIP score:

a new prognostic system for patients with cirrhosis and hepatocellular carcinoma. Hepatology 31:840–845 Colombo M, de Franchis R, Del Ninno E, et al (1991) Hepato-

cellular carcinoma in Italian patients with cirrhosis. N Engl J Med 325:675–680

Craig JR (1997) Fibrolamellar carcinoma: clinical and patho- logic features. In: Okuda K, Tabor E (eds) Liver Cancer.

Churchill Livingston, New York, pp 255–262

Ebara M, Ohto M, Shinagawa T, et al (1986) Natural history of minute hepatocellular carcinoma smaller than three centimeters complicating cirrhosis. A study in 22 patients.

Gastroenterology 90:289–298

Fasani P, Sangiovanni A, De Fazio C, et al (1999) High preva- lence of multinodular hepatocellular carcinoma in patients with cirrhosis due to multiple etiological factors. Hepatol- ogy 29:1704–1707

Franco D, Capussotti L, Smadja C, et al (1990) Resection of hepatocellular carcinomas. Results in 72 European patients with cirrhosis. Gastroenterology 98:733–738

Gibson JB, Sobin LH (1978) Histological typing of tumours of the liver, biliary tract and pancreas. WHO, Geneva Henderson JM, Sherman M, Tavill A, et al (2003) AHPBA/AJCC

consensus conference on staging of hepatocellular carci- noma: consensus statement. HPB 5:143–250

International Working Party (1995) Terminology of nodular hepatocellular lesions. Hepatology 22:983-993

Kim SH, Lim HK, Lee WJ, et al (2000) Needle tract implanta- tion in hepatocellular carcinoma: frequency and CT fi nd- ing after biopsy with a 19.5-gauge automated biopsy gun.

Abdom Imaging 25:246–250

Kojiro M (1997) Pathology of hepatocellular carcinoma. In:

Okuda K, Tabor E, (eds) Liver Cancer. Churchill Livingston, New York, pp 165–187

Kojiro M (1998) Pathology of early hepatocellular carcinoma:

progression from early to advanced. Hepatogastroenterol- ogy 45:1203–1205

Kojiro M (2002) The Evolution of pathologic features of hepa- tocellular carcinoma. In: Zuckerman AJ, Mushahwar IK (eds) Viruses and Liver Cancer. Elsevier, Amsterdam, pp 113–122

Kumada T, Nakano S, Takeda I, et al (1997) Patterns of recur- rence after initial treatment in patients with small hepato- cellular carcinoma. Hepatology 25:87–92

Leung TW, Tang AM, Zee B, et al (2002) Construction of the Chinese University Prognostic Index for hepatocellular carcinoma and comparison with the TNM staging system, the Okuda staging system, and the Cancer of the Liver Ital- ian Program staging system: a study based on 926 patients.

Cancer 94:1760–1769

Llovet JM, Bru C, Bruix J (1999) Prognosis of hepatocellular carcinoma: the BCLC staging classifi cation. Semin Liver Dis 19:329–338

Mazzaferro V, Regalia E, Doci R, et al (1996) Liver transplanta- tion for treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 334:693–699 Oka H, Tamori A, Kuroki T, et al (1994) Prospective study

of α-fetoprotein in cirrhotic patients monitored for development of hepatocellular carcinoma. Hepatology 19:61–66

Okazaki N, Yoshino M, Yoshida T, et al (1989) Evaluation of the prognosis for small hepatocellular carcinoma based on tumor volume doubling time. A preliminary report. Cancer 63:2207–2210

Parkin DM, Bray F, Ferlay J, et al (2001) Estimating the world cancer burden: Globocan 2000. Int J Cancer 94:153–156 Rabe C, Lenz M, Schmitz V, et al (2003). An independent evalu-

ation of modern prognostic scores in a central European cohort of 120 patients with hepatocellular carcinoma. Eur J Gastroenterol Hepatol 15:1305–1315

Sangiovanni A, Del Ninno E, Fasani P, et al (2004). Increased survival of cirrhotic patients with a hepatocellular car- cinoma detected during surveillance. Gastroenterology 126:1005–1014

Sell S (2002) Cellular origin of hepatocellular carcinoma. Sem Cell Develop Biol 13:419–424

Takamori R, Wong LL, Dang C, et al (2000). Needle-tract implantation from hepatocellular cancer: is needle biopsy of the liver always necessary? Liver Transpl 6:67–72 Tsai TJ, Chau GY, Lui WY, et al (2000) Clinical signifi cance of

microscopic tumor venous invasion in patients with resect- able hepatocellular carcinoma. Surgery 127:603-608 Villa E, Moles A, Ferretti I, et al (2000) Natural history of inop-

erable hepatocellular carcinoma: estrogen receptors’ status in the tumor is the strongest prognostic factor for survival.

Hepatology 32:233-238

Villa E, Colantoni A, Camma C, et al (2003) Estrogen recep- tor classifi cation for hepatocellular carcinoma: comparison with clinical staging systems. J Clin Oncol 21:441–446 Vona G, Estepa L, Beroud C, et al (2004) Impact of cytomor-

phological detection of circulating tumor cells in patients with liver cancer. Hepatology 39:792–797

Yuen MF, Cheng CC, Lauder IJ, et al (2000) Early detec-

tion of hepatocellular carcinoma increases the chance

of treatment: Hong Kong experience. Hepatology

31:330–335