Characteristics of Hepatocellular Carcinoma in India: a Retrospective Analysis of 191 Cases

Background: Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide. The outcome of the disease is related to the stage of presentation. A comprehensive analysis of patients with this disease is not available in India. Methods: Retrospective chart review of 246 patients with HCC was done. One hundred ninety-one patients (male 160, female 31; median age 52 years, range 9–85 years) fulfilling diagnostic criteria for HCC adopted by Barcelona-2000 EASL conference were analyzed for clinical, etiological, radiological and cytohistological profile. Results: Underlying cirrhosis was seen in 60% cases with hepatitis B being the most common etiologic agent. HCC caused new onset ascites and recent worsening in three-fourth cases with ascites.


Introduction
Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide with an estimated 500 000 to 1 million new cases per year. 1 The incidence ranges from four cases per 100 000 populations in USA to 150 cases per 100 000 populations in parts of Africa and Asia where HCC is responsible for a large proportion of cancer deaths.A rise in the incidence of mortality from HCC has been observed in different countries. 2Approximately 77% of deaths from HCC occur in developing countries.The prognosis of HCC is dismal with 5-year survival being 1-4%. 3The global distribution of HCC is very variable.According to the ageadjusted HCC incidence per 100 000 population per annum, different geographic regions can be divided into three incidence zones: low (<5), intermediate (between 5 and 15), high (>15). 4ost Asian countries are in intermediate or high incidence zones of HCC.In India, the mean incidence of HCC in four population-based registries is 2.77% for males and 1.38% for females.The prevalence of HCC in India varies from 0.2% to 1.6%. 5,6The geographic model of HCC occurrence is frequently correlated with the Address correspondence to Dr S.K. Sarin, Professor and Head, Department of Gastroenterology, Room No. 201, 2nd Floor, Academic Block, G.B. Pant Hospital, New Delhi, 110002, India.email: sksarin@nda.vsnl.net.in! The Author2008.Published by Oxford University Press on behalf of the Association of Physicians.etiologic factors.Hepatitis B virus (HBV) infection is the most common etiologic factor in high incidence areas, while hepatitis C (HCV) infection is more prevalent in the low incidence areas.7,8 Unlike other low incidence zone, in India HBV is the main etiological factor associated with HCC.[9][10][11][12][13] In the west, majority of HCC are diagnosed incidentally during routine evaluation.However, in India, most of the patients in clinical practice present at an advanced stage ruling out curative treatment in most cases.Despite India being a low incidence zone for HCC, the estimated HCC cases in 2001 would be 12 750.5 However, there is paucity of published literature on profile of HCC patients in India, making formulation of a proper health care strategy difficult.In order to study the characteristics of HCC in India, we comprehensively analyzed these patients especially with regard to their clinical, etiological, radiological and cytohistological profile.

Patients and methods
A retrospective chart review of patients of HCC in the period 1991-2003 at G.B.Pant Hospital, Delhi, India was done.Cases were included if they met EASL diagnostic criteria for HCC. 14Cases were excluded if data record was incomplete to satisfactorily review the case.Records of 246 HCC patients were screened.Fifty-five patients were excluded from the study because of the incomplete records.
The diagnosis of underlying cirrhosis was based on clinical, histological, endoscopic (presence of varices) and radiological documentation [ultrasound (US) and CT scan].Histological evidence was used wherever available.
All clinical, biochemical, serological, radiological and cytohistological details were noted from the case records.

Etiological studies
Patients were considered HBV-related if they had any of the following markers (i) HBsAg, (ii) HBeAb, (iii) IgHBcore (total/IgG) and (iv) serum HBV DNA.HCV was diagnosed if patients were anti-HCV antibody or HCV RNA positive.Alcohol was considered as an etiological factor if patients consumed >80 g/day for 5-years or more.The etiology was considered as unknown if etiological work up was not complete or was negative for HBV, HCV, alcohol or any other etiologic factor namely Budd-Chiari syndrome or metabolic liver disease, etc.

Radiological studies
The tumors were assessed by either ultrasound, CT scan or MRI of the abdomen.The TNM and Okuda staging of HCC were done based on available adequate and appropriate clinical and radiological information.

Statistical analysis
The quantitative data were presented as mean AE SD or median (range).The continuous variables between the two groups were compared by the Mann-Whitney U-test.The Kruskal-Wallis test was used to compare a continuous variable across different stages of tumors.The chi-square test with Yates's correction was used to analyze categorical variables between the two groups and the two-tailed Fisher's exact test where numbers were small.The P-value <0.05 was considered significant.

Results
The baseline characteristics of patients with HCC are presented in Table 1.All pediatric HCC cases (614 year) were HBV positive.The HCV positive patients (median age 60 year; range 33-85) were about a decade older as compared to HBV positive patients (median age 52 year; range 9-81) (P = 0.06) (Table 1).

Clinical features
The clinical features are shown in Table 2.The features of hepatic decompensation were seen in half the patients at first presentation with ascites in 51%, jaundice in 34% and hepatic encephalopathy in 5% patients.New onset ascites was seen in 57% cases while it had recently worsened in 16% patients.
Gastrointestinal bleed including melena was present in 22% patients.The clinical presentation of HCC in most of the cases was similar to the presentation of patients with cirrhosis.Massive hepatomegaly (>5 cm) was seen in more than half the patients (56%) while in 16% of cases liver was not enlarged.The average enlargement of liver was 5.0 AE 2.9 cm.The enlarged liver was hard in two-third cases and was of firm consistency in the rest one-third.Clubbing, hepatic bruit and hypoglycemia were uncommon (Table 2) The biochemical investigations were mildly deranged.Serum alpha feto protein (AFP) was diagnostic (>400 ng/ml) in only 43 of 93 (46%) patients; with normal AFP in 16 of 93 (17%) patients.The median serum AFP value was 320 ng/ml (range 2.5-900 000) (Table 3).The serum AFP elevation in HCC patients with cirrhosis (median 640 ng/ml; range 5-900 000) was significantly higher as compared to those without cirrhosis (median 150 ng/ml; range 2.5-338 160) (P = 0.032).The proportion of patients with normal serum AFP (<20 ng/ml) was similar in both the groups (cirrhotic 14.3% vs. non-cirrhotic 26%; P = ns).However, the serum AFP diagnostic for HCC (>400 ng/ml) was seen in significantly higher proportion of cirrhotic patients (37 of 70; 53%) as compared to non-cirrhotic (6 of 23; 26%) (P = 0.046).

Etiological studies
The etiologic work up was available only in 147 (77%) patients, while in rest of 44 patients (23%) etiology was unknown due to incomplete investigation reports.HBV was the most common viral etiologic agent associated with HCC, observed in 107 of 147 (73%) patients either alone or as cofactor with alcohol or HCV (Table 4).

Radiological studies
Only 100 of 191 cases were analyzed for radiological profile when complete details in US and/or CT scan regarding number, size, distribution and characteristics of liver lesions along with vascular details and metastases were available (Table 5).
The HCC most commonly involved right lobe of liver (48%), followed by bilobar involvement (34%).The left lobe was involved in approximately onefifth of cases.Multicentric HCC involving either one lobe or both lobes were seen in 36% of patients.The average size of HCC was 6.8 AE 3.4 Â 6.1 AE 3.1 cm.Very large tumors (>5 cm) were seen in three-fourth of cases.Small HCC (62 cm lesion) was seen in only 8% of patients approximately.Single lesion was the most common presentation of HCC observed in two-third cases.Three or more lesions were seen in about one-fifth.The heterogeneous (40%) or hypo echoic (38%) lesions were the most common appearance on US observed in 80% of the cases.The hyper echoic lesion was seen in the rest.Similarly, the CT appearance of HCC was hypo dense in 42%, mixed or heterogeneous density in 35% and hyper dense in 23% patients.Vascular invasion of either major branch of splenoportal axis, inferior vena cava (IVC) or hepatic veins (HV) was seen in more than half of the patients.Main trunk of portal vein or its main branches were involved in 51 patients with associated involvement of HV and/or IVC in five patients.Two patients had involvement of HV and/or IVC without portal vein involvement.
Extrahepatic spread of tumor was seen in 13 patients.Two patients had regional periportal lymph node involvement.Metastases involving retroperitoneal lymph nodes were seen in 5 of 11 patients (45%), followed by lung and pleural metastases in 2 of 11 patients (18%).Colon, peritoneum, adrenal and bone were involved in one each.

Histopathological study
Histopathological study of liver lesion was available in 123 of 191 patients (64.4%).Fine needle aspiration (FNA) was available in 118 patients and liver biopsy in 12 patients.Seven patients had undergone both FNA and liver biopsy.Diagnosis of HCC was made based on cytohistology in 108 (88%) patients.Suggestive but equivocal diagnosis on cytohistology was seen in six patients (5%).Cirrhosis with dysplasia was diagnosed in one patient on liver biopsy.In eight patients (6.5%) (7 in FNA group and 1 in liver biopsy group) cytohistology was normal.No major complication or fatality was reported in any of these patients except mild pain at the site of needle puncture in few.More than three-fourth of HCC was in Okuda stage 2 (78%).Only three patients had Okuda stage 1 lesion.The size of HCC lesion was >5 cm in 73% of cases.Hence, HCC was large and very advanced in most of the cases (Table 6).
The serum AFP was not related to different stages of TNM and Okuda (P = ns).

Discussion
This is the first study from Indian subcontinent in which the characteristics of HCC including its clinical, etiological, radiological and pathological profile have been studied.
The incidence of HCC increases with age.The development of HCC is uncommon before 40 years of age in western world.However, the pattern of HCC incidence by age is sometimes dependent on the geographic pattern or on etiologic factors. 4he age distribution of patients with HCC in the present study was similar to other studies in past.Studies from India have shown the maximum incidence of HCC in the fifth to sixth decade. 8,15,16As in this study the HCV-related HCC occurs in older age group as compared to HBV-related HCC. 7,88][19][20][21] The incidence of HCC peaks during fifth and sixth decade as in other studies.The population-based data show a male to female ratio of 3:1-2:1.1. 22However, high preponderance of HCC in males reported in hospital-based data could suggest a gender bias in seeking medical treatment.
The clinical presentation of HCC patients in our study was similar to other studies.However, the incidence of gastrointestinal bleed was high in the present series.This could be due to inclusion of melena in a large proportion of patients.The paraneoplastic syndrome was very rare in this series.Only one patient developed troublesome hypoglycemia.The incidence of paraneoplastic syndromes especially hypoglycemia has been reported in up to 30% cases.The rarity of paraneoplastic syndrome has been seen in other Indian series as well. 9,15,16The symptom duration at diagnosis was mainly 1-3 months like other study from India. 15,16Underlying cirrhosis was seen in almost 60% of our cases.HCC is accompanied by liver cirrhosis in 70-90%. 23The cirrhosis along with HCC is reported in 70-86% of Indian autopsy studies.4,25 The present study is in conformity with the previous studies from India.However, some cases of cirrhosis might have been missed in the present study, as evidence of cirrhosis may not have been strictly looked for in all cases The HBV is the most common etiologic factor in Asian countries.7][28][29][30][31][32] The HBsAg positivity in Indian HCC patients varies from 36% to 74%. 10,12,13,29,33The prevalence of anti-HCV antibody in Indian population varies from 0.3% to 1.8%. 34,35[13][26][27][28][29][30][31][32][33][34][35] Serological evidence of HCV infection in patients with HCC in India is $15%. 9,29Dual infection with or without alcohol was seen in 8% of patients like our previous study. 9The Japanese study has also shown almost similar rate of dual infection in HCC patients. 36India, despite being an intermediate endemic zone for HBV has low incidence of HCC unlike other Asian countries.This phenomenon is akin to low HCC incidence in Greenland Eskimos as compared to Alaskan Eskimos despite similar HBsAg positivity. 37However, in a proportion of patients the etiology was unknown due to incomplete work up or absence of markers for either HBV or HCV and alcohol.This is because of retrospective nature of the study, as patients may not have been aggressively investigated once an advanced lesion was detected ruling out curative therapy in most.However, in an earlier prospective study from our center also, the patients with unknown etiology constituted almost 20% of patients despite aggressive work up including for p53 gene mutation and dietary contamination with aflatoxin B. 9 Such a high proportion of HCC patients without a known etiologic factor may point to the possibility of another hitherto unknown factor or pathogenetic mechanism for HCC in India and needs further study.
The serum AFP estimation is not a good screening for HCC.9][40] Despite a cut off of 100 ng/ml, sensitivity and specificity of AFP was 21% and 93% in one of the study. 41As per the diagnostic criteria adopted by EASL, serum AFP >400 ng/ml was present in only 46% cases.The median AFP level was only 320 ng/ml in the study population despite advanced HCC in three-fourth cases.In another study from India, the prevalence of detectable AFP in patients with HCC was 51%. 42There are some studies which suggest that the production of AFP depends on the size or the degree of differentiation of the hepatoma cells. 43When compared according to either TNM or Okuda staging there was no difference in the serum AFP level to differentiate between early and advanced lesion.Cytohistopathological examinations, including fine needle aspirations and biopsies were undertaken in 123 patients without complication.Mainly FNA was done for the diagnosis of HCC once a lesion was suspected on radiological investigation.The cytohistological examination was positive in 88% of patients and was falsely negative in only 6.5% cases.It was safe and easy to perform.The larger and advanced lesions might have contributed to better results with FNA in the present study.FNA appears to be the best diagnostic modality as it has low false negative rate as compared to serum AFP, once a lesion is detectable on CT scan or US and is diagnostic in around 90% cases as evident in the present study.In a country like ours, a costly investigation like serum AFP with such low sensitivity should be used if cytohistological examination is equivocal in patients with radiological suspicion of HCC.However, in an otherwise unapparent HCC on radiological investigation, screening with serum AFP is the only practical modality available now, despite its limitations.
Like previous studies most of the lesions on US were hypo echoic or heterogeneous in appearance.However, more than half of the patients had vascular invasion, one-third having bilobar distribution and many had metastases.Even when lesions were single they were large enough in most of the cases to rule out curative resection.Another study from a tertiary care center in India has shown that 56% of patients with HCC had tumor size larger than 5 cm and high incidence of vascular invasion (main portal vein in 43% and right portal vein in 30%) with very low resection rate. 44he TNM and Okuda staging of HCC revealed a very high proportion of advanced lesions in this study with expected low survival.Despite almost two-third of patients being cirrhotic the HCC was diagnosed very late.In view of this finding, we need to evolve a better screening program for patients with chronic liver disease in our country.There is a need to aggressively look for HCC by cytohistological methods once there is a radiological suspicion.We have shown, this is a very safe, effective and possibly the best method in patients with radiological apparent lesion.
In conclusion, the HCC in India has few atypical characteristics.The paraneoplastic manifestations are distinctly rare.The HCC is diagnosed very late and presents with vascular invasion or metastases in most of the cases.The low sensitivity of serum AFP despite large lesion and its cost makes fine needle aspiration cytology diagnostic modality of choice in radiological apparent lesion.Since HBV is the most common etiologic agent in our country, the universal vaccination against hepatitis B would prove an effective preventive strategy as in a country like Taiwan.

Table 2
Clinical profile of HCC

Table 5
Radiological profile of HCC based on US and/or CT scan (n = 100)

Table 4
Etiological work up of HCC(n = 191)

Table 6
TNM and Okuda staging of HCC and its correlation with serum AFP