Review of 336 patients with hepatocellular carcinoma at Songklanagarind Hospital

Abstract

AIM: To determine the clinical presentations, survival and prognostic factors of hepatocellular carcinoma (HCC) in Southern Thailand.

METHODS: Retrospective analysis was performed on the 336 hepatocellular carcinoma patients treated at Songklanagarind hospital between 1 January 1991 and 31 January 1999.

RESULTS: Of these 336 patients, 276 were males and 60 were females. The mean age was 54.4 years. The common symptoms and signs were abdominal pain and hepatomegaly. The most common presentation of tumor was a dominant mass with daughter nodules. Portal vein involvement was found in 50% of total. Extra hepatic metastasis was found in 13%, and the lung was the most common site. There were 65.4% with evidence of cirrhosis and half of them were in Child's class B. HBsAg was positive in 72.6%. Regarding Okuda's tumor staging, 15%, 61% and 24% were stage I, II and III, respectively. Overall median survival was 2.1 months (11.5, 2.6 and 0.7 months for stage I, II and III respectively). Treatments of HCC improved patient survival (5.5 months vs 1.6 months for untreated patients). Most common causes of death were hepatic failure. Using multivariate analysis, the prognostic factors identified were tumor staging, alpha-fetoprotein level above 10000 μg·L^-1, extrahepatic metastasis, portal vein thrombosis and treatment.

CONCLUSION: HCC in Thailand is a fatal disease with poor outcome due to late presentation and high prevalence of liver cirrhosis. Early detection and proper management may improve outcome.

Key Words: hepatoma/therapy; biopsy; neoplasm staging; survival analysis; neoplasms metastasis; prognosis; drug therapy

INTRODUCTION

Hepatocellular carcinoma (HCC) is the most common form of primary liver cancer and is the leading cause of cancer death especially among males in South-East Asia including Thailand[1,2]. This may be related to high prevalence of chronic hepatitis B infection (8%-15% in Asia and Africa, and 8%-12% in Thailand) [3-7]. Though many types of treatment have been tried, HCC is still a fatal disease possibly associated with the advanced stage at which the disease is usually diagnosed [2,8-11]. Thus, it remains a serious medical problem in this part of the world. There were many reports of the natural history of HCC in Japan, Mainland China, Southern Africa, Alaskan Eskimos, Taiwan, Italy, Spain and North America, but little information has been published from South -East Asia[9-24]. Therefore, we reviewed 336 HCC patients at Songklanagarind Hospital to describe the clinical presentations and history of known risk factors and determine the survival rate, prognostic factors and the benefit of treatments.

MATERIALS AND METHODS

Patients

The medical records of 336 HCC patients admitted at Songklanagarind hospital between January 1, 1991 and January 31, 1999 were reviewed retrospectively. The diagnosis of HCC was made by liver biopsy or elevated serum alpha-fetoprotein level above 500 μg·L^-1 with radiologic findings suggestive of HCC in patients whose liver biopsy was not available[25]. Data from medical records, including patient demographic, known risk factors, clinical manifestation, abnormal physical findings, laboratory data (complete blood count, coagulogram, renal function test, liver function tests, viral hepatitis serology, serum alpha-fetoprotein level, chest X-ray, ultrasonography, CT scan, liver biopsy and other tissue biopsy if suggested metastasis), survival and treatments modality, were used for analysis. The known risk factors include d alcohol drinking, history of blood transfusion and history of jaundice or viral hepatitis infection or known cases of cirrhosis. Patients were classified into 6 groups based on their clinical presentation: group 1, mass-related symptoms (abdominal pa in or fullness, dyspepsia, palpable mass); group 2, cirrhosis-related symptoms (jaundice, GI bleeding, edema, abdominal enlargement, hepatic encephalopathy); group 3, liver abscess-like symptoms (high fever with abdominal pain and tenderness); group 4, non-specific symptoms (anorexia, nausea, vomiting, malaise, weight loss and anemia); group 5, metastasis symptoms (dyspnea, cough, bone pain and palpable lymph node); and group 6, asymptomatic cases (accidental finding b y routine check-up or complain of other unrelated disease). The abnormal physic al findings included anemia, jaundice, fever, hepatomegaly, splenomegaly, ascites and sign of chronic liver stigmata such as palmar erythema, spider nevi, gynecomastia and superficial dilated vein. A test of viral marker for hepatitis B (HBsAg) was done in most of patients but that for hepatitis C (anti HCV) was not available until 1996. Tumor volume was calculated from ultrasonography or CT scan of the liver by a radiologist. Tumor volume or sum of tumors in instances multiple nodules were expressed as fraction of total liver and subsequence classified into two groups (tumor size ≤ 50%, > 50% of the whole liver). Staging of HCC was made according to Okuda's[9]. Cirrhosis was confirmed by liver biopsy or ultrasono graphy or CT scan and classified by Child- Pugh's (Class A, B or C)[26]. The extrahepatic metastasis was confirmed by histology (incisional biopsy, excisional biopsy, necropsy or autopsy).

Therapy

The treatment of HCC ranged from no treatment, transhepatic artery oily chemoemb olization (TOCE), percutaneous ethanol intralesional injection (PEI), hepatectomy, systemic chemotherapy and multimodality combination chemotherapy.

For systemic chemotherapy before 1995, we used intravenous adriamycin and/or 5-FU injection. Later this was changed to PIAF regimen (cisplatin 80 mg·m^-2 body surface area (BSA) and adriamycin 40 mg·m^-2 BSA at d1, followed by 5-FU 500 mg·m^-2 infusion over 24 h for the following 3 d, with alpha-Interferon 5 mU·m^-2, iv, 3 h after cisplatin and 5-FU every day).

TOCE was performed monthly by super-selective insertion of catheter to the right or left hepatic artery branch feeding the tumor then injection adriamycin 50 mg, lipiodol 8 mL and gelfoam. PEI was performed by using 10 mL of absolute ethanol injected percutaneously under CT-scan guide.

Multimodality combination therapy comprised a combination of several treatments depending on the tumor staging and complication, such as the patients who presented with advanced HCC and had portal vein involvement with lung metastasis, the treatment was started with systemic chemotherapy (PIAF regimen) until no evidence of extrahepatic metastasis remained and was then followed by TOCE and/or intralesional ethanol injection (PEI). Because of the limitation of retrospective study, we were unable to determine the exact outcome or improvement of general condition after treatments, so we determined the outcome by survival analysis.

In patients who were lost to follow-up, we determined the date and cause of death from the population register and personal contact with the family. Patient status was unable to determine in 20% of the patients. For statistical analysis these patients were considered as censored at the date of last contact.

Survival profiles were constructed using the Kaplan-Meier method. Prognostic factors were identified using Cox proportional hazards regression. P value < 0.05 was considered statistically significant.

RESULTS

Of the 336 patients, 276 (82%) were male with a male to female ratio of 4.6:1. The mean age was 54.4 (a range of 20-89) years (54.3 years in mal e and 55 years in female) (Table 1). Diagnosis was confirmed histologically in 273 (72.3%) cases and 63 (18.7%) cases were diagnosed by a combination of elevated serum alpha fetoprotein level above 500 μg·L^-1 and imaging such as ultrasound or CT scan showing a lesion compatible with HCC.

Table 1 HCC patient characteristics.

Patient characters		n	%	Mean	Range
Age (years)	All			54.4	20-89
	Male			54.3	20-89
	Female			55.0	20-81
Sex	Male	276	82.0
	Female	60	18.0
Risk factors	Alcohol drinking	126	38.0
	HBsAg positive (299 sample)	217	72.6
	anti-HCV positive (135 sample)	10	7.4
	Cirrhosis	219	65.2
Liver function test	Total bilirubin (μmol·L-1)			63.8	3.4-752.4
	Direct bilirubin (μmol·L-1)			38.8	0.5-581.4
	Aspatate aminotransferase (U·L-1)			225	17-3890
	Alanine aminotransferase (U·L-1)			97	4-3370
	Alkaline phosphatase (U·L-1)			304	8-2080
	Albumin (g·L-1)			35.6	20-52
	Globulin (g·L-1)			38.2	18-78
Radiologic finding	Solitary type (total)	240	73.0
(n = 329)	with daughter nodules	144	43.8
	without daughter nodules	96	29.2
	Multinodular type	47	14.3
	Diffuse or infiltrative type	42	12.7
Alpha fetoprotein (μg·L-1) (n = 295)					1451102-7990000
	< 10	38	13.0
	10-99	36	12.0
	100-499	33	11.0
	≥ 500	188	64.0
Okuda's staging	Stage I	51	15.0
	Stage II	205	61.0
	Stage III	80	24.0

The most common symptom was mass-related such as abdominal pain, abdominal discomfort, dyspepsia and palpable mass (Table 2). Mean duration of symptoms was 49 d (range < 1 day - 1 year). Among the abnormal physical findings hepatomegaly was the most common, followed by fever and jaundice (Table 2).

Table 2 Presenting symptoms and abnormal physical findings of patients.

Presenting symptoms	n	%
Mass-related symptoms (Abdominal pain or fullness, dyspepsia, palpable mass)	188	56
Cirrhosis-related symptoms (Jaundice, GI bleeding, edema, abdominal enlargement, encephalopathy)	59	17.6
Liver abscess-like symptoms (High fever with acute abdominal pain and tenderness)	45	13.4
Non-specific symptoms (Anorexia, nausea, vomiting, malaise, weight loss, chronic anemia)	26	7.7
Metastasis symptoms (Dyspnea, cough, bone pain, palpable lymph node)	11	3.2
Asymptomatic (Routine checked up or other unrelated disease)	7	2.1
Abnormal physical findings
Hepatomegaly	282	83.9
Fever	185	50.5
Jaundice	143	42.6
Anemia	138	41.1
Ascites	123	36.6
Cachexia	86	25.6
Chronic liver stigmata	86	25.6
Edema	59	17.6
Splenomegaly	42	12.5

Elevation of alkaline phosphatase and serum aspartate aminotransferase (AST) was the most common abnormal finding in the liver function test. Serum bilirubin above 51.3 μmol·L^-1 was found in 30%. The most common radiologic finding was a solitary mass with or without daughter nodules (73%, Table 1). The tumor was located most frequently in the right lobe (53%) followed by both lobes (37%) and left lobe (10%). Tumor volume larger than half of the total liver was found in 71% and portal vein thrombosis or portal vein involvement in 50%. Portal vein thrombosis or involvement was significantly associated with increased tumor volume (P = 0.0005). The mean serum alpha-fetoprotein level was 145110 (range 2-7990000 μg·L^-1) and value s above 500 μg·L^-1 (the cut point for diagnosing HC C) occurred in 64% (Table 1).

The most common risk factor was chronic hepatitis B infection (72.6%), and 65.2% showed evidences of liver cirrhosis (Table 1). Patients with cirrhosis were classified into Child A, B and C in 20%, 55.3% and 24.7% respectively. Cirrhosis was found in 76.2% of alcoholic patients, 66.4% of HBV infected patients an d 90% of HCV infected patients. Okuda's staging distribution was 15%, 61% and 24% for stage I, II and III respectively (Table 1). Spontaneous rupture of HCC was found 11% (4%, 9.3% and 20% in stage I, II and III). At the time of diagnosis, extrahepatic metastasis occurred in 43 cases (13.1%), 13.7%, 10.7% and 18.8% of stage I, II and III respectively. The metastatic sites were lung (76%), lymph node (16%) and bone (7%).

Overall median survival was 2.1 months (Stage I, 11.5 months; Stage II, 2.6 months; and Stage III, 0.7 months) (Table 3, Figure 1). The 1 and 2-year survival rates were 15% and 8% respectively. Treatment of HCC was associated with improvement of patient survival (5.5 vs 1.6 months in non-treated group; P = 0.011) (Table 3, Figure 2). In the non-treated group (n = 245) median survival was 1.6 months (7.7, 1.8 and 0.6 months for stages I, II and III respectively, Table 3, Figure 3). Regarding treatment, patients treated with TOCE, intravenous chemotherapy, multimodality combination therapy and Tamoxifen administration had median survival times of 6.3, 5.33, 17.1 and 3 months, respectively (Table 3). Compared to the non-treated group, patients treated with TOCE, intravenous chemotherapy or combination therapy had significantly better survival (P = 0.0005, 0.011 and 0.007 respectively) whereas survival of the patients treated with Tamoxifen was not significantly different from non-treated (P = 0.86) patients.

Figure 1 Kaplan-Meier survival curve for overall patients. Total (n = 336) median survival of 2.1 months: Okuda's stage I (n = 51), median survival of 11.5 months; Okuda's stage II (n = 205), median survival of 2.6 months; Okuda's stage III (n = 80), median survival of 0.7 months (P = 0.014 and 0.001 respectively).

Figure 2 Kaplan-Meier survival curve in relation to treatment or non-treatment. Treated group (n = 91), median survival of 5.5 months; non-treated group (n = 245), median survival of 1.6 months (P = 0.011).

Table 3 Median survival.

Group of patients	n	Median survival (months)
All	336	2.1
Stage I	51	11.5
Stage II	205	2.6
Stage III	80	0.73
Untreated	245	1.6
Stage I	26	7.7
Stage II	146	1.8
Stage III	73	0.63
Treated	91	5.5
Stage I	25	13.7
Stage II	59	4.2
Stage III	7	1.7
TOCE	44	6.3
Stage I	11	24.3
Stage II	29	5.5
Stage III	4	1.4
Chemotherapy (adriamycin and/or 5-FU)	16	5.2
Multimodality therapy	12	17.1
(PIAF/iv, chemotherapy ± TOCE ± PEI)
Tamoxifen	9	3.0
PIAF regimen chemotherapy	5	*
Hepatectomy	4	$
PEI	1	$

^*Data of PIAF regimen chemotherapy is not completely finished. $It were too small number of patients to evaluated.

Figure 3 Kaplan-Meier survival curve in non-treated group. Total (n = 245), median survival of 1.6 months, Okuda's stage I (n = 26), median survival of 7.7 months; Okuda's stage II (n = 146), median survival of 1.8 months; Okuda's stage III (n = 73), median survival of 0.6 months.

The most common causes of hospital death (n = 54) were hepatic failure, GI bleeding and rupture of tumor (59.2%, 20.4% and 20.4% respectively). Among case s with hepatic failure, sepsis was the most common complication leading to death. Cox proportional hazards model reviewed the following, prognostic factors: Okuda's stage II (P = 0.014, hazard ratio 2.05), stage III (P = 0.001, hazard ratio 3.79), AFP level above 10000 μg·L^-1 (P = 0.001, hazard ratio 2.03), lung metastasis (P = 0.01, hazard ratio 1.93), lymph node metastasis (P = 0.015, hazard ratio 3.76), portal vein involvement or thrombosis (P = 0.0005, hazard ratio 1.79) and treatment (P = 0.011, hazard ratio 0.91).

DISCUSSION

Similar to other studies, we found that HCC was more common in males (a male to female ratio of 4.6:1) because the risk factors such as cirrhosis, chronic HBV infection and alcoholic are more frequently seen in males than females[2,8,12-19,27,28]. In our study, the most common risk factor was chronic HBV infection (72.6%) as HBV is endemic in South-East Asia[5]. The common symptoms were non-specific and included abdominal pain, dyspepsia, jaundice, hepatomegaly, anorexia and weight loss. Clinical jaundice was found in 42.6% and mainly caused by failure of hepatic function due to cirrhosis. The most common type of tumor in our patients was dominant mass with daughter nodules whereas multiple nodules or diffuse lesion are common in Western patients[15,17,18]. The difference in risk factors may explain the variation in tumor characters as chronic HBV infection is the most important risk factors in our region while chronic HCV infection and alcohol drinking are the largest risk factors in Japan and Western countries[3,8,13-18,27,29]. Cirrhosis was found in 65% of our patients and 80% of them were classified as Child's B or C. These may be associated with poor prognosis. In our study, hepatectomy and TOCE were not suitable in more than half of the patients due to portal vein involvement (50%) and advanced liver cirrhosis (24.7%). Extrahepatic metastasis was found in 13% and most of them were located in the lung, probably because of direct drainage into the right heart via the hepatic vein. Elevated serum alpha-fetoprotein above 500 μg·L^-1 was found in only 64% of patients, so this tumor marker was not very sensitive f or diagnosis of HCC in our country. Most our patients had advanced HCC (61% and 24% were stage II and III respectively). The overall median survival in our patients was 2.1 months because majority of our patients had advanced disease with significant liver cirrhosis. In treatment of HCC, hepatectomy and TOCE could improve survival[9,13], but could not be performed because of liver cirrhosis and portal vein involvement. Systemic chemotherapy, arterial infusion and Tamoxifen administration could not improve survival[9,20,24,30-32], whereas data of PIAF regimen chemotherapy showed complete pathological remission, but survival analysis did not[33]. In our study, TOCE and multimodality therapy could improve survival, particularly in patients with stage I and stage II disease as compared with the non-treatment group (24.3% vs 7.7 months in stag e I and 5.5 vs 1.8 months in stage II receiving TOCE and 17.1 months v s 0.63 months in patients receiving multimodality therapy). However, our study is only a retrospective study that had its limitation in comparing survival rate between different groups. The most of our patients died of hepatic failure (60%) as a majority of them had advanced HCC and liver cirrhosis. By multivariate analysis, poor prognosis was associated with advanced stage of the tumor, serum alpha-fetoprotein level > 10000 μg·L^-1, extrahepatic metastasis and portal vein involvement.

ACKNOWLEDGEMENTS

The authors acknowledge the cooperation of the staff of the Registration Unit of Songklanagarind Hospital and of the Official Register of Population for providing patient data and Dr. Allan Geater for English advice.