Screening for hepatocellular carcinoma: the rationale behind

Lik-Fai Cheng 鄭力暉, Eliza P Y Fung 馮寶恩

HK Pract 2005;27:469-474

Summary

Hepatocellular carcinoma (HCC) is the second commonest cancer in men and the ninth commonest cancer in women for 2002. It also accounted for around 11.8% of all cancer deaths in 2002.¹ Hepatitis B, hepatitis C and cirrhosis are the most important risk factors. In Hong Kong, approximately 8-10% of the population are hepatitis B carriers. Only 10-15% of HCC patients are candidates for surgical resection. Non-surgical treatment of HCC includes radiofrequency ablation, alcohol ablation and transcatheter arterial chemoembolization. Numerous studies have proven that the tumour size directly affects the prognosis of patients with HCC and their survival following treatment. The rationale for HCC screening is based on the premise that earlier cancer detection can improve the prognosis and prolong the survival of patients suffering from HCC.

摘要

在2002年，肝癌是男性中第二常見的癌症,而在女性則位列第九，而且在整體癌病死亡中， 約佔11.8%。乙型肝炎、丙型肝炎及肝硬化為最重要的誘因。在香港，大約8-10%的人口為乙肝帶菌者。 只有10-15%的肝癌病人能進行腫瘤切除手術。非手術切除治療包括射頻消除，酒精消除及動脈導管化學栓塞。 很多研究經已證實腫瘤的大小直接影響肝癌病人的預後，以及在治療後的存活率。 肝癌篩選的理念是基於假設愈早發現肝癌能改善病人的預後和延長他們的存活。

Introduction

Hepatocellular carcinoma (HCC) is the most common primary liver cancer. It is the fourth most common cancer in the world.² Age-standardized incidence varies from 3 per 100,000 in North American men to 80 per 100,000 in China.^2,3

In Hong Kong, HCC is the second commonest cancer in men and the ninth commonest cancer in women. The age-standardized rate is 27.2 per 100,000 in male and 6.5 per 100,000 in female.¹

Even though more new techniques and treatment modalities have been developing recently, HCC still carries high mortality and has poor prognosis. Without treatment, the median survival of HCC patients was reported as 8 weeks from symptomatic presentation.⁴ In 2002, a total of 1381 patients died from this cancer, accounting for about 11.8% of all cancer deaths in Hong Kong.¹

Hepatectomy is a potential curative treatment option when the tumour is still small. Unfortunately, most patients with single and small HCC will have no symptoms until the tumours become very large. Late diagnosis of the disease at an untreatable stage precludes most patients from surgery. Therefore early detection of small and asymptomatic HCC at a treatable stage by screening tests is critical. There are specific criteria to be fulfilled before a screening test can be justified: 1) The disease should be common in the population, causing substantial morbidity and mortality; 2) The target or high risk group should be easily identified; 3) The screening tests should be cost effective and highly sensitive; 4) Early disease detection should improve the survival and prognosis.

Prevalence, morbidity and mortality of hepatocellular carcinoma in Hong Kong

According to the Hong Kong Cancer Registry in 2002,¹ the incidence of HCC in Hong Kong was 23.2 per 100,000. There were 1223 new male cases and 343 new female cases, accounting for the second commonest cancer in men and ninth commonest in women. The male to female ratio was about 3.4 to 1.

HCC carries a high mortality. The reported survival rates for untreated symptomatic patients vary from 0% at 4 months to 1% at 2 years.^5-7

In 2002, it caused 1090 deaths in men and 291 deaths in women, representing the second and fourth commonest causes of cancer death in male and female, respectively.

High risk group of hepatocelluar carcinoma

Risk factors of HCC include cirrhosis, chronic hepatitis B and C, carcinogens (e.g. aflatoxin), and inborn errors of metabolism (e.g. A-1 antitrypsin deficiency, Wilson disease, Type-1 glycogen storage disease, haemochromatosis). Cirrhosis, and hepatitis B and C viruses are the most important risk factors. Prospective studies have shown that 10-15% of patients with cirrhosis will develop HCC in Western Europe and USA.⁸

In Hong Kong, approximately 8-10% of the population are hepatitis B carriers. In fact, more than 80% of patients with HCC are found to be hepatitis B carriers. One local study showed that about 92% of symptomatic HCC cases had previous hepatitis B infection.⁹ The cumulative probability of developing HCC in patients with positive hepatitis B surface antigen has been estimated to be 6% in 5 years and 15% in 10 years. The risk of developing HCC is not equal in all infected patients. The hepatitis B carriers with cirrhosis have higher risks than non-cirrhotic patients.

Unlike hepatitis B, the risk of developing HCC in a non-cirrhotic patient with hepatitis C virus is relatively small. However, once cirrhosis is present in patient with hepatitis C virus infection, the risk of developing HCC increases. The cumulative risk of developing HCC in these patients is estimated to be 1.4-3.3% per year.

Ultrasound and alpha-fetoprotein level as screening test for hepatocellular carcinoma

There is no perfect screening tool for HCC. Alpha-fetoprotein (AFP) is a tumour marker for HCC and the level increases with time in the presence of the tumour. However, serum AFP level is usually normal during early stage of HCC. In addition, AFP is not specific for HCC and the level is also elevated in pregnancy and patients with germ-cell tumours.

AFP may also be elevated transiently, persistently or intermittently in patents with viral hepatitis. AFP elevation is most likely due to active hepatitis or seroconversion if it parallels the elevation of transaminases. The diagnostic difficulty, however, will be encountered in differentiating HCC from viral hepatitis when AFP level does not correlate with the alanine transaminase. Therefore AFP level is particularly useful in patients with inactive hepatitis B who have normal liver enzymes and lower hepatitis B virus DNA levels.

The range of AFP levels in hepatitis patients with and without HCC overlaps. The normal reference range of AFP is less than 10 ng/mL. However, there is no well documented guideline to show at what AFP level dedicated investigations for HCC should be triggered.

The sensitivity of AFP level in the detection of HCC depends on the cut-off level. Generally, the higher the cut-off level of AFP, the lower will be the sensitivity of the test. In one study adopting a high cut-off level of 500ng/mL, the sensitivity of AFP in HCC was reported to be 48.6%.¹⁰ In other studies with lower cut-off levels of 15-20ng/mL, the sensitivities varied from 64.3% to 96.9%.^11-13 The specificity of AFP level ranges from 76-91% and the positive predictive value varies from 9-32%.^14-16

Using the AFP level alone for HCC screening is not recommended. Limitations in the sensitivity and specificity of the AFP test lead to the use of ultrasound as an additional method for detection of HCC. Ultrasound examination is cheap, non-invasive, radiation free and easily accessible. However, it is operator dependant and the resolution is limited in patients with high body mass index. Moreover, it is difficult to differentiate small HCCs from regenerative nodules and dysplastic nodules by ultrasound alone.

The sensitivity of ultrasound as a screening test for HCC has been reported to be 71% in non-cirrhotic hepatitis B surface antigen carriers¹⁷ and 78% in cirrhotic patients.¹⁸

Tumours smaller than 1cm are difficult to detect and characterize, either using imaging or biopsy. A single tumour less than 3cm in size, on the other hand, carries a reasonable chance of cure. Sheu et al¹⁹ found the median doubling time of HCC was about 117 days. The most rapidly dividing tumours took 5 months to increase in size from 1 cm to 3 cm.

Because of the doubling time of HCC, 6-month surveillance is a reasonable interval to detect tumours growing from an undetectable to a potentially treatable stage. However, a few studies suggested that annual screening was just as effective.^20,21 Patients with abnormal screening test results should be recalled for further investigations such as CT scan, MRI and liver biopsy, in order to confirm or exclude the diagnosis of HCC. Normal results, on the other hand, cannot totally exclude the presence of the tumour. Regular follow-up is necessary.

Early hepatocellular carcinoma detection improves prognosis and prolong survival

Surgical resection and liver transplant are considered to be the only curative treatment. However 70% to 90% of HCC patients have underlying cirrhosis or chronic hepatitis. Moreover, in many cases the HCCs are multifocal or diffusely infiltrative. As a result, only 10-15% of patients are potential candidates for surgery.

Non-surgical treatments of HCC include radiofrequency (RF) ablation, percutaneous alcohol ablation and transcatheter arterial chemoembolization.

Studies have shown that small tumours (<5cm) are associated with more favourable outcome compared with large tumours after surgery.^22-24 Overall 5 year survival rates of 35-50% have been reported after partial hepatectomy.

A randomized study by Lencioni et al concluded that RF ablation was superior to percutaneous alcohol ablation as first-line treatment of small (less than 5cm) HCC in cirrhotic patients, with respect to local recurrence-free survival rate.²⁵

Studies on RF ablation of HCC have also demonstrated that tumour size is the most important factor for determining the local recurrence of the disease. Tumour foci smaller than 2.5 cm in diameter can be completely treated in 90% of cases, tumours with diameter between 2.5-3.5 cm are completely ablated in 70-90% of cases. For tumour size between 3.5-5.0 cm complete eradication can only be achieved in 50-70% of cases.²⁶ For larger tumours (larger than 5 cm in diameter), less than 50% of cases are likely to be completely ablated.

Transcatheter arterial chemoembolization will be one of the treatment options for unresectable HCC if the tumours are too large or too numerous for RF ablation or percutaneous alcohol ablation. Published studies have shown that responses to chemoembolization and survival are significantly and inversely proportional to tumour size, but not the number of tumour foci.^27,28

Therefore it is reasonable to believe that earlier detection of smaller HCC foci can improve the prognosis and survival.

Evidence on the benefit of hepatocellular carcinoma screening

No prospective randomized control trial has confidently demonstrated that periodic screening for HCC is effective in reducing mortality. Such trials would require thousands of participants, depending on the incidence of HCC in the population. In order to account for the lead time bias, the follow-up periods should be at least 7 to 10 years.

A prospective 16-year population-based cohort study was conducted in Alaska natives to determine the benefit of screening for HCC.¹² 1487 HBSAg-positive carriers were enrolled and serum AFP tests were performed every six months. Those found to have elevated AFP level were further evaluated by ultrasound. The result showed that HCC could be detected at a surgically respectable stage in 70% of cases. There was significantly higher 5- and 10- year tumour free survival in this cohort as comparing with the historical controls.

A prospective randomized study of HCC screening has been performed in Shanghai,²⁹ in which 17,920 HBsAg positive persons were recruited. 8109 were allocated into the screening group and 9711 into the control group. Serum AFP level was checked and ultrasound scan of liver was performed every 6 months for the subjects in the screening group. During a mean follow-up time of 1.2 years, 38 persons with HCC were detected in screening group and 18 in the control group. In the screening group, HCC was diagnosed at a subclinical stage in 76.8% of the patients and 70.6% of those underwent surgical resection. In the control group, none was asymptomatic and candidates for surgical resection. The 1- and 2-year survival rates of those who had resection in the screening group were 88% and 78%, respectively. However, none of the patients with HCC survived more than 1 year in the control group. Due to the short follow up period, the lead time bias in this study precluded a meaningful interpretation of survival benefit.

In Hong Kong, Yuen M F et al³⁰ published a study comparing HCC in asymptomatic patients detected by screening tests (AFP and/or ultrasound) with the tumours in symptomatic patients. They found that the tumour size was significantly smaller in the asymptomatic group compared with that of symptomatic group (3.5 cm vs. 8.1 cm; P <.0001). Bilobar involvement, multifocal tumours, diffuse-type tumours, portal vein infiltration, and distant metastasis were significantly more common in symptomatic patients. Operability and feasibility of treatment by transcatheter arterial chemoembolization in asymptomatic group patients (26.8% and 45.1%, respectively) were significantly better than in symptomatic group (7.9% and 32.3%, P <.0001 and P =.03, respectively). The cumulative survival rate was significantly higher in the asymptomatic group (P <.0001). For those after surgical resection and those after chemoembolization, asymptomatic group patients had a significantly higher cumulative survival rate compared with that of symptomatic group patients (P =.04 and P =.0003, respectively). The authors therefore concluded that screening for HCC can identify tumours at an early stage, resulting in a higher chance of receiving treatment.

Limitations of hepatocellular carcinoma screening

Although serum AFP level and ultrasound are generally recommended for HCC screening, they are no perfect. Not all HCCs are associated with raised serum AFP level. AFP levels are usually normal when the tumours are small. Acute exacerbation of chronic hepatitis, on the other hand, may cause elevation of serum AFP level.

The sensitivity and specificity of AFP level depends on the cut-off value, which there is still no consensus yet. The specificity is sufficient if the cut-off value of serum AFP level is chosen at 500 ng/mL. However, such elevations usually represent advance stage of the diseases.³¹

Ultrasound is operator dependent. The accuracy of the examination will also deteriorate as the body mass index of the subject increases. Although ultrasound detects small lesions down to 1cm, it cannot reliably differentiate HCC from haemangioma, regeneration nodule and dysplastic nodule.

The false positive result of screening tests may lead to over-investigation in some non-malignant liver nodules. The radiation dose and complications of investigations should also be taken into consideration, for example, complications of liver biopsies, include haemorrhage, penetration of viscera, peritonitis, pneumothorax and even death.

Conclusion

HCC is common in Hong Kong, carrying high morbidity and mortality. The aim of HCC screening is to detect the tumour at a treatable stage. Hepatitis B and C virus infections and liver cirrhosis are the major risk factors. The perfect screening tool for HCC does not exist. Ultrasound and serum AFP level provide effective screening tests for HCC. Because of the doubling time of HCC, a 6-month surveillance is generally recommended. A few studies, however, suggested that annual screening might be as effective. Patients with abnormal screening tests should be recalled for further investigations such as CT scan, MRI scan or liver biopsy. Even though no prospective randomized control trial has confidently demonstrated that periodic screening for HCC is effective at reducing the mortality, studies have shown that early HCC detection by screening may improve the chance of receiving treatment and prognosis.

Key messages

1. Hepatocellular carcinoma (HCC) is common in Hong Kong, carrying high mortality and poor prognosis.
2. The aim of hepatocellular screening is to detect the tumour at a treatable stage.
3. Hepatitis B carriers and liver cirrhotic patients are the high risk group and therefore the target population for screening.
4. Ultrasound and serum alpha fetoprotein level at 6 months interval are the commonly recommended screening tests for HCC.
5. Although no prospective randomized control trial has confidently demonstrated that periodic screening for HCC is effective at reducing the mortality. Studies have shown that early HCC detection by screening may improve both the chance of receiving treatment and the prognosis.