Optimal surveillance of hepatocellular carcinoma in patients with chronic viral hepatitis


Prof. Y-F Liaw, Liver Research Unit, Chang Gung Memorial Hospital, 199 Tung Hwa North Road, Taipei, Taiwan. Email: liveryfl@tp.silkera.net

See article in J. Gastroenterol. Hepatol. 2001; 16: 553–9.

Hepatocellular carcinoma (HCC) is one of the leading malignancies in the Asia–Pacific region where chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infection are considered as the major causes.1 Although the prevalence of HBV, HCV infection and HCC in the United States of America is low, there has been an increase in the number of HCC cases over the past 20 years, and over 50% of them are related to either HBV or HCV infection.2

Symptomatic HCC is a dismal disease with a 1-year survival rate usually less than 20%. In patients with symptomatic HCC, the tumors are either too large or too large and extended out (local invasion) or distant metastasis. Therefore, only 10–20% of patients are able to be treated with potentially curative therapy, while the remaining patients can only receive non-surgical therapy, such as a transcatheter arterial embolization or other local ablation treatments with a less satisfactory outcome.3 In contrast, more than 80% of HCC patients are associated with liver cirrhosis upon initial diagnosis, some cases have poor liver function reserve and thus prohibit the patient from therapeutic modality for HCC, or the patient may succumb to hepatic decompensation, per se, before treatment for cancer can be given.

In order to improve the chances of successful treatment and patient survival, great effort has been devoted to detect HCC in the ‘early’ stage when it is ‘small’ and where liver function reserve is good enough for curative therapy. For this specific purpose, repeated evaluations or tests to identify HCC are necessary, particularly in high-risk patients. Therefore, the ideal modality of surveillance should be non-invasive, highly sensitive, acceptable by patients, and not expensive. Serum α-fetoprotein (AFP) and hepatic ultrasonography (US) are widely used as surveillance modalities for high-risk patients, including hemochromatosis and cirrhotic patients without hepatitis viral markers in many Asia–Pacific and European countries. Hepatocellular carcinoma of sizes smaller than 3, 2 and 1 cm can therefore be detected.4–17 In Alaskan native hepatitis B surface antigen (HBsAg) carriers, even semi-annual AFP screening alone was effective in detecting most HCC tumors at a resectable stage.18 With adequate treatment, patients with subclinical HCC have a better survival rate than patients who receive no treatment or those who have symptomatic HCC.7,18 Clearly, the clinical significance of early detection and early treatment of HCC is unequivocal.

In a previous issue of the Journal of Gastroenterology and Hepatology, Tong et al. reports a 7-year prospective surveillance of HCC in 602 patients with chronic HBV or HCV infection by using both AFP and US.19 Thirty-one cases of HCC were detected, with a mean tumor size of 3.5 cm; the largest being 6.5 cm in diameter. One-sixth of cirrhotic patients developed HCC during a mean follow-up period of 3.5 years. The diagnostic sensitivity/specificity of AFP and US are 65/90% and 78/100%, respectively. Tong et al. concluded that a US examination was more accurate than AFP to detect HCC, and suggested using US as the primary modality and using serum AFP as an adjuvant test for the screening and surveillance of HCC.19 Although their conclusions are consistent with what has been accepted, the resectability of HCC in their study was only 13%. The resectability of HCC among large screening programs range from 21,4 47.4,6 5512 to 61%.13 The resectability of HCC in the study conducted by Tong et al. is not only much lower than those in previous studies4,6,12,13 using the same tests, but it is also lower than the study conducted by McMahon et al. that used semiannual AFP screening.18

In evaluating the diagnostic benefit and accuracy of AFP and US for monitoring newly developed HCC, both advantages and limitations should be considered. An AFP assay is easy to perform, less skill dependent, and can be used in a mass survey, as best demonstrated in the study by McMahon et al. in Alaska.18 However, the serum AFP concentration may increase during hepatitis exacerbation,20 pregnancy, and in patients with germinal tumors.18,21 If the target populations are victims of chronic viral infection, hepatic exacerbation is common, and therefore, the false-positive rate of AFP is relatively high. The shortcoming of low specificity of AFP in diagnosing small HCC may be compensated by more specific assays such as the des-γ-carboxy prothrombin and fucosylation index, or lens culinaris agglutinin-reactive fraction of AFP;1,22 however, the cost-effectiveness of these assays requires further evaluation. The diagnostic value of AFP is further downgraded because AFP is normal in a substantial proportion of small HCC.4,5,7,23 In studies conducted in Asia, serum AFP was normal in 366 and 46.2%4 of the patients with HCC upon US detection. In contrast, a high resolution real-time US has a diagnostic accuracy of 90–95% in detecting small HCC.4,6,12 However, the use of US also has some limitations, such as tumors being too small in size or too close to the diaphragm to detect, and other skill-dependent factors such as the experience of the examiner. In addition, the frequency of examination is a critical factor for the effectiveness of surveillance. A short interval may improve the detection of small HCC but may increase the cost, while a longer interval may save the expense, but may miss the chance of early diagnosis. The optimal interval of US examination should consider the cost of examination, the HCC doubling time and the size of HCC that allows for curative therapy to occur. One study conducted in a HBV-endemic area has shown that the calculated median doubling time of HCC was 117 days, and the most rapidly dividing tumor took 5 months to grow from 1 to 3 cm, which is usually still amenable for curative treatment.24 Many of the Asian and European investigators therefore prefer an interval of 6 months3,5,14,16,24 or 3 months in cirrhotic patients.6,8,13 However, the surveillance policy, particularly the interval of examinations, should take high-risk factors into consideration. Patients with elevated AFP levels conceivably require more frequent surveillence.4,6,13 In addition, cirrhotic patients with persistently high serum alanine aminotransferase (ALT) levels, espe-cially those that are HCV related, are also candidates of more frequent HCC surveillance because of their more rapid and frequent cancer development.25 It seems that a 3-month interval is more suitable for higher-risk patients such as those with cirrhosis, high AFP or persistently high ALT levels, excluding acute hepatitis exacerbation. In the study of Tong et al., a total of 1388 US examinations were performed on 602 patients during a mean of 34.5 months, which is equal to 0.8 US examinations per year per patient.19 It is generally accepted that detecting a HCC larger than 3.0 cm or a tumor up to 6.5 cm in diameter is not ‘early’ enough for optimal therapy and many would be too late for optimal therapy.18,19 If we aim to detect ‘early’ or ‘small’ HCC, the diagnostic accuracy at an interval greater than 6 months would be much lower than at a shorter interval. A 6-month interval may be a more optimal choice.

The cost-effectiveness of HCC surveillance may vary among different countries, and in terms of ethnic race, age and cost of examination.26 Although the screening of Alaskan native HBsAg carriers with semiannual AFP to determine in which persons US should be used was effective in detecting most of HCC, the benefits of early detection of HCC in that study seem to be greatest in the younger aged carriers.18 This might be related to the high rate of AFP elevation in younger patients with HCC.27 The AFP and US combination is generally accepted to be superior to either US or AFP alone, and a 6-month interval for the detection of HCC is a rational choice for non-cirrhotic patients,28 while patients with other risk factors such as cirrhosis, high AFP or persistently high ALT levels need to undergo more frequent surveillance.