The population of patients at risk of developing HCC will undergo surveillance through the measurement of AFP levels and evaluation of the liver by USG every 6 months. The population at risk includes patients with liver cirrhosis of any etiology and hepatitis B patients. In the Indonesian National Consensus of the Management of Hepatocellular Carcinoma7 population at risk in developing HCC also includes chronic hepatitis C patients who developed fibrosis, but this population of patients has not yet been included in the current surveillance program in clinical practice.
In our study, we found the sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratio for positive test results of AFP in the surveillance of HCC of AFP (with a cut-off of 10 ng/ml) was 82.6%, 68.9%, 72.6%, 79.8% and 2.77, respectively. Interestingly, this result was in accordance with that of a study conducted in a population of 805 patients of Asian ethnicity by Chan SL, et al18, showing the AFP sensitivity and specificity values of 82.6% with a specificity of 70.4% (with a similar cut-off of 10 ng/ml), with the results of positive predictive values and negative predictive values obtained as 86.6% and 63.6%, respectively. On the other hand, research conducted by Biselli, et al19 in a population of HCC patients in Italy showed an AFP sensitivity with a cut-off of 10 ng/ml was 66.3% with a specificity of 80.6%. It should be of note that the study conducted by Chan SL18 in Asian patients had higher sensitivity levels than that conducted by Biselli, et al.19 Sensitivity of the AFP test in Asian countries, predominantly in developing countries such as in Indonesia, is believed to be higher because the prevalence of HCC with the etiology of hepatitis B tends to be higher—partly due to the lower coverage of hepatitis B immunization in newborns.20
We concluded that with a sensitivity of 82.6% and a specificity of 68.9%, HCC surveillance using AFP test with a cut-off of 10 ng/ml is still useful due to its high sensitivity—as sensitivity levels above 80% still appear to be adequate for screening programs. High specificity, on the other hand, is more useful in establishing a diagnosis, so the specificity of 68.9% is still considered sufficient in the HCC surveillance program. That being said, we still recommend the use of AFP test to also be carried out along with USG to reach a higher sensitivity and specificity level in HCC surveillance.
In our study, it was found that etiology was one of the factors that was statistically significant (p = 0.011) in its probability to cause AFP levels of HCC patients to rise above 10 ng/ml. It can be seen from Table 1 that the highest aetiological prevalence of HCC in the patient population in this study was hepatitis B, which was followed by hepatitis C, and this is as per the current data that shows that there are 400 million patients infected with HBV, 75% of whom are Asian.21 This significant result is also similar to that of a study conducted by Murugavel KG, et al22 which explained that there was a higher proportion of AFP elevation in HCC patients with the etiology of hepatitis B compared with other viral etiologies. Similar to the study conducted by Liu C, et al23, this study showed an increase of AFP level in cases of HCC caused by HBV compared with non-hepatitis cases of HCC. Also, a study conducted by Hann, et al24 showed that an increase in serum AFP in hepatitis B is in line with an increased risk for HCC.
We found that in HCC patients with the etiology of hepatitis B, 88.1% of patients had an increased level of AFP above 10 ng/ml cut-off with an odds ratio (OR) of 5.92, which was statistically significant (p = 0.019). According to Li M, et al25, this tendency of an elevated AFP level in HBV-infected patients was due to the presence of HBV protein (HBx) which could induce AFP receptor regulation, thereby increasing AFP expression in HCC due to HBV infection. Research conducted by Zhang C, et al26 and Yao M, et al27 also showed that HBV co-transcription factors could directly bind to AFP gene promoters, hence increasing its expression.
In the population of HCC patients with hepatitis C aetiology, higher levels of AFP (above cut-off) were also found (with the OR 5.067) compared with the population of non-hepatitis HCC patients, although this OR was not statistically significant. Studies show that a significant increase in AFP was less common in patients with HCC with hepatitis C etiology.28,29 Studies conducted in Egypt, the country known to have the highest prevalence of hepatitis C, showed that the prevalence of an increase in AFP levels above 10 ng/ml in HCC patients with etiology of hepatitis C was proven to be less frequent, occurring only in 11.6% patients.30 These results were similar to those of a research conducted in western countries with a higher prevalence of hepatitis C compared to hepatitis B, wherein AFP elevation above cut-off occurred only in 10%–43% patients.31–33
This study also showed a statistically significant difference in the proportion of HCC patients with cirrhosis and non-cirrhosis, when compared with AFP levels above and below 10 ng/ml (p = 0.016). There were 55.3% (n = 73) patients with cirrhosis in this study. We also found that HCC patients with cirrhosis would have a higher risk of having an elevated AFP level above cut-off (OR 3.508, p = 0.011). These results were similar to those of previous studies, as cirrhosis is one of the main known risk factors for HCC, and is found in 80%–90% of all HCC cases. HCC is also one of the causes of death in patients with cirrhosis of the liver, with the rate of developing HCC in patients with liver cirrhosis per year being 5%.34,35 This causes all patients with cirrhosis with any etiology to be recommended through surveillance of KSH.36,37 Pathological study has shown that patients with chronic liver disease can express AFP without prior development of HCC.38 The study by Harada T, et al39 showed that approximately 40% of all cirrhotic patients will have AFP levels higher than 20 ng/ml. An increase in AFP levels (between 10 and 500 ng/ml, and sometimes up to 1000 ng/ml) can be seen in adult patients with hepatitis or cirrhosis with any aetiology. Also, the frequency of elevated AFP levels (> 10 ng/mL) was reported in 20% cases of chronic hepatitis and 40% cases of cirrhosis.40
This study also found that BCLC stage C (p = 0.002), etiology of hepatitis B KSH (p = 0.005), cirrhosis (p = 0.016) and BCLC B stage (p = 0.019) were all independent predictors of elevated AFP levels above 10 ng/ml. The BCLC C classification, which was considered an advanced stage, classified patients who had already developed tumours invading the portal vein and those who had already developed extrahepatic dissemination, and have clinically shown impairment in their daily activities (ECOG performance status 1–2). The condition of the spread of these tumours, according to types of literature, was related to higher AFP levels, although this was not statistically proven in this study. It has been previously explained that the protein possessed by HBV (HBx) can increase regulation of AFP receptors to increase its production. The co-transcription factor possessed by HBV can also directly bind to AFP gene promoters and increase AFP secretion.26,27 Interestingly, HCC patients with the etiology of hepatitis B had a higher risk of having an elevated AFP level above cut-off when compared with HCC patients with cirrhosis. This was presumably due to other aetiologies outside of hepatitis B that may produce a lower level of AFP.26,27