Page 141 - Read Online

Zacharakis et al. Hepatoma Res 2018;4:65 I http://dx.doi.org/10.20517/2394-5079.2018.76 Page 3 of 15

to 22% while increases the specificity [20,21] . Therefore, the use of AFP in clinical practice is limited by the low
sensitivity at cutoff values maintaining sufficiently high specificity.

Furthermore, it has not escaped our notice that the heterogeneity of AFP cut off values in the literature
could be attributed to several epidemiological factors as the high incidence of HBV infection in Asian
HCC patients and the fatty liver in Western countries. In addition, the control group in most of these
studies includes subjects without HCC instead of subjects with suspected HCC to ensure prevalence rates
comparable to the rates in clinical settings. Finally, several studies suffer from the “verification bias” because
the reference standard (CT scan or MRI) had not been performed in all subjects to exclude tumor presence
in non-HCC cases.

[22]
Although AFP can be used to help define the population at risk of HCC , it has a suboptimal performance
as a serological test for surveillance. The following study supports the use of AFP as a single biomarker for
surveillance of HCC in particular populations or healthcare environments where ultrasound (US) is not
[23]
available . Furthermore, the use of US and AFP levels vs. the use of US alone offered additional detection in
6%-8% of HCC cases. The low specificity of AFP as a biomarker for HCC surveillance could be explained by (1)
the transient rise in AFP levels in patients with cirrhosis reflecting an exacerbation of the hepatitis infection or
in patients with chronic liver disease; and (2) the flares of underlying liver disease such as HBV, HCV or HCC
[24]
development .

Furthermore, AFP shows low sensitivity because it is not overexpressed in all HCC patients. It was found that
[25]
elevated AFP levels were not evident in around 80% of small HCCs . Only about 10%-20% of HCC tumors
at the early stage can present with abnormal AFP serum levels. This observation has been recently found in a
molecular subclass of aggressive HCCs (S2 class, EpCAM positive) [26-28] . In addition, AFP levels may be normal
in up to 40% of patients diagnosed with early HCC. Based on data from the literature, a summary of the
[21]
sensitivity and specificity of AFP is shown in Table 1 .

Another limitation of AFP includes the suboptimal performance in distinguishing intrahepatic
cholangiocarcinoma and HCC. Although AFP (-) was the most sensitive assay for differentiating intrahepatic
cholangiocarcinoma (ICC) from HCC (91.1%), its specificity was significantly lower than other markers such as
[29]
CA242 (+) and carbohydrate antigen (CA) CA19-9 (+) .

This has a critical impact on the outcome of the misdiagnosed patients since surgical resection is generally the
[30]
preferred therapeutic choice for HCC but not ICC . As a result, AFP has been excluded in some guidelines on
HCC surveillance and diagnosis. In fact, the European Association for the Study of the Liver and the European
[31]
Organization for Research and Treatment of Cancer (EASL-EORTC) clinical practice guidelines (CPG)
for HCC screening and diagnosis do not include quantitative measurements of serum AFP and recommend
surveillance by experienced personnel in the at-risk populations using abdominal US every 6 months. However,
[32]
the American Association for the Study of Liver Diseases (AASLD) guidelines recommend surveillance by
the US for cirrhotic adults every 6 months with optional use of AFP due to the poor sensitivity and specificity
of this biomarker. Shorter follow-up interval (every 3-4 months) is recommended in case of any of these
conditions: (1) a nodule of less than one cm has been detected; (2) after liver resection; or (3) after loco-regional
therapy. In contrary, the follow-up 3-6 months of serum AFP was included in the diagnostic algorithm of
hepatic nodules by the Oriental guidelines for HCC management. Therefore, in the most common guidelines,
it is well established and recommended that US should a be part of surveillance and most commonly combined
with AFP.

Furthermore, a recent systematic review on HCC surveillance showed that the most commonly used
[33]
surveillance tests for HCC in cirrhotic patients were US and AFP. Four studies used only US for HCC
surveillance, whereas the rest of studies used the combination of US and AFP at 6-month intervals. The