Page 629 - Read Online

Osho et al. Hepatoma Res 2020;6:55 I http://dx.doi.org/10.20517/2394-5079.2020.42 Page 3 of 15

that poor visualization is in fact associated with lower HCC detection as well as determining the optimal
surveillance strategies in patients with limited visualization. A recent pilot study suggests that repeat
ultrasound in patients with limited visualization (scores B or C) could have sufficient visualization (score A)
[18]
in approximately half of cases; however, validation of these results are needed in larger cohorts .
[19]
Various surveillance intervals have been proposed . The AASLD and EASL recommend semi-annual
[20]
surveillance, which appears reasonable on the basis of the median doubling time of HCC tumors . A
retrospective multicenter study among 649 HCC patients from Italy found patients detected by semi-
annual surveillance had smaller tumor burden and improved survival compared to patients submitted to
annual surveillance (40.3 months vs. 30 months, respectively, P = 0.03) . An RCT evaluated if shorter
[21]
intervals would further improve early detection and survival but found that a 6-month surveillance interval
provided similar early HCC detection compared to a 3-month surveillance interval (79% vs. 70%, P = 0.30) .
[22]
Role of biomarkers for HCC surveillance
Professional societies offer differing guidance regarding the additional value of serum biomarkers
over ultrasound alone for HCC surveillance. The best studied biomarker for HCC surveillance is
[23]
alpha fetoprotein (AFP), which has been validated in all five phases of biomarker development . The
AASLD and Asian Pacific Association for the Study of the Liver (APASL) both recommend ultrasound
[2]
with or without AFP , whereas the EASL recommends ultrasound alone, citing the poor performance
characteristics of AFP [2,3,6,7] . The AASLD and APASL guidelines cite the improvement in sensitivity when
[8]
adding AFP to ultrasound in clinical practice . Various AFP cutoffs have been proposed, with the AASLD
[2]
and APASL recommending AFP cutoffs of 20 and 200 ng/mL, respectively . Notably, AFP levels can be
increased in patients with active hepatitis, and thus, AFP is less accurate in patients with active HCV
infection, whereas AFP has higher sensitivity in other subgroups (e.g., patients with cirrhosis and HIV
infection) [24,25] . In contrast, AFP has better sensitivity in individuals with HBV and HCV, who are either
receiving or have completed antiviral treatments, and therefore, lower threshold values can be established
in using AFP for surveillance in this patient population [26,27] .

Although AFP alone has poor sensitivity for early stage HCC and poor specificity in patients with viral
hepatitis, several studies have suggested a potential benefit of using AFP as an adjunct surveillance test
with ultrasound. A meta-analysis of cohort studies on this topic demonstrated that the combination of
ultrasound and AFP had a significantly higher sensitivity for early stage HCC compared to ultrasound alone
[8]
(63% vs. 45%, respectively) . Although this was associated with a drop in specificity (92% for ultrasound
alone vs. 84% for ultrasound plus AFP), this was not felt to be clinically significant and the diagnostic
odds ratio of the combination remained higher using the two tests together 7(95%CI: 3-15) vs. 8(95%CI:
3-23, respectively). A study by Atiq and colleagues quantified physical harms related to ultrasound with
[28]
or without AFP, as not all false positive lesions may prompt diagnostic evaluation . They found 1 in 4
patients with cirrhosis experience physical harms for false positive or indeterminate surveillance tests,
which are more often related to ultrasound than AFP monitoring - in part related to some patients having
diagnostic evaluation for indeterminate ultrasound results and providers not ordering diagnostic evaluation
in many patients with false positive AFP levels.

Other biomarkers, such as lens culinaris agglutinin-reactive AFP (AFP-L3) and des-gamma-carboxy-
prothrombin (DCP), have been evaluated in phase 2 (case-control) biomarker studies but appear to have
[23]
insufficient performance if used alone . Therefore, there has been increasing interest in biomarker panels,
such as GALAD, which combines AFP, AFP-L3, and DCP with patient age and gender, and has been shown
to have promising performance in large case-control studies [29,30] . In a large multi-national case-control
study with 6,834 patients (2,430 HCC and 4,404 controls), GALAD demonstrated a sensitivity of 60%-
[31]
80% for early HCC detection. A recent study by Yang et al. compared GALAD to ultrasound for HCC