EASL Clinical Practice Guidelines

Diagnosis

Nowadays, early HCC diagnosis is feasible in 30–60% of cases in developed countries and this enables the application of curative treatments. In fact, while tumors less than 2 cm in diameter represented <5% of the cases in the early nineties in Europe, currently they represent up to 30% of cases in Japan. This trend is expected to continue growing in parallel to the wider implementation of surveillance policies in developed countries [102]. However, detection of these minute nodules of ∼2 cm poses a diagnostic challenge as they are difficult to characterize by radiological or pathological examination [[103], [104], [105]].

Proper definition of nodules as pre-neoplastic lesions or early HCC has critical implications. Dysplastic lesions should be followed by regular imaging studies, since at least one-third of them develop a malignant phenotype [[106], [107]]. Conversely, early tumors are treated with potentially curative procedures – albeit expensive – such as resection, transplantation and percutaneous ablation. Thus, there is an urgent need to identify better tools to characterize these lesions. Otherwise, the cost–effectiveness of the recall policies applied within surveillance programs will be significantly undermined.

Non-invasive diagnosis

Accurate diagnosis of small liver nodules is of paramount importance. Until 2000, diagnosis was based on biopsy. This approach had some limitations related to feasibility due to location and risk of complications, such as bleeding or needle-track seeding [108]. In addition, achieving accuracy in differentiating between high-grade dysplastic nodules and early HCCs was complex, since stromal invasion, the most relevant criteria, is difficult to recognize even for an expert pathologist [105]. In 2001, a panel of experts on HCC convened in Barcelona by EASL reported for the first time non-invasive criteria for HCC based on a combination of imaging and laboratory findings [1]. In principle, a unique dynamic radiological behavior (contrast up-take in the arterial phase by CT, MRI, angiography or US) represented the backbone of radiological diagnosis of early HCC. In cirrhotic patients with nodules >2 cm, coincidental findings by two imaging techniques were considered diagnostic, or alternatively, one imaging technique along with AFP levels above 400 ng/ml. In all other circumstances biopsy was mandatory. In 2005, the EASL panel of experts and the American Association for the Study of Liver Diseases (AASLD) guidelines adopted a new HCC radiological hallmark, i.e. contrast uptake in the arterial phase and washout in the venous/late phase [109]. Non-invasive diagnosis was established by one imaging technique in nodules above 2 cm showing the HCC radiological hallmark and two coincidental techniques with nodules of 1–2 cm in diameter (CT, MRI and US-contrast). AFP levels were dropped from the diagnostic scheme [109]. Recent updated AASLD guidelines have proposed that one imaging technique (CT or MRI) showing the HCC radiological hallmark suffices for diagnosing tumors of 1–2 cm in diameter [56].

In order to update the EASL guidelines for non-invasive diagnostic criteria of HCC, two questions are posed. First, what data provides reliable non-invasive diagnostic accuracy for nodules of 1–2 cm in diameter taking into account that the recommendations apply to a wide range of expert physicians and radiologists. And second, what imaging techniques can be used. Regarding the first issue, two prospective studies have shown that using 2 imaging techniques is an approach with high PPV and specificity [[104], [109]]. In one study including 89 consecutive cases of nodules between 0.5 and 2 cm detected within surveillance programs in cirrhotic patients showed that non-invasive criteria are accurate for the diagnosis of HCC, with a specificity of 100% [104]. Unfortunately, such an absolute specificity had the downside of a low sensitivity of 30%, meaning that two-thirds of nodules required pathological confirmation. The other study suggested that the use of a sequential algorithm would maintain an absolute specificity but increase the sensitivity, with significant savings in terms of liver biopsy procedures for nodules of 1–2 cm [110]. A retrospective study reporting diagnostic accuracies of MRI in large series of transplanted patients showed an overall false positive rate exceeding 10% when using one imaging technique [111]. Finally, a recent prospective study, testing the accuracy of imaging techniques in nodules between 1 and 2 cm detected by ultrasound, showed false positive diagnosis – mostly due to high grade dysplastic nodules – above 10% with either 1 or 2 imaging techniques, with a specificity of 81% and 85%, respectively [112]. Hence, the non-invasive diagnosis of 1–2 cm lesions remains a challenging issue, with no unequivocal data in prospective validation studies. While the panel considers incorporating the 1 technique rule in order to have a consistent approach in the field, a more cautious application of this rule is recommended in suboptimal settings, where the technology at disposal or the local skills are not at the high-end level. In these circumstances, we recommended to use two coincidental techniques, since the negative consequences of high rates of false-positive diagnosis offset the benefit. Additional prospective studies to confirm the accuracy of this approach are recommended in order to support a more strong recommendation at the 1A level.

Regarding which imaging techniques should be used, it has to be pointed out the fact that the HCC radiological hallmark is based on the tumor vascular dynamic performance. This limits the usage of US-contrast – since US microbubbles are confined to the intravascular space – as opposed to iodinated contrast-CT or gadolinium-based MR imaging, in which standard contrast agents are rapidly cleared from the blood pool into the extracellular space. A recent study showed that lesions other than HCC, i.e. cholangiocarcinoma, displayed homogeneous contrast uptake at US-contrast followed by washout, i.e. the vascular pattern assumed to represent the hallmark of HCC [113]. Thus, latest generation CT and/or MRI following reported protocols are recommended for non-invasive diagnosis of HCC [114]. On the other hand, recent advances in the use of perfusion CT or MRI with liver-specific contrast agents have not so far provided solid data to support their use as alternate criteria.

It is important to point out that the HCC radiological hallmark only occurs in a small proportion of patients with tiny tumors (1–2 cm) [103], and thus biopsy or tissue biomarkers will be required in most instances. Delaying diagnosis beyond 2 cm leads to increased levels of treatment failure or recurrence, since it is known that satellites and microscopic vascular invasion rise exponentially beyond this size cut-off [101]. Therefore, it is crucial to provide reliable tools for a final diagnosis before the 2 cm cut-off.

Pathological diagnosis

Pathological diagnosis of HCC is based on the definitions of the International Consensus Group for Hepatocellular Neoplasia [115] and is recommended for all nodules occurring in non-cirrhotic livers, and for those cases with inconclusive or atypical imaging appearance in cirrhotic livers. Sensitivity of liver biopsy depends upon location, size and expertise, and might range between 70% and 90% for all tumor sizes. Pathological diagnosis is particularly complex for nodules between 1 and 2 cm [105]. Morphological criteria alone still pose problems for the differential diagnosis of high-grade dysplastic nodules versus early HCC, especially because the pathological hallmark of HCC, stromal invasion, can be absent or difficult to identify in biopsy specimens [105]. In a prospective study, first biopsy was reported positive in ∼60% of cases for tumors less than 2 cm [104]. Thus, a positive tumor biopsy is clinically useful to rule in a diagnosis of HCC, but a negative biopsy does not rule out malignancy. The risk of tumor seeding after liver biopsy is 2.7% with a median time interval between biopsy and seeding of 17 months [116].

Tissue markers might provide a more across-the-board standardized diagnosis of these tumors. Distinct technologies such as genome-wide DNA microarray, qRT-PCR, proteomic and inmunostaining studies have been used in an attempt to identify markers of early diagnosis of HCC. Few studies, however, include a thorough analysis of several markers in a training-validation scheme and with a sufficient number of samples [78]. A study conducted in 128 human samples described a 13-gene signature able to identify HCC lesions with high diagnostic accuracy [117]. Similarly, a three-gene signature (the genes that encode GPC3, LYVE1, and survivin) has been proposed as an accurate molecular tool (>80% accuracy) to discriminate between dysplastic nodules and small HCCs (<2 cm) [118]. The performance of this signature was externally validated in a different set of samples [[118], [119]].

The diagnostic performance of some markers of early HCC identified by genomic studies has been prospectively assessed by immunohistochemistry, a low-cost technique. By examining the tissue, the pathologist can select a representative tumor sample without necrosis or inflammation and define the cell type expressing protein markers and the specific pattern. A promising marker is GPC3, which shows a sensitivity of 68–72% with a specificity superior to 92% [[120], [121]]. Similarly, combinations of different protein markers – HSP70, GPC3, and GS – in 105 hepatocellular nodules performed acceptably (sensitivity and specificity of 72% and 100%, respectively) [120], and were afterwards validated in two larger series [[122], [123]]. The International Consensus Group of Hepatocellular Neoplasia has adopted the recommendation to define a pathological diagnosis of HCC if at least two of these markers are positive [115]. Additional staining can be considered to assess neovascularisation (CD34) or potential progenitor cell origin (Keratin 19, EpCAM) [[101], [105], [124]]. In particular, keratin 19 (K19), a progenitor cell/biliary marker, at a cut-off of 5% of positive tumor cells with immunohistochemistry, has been shown to correlate with poorest outcome [[105], [124], [125]]. Moreover, K19 recognizes biliary features in mixed forms of HCC/cholangiocarcinoma, which are not always detected on hematoxylin–eosin stain.

Assessment of disease extension

Assessment of tumor extension is critical for defining staging and treatment strategy. Several studies with pathological correlation have shown that dynamic contrast-enhanced MRI and 4-phase multidetector CT are the most effective imaging techniques for detecting tumors smaller than 2 cm. However, underestimation of 25–30% is expected even with the best state-of-the-art technology [[126], [127]]. Pre-specified protocols should define the amount and rate of contrast given, the precise individualized timing of the image acquisition and image reconstruction with minimum slice thickness. Lipiodol contrast staining should not be used. Contrast-enhanced ultrasound is unable to compete with CT and MRI in terms of accuracy for detection of lesions. Bone scintigraphy can be used for evaluating bone metastases. PET-based imaging is not accurate to stage early tumors. Pre-operative staging prior to liver transplantation should include abdominal dynamic CT or MRI, chest CT and bone scintigraphy.