Biomarkers for Hepatocellular Carcinoma Diagnosis/Prognosis

Biomarkers for Hepatocellular Carcinoma Diagnosis/Prognosis: Comparison

Please note this is a comparison between Version 1 by Anna Pasetto and Version 2 by Lindsay Dong.

Hepatocellular carcinoma (HCC) causes many deaths worldwide, and current treatments have limitations. Immunotherapies have shown the most promising clinical outcomes for advanced HCC. However, there are many patients with HCC who still respond poorly to these treatments. Circulating biomarkers that can easily be obtained through blood sampling are promising in predicting treatment responses, since they are minimally invasive and enable us to constantly monitor disease progression.

hepatocellular carcinoma (HCC)
immunotherapies
liquid biopsy
circulating tumor DNA (ctDNA)
circulating tumor cells (CTC)

1. Introduction

It has been estimated that more than one million deaths will be attributed to hepatocellular carcinoma (HCC) by 2030 [1], making it one of the deadliest cancer types worldwide. Depending on the stage of HCC, the treatment options can vary. When diagnosed at an early stage, the standard of care options include resection, local ablation or liver transplantation, but the risk of tumor recurrence still remains high [2]. When diagnosed at an intermediate stage, treatment options are limited to transarterial chemoembolization, whereas systemic therapies, such as the multi-kinase inhibitors sorafenib or lenvatinib, have, until recently, been the treatment of choice for late-stage tumors ^[3][4][3,4]. For advanced HCC, as for a few other solid cancers, immunotherapy is one of the most promising and novel treatment approaches. A number of ongoing clinical trials have been reported ^[2][5][2,5] in which various immunotherapies, such as immune checkpoint inhibitors (ICIs), are utilized for the treatment of HCC, either alone or in combination with targeted and/or systemic therapies ^[2][5][2,5].

Despite the great clinical benefit that immunotherapies have offered, there are still many patients who do not respond or respond poorly to this type of treatment. In particular, only about 15–20% of advanced HCC patients respond to ICIs [6]. The reasons for the unsatisfactory clinical responses are not clear. One major area of research is indeed focusing on the identification of biomarkers that can better predict tumor responses to the immunotherapy, in order to improve the clinical outcomes and cover a broader number of cancer patients. Various biomarkers have been shown to predict responses to ICIs including tumoral PD-L1 expression [7] and tumor mutational burden (TMB) [8]. A higher TMB, based on genomic profiling of various tumor biopsies, may reflect a higher likelihood for response to ICIs [9], whereas PD-L1 expression can positively correlate with better responses to anti-PD-L1 therapy ^[10][11][10,11]. In addition, gene expression analysis on HCC adjacent tumor tissues has been able to identify signatures correlated with improved survival [12]. However, tissue biopsies require invasive tumor sampling, therefore making it harder to collect multiple samples and comprehensively track tumor genomic changes throughout the treatment ^[13][14][15][13,14,15]. In particular, HCC is a heterogeneous and molecularly complex cancer type, and conventional tissue biopsies are not able to fully reflect its heterogeneity and thus accurately predict therapy efficacy [16]. In addition, unlike other solid tumors, tissue biopsies for HCC are not frequently available, since diagnosis mainly relies on imaging [17]. Additionally, at a late stage, when the lesions are unresectable, a liver biopsy is usually not recommended for advanced HCC [18], while there is a risk of extrahepatic tumor spread along the needle track [19].

Liquid biopsy, where only a blood sample is taken to analyze circulating tumor cells (CTC) [20] or circulating tumor DNA (ctDNA) [21], can overcome these issues due to its minimally invasive nature. Additionally, it can be used to monitor the disease status systematically [22]. Alpha-fetoprotein (AFP) is one of the first liquid biopsy biomarkers used for the early diagnosis of HCC [23]. However, concerns about its sensitivity and high levels of AFP in non-HCC patients highlight the need to identify more sensitive and reliable biomarkers, which can be used alone or in combination with AFP [16].

2. Liquid Biopsy in HCC

Liquid biopsy refers to all the non-solid biologic materials used for the diagnosis and monitoring of HCC and is mainly based on the detection of ctDNA ^[24][25][26][24,25,26], circulating RNAs (e.g., microRNAs) ^[27][28][29][27,28,29], CTCs ^[20][30][31][20,30,31] and extracellular vesicles (EVs) (e.g., exosomes) ^[32][33][34][32,33,34] (Figure 1). In the following paragraphs, we will focus on two of the most promising liquid biopsy types in HCC, ctDNA and CTCs.

Figure 1. Liquid biopsy in hepatocellular carcinoma (HCC). Illustration of liquid biopsy biomarkers investigated in HCC, including circulating nucleic acids, circulating tumor DNA (ctDNA), circulating RNA (cRNA)/microRNAs (miRNA), extracellular vesicles (EVs)/exosomes and circulating tumor cells (CTC). Created with BioRender.com accessed on 11 August 2021.

2.1. ctDNA in HCC

ctDNA can arise in the bloodstream of cancer patients as a result of tumor cell apoptosis or necrosis [35]. As ctDNA represents the total tumor genome, its role in determining clinical outcomes gains more and more attention, especially in cases of advanced and unresectable HCC in which surgical or other invasive procedures, including tissue biopsy, are not recommended [18]. ctDNA contains cancer-associated molecular characteristics, such as mutational signatures [36], epigenetic changes ^[37][38][37,38] and cancer-derived viral sequences [39], which allow its discrimination from total normal circulating cellular free DNA ^[40][41][42][40,41,42]. Therefore, it could significantly contribute to the improvement in sensitivity for the current diagnostic tools, such as AFP, whose sensitivity remains at an average of 50% among HCC cases [16].

Despite the proven valuable role of ctDNA as a tumor biomarker, it still has some limitations, including the low levels of detection in the early stages, which makes it challenging for the early diagnosis of HCC. Another limitation is the lack of standardized procedures to prepare samples and analyze data [16]. Lastly, this approach is limited by the uncertain ability to capture spatial tumor heterogeneity, which can reflect clonal differences within or across cancer metastatic sites ^[43][48]. This implies that combinational and/or multi-parametric approaches may be needed, in order to increase the sensitivity and specificity of ctDNA as a biomarker for HCC.

2.2. CTCs in HCC

CTCs are also emerging as a promising biomarker for the prediction of HCC treatment efficacies. CTCs arise in the circulation after detachment from primary or metastatic tumor lesions ^[44][49]. They differ from other types of cancer biomarkers as they represent viable tumor cells circulating in the patient’s bloodstream. Therefore, CTCs can also provide comprehensive genetic information about tumor heterogeneity and drug sensitivity [20]. CTCs have been approved by the FDA as diagnostic markers for specific epithelial cancers [16]. However, their diagnostic role in HCC still requires further studies. A widely known CTC biomarker is the epithelial cell adhesion molecule (EpCAM) ^[44][49], a pan-cancer biomarker which has also been observed in HCC patients ^[45][50].

Despite the highly promising role of CTCs as a biomarker for HCC [20], it remains challenging to detect HCC CTCs early and accurately because of the lack of specific markers. Another limitation is that the frequencies of CTCs are usually low in the circulation, especially at the early stages ^[16][20][16,20]. Thus, combinational strategies may be needed, in order to improve the prognostic and diagnostic value of HCC.

3. Liquid Biopsy as a Diagnostic and Prognostic Tool

Liquid biopsy has been explored as a way to monitor cancer prognosis and diagnosis in a non-invasive manner. This technology has shown promising results in early diagnosis ^[46][84], detection of minimal residual disease ^[47][85] and decision making for systemic therapies of different types of cancers, including HCC ^[8][48][49][8,43,86].

Among all liquid biopsy analytes, ctDNA plays an important role in HCC prognosis [17]. ctDNA maintains the same genomic signatures that are present in the matching tumor tissue, allowing for the quantitative and qualitative evaluation of the mutation burden in body fluids ^[50][87]. In this way, ctDNA has been considered as a good biomarker and can be utilized in disease monitoring. The data of ctDNA include quantitative changes, such as differences in the concentration of ctDNA, as well as qualitative changes, such as gene mutations, DNA copy number variations and DNA methylation [16]. Indicatively, studies based on the detection of somatic single-nucleotide mutations and methylation changes in ctDNA could closely correlate with tumor burden over time in HCC patients and could be used to predict recurrence after surgery ^[17][51][52][17,88,89].

As ctDNA represents only a very small proportion of cell-free DNA, very sensitive and reliable detection methods are required. Levels of ctDNA are measured mainly by real-time PCR (RT-PCR) ^[53][93], while digital PCR (dPCR) ^[54][94] or sequencing methods are used for the detection of point mutations ^[55][95]. In addition to TERT and TP53 mutations as the prognostic factors of poor survival ^[56][57][45,47], other mutations have been shown to have prognostic values for HCC. MLH1 mutation was specifically associated with lower survival [1], whereas mutations of genes from the PI3K/mTOR pathway were shown to be the predictors of non-responders to TKI treatments for patients with advanced HCC ^[49][86].

A number of studies have also shown the prognostic values of circulating miRNAs in HCC. Lower survival rates were associated with patients with low levels of miR-1, miR-122, miR-26a, miR-29a and miR-223-3p ^{[58][59][60][61]}[96,97,98,99] or high levels of miR-155, miR-96 and miR-193-5p ^[62][63][100,101]. Furthermore, six additional miRNAs were identified as prognostic factors. Low levels of miR-424-5p or miR- 101-3p and high levels of miR-128, miR-139-5p, miR-382-5p and miR410 were associated with lower survival rates in HCC patients [17]. Alternatively, miRNAs have been studied in association with EVs ^[32][33][34][32,33,34]. In a cohort of 59 HCC patients, a correlation was found between tumor recurrences after liver transplantation and a high level of exosomal miR-718 ^[64][102]. Additionally, high levels of exosomal miR-665 or low levels of exosomal miR-638 and miR-320a were identified as predictors of poor survival ^[65][66][67][103,104,105].

Another cornerstone of liquid biopsy is the isolation and detection of CTCs, which have been described as a useful tool for the prognostication of HCC ^[68][106]. As introduced above, EpCAM-positive CTC cells have been intensively investigated in HCC ^[45][69][50,51]. However, since CTCs can lose their epithelial phenotype through epithelial-to-mesenchymal transition (EMT) in order to survive and metastasize ^[44][49], EpCAM cannot always be considered an optimal biomarker to detect HCC. Alternatively, other phenotypic markers have been explored, such as the hepatocyte-specific asialoglycoprotein receptor (ASGPR) ^[70][67], and the hepatocyte paraffin 1 ^[71][107], or incorporation of several markers simultaneously, as it has extensively been reviewed elsewhere [20]. Most recently, in a prospective study of 80 HCC patients, a multimarker assay combining cell surface markers EpCAM, ASGPR and GPC3 was able to detect HCC CTCs in 97% of the patients with high accuracy. Moreover, a phenotypic variant subset of CTCs was associated with aggressive disease progression and underlying metastasis, therefore highlighting the important implications of CTCs in treatment selection ^[72][108]. Another study showed that the detection of phosphorylated ERK (pERK) and pAkt in CTCs could predict the response to sorafenib efficacy in advanced HCC patients, similarly to tumor tissue biopsy ^[73][59].

4. Liquid Biopsy for Immunotherapy in HCC

The race towards the identification of immunotherapy predictive biomarkers is at the forefront of research in HCC. Among the biomarkers of interest, there are TMB and mutational signatures identified from ctDNA, and PD-L1 expression detected on CTCs ^[74][125]. TMB and PD-L1 expression are considered good predictors in several cancers, but the evidence in liver cancer has not been as established thus far ^[75][117]. In Table 1, we summarize the most recent literature in the field, and we highlight the key findings for each study.

Table 1.

Summary of the key findings in the most recent literature.

Type of Biomarker Analyzed	Key Findings	Reference

Changes in the ctDNA levels	Could significantly correlate with tumor size in cancer patients treated with anti-PD1 drugs and be a valuable prognostic factor of progression-free and overall survival.	^[76]	[126]
Targeted gene analysis of ctDNA	Can be a better option to evaluate TMB prior to immunotherapy in cases of advanced primary liver cancers when tissue biopsy is not recommended.	[18]
Mutational analysis of ctDNA	Could not be associated with response to ICI therapy but only to systemic treatment.	[36]
Levels of ctDNA at baseline	Higher levels of ctDNA at baseline were associated with an increased baseline tumor burden, and ctDNA turned negative in 70%, 27%, 9% and 0% of patients achieving a complete response, partial response, stable disease and disease progression, respectively.	^[77]	[128]
Levels of ctDNA at baseline	Undetectable ctDNA levels during treatment were linked to a longer progression-free survival.
Hyper-mutated ctDNA phenotype	Is associated with a favorable outcome in a cohort of 69 cancer patients with different histologies, including three HCC patients, treated with different immune checkpoint inhibitors.	^[78]	[122]
	Overall response rate, PFS and OS in high-alteration groups were significantly higher than in low-alteration groups.
Detection of Wnt/b-catenin-activating mutations	Wnt/b-catenin-activating mutations in HCC linked to potential tumor immunotherapy resistance in several studies.	^[79]^[80]^[81]^[82]	[129,130,131,132]
Detection of Wnt/b-catenin-activating mutations	Demonstration that liquid biopsy is potentially able to detect Wnt/b-catenin-activating mutations in HCC.	^[75]	[117]
Detection of Wnt/b-catenin-activating mutations	Detection of Wnt/b-catenin pathway-activating mutations might not be sufficient to identify advanced HCC patients with primary resistance to ICI.	^[83]	[119]
Targeted mutational analysis of CTNNB1 p.T41A mutation	ctDNA liquid biopsy managed to reveal mutations that were not detected in single tumor biopsies, thus increasing the detection rate of CTNNB1 mutation in HCC patients.	^[84]	[134]
PD-L1 expression on CTCs	Biomarker to assess ICI-based immunotherapy efficacy of advanced solid tumors.	^[85]	[135]
CTCs expressing PD-L1	PD-L1-positive CTCs are mainly found in advanced stages of disease, and they represent an independent prognostic factor for overall survival. 6 out of 10 patients receiving anti-PD-1 treatment had PD-L1-positive CTCs at baseline, and of these, 5 out of 6 had a favorable treatment response. 4 out of 10 patients receiving anti-PD-1 treatment did not have PD-L1+ CTCs and were non-responders.	^[86]	[136]

5. Limitations and Future Perspectives

Liquid biopsies, from which it is possible to isolate ctDNA and CTCs circulating in the blood stream of cancer patients, have shown promising data as prognostic and diagnostic tools for HCC while, at the same time, also allowing sequential and real-time monitoring of disease status in a minimally invasive manner ^[16][22][87][16,22,138]. This is especially important for patients with advanced and unresectable HCC, when surgical and other invasive procedures are not an option [18]. Interestingly, in some cases, ctDNA has been shown to be superior in identifying mutational signatures that could not be traced in single tumor biopsies ^[84][134] or that could correlate more closely with the tumor load and predict treatment efficacy with higher sensitivity, compared to AFP or imaging in patients with unresectable liver cancer [18]. These data prove the predictive value of analyzing ctDNA, which can add to the existing and well-established diagnostic tools for HCC. However, other studies have not been able to confirm the association of tumoral signature mutations with resistance to ICIs in HCC based on their ctDNA analysis [36]. These conflicting data highlight that larger and more comprehensive clinical studies are required, in order to obtain widely applicable and consistent results, which, at the moment, are missing in the field of liquid biopsy for HCC. CTCs are another important source of biomarkers, and many studies point out their role in the prognosis of HCC ^[69][51]. Importantly, CTCs have also been used to prognosticate responses to ICIs through the expression of PD-L1 in patients with advanced liver cancer ^[86][136]. However, low levels of CTCs in the early stages of HCC and the lack of standardized procedures make it challenging to integrate CTC techniques in the clinical practice for HCC diagnosis ^[69][51].

In this context, factors associated with the individual’s genetic background, the tumor microenvironment and interactions with the host immune system may additionally challenge the selection and evaluation of biomarkers able to predict tumor responses. In order to overcome these issues, integration of multiple biomarkers rather than single analytes as well as combinational approaches based on genomic and proteomic analyses will most probably be able to improve the precision of personalized treatments. Here, the implementation of novel NGS technologies and artificial intelligence might be of great importance to identify genomic and immunologic signatures predictive of treatment responses, as it has encouragingly been shown in recent studies ^{[88][89][90][91]}[73,139,140,141].

Up to the present, ICIs have been the most widely used immune-based approaches in the clinical management of advanced HCC ^[92][110]. Therefore, the majority of cancer immunotherapy predictive biomarkers described thus far refer mainly to responses to ICIs. Other immunotherapy modalities such as CAR-T cell therapies and neoantigen vaccines are currently being tested in ongoing clinical studies for HCC, and the preliminary results are awaited with great interest ^[2][5][2,5].