Current imaging methods and molecular biomarkers for hepatocellular carcinoma diagnosis: a narrative review

Abstract

Background and Objective

Hepatocellular carcinoma (HCC) is one the seventh most prevalent form of cancer in the world and the third highest cause of cancer fatalities in different parts of the world, mostly due to the late diagnosis. and primarily associated with chronic liver diseases. HCC is heavily linked and associated with hepatitis B and C infections, as well as lifestyle factors such as heavy alcohol consumption, obesity, and diabetes. This narrative review critically evaluates current diagnostic approaches for HCC, including imaging modalities, liver biopsy, and established and emerging molecular biomarkers, while highlighting recent advances, existing limitations, and future diagnostic strategies.

Methods

A narrative literature search was conducted using PubMed, Scopus, Google Scholar, and Web of Science to identify relevant English-language publications from 2009 to 2026. Original research articles, systematic reviews, meta-analyses, and clinical practice guidelines related to HCC diagnosis and surveillance were considered. The search terms covered HCC, imaging modalities, biomarkers, GALAD, GAAD, microRNA, and liquid biopsy. Guidelines from the American Association for the Study of Liver Diseases (AASLD), the European Association for the Study of the Liver (EASL), and the Asian Pacific Association for the Study of the Liver (APASL) were also reviewed.

Key Content and Findings

The reviewed evidence indicates that multiphasic computed tomography (CT) and contrast-enhanced magnetic resonance imaging (MRI) remain key imaging modalities for HCC diagnosis. Emerging biomarkers including des-γ-carboxy prothrombin (DCP), fibroblast growth factor 19 (FGF19), various microRNAs (miRNAs), and osteopontin (OPN), circulating microRNAs (miRNAs), show diagnostic potential, particularly when incorporated into multimarker panels or combined with conventional biomarkers such as alpha-fetoprotein (AFP). Artificial intelligence-assisted imaging and multimarker diagnostic models may also support earlier and more accurate detection. However, wider clinical implementation remains limited by insufficient multicentre validation, lack of assay standardization, limited reproducibility, and insufficient cost-effectiveness evidence.

Conclusions

Modern diagnostic methods increasingly incorporate these imaging tools alongside molecular markers to yield a more accurate diagnosis and optimal treatment strategy. While liver biopsy remains the gold standard for histopathological confirmation, its invasive nature and associated risks highlight the urgent need for reliable, non-invasive diagnostic alternatives. This review provides insights into the current strategies to diagnose HCC and with an emphasis on innovative diagnostic strategies.

Introduction

Cancer remains a major cause of death globally and poses a significant challenge to improving life expectancy across all countries. In 2019, the World Health Organization (WHO) reported that cancer is the leading cause of death before the age of 70 years in 112 out of 183 countries and ranked third in an additional 23 nationshttps://www.zotero.org/google-docs/?jvaJrk(1). One of the most prevalent forms of cancer is hepatocellular carcinoma (HCC), the primary type of liver cancer. HCC typically develops in individuals with chronic liver disease (CLD) or cirrhosis. In 2020, there were 906,000 new cases of HCC and 830,000 related deaths globally, making it the fourth most common cause of malignancy-related deaths worldwide and the sixth most common cancer overall. According to the WHO, the global burden of HCC is expected to continue increasing, with annual cases projected to exceed one million by 2025https://www.zotero.org/google-docs/?tlBy4G(2,3). HCC often develops from chronic liver conditions driven by several risk factors. There is a well-established link between HCC and long-term hepatitis B virus (HBV) and hepatitis C virus (HCV) infections. Additional risk factors including hepatitis D virus (HDV) co-infection, excessive alcohol consumption, diabetes mellitus, obesity, metabolic dysfunction-associated steatotic liver disease (MASLD), and smoking, further increase the risk of hepatocarcinogenesis (4,5).

HCC usually develops in patients with chronic liver disease or cirrhosis, regular surveillance has become the cornerstone of early detection. Current international guidelines recommend surveillance every six months for high-risk individuals using abdominal ultrasound (US) with or without serum alpha-fetoprotein (AFP) measurement. The primary aim of surveillance is to detect HCC at an early stage, when curative treatment options remain feasible, and long-term survival can be substantially improved (6). Despite these surveillance strategies, a considerable proportion of patients continue to be diagnosed at intermediate or advanced stages of disease. This is largely attributed to suboptimal surveillance uptake, variability in US performance, and the limited diagnostic sensitivity of AFP, particularly for early-stage tumours (7). Clinical decision-making remains challenging when indeterminate liver nodules are detected, requiring careful selection of appropriate follow-up imaging, biopsy, or continued surveillance. Consequently, improving the accuracy of HCC diagnosis remains an important clinical priority.

Recent updates to the American Association for the Study of Liver Diseases (AASLD) guidance have further refined HCC surveillance and diagnostic pathways by emphasising structured recall strategies following abnormal surveillance findings, assessment of US visualization quality, interpretation of longitudinal AFP trends, and evaluation of lesion growth patterns (8). These recommendations highlight the need for more reliable imaging approaches and complementary biomarkers to improve early detection. In parallel, multi-biomarker prediction models, such as the GALAD score and imaging-integrated models including GAAD, have shown promising diagnostic performance for identifying early-stage HCC and may complement conventional surveillance strategies (9).

The stage at diagnosis plays a major role in affecting overall survival. Patients diagnosed at an early stage are viable candidates for potentially curative treatments, such as surgical resection, ablative therapies, or liver transplantation. In contrast, patients with advanced-stage HCC are usually limited to palliative systemic therapies, which often yield poor response rates in most cases (10). Consequently, the 5-year survival rate surpasses 70% for early-stage patients but drops to less than 5% for those with advanced disease stages (11,12). Unfortunately, many patients are diagnosed at advanced stages because of inadequate surveillance adherence, limitations of current screening modalities, and the lack of effective risk-stratified monitoring strategies (13). The molecular mechanisms underlying hepatocarcinogenesis have been extensively investigated in recent years, leading to the identification of numerous molecular biomarkers with potential diagnostic and prognostic value (10). These advances have enhanced understanding of biomarker specificity, sensitivity, and their potential clinical applications in HCC detection and disease monitoring. In current clinical practice, HCC diagnosis typically relies on a combination of imaging modalities and serum biomarkers, particularly AFP.

In current clinical practice, the diagnosis of HCC is based on a combination of imaging findings, clinical risk assessment, and serum biomarkers, particularly AFP. As described by Brown et al. [2023], AFP remains one of the most widely used serum biomarkers for HCC surveillance and monitoring (14). However, AFP alone has limited sensitivity and specificity, particularly for detecting early-stage HCC, and elevated AFP concentrations may also occur in patients with chronic liver disease in the absence of malignancy. Furthermore, isolated AFP measurements do not reliably correlate with tumour size, vascular invasion, or tumour burden, and current clinical practice increasingly emphasizes longitudinal AFP trends rather than single measurements. Therefore, cross-sectional imaging modalities, including US, computed tomography (CT), and magnetic resonance imaging (MRI), play an essential role in the detection, characterization, and staging of HCC, thereby facilitating appropriate treatment selectionhttps://www.zotero.org/google-docs/?zt03jQ(15).

There are various screening techniques for HCC. Among these, abdominal US remains the recommended first-line surveillance modality because of its affordability, non-invasive nature, widespread availability, and suitability for repeated examinations every six months in high-risk individuals. However, its sensitivity for detecting HCC ranges from 58–70%, with substantially lower sensitivity for early-stage tumours and lesions smaller than 1 cm (16). In addition, the diagnostic performance of US is highly dependent on operator expertise, patient body habitus, liver echogenicity, and the presence of cirrhosis, which may compromise lesion visualization and contribute to missed diagnoses.

A liver biopsy may be necessary when imaging findings are inconclusive, particularly in patients without cirrhosis or when characteristic imaging features of HCC are absent. For tumours larger than 1 cm, the AASLD recommends against biopsy if typical radiological features are demonstrated on appropriate multiphasic imaging examinations. Liver biopsy offers a sensitivity of 65–95%, high specificity, and a high positive predictive value (PPV). However, Omar et al. [2023] reported that liver biopsy is an invasive procedure associated with potential complications, including pain, bleeding, bile leakage, gallbladder injury, and rarely, tumour seeding (17). Consequently, non-invasive imaging-based diagnosis remains the preferred approach whenever diagnostic criteria are fulfilled.

In addition to the conventional use of AFP and imaging tests, their inherent limitations have driven the search for more accurate, sensitive, and cost-effective diagnostic biomarkers for HCC. Recent advances in molecular biology and high-throughput technologies have facilitated the identification of numerous circulating biomarkers with potential applications in early detection, prognosis, and treatment monitoring (18). Examples of promising biomarkers include osteopontin (OPN), fibroblast growth factor-19 (FGF19), des-γ-carboxy prothrombin (DCP), AFP-L3, and various microRNA (miRNA) panels, and multi-biomarker prediction models such as GALAD and GAAD (19,20). According to Wang et al. [2018), preclinical studies emphasize the importance of OPN and miRNA panels for the detection of HCC before clinical symptoms become apparenthttps://www.zotero.org/google-docs/?a1rZw7(21). Similarly, DCP and AFP-L3 have demonstrated complementary diagnostic value to AFP, whereas FGF19 has been associated with tumour growth, disease progression, and therapeutic response (22). Multi-biomarker models such as GALAD, which integrates patient demographics with AFP, AFP-L3, and DCP, have consistently demonstrated superior diagnostic performance compared with AFP alone, particularly for early-stage HCC. More recently, imaging-integrated models such as GAAD have further improved risk stratification by combining biomarker profiles with imaging findings.

The continued validation and clinical implementation of these biomarkers are essential for improving diagnostic accuracy, facilitating earlier detection, predicting disease progression, and supporting individualized treatment strategies for patients with HCC. This narrative review provides a comprehensive overview of current imaging modalities and diagnosis methods for HCC and their limitations. The review also discusses the challenges involved in early diagnosis and biomarker discovery before highlighting strategies and advances in novel diagnostic platforms, biomarker-assisted diagnosis, and personalized medicine that may improve the accuracy of HCC diagnosis, facilitate earlier detection, and ultimately enhance patient outcomes and quality of life. The reported sensitivity and specificity values are presented to describe findings from individual studies rather than as pooled estimates or direct statistical comparisons. Overall, the review aims to identify the limitations of current methods and highlight possible ways to improve the early and accurate diagnosis of HCC. We present this article in accordance with the Narrative Review reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0538/rc).

Methods

A narrative literature search was conducted to identify relevant publications on imaging modalities, biomarkers, and emerging diagnostic strategies for HCC. Electronic databases, including PubMed, Scopus, Google Scholar, and Web of Science, were searched for articles published from 2009 to 2026, with additional landmark studies and guideline documents included where appropriate. The search strategy combined Medical Subject Headings (MeSH) and free-text keywords, including “hepatocellular carcinoma”, “HCC”, “diagnosis”, “surveillance”, “imaging”, “ultrasound”, “computed tomography”, “magnetic resonance imaging”, “biomarkers”, “alpha-fetoprotein”, “AFP-L3”, “des-γ-carboxy prothrombin”, “DCP”, “GALAD”, “GAAD”, “microRNA”, and “liquid biopsy” (Table 1).

Table 1. The search strategy summary.

Items	Specification
Date of search	The literature search was initiated during preparation of the manuscript in approximately January 2025 and updated in 2026 during manuscript revision
Databases and other sources searched	PubMed, Scopus, Google Scholar, and Web of Science. Reference lists of selected publications were also screened for additional relevant studies. Clinical practice guidelines from the AASLD, EASL, and APASL were also reviewed
Search terms used	“hepatocellular carcinoma”, “HCC”, “diagnosis”, “surveillance”, “imaging”, “ultrasound”, “computed tomography”, “magnetic resonance imaging”, “biomarkers”, “alpha-fetoprotein”, “AFP-L3”, “des-γ-carboxy prothrombin”, “DCP”, “GALAD”, “GAAD”, “microRNA”, and “liquid biopsy”
Timeframe	2009–2026
Inclusion and exclusion criteria	Inclusion criteria: English-language original research articles, systematic reviews, meta-analyses, and clinical practice guidelines relevant to HCC diagnosis, surveillance, imaging modalities, biomarker development, diagnostic performance, and clinical implementation. Priority was given to recent, high-quality studies with direct relevance to the objectives of the review. Exclusion criteria: non-English publications, editorials, commentaries, case reports, narrative reviews, and conference abstracts. Publications were excluded if the full-text article was unavailable or if they were duplicate records utilizing an identical patient cohort
Selection process	Study identification and initial selection were conducted by A.R.M.S. based on relevance to the scope of the review and the predefined inclusion criteria. The selected literature was subsequently reviewed by H.A.M.E.S. to assess its relevance and suitability for inclusion. Any uncertainties regarding inclusion were resolved through discussion between the two authors
Additional considerations	Reported sensitivity and specificity values were used descriptively from individual studies rather than pooled or treated as direct statistical comparisons

AASLD, American Association for the Study of the Liver Diseases; APASL, Asian Pacific Association for the Study of the Liver; EASL, European Association for the Study of the Liver; HCC, hepatocellular carcinoma.

English-language original research articles, systematic reviews, meta-analyses, and clinical practice guidelines relevant to HCC diagnosis and surveillance were considered. Reference lists of selected publications were also screened to identify additional relevant studies (Table 1). Clinical practice guidelines published by major professional societies, including the AASLD, the European Association for the Study of the Liver (EASL), and the Asian Pacific Association for the Study of the Liver (APASL), were included because they represent internationally recognised recommendations for HCC surveillance and diagnosis. Priority was given to recent, high-quality studies with direct relevance to imaging techniques, biomarker development, diagnostic performance, and clinical implementation.

Development of HCC and biological samples for diagnosis

HCC develops through a complex and multi-step process, resulting from genetic instability, chronic liver injury, and changes in the tumour microenvironment. Approximately 80% of cases occur in individuals with underlying cirrhosis (12), although the proportion varies depending on the underlying aetiology and geographical region. Increasing evidence also indicates that HCC may develop in patients with non-cirrhotic MASLD, highlighting the evolving epidemiology of HCC. Persistent liver damage frequently caused by HBV or HCV infections, metabolic disorders such as diabetes and non-alcoholic fatty liver disease (NAFLD), excessive alcohol consumption, or prolonged exposure to aflatoxins initiates a cycle of inflammation and oxidative stress (23). This prolonged insult activates hepatic stellate cells, leading to fibrosis and, eventually, cirrhosis, which alters the normal liver architecture and establishes a high-risk environment for malignant transformation. Inflammatory mediators such as interleukin-6 (IL-6), tumour necrosis factor-alpha (TNF-α), and transforming growth factor-beta (TGF-β) (24) play a central role in promoting cellular damage, enhancing fibrosis, and stimulating compensatory hepatocyte proliferation, although the relative contribution of each pathway varies between different HCC aetiologies and remains an active area of investigation. While these molecular mechanisms have been consistently demonstrated in experimental studies, their individual contributions to HCC initiation and progression in clinical settings are often influenced by patient-specific factors and coexisting liver disease. As shown in Figure 1, these processes collectively create a microenvironment conducive to hepatocarcinogenesis. At the molecular level, genetic and epigenetic modifications accumulate over time, disrupting essential regulatory pathways. Mutations in the TERT promoter result in sustained telomerase activity, enabling hepatocytes to bypass senescence (26), whereas TP53 mutation disrupts cell cycle control and apoptotic mechanisms (27). Activation of the Wnt-β-catenin signalling pathway promotes unregulated proliferation, and additional epigenetic changes further destabilize chromosomal integrity, fuelling tumour progression. Together, these molecular events create an environment where hepatocytes lose normal growth constraints, paving the way for malignancy (27). However, the frequency of these molecular alterations differs substantially among patient populations and underlying liver diseases, limiting their current utility as standalone diagnostic biomarkers.

Figure 1.

The mechanism of HCC development, depicting the progression from risk factors to tumour formation (25). HCC, hepatocellular carcinoma.

The tumour microenvironment also undergoes significant changes during tumorigenesis, supporting tumour survival and growth. Angiogenesis is primarily mediated by vascular endothelial growth factor (VEGF), which is upregulated under hypoxic conditions within the tumour to promote neovascularisation and maintain nutrient supply (28). Simultaneously, the tumour exploits immune evasion mechanisms, suppressing anti-tumour immune responses through regulatory T-cell expansion and programmed death-ligand 1 (PD-L1) upregulation, which collectively reduce the effectiveness of immune surveillance (29). In addition, aberrant activation of key survival pathways such as PI3K/AKT and MAPK signalling enhances cell proliferation and resistance to apoptosis, further accelerating tumour growth (29). Although these pathways have been widely reported, much of the supporting evidence originates from preclinical and translational studies. Their diagnostic and prognostic value in routine clinical practice remains under investigation, with ongoing efforts to identify clinically actionable molecular targets. In its advanced stages, HCC is characterised by unchecked hepatocyte proliferation, loss of apoptotic control, and the ability to invade surrounding tissue and metastasize, commonly spreading to the lungs, bones, and lymph nodes (25).

In HCC, diagnosis relies on both invasive and non-invasive biological samples, each offering advantages and limitations in diagnostic accuracy and patient compliance. Among non-invasive methods, blood-based biomarkers, including cell-free DNA (cfDNA) and circulating tumour DNA (ctDNA), have demonstrated significant potential in early detection due to their ability to reflect tumour-specific genetic and epigenetic alterations, collectively known as liquid biopsy (30,31). However, despite encouraging results, the reported diagnostic performance of cfDNA and ctDNA varies considerably across studies because of differences in patient populations, assay platforms, analytical methods, and disease stage at sample collection. Consequently, further standardisation and large-scale prospective validation studies are required before these biomarkers can be routinely implemented in clinical practice.

Similarly, urinary biomarkers, such as microRNAs (miR-93-5p), have been identified as promising indicators of HCC, particularly in HBV-related cases, with high sensitivity and specificity (32). Nevertheless, the majority of published studies involve relatively small cohorts, and external validation across diverse patient populations remains limited, making their diagnostic utility uncertain. While saliva remains a relatively underexplored biological sample in HCC diagnostics, Mariam et al. [2022] reported that saliva-based biomarkers, particularly miRNAs, hold potential as non-invasive tools for detecting HCC (33). A small-scale study found that certain salivary miRNA signatures could distinguish HCC patients from those with cirrhosis, indicating promising candidates for early diagnosis. Metabolomic studies have identified unique salivary metabolite profiles in HCC individuals compared to individuals with chronic liver disease and healthy controls, further supporting the potential use of saliva as a non-invasive diagnostic approach in HCC detection (34). However, evidence supporting salivary biomarkers remains preliminary, with relatively few studies available and limited independent validation. Their clinical applicability therefore requires confirmation through larger multicentre studies before routine adoption can be considered.

Invasive sampling methods, such as liver biopsy as shown in Figure 2, remain the gold standard for HCC diagnosis, enabling direct histopathological examination and molecular characterization of tumour tissues (36). Despite its diagnostic accuracy, liver biopsy is invasive, carries procedural risks and is not routinely required when characteristic imaging features are present according to current international guidelines (37). Consequently, considerable research efforts are focused on developing reliable non-invasive biomarkers that could reduce the need for tissue biopsy while maintaining diagnostic accuracy.

Figure 2.

Non-invasive and invasive biological samples that can be developed as multiple diagnostic avenues for HCC detection and diagnosis (35). HCC, hepatocellular carcinoma.

Blood tests for HCC diagnosis and associated limitations

HCC is the most common primary liver cancer and presents significant diagnostic challenges, particularly in the early stages of the disease. In clinical practice, blood biomarkers play a key role in the screening and surveillance of HCC, especially among high-risk individuals with chronic liver diseases (38). The most extensively utilised biomarker is AFP; however, its diagnostic utility is compromised by limited sensitivity and specificity (39). The diagnostic performance of AFP also varies considerably across studies, largely because of differences in patient populations, tumour stage, underlying liver disease, and the AFP cut-off values used, making direct comparison between studies challenging (40). Consequently, although AFP remains the most widely used serum biomarker in clinical practice, it is no longer recommended as a standalone diagnostic test by major international guidelines. To improve diagnostic performance, additional biomarkers such as AFP-L3, a glycoform of AFP, and DCP have been investigated, with several studies reporting improved specificity, particularly when these biomarkers are used in combination rather than individually (39). However, evidence regarding their incremental diagnostic value remains inconsistent, as reported sensitivities and specificities differ across studies owing to variations in study design, assay methodology, disease aetiology, and the proportion of early-stage HCC included in the study cohorts. Furthermore, many validation studies have been conducted in HBV-endemic populations, which may limit the generalisability of their findings to other patient populations.

Despite this, blood tests alone are insufficient to establish a definitive HCC diagnosis due to potential false positives in patients with benign liver conditions and false negatives in cases where tumours do not secrete these biomarkers (41). These limitations are particularly relevant for early-stage HCC, where biomarker concentrations may remain below diagnostic thresholds despite the presence of malignancy. Consequently, blood-based testing serves primarily as an adjunct to screening and surveillance, necessitating confirmation through imaging modalities like US, CT, or MRI (42). Current evidence therefore supports a multimodal diagnostic approach that integrates serum biomarkers with imaging findings and clinical risk factors to maximise diagnostic accuracy, rather than relying on any single biomarker alone (43). Ongoing research is focused on identifying novel biomarker panels and multimarker algorithms that can overcome the limitations of conventional blood tests and improve the early detection of HCC.

AFP

A widely used biomarker for HCC diagnosis is AFP, though its accuracy depends on the threshold selected for clinical interpretation. A systematic review and meta-analysis found that an AFP cutoff of 20 ng/mL has a sensitivity of 60% and a specificity of 86%. However, raising the threshold to 200 ng/mL increases specificity to 99% but lowers sensitivity to 31% (44). This illustrates the trade-off between sensitivity and specificity when selecting AFP cut-off values. Lower thresholds are generally preferred for surveillance because they improve the detection of early-stage HCC, whereas higher thresholds reduce false-positive results but may miss a substantial proportion of patients with early disease. Consequently, the optimal AFP threshold remains dependent on the intended clinical application and patient population.

As a result of these limitations, complementary biomarkers have been explored to improve diagnostic accuracy. The AFP-L3 fraction, which is a glycoform of AFP, is more specific for HCC, especially when interpreted in conjunction with total AFP levels. Similarly, DCP has demonstrated improved diagnostic performance when used alongside AFP and AFP-L3, with several studies reporting higher sensitivity and specificity than AFP alone (40). However, the reported diagnostic performance of AFP-L3 and DCP remains variable across studies, reflecting differences in assay methodology, disease stage, underlying liver disease, and the demographic characteristics of the study populations. In particular, many studies have been conducted in HBV-endemic regions, and further validation in more diverse populations is required before these biomarkers can be universally adopted.

Newer methods that combine the neutrophil-to- lymphocyte ratio (NLR) with AFP have demonstrated better diagnostic performance than AFP alone Although these findings are encouraging, current evidence is derived predominantly from retrospective studies with relatively small sample sizes, and prospective multicentre validation is still required before routine clinical implementation. In contrast, AFP continues to present diagnostic challenges, as approximately 15–30% of patients with advanced HCC exhibit normal AFP levels, leading to potential false negatives (45). Moreover, elevated AFP concentrations are not exclusive to HCC, as chronic liver diseases and other malignancies may also cause abnormal levels, contributing to false-positive results (40). Additionally, inconsistencies in AFP measurement techniques across laboratories can further impact diagnostic accuracy (46). These limitations have contributed to a shift in clinical practice towards interpreting serial AFP measurements together with imaging findings and patient risk factors, rather than relying on isolated AFP values.

Furthermore, the clinical application of AFP differs across healthcare settings depending on the availability of diagnostic resources. AFP testing is often conducted in primary care or smaller diagnostic centres using enzyme-linked immunosorbent assay (ELISA) techniques due to their ease of use and cost-effectiveness (40). These facilities often rely on AFP measurements as initial screening tools, especially in regions where advanced imaging modalities may not be readily accessible. However, the specificity and sensitivity of AFP alone are limited, necessitating further diagnostic evaluations for accurate HCC detection (47). This reliance on AFP may contribute to variability in diagnostic performance across healthcare systems, particularly in low-resource settings where access to CT or MRI is restricted. In contrast, specialized hospitals and oncology centres adopt a more comprehensive diagnostic approach. AFP testing is integrated with advanced imaging techniques to enhance diagnostic accuracy. These institutions often have access to sophisticated laboratory infrastructure, enabling the assessment of additional biomarkers such as AFP-L3 and DCP, which have shown improved specificity for HCC detection. Combining multiple serum biomarkers with imaging studies provides greater diagnostic confidence and improves patient stratification, allowing more timely and appropriate therapeutic decision-making (48). Nevertheless, the cost-effectiveness and accessibility of these multimarker approaches remain important considerations, particularly in resource-limited healthcare settings, and further prospective studies are needed to establish their role in routine clinical practice.

Imaging tests for HCC diagnosis and associated limitations

Precise diagnosis and staging are essential for selecting the most appropriate treatment strategy and improving clinical outcomes in patients with HCC. Liver cancers are typically detected through imaging techniques (Table 2), and their characteristics are confirmed through tissue biopsy for histological examination when imaging findings are inconclusive or atypical (55,56). However, current international guidelines increasingly support non-invasive imaging-based diagnosis in patients with cirrhosis who demonstrate characteristic radiological features of HCC, thereby reducing the need for biopsy in many clinical scenarios (57).

Table 2. An overview of the different imaging devices and their clinical applications.

Imaging device	Development	Indications for HCC patients	Sensitivity and specificity	Contraindications	Advantages/application	Disadvantages/limitation	Reference
US	Uses sound waves for imaging	First-line for surveillance in high-risk patients	Sensitivity: 40–81%, specificity: 80–100%	None absolute; obesity & gas can reduce quality	Non-invasive, cheap, widely available	Operator-dependent, limited for small lesions (<1 cm)	(49)
CEUS	Uses microbubble contrast for enhanced imaging	Characterising lesions & monitoring treatment	Sensitivity: 92%, specificity: 87%	Contrast allergy, severe heart/lung conditions	No radiation, real-time imaging	Operator-dependent, limited availability	(50,51)
CT	X-ray-based cross-sectional imaging	Diagnosis, staging, and surgical planning	Sensitivity: 77.5%, specificity: 91.3%	Severe allergy to contrast, renal impairment, pregnancy (radiation)	Detailed anatomy, good for tumour vascularity	Radiation exposure, contrast nephrotoxicity, less effective for small tumours	(52,53)
MRI	Magnetic fields & radio waves for high-resolution images	Preferred for detailed lesion characterisation	Sensitivity: 97%, specificity: 100%	Metal implants, claustrophobia, renal impairment	No radiation and non-invasive	Expensive, long scan time, limited availability	(54)

CEUS, contrast-enhanced US; CT, computed tomography; HCC, hepatocellular carcinoma; MRI, magnetic resonance imaging; US, ultrasound.

Conventional imaging techniques for HCC include ultrasonography, CT, and MRI, which offer detailed assessment of hepatic lesions by identifying their location, size, and characteristics (56). Each imaging modality possesses distinct strengths and limitations, and their diagnostic performance is influenced by factors including tumour size, disease stage, liver background, imaging protocol, and operator expertise. Consequently, reported diagnostic accuracy varies considerably across studies, making direct comparisons challenging. Conversely, contrast-enhanced US (CEUS) has emerged as a valuable adjunctive imaging modality, particularly for the characterisation of small or indeterminate focal liver lesions, although its diagnostic performance is highly dependent on operator experience and the availability of US contrast agents (58). Moreover, these imaging modalities play complementary roles throughout the clinical management of HCC, extending beyond diagnosis to include disease surveillance, tumour staging, treatment planning, assessment of therapeutic response, and post-treatment follow-up. While advances in imaging technology have substantially improved lesion detection and characterisation, no single modality is universally superior in all clinical situations. Therefore, imaging findings should be interpreted alongside clinical history, laboratory investigations, and serum biomarkers to optimise diagnostic accuracy and guide patient management.

US

US is a medical imaging technique that uses high-frequency sound waves to capture the body’s internal organs in real time as the waves bounce off tissues and organs (59,60). It is widely used for the diagnosis and monitoring of various medical conditions and remains the recommended first-line imaging modality for HCC surveillance in high-risk patients because of its non-invasive nature, low cost, widespread availability, and lack of ionising radiation (61). Consequently, US plays a central role in routine HCC surveillance programmes. The detection and characterization of hepatic lesions using US has shown a high level of accuracy (16). Hu et al. reported that US has sensitivity ranging from 40-81% and a specificity of 80–100% (49). However, the wide variation in reported diagnostic performance reflects differences in study design, patient populations, operator expertise, surveillance versus diagnostic settings, and lesion characteristics. In particular, US demonstrates lower sensitivity for detecting early-stage HCC and small lesions (<2 cm), which are the primary targets of surveillance programmes. Although US is a valuable screening modality, it has limited sensitivity for detecting smaller HCC; therefore, more sensitive imaging techniques, such as CEUS, have been developed to improve lesion characterisation (62).

CEUS is an advanced US technique that uses microbubble-based US contrast agents (USCAs) to give a more comprehensive assessment of liver vascularity and parenchymal enhancement patterns (63). USCAs, which consist of 1 to 10 μm gas bubbles filled with low-solubility gases, are injected intravenously, pass through the pulmonary circulation, and act as intravascular blood markers (50). This enables real-time visualisation of tumour vascularisation and microvascular blood flow, facilitating the characterisation of focal liver lesions, including HCC. High-risk screening of HCC as well as the assessment of suspected liver nodules is also widely performed using CEUS (50). CEUS has been reported to achieve a sensitivity of 92% and specificity of 87% in diagnosing HCC (51). Nevertheless, these values should be interpreted with caution, as diagnostic performance varies considerably among studies depending on lesion size, patient selection, imaging protocols, and operator experience. CEUS has demonstrated particular value in the evaluation of indeterminate liver lesions detected on conventional US and in differentiating between benign from malignant liver tumours (14,64). Several studies have also reported improved sensitivity for detecting hypervascular lesions compared with conventional US, particularly in experienced centres. However, most supporting evidence originates from specialised institutions, and its performance may not be directly generalisable to routine clinical practice.

Despite its benefits in HCC detection and characterisation, CEUS also has limitations. Its sensitivity decreases markedly for tumours measuring less than 5 mm (65) and remains suboptimal for lesions smaller than 1 cm (66). Hence, CEUS should be used when conventional US reveals a lesion measuring more than 1 cm (49). This limitation is not unique to CEUS and reflects the inherent challenge of detecting sub centimetre HCC across all currently available imaging modalities. Another limitation is that the diagnostic accuracy of CEUS depends heavily on operator expertise and experience, highlighting the need for standardised training and quality assurance programmes (67). Additionally, US beam penetration is reduced in obese patients and those with fatty liver disease because increased hepatic echogenicity limits visualisation of deeper liver structures, thereby reducing diagnostic performance (68). These patient-related factors contribute to variability in CEUS performance reported across different clinical studies. There are also few absolute contraindications, such as severe coronary artery disease and pulmonary hypertension, and although USCAs are generally considered safe, adverse reactions have been reported in a small proportion of patients (69,70).

In 2023, Lee and colleagues reported that, following the approval of CEUS for imaging focal liver lesions, accumulating clinical evidence demonstrated improved sensitivity for HCC detection compared with conventional US. As a result, CEUS has been incorporated into several clinical guidelines as a second-line diagnostic tool for HCC (1,64,66). Current international guidelines generally recommend CEUS as a complementary imaging modality rather than a replacement for multiphasic CT or MRI because it provides excellent real-time vascular assessment but is limited by operator dependency and the inability to comprehensively evaluate the entire liver and detect extrahepatic disease (71). Therefore, CEUS is particularly valuable for characterising indeterminate focal liver lesions identified during surveillance and may improve diagnostic confidence when used alongside other imaging modalities. Future prospective multicentre studies are required to standardise CEUS protocols and further validate its diagnostic performance across diverse patient populations and healthcare settings.

CT scan

CT scan refers to a radiological imaging process that enables cross-sectional visualisation of internal body structures using X-rays (52). During the CT scan process, an X-ray tube rotates around the patient to generate computer-generated cross-sectional images of the body (53). Multiphasic contrast-enhanced CT has become one of the cornerstone imaging modalities for the diagnosis and staging of HCC owing to its high spatial resolution, rapid imaging acquisition, and ability to provide detailed information about the arterial and venous phases of liver perfusion. This technology is exceptionally effective in characterizing the nature of liver lesions and confirming an HCC diagnosis. The diagnostic capability of CT is based on its ability to identify the characteristic vascular enhancement pattern of HCC, including arterial phase hyperenhancement followed by washout during the portal venous or delayed phases, which forms the basis of non-invasive HCC diagnosis in current international guidelines. The distinct blood flow and extracellular volume in HCC compared to normal liver tissue create key imaging features that help in its diagnosis (72).

Besides, CT demonstrates strong diagnostic accuracy for HCC across different disease stages, with a reported sensitivity of 77.5% and specificity of 91.3% (73). However, these diagnostic estimates should be interpreted cautiously because reported sensitivities and specificities vary considerably among studies owing to differences in lesion size, imaging protocols, contrast timing, patient characteristics, and reference standards used for diagnosis. In general, CT demonstrates higher diagnostic accuracy for lesions larger than 2 cm, whereas its sensitivity decreases for small or early-stage HCC, which remains a common limitation across cross-sectional imaging modalities (74). Several meta-analyses have similarly reported that although CT provides excellent specificity, MRI generally offers superior sensitivity for detecting small HCC lesions, particularly when hepatobiliary contrast agents are used (75).

The finding makes CT a highly effective tool for identifying and characterizing HCC. CT scan is often favoured over MRI in cases where patients are claustrophobic or heavily built, as CT examinations are shorter and do not require prolonged placement within an enclosed scanner (76). Furthermore, CT is more widely available than MRI in many healthcare systems, making it an important diagnostic modality, particularly in emergency settings and institutions with limited MRI accessibility. Patients can be uncooperative or may experience difficulty remaining still during prolonged imaging examinations; in these situations, CT is quicker, more comfortable, and less affected by motion artifacts than MRI and can be employed for the rapid detection and management of cirrhotic patients who are at elevated risk for developing HCC (77).

Despite these advantages, CT has several important limitations. The use of ionising radiation makes repeated examinations less desirable, particularly in patients undergoing long-term surveillance. In addition, iodinated contrast agents carry the risk of contrast-induced nephropathy and allergic reactions, which may limit their use in patients with impaired renal function or previous contrast hypersensitivity (78). These factors should be carefully considered when selecting the most appropriate imaging modality for individual patients.

Current evidence supports CT as an essential imaging modality for confirming HCC diagnosis, evaluating tumour extent, and staging disease prior to treatment. However, it is generally not recommended as a primary surveillance tool because of radiation exposure and contrast-related risks. Instead, CT is most performed following abnormal surveillance findings on US or when further lesion characterisation is required (79). Future advances in CT technology, including dual-energy CT, photon-counting CT, and artificial intelligence-assisted image analysis, may further improve lesion detection and diagnostic accuracy while reducing radiation exposure (80).

MRI scan

MRI is a non-invasive medical imaging tool that creates high-resolution 3D images of the organs and tissues of the body by utilising a strong magnetic field and computer-generated radio waves (81). MRI uses magnets to produce strong magnetic field, which aligns hydrogen nuclei within the body before radiofrequency pulses generate detailed tissue images (81). Owing to its superior soft tissue contrast and multiparametric imaging capability, MRI is widely regarded as one of the most accurate imaging modalities for the detection and characterisation of HCC, particularly in patients with indeterminate liver lesions (82). It has been stated that gadoxetic acid-enhanced MRI significantly improves the detection of small liver lesions, including early-stage HCC and high-grade dysplastic nodules (54). Studies reported that contrast-enhanced MRI (CE-MRI) was more accurate than dynamic CT and traditional MRI for detecting tumours smaller than 2 cm, with a specificity of 100% and a sensitivity of 97%, although sensitivity for lesions smaller than 1cm can be reduced (83,84). The improved performance of gadoxetic acid-enhanced MRI is largely attributed to the additional hepatobiliary phase, which enhances lesion conspicuity and facilitates the detection of early HCC that may not demonstrate typical vascular enhancement patterns on conventional imaging. This makes MRI a useful tool for diagnosing and monitoring HCC, particularly for identifying small lesions, as it provides detailed tissue imaging characterisation and vascular assessment. Nevertheless, the reported diagnostic performance of MRI varies across studies because of differences in MRI protocols, contrast agents, imaging sequences, patient selection, and reference standards. Consequently, direct comparison of sensitivity and specificity between studies should be interpreted with caution. Recent meta-analyses have shown that CE-MRI provides strong overall sensitivity and specificity in detecting HCC nodules However, its sensitivity decreases for smaller lesions, particularly those under 2 cm, and is significantly reduced for subcentimetre lesions, which may limit early detection in certain cases (85,86). This limitation remains a common challenge across all currently available imaging modalities and reflects the difficulty of distinguishing very early HCC from regenerative or dysplastic nodules in cirrhotic livers. Despite this, MRI consistently demonstrates higher sensitivity than CT for the detection of small HCC lesions in most comparative studies, supporting its role as the preferred problem-solving modality when US or CT findings are inconclusive (87).

Despite its diagnostic advantages, MRI has notable limitations. MRI is less accessible in areas with limited resources since it is more costly and less generally available than CT or US (88). These economic and logistical considerations may limit its routine use for surveillance, particularly in low- and middle-income countries where US remains the primary screening modality. Additionally, MRI requires patients to remain still in an enclosed area for long periods, making it challenging for patients who suffer from claustrophobia, who may experience significant discomfort during the examination. This could lead to motion artifacts, reducing image quality, thereby limiting image interpretation and reducing diagnostic accuracy (76,89). Furthermore, MRI may be contraindicated in patients with certain implanted medical devices or severe renal impairment when gadolinium-based contrast agents are required, further restricting its applicability in selected patient populations.

Overall, current evidence supports MRI as one of the most sensitive imaging modalities for the diagnosis and characterisation of HCC, particularly for small or indeterminate lesions. However, its higher cost, longer examination time, and limited accessibility necessitate careful patient selection. In current clinical practice, MRI is most commonly used for diagnostic confirmation, lesion characterisation, and pre-treatment staging rather than routine surveillance. Future developments, including abbreviated MRI protocols, quantitative imaging biomarkers, and artificial intelligence-assisted image analysis, may improve the accessibility, efficiency, and diagnostic performance of MRI in HCC management (90).

Liver biopsy and associated limitations

Liver biopsy continues to be the reference standard diagnostic technique for assessing various hepatic conditions (91). It is warranted when critical clinical insights regarding a patient’s diagnosis, prognosis, or treatment strategy cannot be acquired through less invasive methods or when required for research. This procedure has evolved into a reliable and indispensable tool for hepatologists in evaluating diverse liver diseases (92). Different approaches, such as percutaneous, transvenous, laparoscopic, and plugged biopsy techniques, are available, with each method chosen based on patient-specific considerations like bleeding risk or the necessity for precision sampling. Although non-invasive alternatives exist, liver biopsy remains invaluable due to its capacity to provide comprehensive histopathological analysis allowing direct assessment of tumour morphology, tumour differentiation, vascular invasion, and molecular characteristics that cannot be fully evaluated using imaging alone. These histopathological findings are particularly valuable for confirming atypical lesions and supporting precision medicine approaches, including molecular profiling and biomarker discovery (93).

While advances in imaging tools like CT and MRI can identify liver lesions and have decreased the demand for biopsies (94,95), current international guidelines increasingly support imaging-based diagnosis for patients with cirrhosis who demonstrate characteristic imaging features of HCC, thereby avoiding unnecessary invasive procedures. Consequently, the role of liver biopsy has shifted from routine diagnosis towards selected clinical scenarios where imaging findings are inconclusive or atypical. Nevertheless, evidence suggests that the diagnostic accuracy of biopsy is not absolute. Histopathological interpretation of HCC remains challenging because tumour heterogeneity, small sample size, and overlapping morphological features between early HCC, dysplastic nodules, and regenerative nodules may affect diagnostic accuracy. These factors contribute to sampling error and may lead to false-negative findings, particularly in patients with cirrhotic livers (96,97).

In addition, liver biopsy is associated with procedure-related complications such as bleeding, infection, and pain, with mild post-procedural discomfort reported in up to 84% of patients (92). Although major complications are uncommon, their potential occurrence necessitates careful patient selection and risk-benefit assessment before biopsy is performed. The risk of sampling error, where a small tissue sample may not fully represent the heterogeneity of the tumour, and the potential for false-negative results, particularly in cirrhotic livers where distinguishing between HCC and benign regenerative nodules can be difficult (98). Furthermore, the invasive nature of the procedure exposes cancer cells to the risk of spreading along the biopsy tract despite the incidence of tumour seeding being relatively low according to contemporary studies (99). The reported risk of needle tract seeding varies among studies because of differences in biopsy techniques, follow-up duration, and patient populations, although improvements in biopsy procedures have reduced this complication considerably over time.

Despite these challenges, liver biopsy remains essential in cases where imaging findings are inconclusive. A histopathological assessment may be required, particularly in non-cirrhotic individuals suspected of having HCC or when radiological findings fail to provide a definitive diagnosis. According to the AASLD, biopsy is not advised for lesions exceeding 1 cm if two independent imaging modalities produce consistent results indicative of HCC (10,100). Similarly, other international guidelines advocate a selective rather than routine use of liver biopsy, reflecting the increasing diagnostic accuracy of multiphasic CT and MRI. Therefore, current evidence supports liver biopsy as a complementary diagnostic tool reserved for diagnostically challenging cases rather than the primary method for HCC diagnosis. Future research should focus on integrating histopathological findings with emerging molecular biomarkers and liquid biopsy approaches to improve diagnostic precision while minimising the need for invasive tissue sampling.

Emerging and potential diagnostic molecular biomarkers for HCC

HCC continues to pose a major global health challenge, with high mortality rates due to the frequent diagnosis of patients at advanced disease stages. Consequently, the identification of reliable molecular biomarkers capable of improving early detection has become a major focus of HCC research. Although AFP remains the most widely used serum biomarker, its limited sensitivity for early-stage HCC and suboptimal specificity have driven the search for novel biomarkers that may complement existing diagnostic approaches rather than replace them (101). Increasing evidence suggests that combining multiple biomarkers with clinical risk factors and imaging findings may provide greater diagnostic accuracy than reliance on a single biomarker alone (102).

Among the promising candidates is FGF19, a key regulator of bile acid metabolism, which is frequently elevated and overexpressed in HCC (21). Another potential marker is OPN, a glycoprotein associated with inflammation and tumour development, which has shown promise as a serum biomarker, with elevated levels correlating with HCC progression (98). Another well-established marker, DCP, is produced due to defective vitamin K-dependent carboxylation in malignant liver cells and has been widely used in HCC diagnosis (103). Additionally, miRNAs, small non-coding RNAs that regulate gene expression, have been increasingly recognized for their role in HCC, with specific miRNA signatures serving as potential non-invasive diagnostic tools (99).

Although these biomarkers have demonstrated encouraging diagnostic performance in individual studies, their reported sensitivity and specificity vary considerably because of differences in patient populations, disease aetiology, tumour stage, laboratory methodologies, and study design. Furthermore, many biomarkers have been evaluated predominantly in retrospective studies or geographically restricted cohorts, limiting the generalisability of the available evidence. Consequently, large prospective multicentre validation studies remain necessary before these biomarkers can be incorporated into routine clinical practice.

Recent research has also shifted from evaluating individual biomarkers towards developing multi-marker diagnostic models, such as GALAD and GAAD, which integrate serum biomarkers with demographic or clinical variables to improve diagnostic performance, particularly for early-stage HCC (9). These approaches have demonstrated promising results across multiple validation studies and highlight the growing trend towards biomarker panels rather than reliance on single biomarkers. These biomarkers are reviewed individually in the following sections to critically evaluate their biological significance, diagnostic performance, current clinical evidence, and limitations. Collectively, they represent promising candidates for improving early HCC detection and may contribute to more personalised diagnostic strategies when integrated with imaging findings and clinical risk assessment.

FGF19

FGF19 acts as a potential molecular biomarker, particularly in HCC. This hormone, primarily synthesised in the small intestine, regulates bile acid metabolism in the liver. However, overexpression of FGF19 has been associated with hepatocarcinogenesis and tumorigenesis in the liver (22). Overexpression of this ligand can activate the FGF receptor 4 (FGFR4) pathway, thereby promoting tumour progression, cell survival, and angiogenesis by stimulating the formation of new blood vessels that supply nutrients and oxygen to the tumour, facilitating tumour growth, and increasing its metastatic potential (100). According to Kanzaki et al. [2021], FGF19 has emerged as a promising biomarker for monitoring tumour progression and demonstrated moderate to high sensitivity for HCC detection (103). However, the reported diagnostic performance varies across studies because of differences in patient populations, disease stage, assay methodology, and sample size. Consequently, the current evidence should be interpreted cautiously until larger prospective validation studies become available. Clinical studies evaluating FGF19 remain relatively limited compared with established biomarkers such as AFP and DCP.

One study developed laboratory tests for FGF19 using evidence from clinical trials and reported that FGF19 was expressed in 27% of the 395 tumour samples analysed (104). The immunohistochemistry results for FGF19 aligned with its levels in liver cancer tissues. FGF19 was found to be overexpressed in HCC tumour samples, and FGF19-positive tissues were also associated with high levels of AFP and macrovascular invasion, whereas FGF19-negative tissues showed minimal disease progression and low to undetectable AFP levels (105). These findings suggest that FGF19 expression may be associated with more aggressive tumour behaviour. However, because these observations were derived from a single study cohort, independent validation in larger and more diverse populations is required before FGF19 can be considered a reliable prognostic biomarker.

A number of preclinical observations have repeatedly demonstrated that the complex of FGF19-FGFR4 signalling axis plays an important role in HCC progression, particularly when aberrantly activated (106-108). FGF19 is also considered a promising therapeutic target for the treatment of HCC, with encouraging findings reported from both in vitro and in vivo studies (109). Additionally, dysregulation of FGFR4 and its association with TGF-β1 have further highlighted FGFR4 as a potential therapeutic target, particularly in HCC cases characterised by invasiveness and metastasis (106,107). Nevertheless, much of the mechanistic evidence supporting the FGF19-FGFR4 pathway originates from experimental models, and translation into routine clinical practice remains limited. Further clinical studies are needed to determine whether inhibition of this pathway leads to meaningful improvements in patient outcomes. Studies have shown a strong correlation between tumour progression and FGF19 expression in HCC patients, proposing its potential as a prognostic biomarker for disease severity and clinical outcomes, with elevated FGF19 levels being associated with poorer prognosis, reinforcing its role in tumour aggressiveness (106). Desnoyers et al. [2007] provided experimental evidence supporting the oncogenic role of FGF19 by showing that targeting FGF19 signalling effectively inhibited tumour growth in both colon cancer xenografts and FGF19 transgenic HCC models (108). Although these findings provide important proof-of-concept evidence, they were generated using preclinical models and therefore cannot be directly extrapolated to clinical practice. Prospective clinical trials are required to establish whether therapeutic targeting of FGF19 produces similar benefits in patients with HCC.

Together, these studies underscore the dual significance of FGF19 as both a prognostic marker and a potential therapeutic target in HCC management (109). Overall, the available evidence indicates that FGF19 is a biologically plausible biomarker with potential diagnostic, prognostic, and therapeutic value. However, compared with established biomarkers, the current clinical evidence remains relatively limited and heterogeneous. Standardisation of detection methods, larger prospective multicentre studies, and external validation across different HCC aetiologies are required before FGF19 can be incorporated into routine clinical practice. Furthermore, future studies should investigate whether combining FGF19 with established biomarkers, imaging findings, or multi-marker models such as GALAD could further improve the early detection of HCC.

OPN

OPN expression is closely associated with diseases characterised by inflammation, angiogenesis, fibrosis, and carcinogenesis. OPN is a phosphorylated glycoprotein involved in wound healing, immune regulation, and tissue remodelling, and its dysregulated expression has been implicated in HCC development and progression (110). Because OPN is released into the circulation, it has attracted considerable interest as a potential non-invasive serum biomarker for HCC diagnosis and prognosis (111). In a study conducted by Abdel-Hafiz et al. [2018], serum OPN levels were significantly higher in HCV-associated HCC patients than in the control group, suggesting that OPN may be a useful biomarker in patients with chronic HCV-related cirrhosis (112). However, these findings were obtained from a relatively specific patient population, and additional validation in cohorts with different HCC aetiologies, including HBV infection and MASLD, is required to determine the broader clinical applicability of OPN.

Many cancers, including lung, breast, colorectal, gastric, ovarian, papillary thyroid carcinoma, melanoma, pleural mesothelioma, and HCC, demonstrate elevated OPN expression (20,113). Therefore, despite its biological relevance, OPN lacks tumour specificity when used as an isolated biomarker. This represents one of its major limitations, as elevated OPN concentrations may reflect malignant or inflammatory processes rather than HCC specifically. Consequently, OPN is frequently assessed in conjunction with other biomarkers, especially AFP, to improve diagnostic accuracy, with several studies reporting improved sensitivity compared with AFP alone (114). Moreover, OPN expression is strongly associated with fibrosis and chronic liver inflammation, both of which are critical risk factors in hepatocarcinogenesis (115).

The study conducted by Ge et al. [2015] evaluated the combined diagnostic value of AFP and OPN in 390 participants, comprising 89 HCC patients, 36 with liver cirrhosis, 65 with chronic hepatitis B, and 200 healthy controls. The study reported that combining these biomarkers increased the sensitivity and specificity of HCC diagnosis compared with AFP alone (116). These findings support the growing consensus that multimarker approaches are likely to provide greater diagnostic accuracy than reliance on individual biomarkers. Nevertheless, the incremental benefit of adding OPN to existing diagnostic algorithms requires further validation in larger prospective studies before routine clinical implementation can be recommended.

In a retrospective cohort study conducted by Wu et al. [2025], serum OPN levels were significantly elevated in patients with HCC compared with patients with other chronic liver diseases and were consistently associated with poorer clinical outcomes. Furthermore, elevated OPN levels were associated with improved responses to anti-PD-L1 immunotherapy, suggesting that OPN may have prognostic as well as predictive value (117). However, these findings require confirmation in prospective studies before OPN can be considered a reliable predictive biomarker for treatment response. A meta-analysis of 12 studies concluded that the combination of AFP and OPN achieved superior pooled diagnostic sensitivity compared with AFP alone, indicating that OPN may serve as a valuable complementary biomarker rather than a replacement for AFP (118). OPN has been extensively investigated in both preclinical and clinical studies, which consistently support its role as a key driver of HCC progression and metastasis. The available evidence suggests that OPN enhances tumour cell migration and invasion, facilitating the spread of HCC within the liver and to distant organs (119). Beyond tumour progression, OPN has been implicated in obesity-related inflammation and MASLD, suggesting that it may contribute to hepatocarcinogenesis through multiple biological pathways (119,120). Experimental studies have further demonstrated that OPN deficiency reduces liver inflammation and tumour burden, supporting its potential as a therapeutic target. However, these findings are predominantly based on animal models, and their clinical significance remains to be established.

A clinical trial conducted on a Syrian cohort demonstrated that plasma OPN levels were remarkably elevated in HCC patients compared to those with chronic liver diseases and healthy controls, suggesting its potential as a biomarker for distinguishing malignant and non-malignant liver conditions (121). A study evaluating OPN in combination with glypican-3 (GPC3) concluded that a dual biomarker approach enhanced diagnostic accuracy, outperforming AFP alone (122). Another study demonstrated that patients with HCV-related cirrhosis exhibited markedly higher plasma OPN levels upon progression to HCC, indicating its potential utility for surveillance in high-risk populations (112). Similarly, a European cohort study using pre-diagnostic OPN levels found that elevated circulating OPN concentrations were associated with an increased incidence of HCC, suggesting that OPN may have value for predicting early tumour development (123).

Collectively, the available evidence indicates that OPN is one of the most extensively investigated emerging biomarkers for HCC. Its biological role in inflammation, fibrosis, and tumour progression provides strong mechanistic support for its clinical application. However, its lack of tumour specificity and variability across different patient populations currently limit its use as a standalone diagnostic biomarker. Future research should focus on prospective multicentre validation studies and the integration of OPN into multimarker models, such as combinations with AFP, GPC3, or GALAD-based algorithms, to determine its optimal role in routine clinical practice.

DCP

Among the potential biomarkers for HCC, DCP, also known as protein, induced by vitamin K absence or antagonist-II (PIVKA-II), is an abnormal form of the coagulation protein prothrombin. The formation of DCP results from impaired post-translational γ-carboxylation, leading to the production of an abnormal prothrombin molecule by malignant hepatocytes (23). Unlike AFP, sustained DCP elevation has been associated with an increased risk of disease progression and poorer clinical outcomes in patients with unresectable HCC (124). This characteristic has made DCP an attractive complementary biomarker for improving diagnostic accuracy, particularly in patients with normal AFP concentrations.

It has been stated that HCC produces DCP without interacting with other coagulation factors. DCP has been observed as a specific marker for the diagnosis of HCC, with high plasma DCP levels reported in 44–81% of HCC patients (125). However, the wide range of reported positivity rates reflects substantial heterogeneity among studies, including differences in disease stage, assay methodology, patient populations, and diagnostic thresholds. Therefore, comparisons between studies should be interpreted cautiously. Individuals with elevated DCP levels are more likely to experience tumour recurrence, vascular invasion, and metastasis than individuals with low DCP levels, with enzyme immunoassay studies demonstrating that higher plasma DCP concentrations are associated with larger tumour burden (126,127). These findings suggest that DCP may have prognostic as well as diagnostic value.

DCP has also been found to actively contribute to HCC progression by promoting tumour proliferation, invasion, metastasis, and angiogenesis through several signalling pathways, including MAPK-related mechanisms (128). These enzymes degrade the extracellular matrix, facilitating tumour invasion and the spread of cancerous cells. Specifically, the increased activity of MMPs enables HCC cells to penetrate surrounding tissues, thereby promoting an aggressive cancer phenotype. Another study by Zhang et al. [2014] emphasised that DCP functions not just as a biomarker but also as an active contributor to tumour growth, promoting tumour cell proliferation by influencing various intracellular signalling pathways associated with hepatocarcinogenesis (129). Although these mechanistic studies provide strong biological support from the role of DCP in HCC progression, much of this evidence is derived from experimental models, and further clinical studies are required to determine the therapeutic implications of targeting DCP-related pathways. In addition, DCP has demonstrated higher diagnostic accuracy than AFP in several comparative studies. A multicentre retrospective study conducted by Choi [2013] reported an area under the receiver operating characteristic curve (AUC) of 0.939 for DCP, significantly greater than the 0.817 reported for AFP (P<0.05) (130). While these findings suggest that DCP may outperform AFP in certain patient populations, the retrospective nature of the study limits the strength of the evidence, highlighting the need for prospective validation. In another study, the combined use of AFP, DCP and AFP-L3 achieved an AUC of 0.895, demonstrating improved diagnostic performance compared with individual biomarkers (131). These findings support the growing consensus that multi-marker panels provide superior diagnostic performance by compensating for the limitations of individual biomarkers.

A meta-analysis also demonstrated the diagnostic capabilities of AFP-L3, with a pooled sensitivity of 48.3% and specificity of 92.9% for HCC diagnosis (132). Similarly, Chen et al. [2023] reported that combining DCP and AFP significantly improved prognosis or diagnosis (ROC) AUC values for HCC detection. Specifically, the sensitivity increased to 81.95% and specificity to 89.96% when both biomarkers were combined, compared with lower values obtained using DCP alone (133). These findings further support the complementary role of DCP in multi-marker diagnostic strategies rather than its use as an isolated biomarker.

Overall, the current evidence indicates that DCP is one of the most clinically validated emerging biomarkers for HCC and demonstrates greater specificity than AFP, particularly for identifying aggressive tumours associated with vascular invasion and poorer prognosis (134). Nevertheless, variability in assay platforms, diagnostic cut-off values, and patient populations has contributed to inconsistent diagnostic performance across studies. Consequently, standardisation of DCP measurement and further prospective multicentre validation studies are required before its widespread adoption as a standalone diagnostic biomarker. Future research should focus on integrating DCP with AFP, AFP-L3, imaging findings, and validated prediction models such as GALAD to maximise diagnostic accuracy for early-stage HCC.

miRNAs panels

miRNAs are small non-coding RNA (ncRNA) molecules (approximately 22 nucleotides in length) that regulate gene expression and are involved in biological processes, such as cell proliferation, differentiation, apoptosis, and maintenance of cellular homeostasis (135). Dysregulation of miRNA expression has been implicated in the initiation and progression of numerous cancers, including HCC, making circulating miRNAs attractive candidates for non-invasive cancer diagnosis. Unlike conventional serum biomarkers, miRNAs can be detected in multiple body fluids, including blood, urine, and saliva, where they exhibit relatively high stability despite exposure to RNases. This biological stability has generated considerable interest in their application as liquid biopsy biomarkers for HCC (21,136).

Several miRNAs, such as miR-21, have been extensively investigated because of their consistent dysregulation across multiple malignancies. In particular, miR-21 exhibits consistently increased expression in HCC, as well as in several other cancer types, suggesting that although it is highly sensitive, its tumour specificity may be limited (136,137). Consequently, miR-21 alone is unlikely to provide sufficient diagnostic accuracy for HCC and is more appropriately used as part of a biomarker panel.

More importantly, other circulating miRNAs, including miR-1972, miR-193a-5p, miR-214-3p, and miR-365a-3p, appear to demonstrate greater specificity for HCC, as their expression levels remain relatively unchanged in healthy individuals, chronic hepatitis B (CHB), and cirrhotic patients but are significantly elevated in HCC. Circulating miRNAs were evaluated as diagnostic and prognostic biomarkers for HCC, identifying these four miRNAs as significantly elevated in HCC patients but not in individuals with CHB or cirrhosis (138). Among these, miR-1972 demonstrated the highest diagnostic performance, while miR-193a-5p also showed consistent overexpression in HCC compared with non-HCC liver disease controls (138). These findings suggest that disease-specific miRNA signatures may offer greater diagnostic specificity than individual miRNAs. However, the study was conducted in a single patient cohort, and external validation is required before these biomarkers can be generalised to broader populations.

In contrast, miR-214-3p and miR-365a-3p, while elevated in the plasma, have previously been reported as downregulated in tumour tissues, indicating that circulating and tissue-derived miRNA profiles may differ because of complex biological mechanisms governing miRNA release and transport (138). This discrepancy highlights one of the major challenges in miRNA research and complicates direct interpretation of circulating biomarker studies.

Besides, miR-21 functions as a well-characterised oncomiR, promoting tumorigenesis through downregulation of tumour suppressor genes, including programmed cell death 4 (PDCD4), phosphatase and tensin homolog (PTEN), and tissue inhibitor of metalloproteinases-3 (TIMP3), thereby increasing metastatic potential and apoptosis and inhibiting apoptosis (136,138). miR-21 has demonstrated a specificity of 83.3% and sensitivity of 61.1% (AUC =0.773) for differentiating HCC from chronic hepatitis patients (139). Additionally, when distinguishing HCC from healthy volunteers, miR-21 exhibited 87.3% sensitivity and 92.0% specificity, with improved diagnostic performance observed when combined with AFP (139). Although these results are encouraging, diagnostic performance varies considerably across studies because of differences in patient ethnicity, disease aetiology, laboratory protocols, and miRNA quantification methods. Consequently, standardisation remains a major barrier to clinical implementation.

miR-21 is also highly elevated in HBV-associated HCC patients and has been shown to suppress the tumour suppressor protein SASH1, thereby promoting HBV-driven oncogenic signalling and reducing apoptosis induced by sorafenib. Inhibition of miR-21 restores SASH1 expression and improves sensitivity to sorafenib, suggesting that miR-21 may possess therapeutic as well as diagnostic significance (140). In a study conducted by Hoa et al. [2025] involving 450 Vietnamese patients, expression of miR-21-5p, miR-34a-5p, and miR-30d-3p was significantly elevated and demonstrated potential for early HCC detection when combined with AFP. Notably, miR-21-5p exhibited the greatest increase (5.43-fold) and showed promise in distinguishing HBV-related HCC from CHB (141). Although these findings support the complementary role of miRNAs alongside AFP, further validation across different geographical regions and HCC aetiologies is required before routine clinical application.

The measurement of specific miRNAs has also been employed in urine, offering a completely non-invasive approach for early HCC detection. Studies involving patients with HCV infection identified several urinary miRNAs with altered expression patterns, including upregulation of miR-625, miR-532, and miR-618, together with downregulation of miR-516-5p and miR-650 (142). These findings highlight the potential of urinary miRNAs as convenient surveillance biomarkers for high-risk populations. However, current evidence remains limited to relatively small observational studies, and further validation is required before urinary miRNA testing can be incorporated into clinical surveillance programmes.

The overexpression of miR-500a has also been recognised in HCC, with elevated expression reported in nearly half of HCC cases. miR-500a promotes tumour proliferation, invasion, and metastasis by suppressing the tumour suppressor gene PTEN (143). Its presence within macrophage-derived exosomes further suggests that exosomal miRNAs may represent another promising class of liquid biopsy biomarkers for HCC diagnosis and disease monitoring (22).

Overall, circulating miRNAs represent one of the most promising emerging biomarker classes for HCC because they are biologically stable, minimally invasive, and closely associated with molecular mechanisms underlying hepatocarcinogenesis. Nevertheless, considerable heterogeneity exists across published studies owing to differences in sample type (serum, plasma, urine, or saliva), RNA extraction methods, detection platforms, normalisation strategies, and patient populations. As a result, few individual miRNAs have demonstrated sufficiently consistent performance for routine clinical use. Current evidence increasingly favours the use of multi-miRNA panels integrated with established with biomarkers such as AFP and DCP, together with clinical risk factors, to improve diagnostic accuracy (141). Large prospective multicentre studies and international standardisation of analytics methods remain essential before miRNA-based diagnostics can be implemented in routine HCC surveillance and diagnosis.

Challenges of turning emerging biomarkers into clinical use for HCC diagnosis

HCC is a highly heterogeneous disease, with significant genetic and molecular diversity both within individual tumours and across different patients (144). Chronic liver disease, diabetes, viral hepatitis, alcohol use, and MASLD are some of the conditions that contribute to HCC heterogeneity (145). Finding biomarkers that accurately reflect the underlying biology of the diseases and are consistently expressed across these diverse patient populations remains a major challenge (146). This biological heterogeneity partly explains why many biomarkers demonstrate encouraging performance in discovery studies but fail to achieve similar diagnostic accuracy during external validation. Consequently, reliance on a single molecular biomarker is unlikely to provide adequate diagnostic performance across the full spectrum of HCC patients. Usage of a single molecular marker for diagnosis and disease monitoring may be challenging due to the unpredictability of HCC, which includes evolving tumours and emerging resistance mechanisms (147).

A major problem is the suboptimal diagnostic performance of commonly used markers. For instance, studies evaluating AFP have shown that its sensitivity is particularly low in early-stage HCC, with up to 80% of small tumours not exhibiting elevated levels, while its specificity is also reduced because AFP concentrations may increase in patients with cirrhosis or chronic hepatitis (146). Although newer biomarkers such as DCP, OPN, FGF19, and circulating miRNAs have demonstrated improved performance in selected studies, none has consistently achieved sufficient sensitivity and specificity to replace AFP or imaging in routine clinical practice. Instead, increasing evidence supports combining multiple biomarkers with imaging and clinical risk factors to improve overall diagnostic accuracy.

The lack of large‐scale prospective validation studies remains one of the greatest barriers to the clinical translation of emerging biomarkers. Many published studies are retrospective, single-centre investigations with relatively small sample sizes and geographically restricted populations, limiting the generalisability of their findings. A large number of samples from multiple centres is required to validate the reliability of these biomarkers (147). In addition, studies have demonstrated substantial intratumoural heterogeneity within HCC, where different regions of the same tumour, or even multiple nodules in a single patient, may display distinct genetic profiles, with approximately 36% of patients harbouring tumours of different clonal origins (144,148). This heterogeneity further complicates biomarker discovery because tissue obtained from a single biopsy may not accurately represent the molecular characteristics of the entire tumour.

The accessibility and cost-effectiveness of incorporating these biological markers into current clinical practice, including routine surveillance programmes, must also be carefully considered, particularly in resource-limited settings (149). Other studies have similarly highlighted the limitations associated with implementing emerging biomarkers in HCC diagnosis. For example, Aralica et al. [2024] conducted a meta-analysis demonstrating that US alone has limited sensitivity for early HCC detection, and although combining US with AFP improves detection rates, diagnostic performance remains inadequate for reliable early diagnosis (150). Similarly, Fares et al. [2024] reported variability in AFP performance depending on patient population, particularly among individuals with cirrhosis, further emphasising its limited specificity (151). These findings reinforce the need for multimodal diagnostic strategies rather than dependence on any single biomarker or imaging modality. There are also major obstacles related to analytical standardisation and reproducibility. Differences in sample type, specimen handling, laboratory protocols, detection platforms, assay cut-off values, and data processing methods contribute to substantial inter-laboratory variability, making direct comparison between studies difficult (18). These methodological differences are one of the principal reasons why reported diagnostic performance varies considerably across the literature. In addition, studies focusing on biomarker panels have suggested improved diagnostic accuracy but also highlighted translational challenges. For example, Piñero et al. [2020] developed the GALAD score, which combines AFP, AFP-L3, DCP, and clinical variables, demonstrating high sensitivity and specificity; however, its widespread clinical adoption remains limited because further external validation across diverse ethnic groups, disease aetiologies, and healthcare systems is still required (152).

Regulatory approval represents another important challenge. New biomarkers must demonstrate not only analytical validity but also clear clinical utility before they can be incorporated into routine patient care. Regulatory agencies such as the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA) require robust evidence demonstrating safety, reliability, reproducibility, and clinical benefit before approving novel diagnostic tests (153). Importantly, high diagnostic accuracy alone is insufficient if the biomarker does not improve clinical decision-making or patient outcomes.

Researchers must also address assay standardisation, reproducibility, and quality control. Beyond regulatory approval, widespread adoption depends on healthcare provider and patient acceptance, requiring education regarding the benefits and limitations of emerging biomarkers. Cost-effectiveness and accessibility are also critical for successful clinical integration (154). Overall, current evidence suggests that the greatest opportunity lies not in identifying a single superior biomarker but in integrating molecular biomarkers with imaging findings, clinical risk factors, and validated prediction models such as GALAD. Future multicentre prospective studies employing standardised analytical protocols and diverse patient populations will be essential to establish the clinical utility of these emerging biomarkers and facilitate their incorporation into international HCC surveillance guidelines.

Strategies to overcome the challenges in HCC diagnosis

There have been several strategies proposed to address the limitations associated with the diagnosis of HCC. Firstly, the use of multiple serum biomarkers instead of relying on a single biomarker, such as AFP, has been proposed. It has been demonstrated that the combination of several biomarkers, including AFP, AFP-L3, and DCP, significantly improves the specificity and sensitivity for the early detection of HCC (131,155). This approach compensates for the limitations of individual biomarkers, as each reflects different aspects of tumour biology. Consequently, multi-marker strategies are increasingly regarded as a more robust diagnostic approach than reliance on AFP alone.

Although AFP remains the most widely used biomarker, its sensitivity for early-stage HCC is suboptimal. Liao et al. [2026] demonstrated that combining AFP with AFP-L3 and DCP significantly improved diagnostic accuracy. In a cohort of 578 patients, the combined use of AFP, AFP-L3, and DCP achieved a higher diagnostic performance (AUC =0.895) compared with AFP alone, highlighting the benefit of multi-marker strategies for early detection and recurrence prediction (131). Similarly, multiple studies have confirmed that combinations such as AFP + DCP + AFP-L3 consistently outperform single biomarkers in terms of sensitivity and specificity (156-158). Nevertheless, although these findings are encouraging, differences in assay methodologies, diagnostic thresholds, and patient populations continue to limit direct comparison across studies. Large prospective multicentre studies are therefore required before these biomarker combinations can be universally adopted.

Novel biomarkers such as GPC3, golgi protein-73 (GP73), and circulating miRNAs such as miR-625, miR-532, and miR-618 have also been proposed as useful tools for the early diagnosis and prognosis of HCC (18). A systematic review and meta-analysis including over 6,000 participants demonstrated that GPC3 has good diagnostic accuracy and that its combination with AFP or GP73 further enhances detection performance, particularly in distinguishing HCC from cirrhosis (159). Similarly, a diagnostic meta-analysis by Zhao et al. [2020] reported that the combined use of GP73, GPC3, and AFP achieved a pooled sensitivity of 0.91, specificity of 0.84, and an AUC of 0.95, indicating excellent diagnostic capability (160). Furthermore, earlier clinical studies have shown that GP73 may exhibit higher sensitivity than AFP, particularly for detecting small HCC lesions, while combining GP73 with AFP significantly improved diagnostic performance (AUC ≈0.93) (131,161). Although these biomarkers demonstrate encouraging results, many have not yet undergone sufficient prospective external validation across different HCC aetiologies and geographical populations. Consequently, their routine clinical implementation remains limited despite promising diagnostic performance.

Molecular biomarkers, particularly circulating miRNAs, have also provided new opportunities for non-invasive diagnosis. Studies have shown that several miRNAs, including miR-21, miR-122, and miR-224, are dysregulated in HCC and can serve as diagnostic and prognostic indicators (136). These molecules are highly stable in circulation and have been reported to achieve high diagnostic accuracy when used as multi-miRNA panels (162). However, considerable variation in sample processing, RNA extraction methods, normalisation strategies, and analytical platforms remains a significant obstacle to their widespread clinical adoption. International standardisation of laboratory methodologies will therefore be essential before miRNA-based diagnostics can be routinely implemented (163). In addition, many studies remain small, retrospective, or lack prospective multicentre validation and regulatory qualification. Similar concerns apply to AI, radiomics, and multi-omics approaches, which still require external validation, standardised data collection, explainable models, workflow integration, and evidence of clinical benefit before routine implementation (164).

Advances in imaging techniques, including multiphasic CT, contrast-enhanced MRI, and CEUS, have improved the detection and characterisation of HCC. More recently, Artificial intelligence-assisted radiomics has emerged as a promising approach for improving lesion detection, image interpretation, and risk stratification by extracting quantitative imaging features beyond conventional radiological assessment. It has also improved the accuracy of HCC diagnosis, particularly for small or indeterminate lesions (165). Despite these advances, most AI models remain at the validation stage, and concerns regarding model generalisability, external validation, explainability, and regulatory approval continue to limit widespread clinical implementation.

In addition, several predictive models and scoring systems, including the GALAD score, have been proposed for the early detection of HCC in high-risk populations, particularly patients with cirrhosis (166,167). Among these, GALAD currently possesses one of the strongest evidence-bases because it integrates demographic variables with multiple serum biomarkers, thereby reducing the limitations associated with individual biomarkers. The platelet, age, gender, hepatitis B (PAGE-B) score, originally developed to predict HCC risk in patients with hepatitis B, was recently modified by incorporating serum albumin, resulting in improved predictive performance (168). Another model, the ADRESS-HCC model, was developed to predict the one-year risk of HCC in patients with cirrhosis. Among the many prognostic scoring indexes, Toronto Hepatocellular Carcinoma Risk Index (THRI) remains one of the extensively validated tools for risk stratification (169).

Machine learning (ML) models have further expanded opportunities for personalised risk prediction. Xu et al. [2024] developed an XGBoost prediction model using clinical data from 6,980 patients, achieving an AUC of 0.746 for HCC risk prediction (170). Similarly, ML models incorporating longitudinal sustained virological response data have demonstrated potential for predicting HCC risk among patients with HCV-related cirrhosis, thereby supporting personalised surveillance strategies (171). Although these predictive models demonstrate considerable promise, most remain dependent on retrospective datasets and require prospective multicentre validation before integration into routine clinical practice. Furthermore, successful implementation will require seamless integration with electronic health records and careful evaluation of cost-effectiveness, clinical utility, and healthcare workflow.

The clinical adoption of biomarker-based surveillance also depends on economic feasibility. Although some published studies suggest that multi-marker algorithms may be cost-effective, or potentially cost saving, in selected patients with cirrhosis (156), these findings may not apply equally across all healthcare systems. Their economic value depends on assay costs, testing frequency, local imaging costs, false-positive rates, and the need for follow-up CT or MRI. Therefore, moving completely away from traditional ultrasound-based surveillance may not be economically realistic, particularly in resource-limited settings. A more practical strategy may be to use multimarker algorithms alongside ultrasound or in selected high-risk patients with poor ultrasound visualisation

When the available approaches are considered together, each has a different role in HCC surveillance and diagnosis. Ultrasound remains the most practical first-line surveillance method because it is non-invasive, widely available, and relatively inexpensive. However, its performance may be reduced in patients with obesity, MASLD, advanced cirrhosis, or poor liver visualisation (172). Single biomarkers such as AFP are simple to measure and inexpensive, but their sensitivity and specificity are limited when they are used alone. In comparison, multi-marker algorithms such as GALAD may improve risk assessment by combining several biomarkers with clinical variables. Nevertheless, these algorithms require additional laboratory testing, standardised assays, external validation, and clear diagnostic thresholds. Therefore, they are currently more realistic as complementary or risk-based tools rather than complete replacements for ultrasound (145).

Overall, the available evidence suggests that no single strategy is likely to overcome the diagnostic challenges associated with HCC. Instead, the most realistic approach is an integrated d and risk-based pathway in which ultrasound remains the main surveillance tool, while serum biomarkers, advanced imaging techniques, AI , and validated risk prediction models (164,173) are used to support decision making in selected patients Future research should prioritise large prospective multicentre studies, standardised laboratory protocols, and cost-effectiveness analyses to determine how these complementary technologies can be optimally incorporated into routine HCC surveillance

Conclusions

Liver cancer, particularly HCC, remains a significant global health concern due to its high mortality rate and frequent late-stage diagnosis. Due to the heterogeneous biology of HCC, underlying liver disease, and diverse patient populations, the diagnosis process remains complicated and challenging. Although current diagnostic approaches combining imaging and serum biomarkers have substantially improved HCC detection, each modality has inherent limitations that reduce diagnostic accuracy, particularly for early-stage disease. Conventional diagnostic methods, including liver biopsy, also remain invasive and are associated with complications such as bleeding, infection, sampling error, and tumour seeding. Therefore, there is an increasing need for non-invasive diagnostic strategies with an emphasis on early diagnosis. Improving surveillance and detecting HCC at an earlier stage remain essential because delayed diagnosis limits access to potentially curative therapies and contributes to poor survival outcomes.

Liver biopsy has traditionally remained the gold standard for diagnosis; however, its invasive nature places patients at risk of bleeding, infection, and tumour seeding. As a result, there has been an increasing shift toward non-invasive diagnostic methods, including advanced imaging techniques and molecular biomarkers. Among these, emerging biomarkers such as FGF19, DCP, OPN, and circulating miRNAs have demonstrated promising diagnostic and prognostic potential, particularly when incorporated into multi-marker panels rather than used individually. Similarly, advances in CEUS, multiphasic CT, and gadoxetic acid-enhanced MRI have improved lesion detection and characterisation, especially when integrated with biomarker assessment.

However, despite encouraging diagnostic performance reported in numerous studies, relatively few biomarkers have progressed into routine clinical practice. The major barriers remain limited prospective multicentre validation, biological heterogeneity, assay standardisation, cost-effectiveness, and regulatory approval.

Future progress in HCC diagnosis is therefore likely to depend on integrated multimodal approaches that combine advanced imaging, liquid biopsy, validated biomarker panels, clinical risk prediction models such as GALAD, and artificial intelligence-driven image analysis. ML and deep learning algorithms also offer considerable potential to improve personalised risk prediction, surveillance, and clinical decision-making, although prospective validation and integration into routine healthcare systems remain necessary. Additionally, next-generation sequencing (NGS) and proteomic technologies are expected to further accelerate biomarker discovery, supporting more precise and personalised diagnostic strategies.

In conclusion, substantial progress has been made in the development of non-invasive diagnostic approaches for HCC. Nevertheless, future research should prioritise large prospective multicentre studies, standardised analytical methodologies, and artificial intelligence-assisted diagnostic tools into routine clinical practice. Such integrated strategies have the greatest potential to improve early detection, optimise treatment selection, and ultimately enhance survival and quality of life for patients with HCC.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.