Integrating molecular and immune biomarkers for precision therapy in hepatitis B: Associated hepatocellular carcinoma

Abstract

In this editorial, we comment on the article by Wang et al, which investigates molecular and immune biomarkers predictive of response to sintilimab plus lenvatinib in hepatitis B virus-associated hepatocellular carcinoma (HCC). Yet, despite remarkable progress with immune-checkpoint and anti-angiogenic combinations, the biological heterogeneity of HCC continues to limit durable responses and individualized care. By integrating high-resolution transcriptomic, exomic, and immune-cell-profiling data, Wang et al identified a coherent triad - elevated LINC01554 expression, enrichment of CD4+ central-memory T cells, and solitary-tumour morphology - that independently predicted prolonged progression-free survival. This constellation links tumour-intrinsic transcriptional restraint, adaptive immune competence, and anatomical containment, illustrating how multi-omic profiling can clarify determinants of therapeutic benefit. These insights signify a shift from empiricism to biologically guided therapy, providing a scaffold for biologic stratification, longitudinal response monitoring, and rational sequencing of immunotherapeutic and anti-angiogenic agents. Collectively, they redefine HCC as a dynamic biological ecosystem rather than a uniform malignancy and highlight the imperative to embed multi-omic biomarker platforms within future clinical-trial design - marking a decisive step toward precision hepatology in inflammation-driven cancers.

Key Words: Hepatitis B virus-associated hepatocellular carcinoma; LINC01554; CD4+ central memory T cells; Sintilimab; Lenvatinib; Multi-omic biomarkers; Immune checkpoint inhibitors; Precision hepatology

Core Tip: Hepatitis B virus-associated hepatocellular carcinoma exemplifies the complexity of inflammation-driven oncogenesis. The integration of molecular and immune biomarkers enables a more precise prediction of therapeutic response. Emerging multiomic evidence identifies a triad of high LINC01554 expression, increased CD4+ central memory T cells, and solitary tumour morphology that signifies restrained oncogenic signalling, preserved immune competence, and limited clonal heterogeneity. Recognition of this molecular-immune-anatomical interplay provides a biologic framework for patient selection, response monitoring, and rational sequencing of sintilimab and lenvatinib, representing a decisive step toward scientifically grounded precision hepatology in viral hepatocellular carcinoma.

INTRODUCTION

Hepatocellular carcinoma (HCC) remains one of the leading causes of cancer-related mortality worldwide, with a disproportionate burden in regions where chronic hepatitis B virus (HBV) infection is highly prevalent and access to surveillance and early diagnosis is variable[1-3]. Even in settings with well-organised surveillance, a substantial proportion of patients with HBV-associated HCC present at intermediate or advanced stages, when curative therapies are not immediately feasible and systemic or locoregional approaches dominate management[2].

Over the past decade, systemic treatment for unresectable HCC has evolved from single-agent multikinase inhibitors such as sorafenib and lenvatinib to a more complex armamentarium that includes immune checkpoint inhibitors (ICIs) and ICI-based combinations[4-6]. Lenvatinib has demonstrated non-inferiority to sorafenib in the first-line setting, while multiple second-line agents - including regorafenib, cabozantinib, and ramucirumab - have incrementally extended survival in selected populations[6].

Despite these advances, clinical experience and trial data converge on a sobering reality: Responses to ICI-based and tyrosine kinase inhibitor (TKI)-based therapies are heterogeneous, and durable benefit is limited to a subset of patients. In most series, objective responses range from 20% to 35%, and even within responders, progression-free survival (PFS) is highly variable[5-7]. This variability reflects the interplay of tumour-intrinsic oncogenic programmes, the immunological composition and organisation of the tumour microenvironment, and classical clinical determinants such as tumour burden, vascular invasion, and liver function[8-11].

The integrative biomarker analysis by Wang et al[1] in the World Journal of Hepatology takes an important step toward rationalising and refining the use of sintilimab plus lenvatinib in HBV-associated HCC through integrative biomarker discovery. Building on a prospective conversion-therapy trial of sintilimab plus lenvatinib in intermediate and locally advanced HCC, which demonstrated encouraging objective responses and a meaningful surgical conversion rate, the authors performed multi-omic profiling of baseline tumour biopsies from patients with predominantly HBV-related disease[1,12].

Using RNA sequencing, immune-cell deconvolution, and whole-exome sequencing, they identified a triad of features - high LINC01554 expression, increased CD4+ central memory T-cell (Tcm) infiltration, and solitary tumour morphology - that independently predicted prolonged PFS under sintilimab-lenvatinib therapy[12-16]. This editorial situates those observations within the broader evolution of precision hepatology, explores their mechanistic plausibility, and proposes a conceptual framework for integrating molecular and immune biomarkers into future clinical decision-making for HBV-associated HCC[17-20].

THE EVOLVING THERAPEUTIC CONTEXT: DUAL PROGRAMMED CELL DEATH PROTEIN 1 AND VASCULAR ENDOTHELIAL GROWTH FACTOR-PATHWAY BLOCKADE

The rationale for combining programmed cell death protein 1 (PD-1)/programmed death ligand 1 blockade with inhibition of the vascular endothelial growth factor (VEGF) axis in HCC rests on both clinical and mechanistic foundations[6,13]. Sorafenib and lenvatinib - agents with potent anti-angiogenic activity - established VEGF signalling as a central therapeutic target in advanced HCC[10]. Parallel work with ICIs such as nivolumab and pembrolizumab showed that PD-1 inhibition can induce durable responses in a minority of patients, but single-agent activity was insufficient to transform outcomes at the population level[19,20].

Translational analyses from the pivotal trials have further shown that pre-existing T-cell-inflamed phenotypes, high CD8+ T-cell density, and VEGF-related signatures influence the balance of benefit between atezolizumab monotherapy, atezolizumab-bevacizumab, and sorafenib[6,13].

Lenvatinib, a multikinase inhibitor targeting VEGFR1-3, fibroblast growth factor receptor 1-4, platelet-derived growth factor receptor-alpha, RET, and KIT, possesses both anti-angiogenic and immunomodulatory properties that may potentiate checkpoint blockade[10].

Sintilimab, a humanised immunoglobulin G4 anti-PD-1 antibody, has similarly shown clinically meaningful activity in multiple tumour types, and its combination with bevacizumab or lenvatinib has yielded improved outcomes in HBV-rich Asian HCC cohorts[7,8].

The phase II trial by Wang et al[1] evaluating sintilimab plus lenvatinib as conversion therapy for intermediate or locally advanced HCC reported an objective response rate of 36% and a disease-control rate exceeding 90%, with a notable proportion of patients successfully converted to resection or ablation. The subsequent integrative biomarker analysis in HBV-associated HCC, now in press, builds directly on this clinical foundation and seeks to explain why some tumours are amenable to durable control and conversion while others progress despite ostensibly similar treatment[15].

OVERVIEW OF THE BIOMARKER STUDY

In the cohort, Wang et al[1] analysed tumour biopsies from patients with HBV-associated HCC treated with sintilimab plus lenvatinib. All patients had unresectable disease at baseline but were considered potential candidates for conversion to curative therapy[1]. Treatment response was assessed with modified RECIST[13].

The authors undertook a multi-layered analysis: (1) Transcriptomics (RNA sequencing): Differentially expressed genes were identified between responders and non-responders. Among these differentially expressed genes, the long intergenic noncoding RNA LINC01554 and WHRN (whirlin) were significantly associated with longer PFS; (2) Immune microenvironment (xCell-based deconvolution): Responding tumours exhibited higher inferred abundance of B-cell subsets - including pro-B cells, class-switched memory B cells, and plasma cells - whereas the T-cell landscape was more complex, with CD4+ Tcm cells showing a borderline association with improved outcomes in univariable analysis; and (3) Genomic landscape (whole-exome sequencing): The mutational profile included recurrent TP53 and APC alterations typical of HCC. FANCD2 mutations were enriched among non-responders, and CUX1 mutations correlated with shorter PFS, although neither showed robust prognostic significance in external validation datasets.

When clinical, molecular, and immune variables were entered into multivariable Cox regression, three independent predictors of prolonged PFS emerged: (1) Solitary tumour; (2) High LINC01554 expression; and (3) Elevated CD4+ Tcm infiltration. Together, these variables define a multi-dimensional biomarker constellation that aligns with broader principles emerging from the immuno-oncology of HCC.

LINC01554 AND TUMOUR-INTRINSIC RESTRAINT: A MOLECULAR ANCHOR

Among the identified features, LINC01554 stands out as a particularly compelling molecular anchor for a more favourable tumour phenotype. Long noncoding RNAs have emerged as important regulators of hepatocarcinogenesis, influencing chromatin organisation, transcriptional networks, and key oncogenic pathways. Comprehensive bioinformatic and experimental work by Li et al[15] has shown that LINC01554 is downregulated in HCC compared with non-tumour liver, and that low expression correlates with larger tumour size, portal vein invasion, higher TNM stage, and poorer overall survival.

Functionally, LINC01554 overexpression inhibits proliferation, migration, and invasion in HCC cell lines, promotes G0/G1 arrest, and attenuates epithelial-mesenchymal transition (increased E-cadherin and ZO-1, decreased N-cadherin and vimentin). In the sintilimab-lenvatinib cohort, high baseline LINC01554 expression was enriched among responders and independently associated with longer PFS[15,20].

A coherent picture thus emerges: Maintenance of LINC01554 expression marks a subset of HBV-associated HCCs in which key oncogenic pathways remain partially restrained, and in which anti-angiogenic blockade plus PD-1 inhibition can more effectively tip the balance toward durable control. Conversely, LINC01554-low tumours likely represent a more plastic, pathway-redundant state capable of rapid adaptation and early escape. These observations do not yet justify clinical testing of LINC01554 as a stand-alone biomarker; however, they provide a compelling rationale for prospective validation and for deeper mechanistic dissection in preclinical models that combine lenvatinib or other VEGF-pathway inhibitors with ICIs[4,5,8,10].

CD4+ CENTRAL MEMORY T CELLS AND THE IMMUNE CONTEXTURE OF BENEFIT

The second pillar of the biomarker triad, CD4+ central memory T cells, highlights the importance of immune contexture in determining response to systemic therapy. Central memory T cells occupy an intermediate state between naïve and effector pools, endowed with proliferative capacity, rapid cytokine production, and the ability to differentiate into cytotoxic or helper subsets upon antigen re-encounter.

In the Wang et al’s analysis[1], higher inferred CD4+ Tcm abundance at baseline independently predicted longer PFS after adjustment for clinical and genomic features. This finding is consistent with broader immuno-oncology data in HCC showing that inflamed tumours - characterised by CD8+ and CD4+ effector and memory cells, activated dendritic cells, and lower regulatory T-cell predominance - tend to respond better to ICI-containing regimens[1,10,17].

Notably, the global immune pattern in the sintilimab-lenvatinib cohort also pointed to an enrichment of B-cell subsets in responders, including pro-B cells, class-switched memory B cells, and plasma cells. This echoes high-impact work in melanoma and other solid tumours demonstrating that tumour-associated B cells and tertiary lymphoid structures are closely linked to ICI responsiveness, with evidence of clonal B-cell expansion, germinal centre formation, and active humoral immunity within tertiary lymphoid structures[18,19].

Taken together, these findings support a model in which baseline immune ecosystem quality - defined by an intact CD4+ memory compartment, supportive B-cell architecture, and limited immunosuppressive skewing - creates permissive conditions for PD-1 blockade to generate durable anti-tumour immunity. In contrast, tumours with depleted memory pools, dominant regulatory or myeloid suppressor populations, or profound T-cell exhaustion are more likely to exhibit primary resistance[10,17-19].

SOLITARY TUMOUR MORPHOLOGY: CLINICAL ANATOMY AS A BIOMARKER LAYER

The third component of the triad - solitary tumour morphology - underscores the continued relevance of conventional clinical and radiological features in a precision-medicine framework. Multifocal tumours, particularly in HBV-related cirrhosis, may arise from multicentric carcinogenesis on a background of field cancerisation and chronic necroinflammation, yielding a more heterogeneous and potentially aggressive disease biology[2,3].

In the sintilimab-lenvatinib biomarker analysis, solitary tumours were significantly associated with longer PFS and remained an independent predictor in multivariable models alongside LINC01554 and CD4+ Tcm. This reinforces the principle that anatomical disease burden and distribution should not be eclipsed by molecular data, but rather integrated into a layered risk model. For conversion therapy in particular, tumour number and distribution have direct implications for resectability, ablation feasibility, and the likelihood of achieving durable locoregional control once systemic therapy has induced regression[4,5,8,14].

TOWARD AN INTEGRATED BIOMARKER FRAMEWORK

The distinctive strength of recent multi-omic analyses in HBV-associated HCC lies not in identifying a single determinant of response but in demonstrating the convergence of three orthogonal biological domains: (1) Tumour-intrinsic transcriptional restraint - characterised by elevated LINC01554 expression and partial suppression of phosphoinositide 3-kinase/protein kinase B and Wnt/β-catenin pathways[15,20]; (2) Supportive immune contexture - defined by increased CD4+ central-Tcm infiltration and enrichment of B-cell subsets forming tertiary lymphoid-like structures[10,17-19]; and (3) Favourable anatomical phenotype - represented by solitary-tumour morphology indicating a more tractable disease architecture[2,3,14].

These dimensions can be conceptualised as interacting axes that together determine the “biological temperature” of the tumour and its sensitivity to combined PD-1 and VEGF-pathway inhibition. When all three axes align favourably, the likelihood of durable response and conversion to resection or ablation increases; when one or more are unfavourable, the balance shifts toward resistance. Key features of each biomarker domain and their putative clinical implications are summarised in Table 1, and their inter-relationships are illustrated in Figure 1.

Figure 1 Tri-axial biomarker model for response to sintilimab plus lenvatinib in hepatitis B virus-associated hepatocellular carcinoma. This schematic illustrates how molecular (LINC01554 expression), immune (CD4+ central memory T cell infiltration), and anatomical (tumour morphology) axes interact to define therapeutic outcomes. Tumours characterised by high LINC01554, high CD4+ memory T-cell, and solitary morphology cluster within a favourable response zone, whereas loss of these features such as reduced LINC01554 expression, declining CD4+ memory T-cell, or increased tumour number progressively shifts the disease toward a resistance zone, reflecting transition from biological sensitivity to refractoriness.

Table 1 Key biomarker domains and their putative clinical implications in hepatitis B virus-associated hepatocellular carcinoma treated with sintilimab + lenvatinib.

Domain	Feature/marker	Direction associated with longer progression-free survival	Biologic/mechanistic implication	Putative clinical relevance
Molecular (long noncoding RNAs)	High LINC01554 expression	Favourable	Tumour-suppressive activity; restrains Wnt/β-catenin and phosphoinositide 3-kinase/protein kinase B pathways	Candidate biomarker for selecting patients likely to respond to PD-1/tyrosine kinase inhibitor combinations
Immune	High CD4+ central memory T-cell infiltration	Favourable	Preserved adaptive memory enabling sustained immune reactivation after PD-1 blockade	Supports durable anti-tumour immunity and prolonged disease control
Immune (B-cell axis)	Enrichment of B-cell subsets (pro-B, class-switched memory, plasma cells)	Observed in responders	Tertiary lymphoid structure-like microenvironment	May augment CD4+/CD8+ T-cell coordination and checkpoint efficacy
Genomic	CUX1 mutation	Unfavourable	Transcriptional dysregulation; possible link to genomic instability	Potential prognostic marker of poor outcome; needs independent validation
Genomic (DNA repair)	FANCD2 mutation	Enriched in non-responders	Defective DNA repair and replication stress	May signify intrinsic therapy resistance or need for alternative pathway targeting
Clinical/anatomical	Solitary tumour morphology	Favourable	Lower clonal heterogeneity and contained microenvironment	Higher likelihood of durable response and successful conversion to resection or ablation

PD-1: Programmed cell death protein 1.

FROM DISCOVERY TO PRACTICE: EMBEDDING MULTIOMIC DATA INTO CLINICAL DECISION-MAKING

A structured framework can be developed to integrate molecular and immune biomarkers into therapeutic decision-making for HBV-associated HCC. This approach demonstrates how multidimensional biomarker data can be layered onto established clinical pathways to refine decisions and personalise therapy[9,10].

POSITIONING THE TRIAD IN THE BROADER IMMUNO-ONCOLOGY LANDSCAPE

The findings by Wang et al[1] align with several broader themes in the immuno-oncology of HCC. Integrated analyses from atezolizumab-bevacizumab trials have emphasised that pre-existing T-cell inflamed phenotypes, high programmed death ligand 1 expression, and intratumoural CD8+ T-cell density correlate with improved outcomes on combination therapy, whereas an abundance of regulatory T cells, oncofetal signatures (e.g., glypican 3, alpha fetoprotein), and myeloid inflammation portend inferior benefit[5,10].

HBV-driven HCC often arises in a background of chronic antigen exposure and immune exhaustion, yet can harbour immunologically active niches that are potentially reprogrammable with ICIs and anti-angiogenic agents[2,3,9,10,17].

In this context[10,15], the LINC01554 - CD4+ Tcm - solitary tumour triad can be viewed as a prototype integrated signature for HBV-associated HCC treated with PD-1/TKI combinations[16-20]. Its components are conceptually consistent with three cross-cutting principles: Tumours with preserved tumour-suppressive circuitry and less chaotic signalling are more amenable to durable control.

A resilient and adaptable immune network - especially memory T cells supported by B-cell-rich structures - facilitates sustained responses to checkpoint blockade. Lower anatomical complexity and clonal dispersion enhance both systemic and locoregional treatment success.

HBV itself adds further complexity, given the influence of viral integration, chronic inflammation, and viral antigen expression on both tumour biology and immune surveillance. Any biomarker framework applied to HBV-associated HCC must therefore be interpreted through this aetiological lens.

IMPLICATIONS FOR FUTURE TRIAL DESIGN AND TRANSLATIONAL RESEARCH

The immediate implication of the sintilimab-lenvatinib biomarker study is not that clinicians should routinely order LINC01554 or Tcm assays today, but rather that future clinical trials evaluating PD-1/TKI or ICI/anti-VEGF combinations should be designed with integrated biomarker platforms from the outset[4,5,8,10].

LIMITATIONS

Several important caveats must temper interpretation of the biomarker triad: (1) Sample size and single-centre design: The study included 33 patients from a single institution, raising the risk of overfitting and limiting external validity; (2) Retrospective biomarker analysis: Although embedded within a prospective treatment trial, the biomarker work was retrospective, and thresholds were optimised internally rather than pre-specified; and (3) Limited external validation at the integrative level: While individual genes such as LINC01554 correlate with survival in independent datasets, the specific combination of high LINC01554, high CD4+ Tcm, and solitary tumour has not yet been tested prospectively or in other therapeutic contexts[15].

Therefore, the proposed framework should be regarded as hypothesis-generating, offering a conceptual scaffold for future research rather than a ready-to-apply clinical tool. Premature adoption of any one component as a stand-alone decision-maker would be inappropriate.

CONCLUSION

The study by Wang et al[1] illustrates how multi-omic integration can begin to move HBV-associated HCC from empiric treatment toward biologically guided therapy. High LINC01554 expression, increased CD4+ central Tcm infiltration, and solitary tumour morphology together delineate a subset of patients more likely to achieve durable disease control with sintilimab plus lenvatinib, and more likely to be converted to potentially curative interventions. For clinicians, these data support a stance of cautious optimism: PD-1/TKI combinations can deliver striking outcomes in selected patients, particularly when used with a clear intent to convert to resection or ablation. For translational investigators, they highlight the necessity of embedding robust molecular and immune profiling into future trials, and of viewing HCC not as a homogeneous entity but as a collection of biologically distinct subtypes shaped by viral aetiology, tumour genomics, immune architecture, and clinical anatomy.