Portal Vein Tumor Thrombus in Hepatocellular Carcinoma: An Immune Sanctuary Driving Immune Exclusion and Therapeutic Resistance

Abstract

Background

Portal vein tumor thrombus (PVTT) is the most common form of macrovascular invasion in hepatocellular carcinoma (HCC) and is associated with poor prognosis, rapid intrahepatic dissemination, and limited responsiveness to systemic therapy. Increasing evidence indicates that PVTT is not merely an intravascular extension of the primary tumor, but rather a biologically distinct and highly adaptive immunosuppressive niche that contributes to immune escape and therapeutic resistance.

Summary

This review summarizes current genomic, transcriptomic, single-cell, and spatially resolved evidence regarding the immune landscape of PVTT in HCC. We discuss how early clonal divergence, hypoxia-associated metabolic rewiring, and stromal-vascular remodeling cooperatively shape the PVTT microenvironment. PVTT is characterized by enrichment of suppressive myeloid populations, activation of cancer-associated fibroblast programs, extracellular matrix accumulation, endothelial barriers to immune-cell trafficking, and progressive dysfunction or exclusion of effector lymphocytes. Together, these features establish a strongly immune-excluded and immunosuppressive ecosystem, providing a mechanistic basis for the limited efficacy of immune checkpoint blockade in patients with HCC-PVTT. We further highlight the translational implications of this framework for biomarker development, patient stratification, and rational combination strategies targeting suppressive myeloid and stromal programs, vascular barriers, and metabolic-immune crosstalk.

Key Messages

PVTT should be regarded as a distinct immunosuppressive niche rather than a passive vascular extension of the primary tumor. Its immune landscape is shaped by coordinated interactions among tumor evolution, metabolic adaptation, stromal remodeling, and immune dysfunction. A better mechanistic understanding of PVTT may support biomarker-guided stratification and the development of more effective precision therapies for HCC with macrovascular invasion.

Plain Language Summary

Portal vein tumor thrombus (PVTT) happens when liver cancer grows into the portal vein, a major blood vessel of the liver. In hepatocellular carcinoma, PVTT is linked to faster disease spread, worse liver function, and poorer survival. Patients with PVTT also often benefit less from immunotherapy, which is a treatment that helps the immune system attack cancer. This review explains why PVTT may be harder to treat than the main liver tumor. We describe evidence showing that PVTT is not just an extension of the original tumor. Instead, it seems to behave like a separate cancer-supporting environment. In this environment, cancer cells can change their metabolism, build physical barriers around themselves, and interact with nearby supportive cells in ways that weaken the body’s immune response. As a result, immune cells may have trouble entering the tumor area or may become less effective even if they do reach it. We also summarize how these findings may help with future treatment. A better understanding of PVTT could support the development of better biomarkers, improve patient selection for therapy, and guide combination treatments that target not only cancer cells but also the surrounding immune-suppressive environment.

Introduction

Hepatocellular carcinoma (HCC) remains among the leading causes of cancer-related mortality worldwide [1, 2]. Among the hallmarks of advanced disease, portal vein tumor thrombus (PVTT) represents the most common form of macrovascular invasion and a major determinant of poor prognosis. PVTT may be present at initial diagnosis or arise during disease progression, and its occurrence is associated with rapid intrahepatic dissemination, deterioration of liver function, and markedly shortened survival [3, 4]. In the absence of effective treatment, the median survival is only approximately 2.7–4 months [5]. The prognostic impact of PVTT is highlighted by a large cohort of 1,317 patients with HCC, in which median overall survival was 35.7 months in patients without PVTT but only 7.2 months in those with PVTT. Survival decreased stepwise with greater PVTT extent, from 14.6 months in Vp1 to 4.8 months in Vp4 disease, underscoring PVTT as a clinically aggressive phenotype rather than a purely anatomical finding [6]. A further diagnostic caveat is the distinction between bland portal-vein thrombosis and viable PVTT. This distinction is particularly relevant in PET-negative cases because the absence of FDG uptake does not reliably exclude tumor thrombus, given the variable FDG avidity and necrotic components of HCC [7]. Therefore, PVTT diagnosis should not rely on PET/CT alone but should integrate contrast-enhanced CT or MRI features, including thrombus enhancement, venous expansion, neovascularity, and contiguity with the primary HCC lesion. The A-VENA criteria, which combine AFP elevation with venous expansion, thrombus enhancement, neovascularity, and adjacency to HCC, provide a practical noninvasive framework for distinguishing a tumor from bland portal-vein thrombus [8]. When imaging remains equivocal and treatment allocation would be altered, short-interval reassessment or tissue confirmation may be considered. Although immune checkpoint inhibitors (ICIs) have reshaped the systemic treatment landscape for advanced HCC, their clinical benefit in patients with PVTT remains substantially attenuated [5, 9], suggesting that this setting harbors additional mechanisms of immune escape and therapeutic resistance beyond those observed in the primary tumor. Conceptually, this phenomenon may be viewed as a localized manifestation of both primary and acquired resistance to cancer immunotherapy [9, 10].

Accumulating evidence indicates that PVTT should not be regarded merely as a passive intravascular extension or metastatic deposit of the primary lesion. In this review, we define PVTT as an “immune sanctuary” in the sense of an anatomically compartmentalized and functionally immune-restrictive intravascular niche, rather than a classical immune-privileged site. Unlike the primary tumor microenvironment, PVTT develops within the portal venous lumen and is shaped by hemodynamic stress, endothelial interaction, thrombo-inflammatory remodeling, hypoxia, and stromal-myeloid reprogramming. These PVTT-specific pressures converge to limit effector immune-cell trafficking and cytotoxic function, resulting in a more immune-excluded and immunosuppressive microenvironment than that of the matched primary tumor [11]. Emerging genomic, transcriptomic, single-cell, and spatially resolved studies suggest that PVTT is characterized by early clonal divergence, profound metabolic rewiring, enrichment of suppressive myeloid populations, stromal activation, extracellular matrix (ECM) remodeling, endothelial dysfunction, and progressive exhaustion or exclusion of effector lymphocytes. Together, these features establish a more strongly immunosuppressive and immune-excluded microenvironment than the matched primary tumor has, thus providing a plausible biological basis for the limited efficacy of immunotherapy in patients with HCC-PVTT [12, 13]. This conceptual framework is summarized in Figure 1, with the primary tumor shown in the left panel and the PVTT immune sanctuary shown in the right panel.

Fig. 1.

The PVTT as a distinct immunosuppressive niche. This diagram illustrates the evolution from the primary HCC tumor to the PVTT. The primary tumor (left) maintains a microenvironment where ICI therapy can have some effect, with effector T-cells present. Following clonal evolution and divergence, the PVTT (right) forms a unique biological entity within the portal vein. This niche is characterized by a profoundly immunosuppressive tumor microenvironment, densely populated by TAMs and CAFs, which leads to T-cell exhaustion and robust resistance to ICI therapy. HCC, hepatocellular carcinoma; PVTT, portal vein tumor thrombus; ICI, immune checkpoint inhibitor; TAM, tumor-associated macrophage; CAF, cancer-associated fibroblast.

Despite growing interest in PVTT biology, important gaps remain. Current evidence across clonal evolution, metabolic adaptation, stromal remodeling, and immune dysfunction is still fragmented and has not been integrated into a clinically meaningful framework. In addition, PVTT-specific findings are often difficult to distinguish from broader inferences based on general HCC immunobiology. In this review, we synthesize multi-omics, single-cell, and spatial evidence to show how clonal evolution, hypoxia-associated metabolic rewiring, and stromal-vascular remodeling shape the PVTT immune niche. We further discuss the implications of this framework for biomarker discovery, patient stratification, and rational therapeutic strategies in HCC-PVTT.

The Clinical Challenge of HCC-PVTT and Fundamentals of Cancer Immunotherapy

HCC constitutes a major global public health burden. Its diverse etiologies and pronounced molecular heterogeneity contribute to the fact that a large proportion of patients are diagnosed at an advanced stage [14, 15]. Hepatocarcinogenesis is tightly linked to chronic liver injury, with the principal risk factors including viral hepatitis, alcohol-related liver disease, and metabolic dysfunction-associated steatotic liver disease [1]. Owing to its clinically silent early course and the limitations of current surveillance strategies, many patients have already missed the window for curative treatment at the time of diagnosis [16]. In addition, complex molecular derangements together with the liver’s intrinsically tolerogenic immune milieu further fuel tumor progression, promote therapeutic resistance, and attenuate antitumor immune responses. The development of PVTT can abruptly alter the clinical trajectory of HCC. As the predominant manifestation of macrovascular invasion, PVTT typically signifies transition to advanced-stage disease and is closely associated with a dismal prognosis [17, 18], as illustrated in the clinical-progression module of Figure 2.

Fig. 2.

The pathological progression of HCC to advanced disease with PVTT. PVTT formation represents a critical stage transition, leading to intrahepatic spread and compromised liver function, which is associated with a dismal prognosis. HCC, hepatocellular carcinoma; PVTT, portal vein tumor thrombus.

PVTT is detected in approximately 10–60% of patients with HCC and represents a key adverse prognostic determinant indicative of a highly aggressive tumor phenotype. Its development not only facilitates intrahepatic dissemination but also exacerbates portal hypertension and precipitates hepatic decompensation, thereby driving disease progression to Barcelona Clinic Liver Cancer (BCLC) stage C. In the absence of effective treatment, the median survival of this patient population is only ∼2.7–4 months [1, 19, 20].

Although cancer immunotherapy – particularly ICIs – has reshaped the therapeutic landscape of multiple solid malignancies by reinvigorating antitumor immunity through blockade of the PD-1/PD-L1 axis [21, 22], clinical benefit remains markedly constrained in patients with HCC complicated by portal vein tumor thrombus (HCC-PVTT). This limited efficacy points to deep-seated, spatially localized mechanisms of immunosuppression and therapeutic resistance within the PVTT niche.

Clinical data support the concept of relative, rather than absolute, therapeutic resistance in PVTT. In the exploratory IMbrave150 Vp4 analysis, patients with main-trunk and/or contralateral portal-vein invasion had shorter outcomes than those without Vp4 disease. Among patients receiving atezolizumab plus bevacizumab, the confirmed objective response rate was 23% in the Vp4 subgroup compared with 31% in patients without Vp4 involvement; median OS and PFS were also shorter in the Vp4 subgroup, although benefit over sorafenib was preserved [23]. Real-world PD-1 inhibitor data provide a similar nuance: patients with macrovascular tumor thrombi had a RECIST ORR comparable to those without tumor thrombi (17.6% vs. 11.8%), but their median OS was shorter (8.9 vs. 12.1 months), and vascular responses of the thrombus itself were observed [24]. Thus, PVTT should be considered a high-risk and relatively ICI-resistant state, rather than a uniformly ICI-refractory lesion.

The PVTT as a Distinct Biological Niche: Genetic, Metabolic, and Immunological Divergence

Biological Initiation, Clonal Evolution, and Genetic Heterogeneity

Although clinically overt PVTT is diagnosed as macrovascular invasion, its biological initiation likely begins long before a radiologically visible portal-vein filling defect emerges. Rather than representing a simple mechanical extension of the primary tumor into the vascular lumen, PVTT formation appears to reflect a multistep evolutionary process in which aggressive tumor clones acquire vascular-invasive competence within the primary lesion and are subsequently selected within the portal venous milieu. In HCC, recurrent trunk-level alterations, including TERT promoter activation and disruption of the TP53 pathway, may establish a permissive clonal background for invasive evolution [25, 26]. However, current evidence does not support any single trunk mutation as a PVTT-specific determinant. Instead, PVTT development is more plausibly driven by the convergence of early clonal selection, epithelial-mesenchymal transition (EMT) or partial-EMT plasticity, stem-like programs, hypoxia-induced angiogenesis, extracellular-matrix remodeling, endothelial interaction, and coagulation-related adaptation [27, 28]. These coordinated programs enable tumor cells to detach from the primary lesion, invade endothelial-lined vascular spaces, withstand intravascular shear stress, interact with the portal-vein endothelium, and ultimately establish a macroscopic intravascular tumor niche.

Microvascular invasion (MVI) provides an important pathological window into this vascular-invasive trajectory, but its relationship with PVTT should be interpreted with caution. MVI reflects microscopic tumor nests within endothelial-lined vascular channels, whereas PVTT represents established tumor colonization of the portal venous lumen. Thus, MVI and PVTT may be viewed as overlapping manifestations within the vascular invasion spectrum of HCC, rather than as strictly sequential or mutually exclusive entities [28, 29]. Surgical cohorts of HCC with PVTT have shown that concomitant MVI is independently associated with inferior overall and recurrence-free survival, supporting its value as a marker of aggressive vascular biology and as a potential component of PVTT risk stratification [29]. Nevertheless, the observation that a subset of PVTT patients lacks detectable MVI suggests that MVI is not an obligatory precursor of PVTT, nor do all PVTT lesions necessarily pass through a pathologically captured MVI stage [29]. Progression from a vascular-invasive phenotype to macroscopic PVTT likely requires additional portal-vein-specific adaptations, including hemodynamic selection, endothelial reprogramming, thrombo-inflammatory remodeling, immune evasion, and stromal support.

From an evolutionary perspective, paired analyses of primary HCC and PVTT further suggest that PVTT usually shares a common ancestral clone with the primary tumor but may diverge early and undergo branched evolution under the selective pressures of the portal venous environment [30]. This model reconciles two clinically relevant observations: PVTT retains molecular continuity with the primary tumor, yet often displays distinct transcriptomic, metabolic, immune, and therapeutic-resistance features. Therefore, PVTT should be understood not simply as the terminal morphological consequence of vascular invasion, but as a biologically selected and locally adapted intravascular ecosystem. Defining the molecular events that connect trunk genomic alterations, EMT-associated plasticity, MVI, and portal-vein niche adaptation may provide a rational framework for earlier prediction and mechanism-guided intervention in HCC patients at high risk of PVTT.

PVTT should be regarded as a tumor entity with partially autonomous biological features. Its pronounced therapeutic refractoriness appears to arise from evolutionary trajectories that diverge from those of the primary lesion, thereby generating a distinct genetic and transcriptomic landscape. Notably, these characteristics are more consistent with relatively independent clonal evolution following early metastatic seeding, rather than being a mere consequence of late-stage spatial segregation during disease progression. The resulting intratumoral heterogeneity provides a critical substrate for the highly invasive phenotype and immune-evasive capacity of PVTT [31].

Evidence for Early Seeding and Divergent Clonal Evolution

High-resolution genomic profiling indicates that, while PVTT and the matched primary HCC share a common clonal origin, they often diverge early during tumorigenesis and subsequently follow a branched evolutionary course characterized by the ongoing acquisition of PVTT-private alterations [30, 31]. Phylogenetic reconstructions further support that both the primary lesion and PVTT arise from the same ancestral trunk, yet undergo early bifurcation rather than representing a terminal, late-stage outgrowth. Collectively, these findings suggest that PVTT formation frequently reflects an early metastatic dissemination event within the continuum of HCC progression, conforming to a branched evolution model that is tightly linked to distinct therapeutic challenges [30].

Early dissemination may contribute to PVTT heterogeneity through founder selection followed by lesion-specific evolution [30]. Only selected subclones with vascular-invasive, EMT-like, stem-like, or immune-evasive traits may successfully seed the portal venous compartment. Once established, these founder populations are exposed to niche-specific pressures, including shear stress, hypoxia, coagulation, endothelial interaction, and intravascular immune surveillance, which can drive PVTT-private alterations and divergent transcriptional and stromal-immune states [30, 32]. Thus, PVTT heterogeneity should not be viewed merely as spatial sampling variation from the primary tumor, but as ecological divergence after early dissemination. This model explains how PVTT can retain clonal continuity with the primary lesion while acquiring partially autonomous biological features associated with immune exclusion and therapeutic resistance. Early metastatic dissemination can generate substantial intratumoral heterogeneity, a hallmark of aggressive cancers that is closely linked to therapeutic resistance and unfavorable outcomes [31, 33]. Circulating tumor cells (CTCs) are widely considered a key biological conduit for this process. Shed from the primary tumor, CTCs adapt to the hemodynamic and immunologic stresses of the circulation and contribute to systemic dissemination as well as macrovascular invasion, including the development of PVTT [33, 34]. Marked heterogeneity of CTCs along the EMT continuum and in stem-like traits may enable selected subclones to depart the primary lesion and seed a founder population within the portal venous system. Consistent with this model, clinical studies have shown that preoperative CTC counts are significantly higher in HCC patients with PVTT than in those without PVTT, and that elevated preoperative CTC levels are associated with inferior disease-free and overall survival in patients with HCC-PVTT [35]. More importantly, mesenchymal CTC positivity, rather than total CTC burden alone, has been associated with MVI, PVTT, high AFP levels, advanced BCLC stage, and poor overall survival, suggesting that EMT-shifted CTC subtypes may serve as potential pre-radiological risk biomarkers for PVTT development [36]. Nevertheless, whether mesenchymal CTCs can reliably predict radiologically occult PVTT onset remains to be determined in prospective longitudinal cohorts. However, the low abundance of CTCs and their profound heterogeneity pose major technical barriers to robust detection and mechanistic dissection [34]. Therefore, CTCs should be interpreted as probabilistic risk biomarkers rather than deterministic prerequisites for PVTT formation. CTC positivity may indicate an increased probability of vascular invasion, but CTC negativity does not exclude existing or future PVTT. In the propensity-score-matched cohort mentioned above, although the proportion of patients with elevated CTC counts was markedly higher in the PVTT group than in the non-PVTT group, a substantial fraction of PVTT patients remained below the predefined CTC-positive threshold [35]. This limitation may reflect both biological heterogeneity and technical constraints, particularly because epithelial-marker-dependent platforms may miss EMT-shifted or EpCAM-low HCC cells. Future studies should evaluate mesenchymal CTCs longitudinally before radiological PVTT onset and integrate them with imaging, AFP, DCP/PIVKA-II, radiomics, and spatial immune biomarkers.

Driver Alterations as a Bridge between Clonal Evolution and TIME Remodeling

Driver alterations may help translate branched evolution into a PVTT-specific immunosuppressive phenotype. TERT promoter activation, an early and frequent event in hepatocarcinogenesis, may provide a permissive trunk background for clonal persistence, diversification, and immune editing [37]. TP53 pathway disruption is more directly linked to immune remodeling, being associated with attenuated immune-response programs, enrichment of regulatory T cells (Tregs) and macrophage populations, and increased expression of inhibitory checkpoints such as CTLA-4, PD-1, and TIM-3 [38]. In parallel, CTNNB1/Wnt-β-catenin activation can promote immune exclusion by reducing chemokine expression and limiting immune-cell infiltration, a phenotype associated with primary resistance to anti-PD-1 therapy [39]. Under portal-vein-specific selection pressures, including hypoxia, endothelial interaction, thrombo-inflammation, and stromal remodeling, these driver-associated immune states may be amplified into the tumor-associated macrophage (TAM)-rich, T-cell-excluded, and checkpoint-reinforced niche characteristic of PVTT.

Transcriptomic Reprogramming Drives the Formation of an Immunosuppressive PVTT Niche

PVTT appears to acquire a stable and PVTT-specific transcriptomic program through relatively independent clonal evolution. Compared with the matched primary tumor, PVTT shows marked transcriptomic divergence with enrichment of pathways involved in cell-cycle progression, invasion, and metabolic reprogramming [40]. This divergence is characterized by the aberrant upregulation of oncogenic drivers such as KIF11, PRR11, and RACGAP1, accompanied by disruption of coding RNA-non-coding RNA regulatory circuitry (e.g., miR-3607), which is ultimately mirrored at the proteomic level [41]. The resulting genomic and transcriptomic heterogeneity not only underpins the highly aggressive phenotype of PVTT but also directly contributes to the emergence of an immunosuppressive tumor microenvironment. Multi-omics and single-cell studies further suggest that specific clonal genotypes are tightly coupled to an immune-excluded state, typified by enrichment of M2-like TAMs alongside a pronounced paucity of cytotoxic CD8⁺ T cells, thereby creating an “immune sanctuary” that is highly refractory to T cell-mediated killing and systemic immunotherapy [4, 30, 32].

Metabolic Reprogramming and Its Immunosuppressive Consequences

Beyond genetic divergence, PVTT acquires distinctive biological properties through profound metabolic reprogramming, which rewires cellular energetics to sustain rapid proliferation while actively fostering an immunosuppressive tumor microenvironment [42]. Importantly, this metabolic remodeling is unlikely to represent a merely passive adaptation to intravascular nutrient constraints; rather, it constitutes a key, tumor-intrinsic mechanism that promotes the highly invasive phenotype of PVTT and underpins therapeutic refractoriness.

PVTT Metabolic Activity as an Independent Prognostic Marker

Clinical imaging studies indicate that the metabolic activity of PVTT can be quantitatively assessed using modalities such as ^18 F-FDG PET/CT, and that the intensity of glucose uptake serves as an independent prognostic factor for survival. These observations support the notion that PVTT is not merely a passive extension of the primary lesion but rather a highly aggressive focus with partially autonomous biological behavior [43]. In particular, the PVTT-to-liver uptake ratio, derived from FDG-PET, robustly predicts inferior pro-gression-free survival and overall survival in patients with advanced HCC, providing clinical evidence for the hypermetabolic and invasive nature of PVTT [43, 44]. Complementary quantitative imaging metrics – such as hepatobiliary phase-enhanced MRI – further refine functional characterization across distinct lesions, underscoring the need to elucidate the molecular determinants of enhanced glycolysis in PVTT and their mechanistic links to aggressive clinical behavior [45].

ATP Citrate Lyase-Driven Lipogenesis as a Core Metabolic Engine

At the molecular level, the hypermetabolic phenotype of PVTT has been linked to upregulation of ATP citrate lyase (ACLY), which couples enhanced glycolysis to de novo lipogenesis and supplies the structural lipid precursors required for rapid tumor growth. This ACLY-centered lipid metabolic rewiring redirects glucose-derived carbon flux from ATP production toward biomass synthesis, thereby conferring a proliferative advantage and supporting the invasive phenotype of PVTT [42, 46]. Moreover, under stress conditions such as hypoxia, PVTT may further enhance intravascular fitness by modulating the expression of fatty-acid up-take-associated molecules (e.g., SLC27A2), reducing lipid peroxidation and promoting cellular survival, which reinforces adaptive advantages within the vascular microenvironment [47].

Lipid Accumulation and Hypoxia as Drivers of Immune Evasion

ACLY-driven lipogenesis frequently coexists with a locally hypoxic milieu. Together, these factors promote lipid accumulation that directly suppresses cytotoxic T-cell effector function and skews myeloid cells toward pro-tumorigenic polarization, thereby actively shaping an immunosuppressive microenvironment. Elevated ACLY expression and the resulting lipid-enriched niche are closely associated with immune checkpoint upregulation, while lipid overload can disrupt lipid-raft organization within immune-cell membranes, weakening T-cell receptor signal transduction [48]. These effects are further amplified under hypoxia: stabilization of hypoxia-inducible factor-1α (HIF-1α) can enhance immunosuppression by recruiting Treg cells, promoting the accumulation of metabolites such as lactate, and driving the release of extracellular vesicles carrying immunosuppressive microRNAs. Beyond lipid accumulation, hypoxia-driven glycolysis may impose additional metabolic checkpoints that directly restrain T-cell immunity in PVTT [49]. In the compact intravascular PVTT niche, lactate accumulation and extracellular acidosis can impair effector T-cell activation, cytokine production, and cytotoxicity, while favoring immunosuppressive Treg and macrophage programs [49, 50]. In parallel, suppressive myeloid cells may deplete arginine through ARG1/iNOS-related pathways, whereas IDO/TDO-mediated tryptophan catabolism can generate kynurenine metabolites, thereby limiting T-cell proliferation and promoting anergy or Treg differentiation [51]. These lactate- and amino-acid-dependent constraints may cooperate with immune-checkpoint signaling to sustain T-cell dysfunction and immune exclusion in PVTT. Ultimately, this coordinated program fosters an “immune sanctuary” dominated by M2-like TAMs and metabolically reprogrammed cancer-associated fibroblasts (CAFs) [52].

Comparative Immunophenotyping of the PVTT vs. Primary Tumor

Compared with primary HCC, PVTT constitutes a partially autonomous and more tightly organized immunosuppressive niche. It is characterized by marked enrichment of suppressive myeloid populations, profound lymphocyte exhaustion, and an activated stromal scaffold that together establish multilayered barriers to antitumor immunity. Single-cell and spatial transcriptomic studies further reveal that the intravascular milieu of PVTT sustains a distinctive “cellular neighborhood” centered on immune evasion, with a substantially stronger immunosuppressive tone than that observed in the primary tumor [53]. This immunoarchitectural divergence provides a biological rationale for the poor responsiveness of PVTT to immunotherapy.

Enrichment of Suppressive Myeloid and Exhausted Lymphoid Infiltrates

The immune microenvironment of PVTT is markedly skewed toward a pro-tumor, “non-inflamed” state, characterized by a pronounced accumulation of immunosuppressive myeloid cells – most notably TAMs – at densities exceeding those in paired primary tumors [54]. Among these, C5aR⁺ TAMs represent a key subset that is highly enriched in PVTT, exhibiting a predominantly M2-like polarization and forming spatially organized immunosuppressive clusters. Through coordinated inhibitory receptor-ligand interactions and cytokine secretion, these clusters dampen cytotoxic T-cell activity [11]. In parallel, chemokine axes such as CCL20–CCR6 further promote the recruitment and retention of TAMs and Tregs, thereby reinforcing a multilayered immunosuppressive cellular network [55, 56].

In parallel with heightened myeloid immunosuppression, the lymphoid compartment within PVTT is characterized by a marked numerical decline and functional impairment across multiple subsets, including T cells and natural killer cells. Cytotoxic CD8⁺ T cells in PVTT are frequently physically excluded and/or driven into a state of profound dysfunction, typified by high expression of multiple co-inhibitory receptors, resulting in substantially compromised antitumor activity [57]. The paucity of effector lymphocytes creates a local “immune void” that favors sustained tumor outgrowth and promotes refractoriness to immunotherapy, as illustrated in the PVTT immune-exclusion region of Figure 3.

Fig. 3.

Comparative immunophenotyping reveals the PVTT as a distinct immunosuppressive niche. Compared to the primary tumor, the PVTT is characterized by an enrichment of suppressive myeloid cells (M2 TAMs, Tregs), activated stroma (CAFs), and a depletion of functional cytotoxic CD8+ T-cells, often lacking organized TLSs. PVTT, portal vein tumor thrombus; TAM, tumor-associated macrophage; Treg, regulatory T cell; CAF, cancer-associated fibroblast; TLS, tertiary lymphoid structure.

Beyond local immunosuppression, the PVTT microenvironment may also play a pivotal role in promoting systemic disease progression. PVTT has been proposed to function as a hub for the formation of CTC-white blood cell clusters and to foster the emergence of specialized populations such as erythroid-differentiated myeloid cells, thereby compromising vascular integrity and facilitating metastatic dissemination [58]. Notably, CTC-white blood cell aggregates are closely associated with enhanced metastatic competence and poor clinical outcomes, further implicating PVTT as a critical nexus linking local immune evasion to systemic spread.

Aberrant Stromal Scaffolding and Disorganized Lymphoid Structures

The intravascular growth pattern of PVTT fosters a distinctive stromal architecture characterized by activated CAFs and functionally aberrant vasculature, which physically excludes effector immune cells and promotes an immune-excluded phenotype [59]. Spatial transcriptomic analyses suggest that ligand-receptor programs such as SPP1-CD44 and MDK-SDC2 facilitate CAF-driven assembly of a dense, fibrotic matrix, strengthening tumor-stroma coupling while impeding cytotoxic T-cell infiltration [40, 53, 56]. This structural barrier, compounded by inadequate vascular perfusion, can markedly limit the delivery and efficacy of systemic immunotherapy and represents a key challenge that combination regimens incorporating anti-angiogenic strategies must address.

In contrast to primary HCC, where mature tertiary lymphoid structures (TLS) are frequently observed and are associated with favorable antitumor immune responses, TLS in PVTT are often absent or remain developmentally immature. Even when lymphoid aggregates are present, they typically lack high endothelial venules and well-demarcated B- and T-cell zones, limiting their capacity to initiate and amplify local immunity [60]. The persistent absence or architectural disarray of TLS further reinforces PVTT as an immunologically “cold” niche.

Intensified Molecular Signatures of Immune Exhaustion and Suppression

Multiple immune checkpoint molecules and suppressive pathways are markedly upregulated within the PVTT microenvironment, extending well beyond the PD-1/PD-L1 axis and converging on a state of profound T-cell exhaustion that confers intrinsic refractoriness to single-agent checkpoint blockade. Transcriptomic profiling and immunohistochemical analyses consistently demonstrate elevated expression of TIM-3, LAG-3, VISTA, and the exhaustion-associated transcription factor TOX in PVTT-infiltrating lymphocytes, collectively constituting a robust immunosuppressive network [61]. This molecular signature provides a mechanistic explanation for the limited capacity of PD-1/PD-L1 monotherapy to reinvigorate T-cell function and offers a rationale for multi-target, combinatorial immunotherapeutic strategies.

In addition, a hostile metabolic milieu – shaped by hypoxia and the accumulation of immunosuppressive metabolites – further entrenches this state of profound T-cell exhaustion. The highly compact cellular architecture of PVTT predisposes to pronounced hypoxia, which not only promotes invasive tumor progression but also directly compromises the effector functions of T cells and natural killer cells [62]. Under hypoxic conditions, aberrant activation of the adenosine pathway suppresses cytotoxic immunity via extracellular adenosine signaling through A2A/A2B receptors. In parallel, tumor-derived extracellular vesicles carrying immunosuppressive microRNAs can further drive M2-like macrophage polarization [63]. These layered metabolic constraints act in concert with immune checkpoint signaling to “lock in” PVTT as a prototypical exhaustion-dominant microenvironment, providing a mechanistic basis for its stubborn therapeutic refractoriness and its reliance on integrated, multi-modality treatment strategies.

Architects of Immunosuppression: Cellular Drivers and Molecular Pathways in PVTT

TAMs: Central Orchestrators of the Immunosuppressive Microenvironment

TAMs constitute a central pillar of the immunosuppressive microenvironment in HCC and are highly enriched in PVTT. Acting as pivotal orchestrators of local immune regulation, TAMs promote tumor progression and therapeutic refractoriness [64]. Notably, TAMs display substantial heterogeneity and plasticity, enabling them to sense and integrate complex cues derived from both tumor cells and the stromal compartment, thereby remodeling the PVTT niche and fostering angiogenesis and immune evasion [65, 66]. Accordingly, a systematic dissection of TAM origins, polarization states, and the molecular programs governing their functional remodeling is essential for defining the immunosuppressive nature of PVTT and for developing rational, macro-phage-targeted therapeutic strategies.

Heterogeneity and Polarization: Diverse TAM Identities in PVTT

The immunosuppressive activity of TAMs in PVTT is rooted in their pronounced cellular heterogeneity. TAMs may originate from liver-resident macrophages (e.g., Kupffer cells) and/or from circulating monocytes recruited via chemotactic axes such as CCL2–CCR2 and CSF1–CSF1R. Upon “education” by the tumor microenvironment, they undergo polarization along a continuous activation spectrum and, in PVTT, are overall skewed toward a pro-tumorigenic, M2-like state [64]. This polarization is actively driven by tumor-derived cues. For example, CCL20 released by hypoxic HCC cells and signaling through the C5a–C5aR axis can promote the recruitment and functional reinforcement of immunosuppressive TAMs, thereby directly impairing CD8⁺ T-cell effector activity and correlating with adverse clinical outcomes [11, 27].

With advances in single-cell and spatial transcriptomics, the field has moved beyond the oversimplified M1/M2 paradigm and uncovered a finer functional stratification of TAMs within PVTT. Distinct sub-sets – such as tumor pro-gression-associated SPP1⁺ macrophages, perivascular TAMs, and lipid-associated macrophages – are enriched within specific spatial compartments and cooperate to promote angiogenesis, immune exclusion, and the maintenance of an immunosuppressive niche [53, 54, 67]. This highly complex and plastic landscape underscores a major challenge for TAM-targeted therapies: effective intervention must contend with the coexistence of multiple functionally divergent yet convergent macrophage programs that collectively drive immune escape.

The TAM Receptor Axis: A Direct Molecular Switch for Immunosuppression

The TAM receptor tyrosine kinase family (Tyro3, Axl, and MerTK) constitutes a pivotal immunoregulatory axis that directly promotes immune escape and resistance to ICIs by restraining effector T-cell activity and shaping an immune-excluded tumor microenvironment [55]. In myeloid cells, this pathway is engaged by the ligands Gas6 and Protein S, triggering an anti-inflammatory program characterized by induction of immunosuppressive mediators such as IL-10 and TGF-β while dampening pro-inflammatory signaling. Collectively, these effects attenuate T-cell priming and activation and favor the establishment of an immunologically “cold” state in PVTT [68].

Sustained activation of TAM receptor signaling provides a non-redundant immunosuppressive program that can drive both primary and acquired resistance to ICIs; notably, it can maintain T-cell dysfunction even after PD-1/PD-L1 blockade [55]. Moreover, TAMs can further amplify immunosuppression through convergent mechanisms such as purinergic signaling. For example, tumor-derived exosomal circTMEM181 has been reported to upregulate CD39 expression, thereby enhancing adenosine-mediated suppression and promoting resistance to anti-PD-1 therapy [63]. Collectively, these findings position the TAM receptor axis and its cooperative immunometabolic circuits as actionable vulnerabilities for overcoming immunotherapy refractoriness in PVTT.

Metabolic Drivers of Pro-Tumorigenic TAM Functions

The hypoxic and nutrient-deprived milieu within PVTT induces a lipid metabolism-centered metabolic rewiring of TAMs. Rather than a purely passive adaptation, this reprogramming actively drives M2-like polarization and reinforces macro-phage-mediated immunosuppression. Mechanistically, TAMs shift from an M1-associated, glycolysis-dominant state toward a metabolic program favoring fatty-acid oxidation (FAO) and oxidative phosphorylation (OXPHOS) – a hallmark of M2 polarization that has been closely linked to immunotherapy resistance in HCC [46].

Tumor cells can actively orchestrate metabolic rewiring in TAMs. For instance, under hypoxic stress, downregulation of the fatty-acid transporter SLC27A2 in HCC cells can increase the availability of long-chain fatty acids for macrophage uptake, thereby activating a pro-tumorigenic PPARγ-dependent transcriptional program in TAMs [47]. The pronounced hypoxia within PVTT further stabilizes hypoxia-inducible factor-1α (HIF-1α), which cooperatively reinforces protumoral metabolic states and functional phenotypes in TAMs [27].

At the molecular level, key metabolic enzymes such as ACLY couple glucose utilization to lipid biosynthesis, thereby establishing a lipid-enriched, immunosuppressive niche that favors M2-like polarization while compromising T-cell function, and is associated with adverse clinical outcomes [20]. Accordingly, therapeutically targeting TAM dependence on fatty-acid oxidation/OXPHOS-centered programs may rewire macrophage states, shifting them from pro-tumorigenic toward antitumor phenotypes and thereby restoring effective antitumor immunity [42].

TAM-Driven Structural and Vesicle-Mediated Orchestration of the PVTT Immunosuppressive Niche

Beyond directly suppressing T-cell responses, TAMs act as central coordinators of the PVTT niche through bidirectional crosstalk with tumor cells, fibroblasts, and the vascular compartment, thereby promoting angiogenesis and metastatic dissemination while reinforcing immunosuppressive barriers [69], as shown in the TAM-mediated crosstalk and extracellular-vesicle signaling regions of Figure 4. For example, TREM1⁺ TAMs can induce EMT in HCC cells via secretion of CCL7, whereas perivascular TAMs support angiogenesis and tumor recurrence through VEGF-A and engage in intimate structural and metabolic interactions with vascular cells [70, 71]. In addition, the integrity of lipid-raft microdomains in the TAM plasma membrane is closely linked to M2 polarization and the recruitment of Treg cells, providing a direct molecular bridge between metabolic state and immunoregulatory function [48].

Fig. 4.

TAM function as central hubs in the PVTT. TAM orchestrate immunosuppression in PVTT through direct T-cell inhibition, EMT-related tumor crosstalk, C–C motif chemokine ligand 7 (CCL7)-mediated signaling, vascular endothelial growth factor A (VEGF-A)-associated angiogenic signaling, FAO/oxidative phosphorylation (OXPHOS)-linked metabolic reprogramming, and complex intercellular communication via secreted factors, EVs, interleukin-10 (IL-10), and transforming growth factor-β (TGF-β). PVTT, portal vein tumor thrombus; TAM, tumor-associated macrophage; EV, extracellular vesicle; EMT, epithelial-mesenchymal transition; CCL7, C–C motif chemokine ligand 7; VEGF-A, vascular endothelial growth factor A; FAO, fatty acid oxidation; OXPHOS, oxidative phosphorylation; IL-10, interleukin-10; TGF-β, transforming growth factor-β.

TAMs also mediate efficient intercellular communication within the tumor microenvironment through the release of extracellular vesicles (EVs), actively propagating immunosuppressive cues to neighboring cellular compartments. Under hypoxic stress, TAMs can secrete exosomes enriched in specific microRNAs (e.g., miR-155-5p and miR-21), which have been implicated in promoting tumor progression, activating STAT3 signaling, and inducing CAF activation, thereby enhancing angiogenesis and therapeutic resistance [72]. This bidirectional EV-based circuitry is further shaped by tumor genetic heterogeneity, enabling TAMs to coordinate immunosuppressive programs beyond their immediate vicinity and to consolidate a resilient immune-suppressive ecosystem – reinforcing their role as key architects of immune escape in PVTT.

CAFs and Stromal Reinforcement of Immune Exclusion

CAFs, together with TAMs, form a core structural and functional backbone of the immunosuppressive defense system in PVTT, promoting immune escape and therapeutic resistance by constructing a dense stromal barrier [73]. Under dynamic recruitment and reprogramming cues from tumor cells, CAFs actively remodel the tumor microenvironment and assemble a defensive network that exerts both structural constraints and biochemical suppression. Accordingly, dissecting the mechanisms governing CAF activation and their functional heterogeneity is critical for understanding the immune-excluded phenotype of PVTT and for identifying rational, stroma-directed therapeutic strategies.

Tumor-Derived Signals Driving CAF Recruitment and Activation

HCC cells initiate stromal remodeling through the secretion of paracrine factors. For example, epidermal growth factor-like domain 7 (EGFL7) has been reported to actively recruit quiescent fibroblasts and, via defined signaling cascades, drive their conversion into pro-metastatic CAFs. Thus, CAF emergence is best viewed as a direct consequence of coordinated molecular crosstalk between tumor cells and microenvironmental components, rather than a spontaneous process [74, 75].

Mechanistically, tumor-derived EGFL7 binds αvβ3 integrin on quiescent fibroblasts – primarily originating from hepatic stellate cells – thereby activating the FAK/AKT signaling cascade and promoting their transdifferentiation toward an activated CAF phenotype [74]. In addition, tumor-fibroblast feed-forward loops, including SPP1–CD44 and CCN2/TGF-β circuits, further reinforce the stabilization of a pro-metastatic niche [40].

The Architectural and Biochemical Barrier Created by CAFs

Activated CAFs reinforce immune exclusion through two mechanistically distinct but synergistic processes: ECM-mediated physical restriction and CAF-derived biochemical inhibition. ECM accumulation generates a dense physical scaffold that restricts T-cell infiltration, whereas CAF-derived cytokines, chemokines, and extracellular vesicles functionally suppress antitumor immunity, as shown in the physical-barrier and biochemical-suppression modules of Figure 5.

Fig. 5.

CAFs create a formidable defensive barrier through a dual mechanism. This barrier consists of the deposition of a dense ECM that physically excludes T-cells, and the secretion of a suppressive biochemical milieu containing cytokines and extracellular vesicles. CAF, cancer-associated fibroblast; ECM, extracellular matrix; EV, extracellular vesicle; TGF-β, transforming growth factor-β; CXCL12, C–X-C motif chemokine ligand 12.

By synthesizing and remodeling the ECM, CAFs increase stromal density and stiffness, thereby fostering an immune-excluded phenotype. This process can be regulated by CAF-derived factors such as decorin and is further amplified through TAM–CAF crosstalk and signaling via the TAM receptor axis [76, 77].

CAFs further enforce immunosuppression by secreting mediators such as TGF-β and CXCL12, which directly impair T-cell effector functions and promote the recruitment of Tregs; these effects are markedly potentiated under hypoxia via stabilization of hypoxia-inducible factor-1α (HIF-1α) [78]. In addition, CAF-derived extracellular vesicles carrying cargos such as miR-21 can promote tumor drug resistance and modulate immune-cell behavior, thereby extending immunosuppressive signaling across the microenvironment [79].

Importantly, ECM accumulation and CAF activation may reinforce each other in a feed-forward manner: matrix stiffening and hypoxia can sustain CAF activation, whereas activated CAFs further promote ECM deposition and release immunosuppressive mediators, thereby consolidating a stromal barrier that is both physically restrictive and biochemically suppressive.

Functional Heterogeneity: The Challenge of Targeting a Diverse Stromal Landscape

CAFs do not constitute a single uniform population; rather, they comprise multiple phenotypically and functionally distinct subsets whose roles in tumor progression are highly context dependent [80]. Single-cell studies have broadly categorized CAFs into several major states, including myofibroblastic CAFs (myCAFs) that generate dense ECM, inflammatory CAFs that secrete a wide range of cytokines, antigen-presenting CAFs with antigen pre-sentation-related features, and vascular CAFs (vCAFs) that modulate angiogenesis. These subsets can exert specialized – and at times even opposing – functions across different tumor settings [52, 80, 81].

The pronounced heterogeneity of CAFs, the absence of universally specific markers, and their intricate crosstalk with other microenvironmental compartments collectively pose substantial challenges for stroma-targeted therapies. Intervening against a single CAF subset may be insufficient – and in some contexts even detrimental – which likely contributes to the limited clinical efficacy observed with several stromal-targeting approaches. These considerations underscore the need for precision strategies that selectively target the pro-tumorigenic CAF programs driving immune evasion within PVTT.

Functional Impairment of Effector T Cells and Other Suppressive Lineages

T-cell dysfunction in PVTT is not attributable to a single pathway; rather, it reflects a composite failure driven by T cell-intrinsic exhaustion in concert with multiple extrinsic suppressive forces. Although effector T cells are the principal executors of antitumor immunity, the PVTT microenvironment imposes persistent, multi-layered inhibitory programs that continuously erode T-cell priming, activation, and effector function. This suppressive state is further amplified through coordinated actions of pro-tumor immune populations – most notably Tregs and myeloid-derived suppressor cells (MDSCs) – which can be recruited and expanded in response to tumor-derived cues, including extracellular vesicle-mediated signaling [65, 82].

Phenotypic and Functional Exhaustion of Effector T Cells

Effector T cells in PVTT are broadly driven into an exhausted state, characterized by the co-expression of multiple inhibitory receptors and closely associated with loss of cytotoxic function and adverse clinical outcomes. Importantly, the determinant of effective antitumor immunity is not T-cell abundance per se, but rather the functional state of CD8⁺ T cells, instead of their infiltration density [57]. This suggests that the central immunological defect in PVTT is not simply a paucity of T cells, but profound functional inactivation within the thrombus niche. Such dysfunction is typically imposed by chronic antigen stimulation coupled with sustained inhibitory signaling that reinforces T-cell exhaustion. Beyond PD-1, PVTT-infiltrating T cells frequently co-express co-inhibitory receptors such as TIM-3 and LAG-3; their convergent signaling markedly suppresses cytokine production and cytolytic activity, driving deep hyporesponsiveness and facilitating tumor immune escape.

Active Immunosuppression Mediated by Regulatory and Myeloid Lineages

Tregs and MDSCs are markedly enriched within the PVTT microenvironment, where they are actively recruited through defined signaling programs and continuously impose and sustain effector T-cell exhaustion. High infiltration of Foxp3⁺ Tregs is a characteristic feature of PVTT and is closely associated with disease progression and poor clinical outcomes. Their accumulation has been linked to recruitment pathways such as the CCL5–TGF-β1–p38–MAX axis, which perpetuates and strengthens the local immunosuppressive barrier [83, 84].

Concurrently, MDSCs potently inhibit T-cell activity through mechanisms such as depletion of essential amino acids and the generation of reactive oxygen species [65]. Acting in concert, Tregs and MDSCs assemble a multilayered and durable immunosuppressive network that broadly attenuates antitumor immunity, enabling PVTT to persistently evade immune surveillance.

Stromal and Endothelial Reinforcement of the Immunosuppressive Niche

In PVTT, hypoxia-conditioned CAFs provide critical physical and biochemical support for immunosuppression by depositing a dense ECM and establishing a metabolically hostile milieu. Specific CAF pro-grams – such as inflammatory CAFs – directly contribute to immune exclusion and reinforce immune escape through the secretion of inhibitory cytokines and chemokines [85]. In parallel, dysfunctional tumor endothelial cells, which lack key adhesion molecules required for T-cell extravasation, impose a vascular-level constraint that limits immune-cell entry into the tumor core [86]. The native portal-vein wall may also actively participate in PVTT immune exclusion. Rather than serving as a passive scaffold, portal-vein endothelial cells exposed to tumor-derived inflammatory, angiogenic, and coagulation-related signals may acquire tumor-endothelial features that restrict immune-cell trafficking. In other tumor models, tumor-associated endothelium can express FasL in response to VEGF-A, IL-10, and PGE2, selectively eliminating effector CD8+ T cells while sparing Tregs, thereby forming a functional immune barrier [87]. In HCC-PVTT, erythroid-transdifferentiated myeloid cells have also been shown to compromise vascular endothelial integrity, aggravate coagulation, and promote PVTT formation [58]. These findings support a vessel-wall-centered model in which malignant clones, reprogrammed portal-vein endothelium, and stromal-myeloid populations jointly establish an immune-restrictive intravascular niche. Additional stromal remodeling by immune populations, including SPP1⁺ TAMs, further aggravates immune exclusion, collectively “locking in” PVTT as a profoundly immunologically cold and suppressive niche [67].

In summary, the immunosuppressive niche of PVTT is co-constructed by multiple cellular compartments – including TAMs, CAFs, MDSCs/Tregs, and exhausted T cells – and is sustained by convergent mechanisms. These include chemokine-driven recruitment networks, multi-node immune checkpoint signaling, ECM remodeling with vascular/endothelial barriers, and hypoxia-coupled metabolic reprogramming, which together enforce immune exclusion and progressive immune dysfunction, as summarized across the immune-sanctuary, intrinsic-resistance, and adaptive-resistance modules of Figure 6.

Fig. 6.

A hierarchical classification of the mechanisms of immunotherapy resistance in HCC-PVTT. The framework is organized into three primary axes: the function of the PVTT as an ‘immune sanctuary’ that creates a local-systemic dichotomy; the intrinsic resistance conferred by a suppressive stromal and cellular microenvironment; and the process of adaptive resistance driven by the upregulation of alternative immune checkpoints and metabolic reprogramming. HCC, hepatocellular carcinoma; PVTT, portal vein tumor thrombus; HCC-PVTT, hepatocellular carcinoma with portal vein tumor thrombus.

Conclusions and Future Perspectives

Collectively, current evidence supports the view that PVTT is not simply a passive extension of the primary tumor, but a relatively autonomous and highly adaptive immunosuppressive niche in HCC. Its establishment appears to be driven by early clonal divergence and reinforced by hypoxia-associated metabolic reprogramming, myeloid- and fibroblast-dominated stromal remodeling, ECM deposition, and endothelial dysfunction. These coordinated processes converge to restrict immune-cell trafficking, amplify T-cell exhaustion, and sustain an immune-excluded phenotype, thereby providing a mechanistic explanation for the limited efficacy of immune checkpoint blockade in HCC with PVTT.

From a translational perspective, this conceptual framework suggests that PVTT should be considered separately from the primary tumor when evaluating therapeutic vulnerability, biomarker development, and treatment stratification. In particular, approaches aimed at reprogramming suppressive macrophage and fibroblast states, normalizing vascular and stromal barriers, and interrupting hypoxia-driven metabolic-immune crosstalk may be required to restore immune responsiveness in this setting. Future studies should move beyond static comparisons between primary tumors and established PVTT and instead define how the PVTT microenvironment evolves under ICI-based therapeutic pressure. When clinically feasible, longitudinal paired sampling of primary tumors, PVTT, blood, and imaging data at baseline, early on-treatment, response, and progression should be prioritized. Integrated single-cell and spatial multi-omics approaches – including scRNA-seq, T-cell receptor/BCR profiling, spatial transcriptomics, multiplex immunofluorescence, and spatial pro-teomics – could help trace clonal evolution, antigen-presentation remodeling, endothelial-barrier conversion, myeloid polarization, T-cell clonal expansion or exhaustion, and stromal reorganization during treatment [4, 11]. Because repeated PVTT biopsy is often difficult, serial CTC phenotyping, mesenchymal-CTC quantification, ctDNA-based clonal tracking, AFP/PIVKA-II kinetics, contrast-enhanced imaging, and radiomics should be incorporated as complementary noninvasive readouts [35]. Such longitudinal frameworks may distinguish reversible immune exclusion from fixed stromal-vascular resistance and guide rational combinations of ICIs with anti-angiogenic therapy, locoregional therapy, myeloid-targeting agents, or stromal-remodeling strategies.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

This work was supported by General Program of the Development Center for Medical Science & Technology, National Health Commission of the People’s Republic of China (WKZX2023CX020004), the National Natural Science Foundation of China (82172693). The funder had no role in the design, data collection, data analysis, and reporting of this study.

Author Contributions

Xingfei Li, Qian Cheng, and Jie Gao conceived the review and designed the manuscript framework. Xingfei Li and Qian Cheng performed the literature search, analyzed the literature, and drafted the manuscript. They contributed equally to this work. Jiye Zhu, Zhigao Yuan, Shaowei Mu, and Tong Zhou contributed to literature screening, data interpretation, and critical revision of the manuscript. Jie Gao supervised the study and critically revised the manuscript for important intellectual content. All authors approved the final version of the manuscript.

Funding Statement

This work was supported by General Program of the Development Center for Medical Science & Technology, National Health Commission of the People’s Republic of China (WKZX2023CX020004), the National Natural Science Foundation of China (82172693). The funder had no role in the design, data collection, data analysis, and reporting of this study.