Multimodal clinical parameters-based immune status associated with the prognosis in patients with hepatocellular carcinoma

Abstract

Hepatocellular carcinoma presents with three distinct immune phenotypes, including immune-desert, immune-excluded, and immune-inflamed, indicating various treatment responses and prognostic outcomes. The clinical application of multi-omics parameters is still restricted by the expensive and less accessible assays, although they accurately reflect immune status. A comprehensive evaluation framework based on “easy-to-obtain” multi-model clinical parameters is urgently required, incorporating clinical features to establish baseline patient profiles and disease staging; routine blood tests assessing systemic metabolic and functional status; immune cell subsets quantifying subcluster dynamics; imaging features delineating tumor morphology, spatial configuration, and perilesional anatomical relationships; immunohistochemical markers positioning qualitative and quantitative detection of tumor antigens from the cellular and molecular level. This integrated phenomic approach aims to improve prognostic stratification and clinical decision-making in hepatocellular carcinoma management conveniently and practically.

Key Words: Hepatocellular carcinoma; Immune status; Phenotype; Multimodal parameters; Prognosis

Core Tip: The immune status of hepatocellular carcinoma patients can be comprehensively assessed using multimodal phenomics analyses, which include clinical features to provide general patient information, routine blood tests to evaluate metabolism and function, immune cell subsets to analyze immune system composition, imaging to visualize tumor morphology and anatomical relationships, and immunohistochemical markers to detect antigens in tumor tissue. By integrating these parameters through multimodal phenomics analyses, we aim to enhance the evaluation of prognosis and clinical outcomes for hepatocellular carcinoma patients.

INTRODUCTION

Hepatocellular carcinoma (HCC) is often diagnosed at an advanced stage with high mortality[1]. Survival rates are poor with conventional sorafenib therapy, but combining interventional therapies and immunotherapy has improved outcomes[2]. The immune system plays a key role in HCC development, resistance, and progression. Tumor cells create an inhibitory immune microenvironment, altering immune cell function and reducing immune response. Additionally, immune cell metabolism regulation is crucial for anti-tumor immunity. The liver’s microenvironment, influenced by bacterial components and foreign antigens, exhibits immunosuppression[3], contributing to HCC progression and immune escape. Recently, immunotherapy has gained prominence, with immune checkpoint inhibition demonstrating clinical benefits across various tumor types. Routine blood tests and patient clinical characteristics were utilized to accurately predict the efficacy of immune checkpoint inhibition therapy in cancer patients[4], providing valuable insights and direction for future research (Figure 1).

Figure 1 Exploring the relationship between immune status and prognosis of hepatocellular carcinoma with multimodal clinical parameters. Hepatocellular carcinoma patients present with three distinct immune profiles (“immune-desert”, “immune-excluded”, and “immune-inflamed”), corresponding to divergent prognostic outcomes (poor vs good). Clinical features, routine blood test, immune cell subsets, imaging features and immunochemical markers were used to indicate the different immune status of hepatocellular carcinoma patients, and collected and analyzed their prognostic effects. HCC: Hepatocellular carcinoma; ECOG-PS: Eastern Cooperative Oncology Group-performance status; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; ALP: Alkaline phosphatase; APAR: Alkaline phosphatase to albumin ratio; GGT: Gamma-glutamyl transferase; LDH: Lactate dehydrogenase; RBC: Red blood cell; WBC: White blood cell; LCR: Lymphocyte-to-C-reactive protein ratio; NLR: Neutrophil-to-lymphocyte ratio; PLT: Platelet; PLR: Platelet-to-lymphocyte ratio; CRP: C-reactive protein; IL: Interleukin; TGF-β: Transforming growth factor-β; DC: Dendritic cell; TAM: Tumor-associated macrophage; MDSC: Myeloid-derived suppressor cell; NK: Natural killer cell; Treg: Regulatory T cell; Th: T helper cell; NKT: Natural killer T cell; CT: Computed tomography; MR: Magnetic resonance; DSA: Digital subtraction angiography; MVI: Microvascular invasion; VLP: Vascular invasion-like pattern; AFP: Alpha-fetoprotein; AFP-L3: Alpha-fetoprotein L3 fraction; DCP: Des-gamma-carboxy prothrombin; CK19: Cytokeratin 19; GP73: Golgi protein 73; PD1: Programmed cell death protein 1; PD-L1: Programmed death-ligand 1; PD-L2: Programmed death-ligand 2; MMPs: Matrix metalloproteinases; CTLA-4: Cytotoxic T-lymphocyte-associated protein 4; TIM-3: T-cell immunoglobulin and mucin-domain containing-3; LAG-3: Lymphocyte-activation gene 3.

IMMUNE PHENOTYPE ASSESSMENT

The “immune-desert” phenotype describes a tumor microenvironment devoid of functional immune infiltration, particularly cytotoxic T cells, representing a dynamic and complex immunosuppressive state. According to Jia et al[5], the risk of recurrence was significantly reduced in patients with high immune infiltration, and the prediction accuracy was 82.2%. The “immune-excluded” phenotype refers to a tumor microenvironment that exhibits an immunosuppressive state due to reduced infiltration of immune cells in the tumor tissue or uneven distribution. Based on the study of construction and validation of immune risk prognostic models by Chen et al[6], such patients tend to have poor prognosis, high recurrence rate, and poor response to immunotherapy. Consequently, nutritional immune status must be evaluated with other biomarkers to develop more optimal treatment strategies. The “immune-inflamed” phenotype means that the degree of immune infiltration is between cold and hot, and there can be local immune activation; however, the immune system is suppressed. Careful consideration of their genotype and metabolic status is required for such patients to evaluate the prognostic effect. The formulation of treatment plans must be more targeted due to the greater significance of the individualized effect.

CLINICAL FEATURES

Sex

Compared with males, female HCC patients typically present with unifocal, smaller, well-differentiated tumors at earlier tumor-node-metastasis stages[7,8]. Nevola et al[9] reported that women often experience longer overall survival (OS) and disease-free survival, with lower recurrence and mortality rates, underscoring sex-based clinical and prognostic variations.

Age

Elderly patients often present with comorbidities such as cirrhosis and cardiovascular disease, leading to reduced tolerance to surgery and chemotherapy. This affects treatment effectiveness and results in an overall less ideal prognosis.

Eastern Cooperative Oncology Group-performance status

The Eastern Cooperative Oncology Group (ECOG) performance status score determines the physical condition and activity levels of patients with cancer, facilitating treatment planning. ECOG can influence time to disease progression in quality of life and ECOG 0 vs > 0 predicted better OS in multivariate analysis[10,11].

Therapeutic history

Early-stage patients treated with resection, transplantation, or ablation can achieve a median survival of over 60 months (vs 36 months untreated). Intermediate-stage patients with preserved liver function reach 26 months with chemoembolization, while advanced-stage patients gain 3-month survival benefit from sorafenib, as reported by Vogel et al[12].

Muscle and fat status

The prognosis of HCC resection may be predicted by preoperative muscle mass, although research on this topic is limited. Reductions in muscle quantity and quality are recognized surgical risk factors[13]. It’s found by Hamaguchi et al[14] that high intramuscular adipose tissue significantly reduced OS and recurrence-free survival (RFS), elevating post-hepatectomy mortality and recurrence.

ROUTINE BLOOD TESTS

Routine blood tests reflect systemic metabolic and functional state at a molecular level and serve as independent risk factors for HCC recurrence and survival. In this section, we will review several survival predictors of routine blood tests in HCC and their prediction performance in existing studies (Table 1).

Table 1 Survival predictors of routine blood tests in hepatocellular carcinoma.

Ref.	Patients	Simple size	Variables	Threshold	Prediction performance
Liu et al[16]	HCC treated with hepatectomy	315	AST (U/L)		The median (Q1, Q3) data are 34 (27.5, 47.5) of group no recurrence, 40 (30, 50.75) of group recurrence, P = 0.041
Liu et al[16]	HCC treated with hepatectomy	315	ALT (U/L)		The median (Q1, Q3) data are 36 (24, 48) of group no recurrence, 40 (30, 56.75) of group recurrence, P = 0.042
Su et al[15]	HCC	2327	ALP (U/L)	172	The median OS in the high ALP group was significantly shorter than in the low ALP group (7.7 months vs 55.4 months)
Zhang et al[20]	HCC treated with hepatectomy	414	GGT (mg/L)	48.5	Kaplan-Meier analysis showed patients with GGT < 48.5 had significantly longer overall survival (P = 0.000)
Zhang et al[20]	HCC treated with LT	155	LDH (U/L)	80.5	Kaplan-Meier analysis showed high LDH levels significantly correlated with poor RFS (P = 0.001) and OS (P = 0.008)
Carr et al[21]	HCC	995	CRP (mg/L)	10, 50	With increase in CRP grouping, significant trend for increase in each of AFP, MTD and percent PVT parameters were found
Utsumi et al[22]	HCC	173	LCR	9500	Kaplan-Meier analysis showed high LCR group patients had longer RFS and OS compared to the low LCR group
Wen et al[24]	HCC	126	NLR	3.867	Median survival time was higher in patients with NLR (not achieved vs 18 months, P = 0.014)
Pang et al[18]	HCC treated by hepatic resection	172	PLT (/L)	148 × 109	Kaplan-Meier curves stratified by PLT (P = 0.021) showed significantly higher recurrence rates in non-cirrhotic patients with PLT ≥ 148 × 109/L
Ismael et al[25]	HCC treated with LT	212	PLR	150	PLR ≥ 150 to be a strong predictor of worse 5-year OS (40.2% vs 79.4%) and RFS (70.2% vs 95.9%) in HCC patients
Wang et al[26]	HCC treated with TACE	53	IL-2 (pg/mL)	0.025	P value of time to progression equals to 0.031
Wang et al[26]	HCC treated with TACE	53	IL-6 (pg/mL)	4.400	P value of time to progression equals to 0.013; P value of OS equals to 0.007
Peng et al[27]	HCC	423	TGF-β1	High/low	HCC patients with higher TGF-β1 expression had a shorter OS compared to those with lower expression (HR = 1.417, 95%CI: 1.014-1.979, P = 0.0411)
Müller et al[28]	HCC treated with TACE	280	PNI	37.59	Low PNI patients had shorter median OS than high PNI patients (7.5 months vs 21.4 months, P < 0.001)
Li et al[30]	HCC	1334	CONUT	8	Postoperative CONUTS ≥ 8 was an independent risk factor for complication III-V (OR = 2.054, 95%CI: 1.371-3.078, P < 0.001)

HCC: Hepatocellular carcinoma; LT: Liver transplant; TACE: Transarterial chemoembolization; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; ALP: Alkaline phosphatase; GGT: Gamma-glutamyl transpeptidase; LDH: Lactate dehydrogenase; CRP: C-reactive protein; LCR: Lymphocyte-to-C-reactive protein ratio; NLR: Neutrophil-to-lymphocyte ratio; PLT: Platelet count; PLR: Platelet-to-lymphocyte ratio; IL-2: Interleukin-2; IL-6: Interleukin-6; TGF-β1: Transforming growth factor β1; PNI: Prognostic nutritive index; CONUT: Controlling nutritional status; OS: Overall survival; RFS: Recurrence-free survival; AFP: Alpha-fetoprotein; MTD: Maximum tumor dimension; PVT: Portal vein thrombosis; HR: Hazard ratio; CI: Confidence interval; OR: Odds ratio.

Liver function indicators

Aspartate aminotransferase (AST) is an enzyme that reflects liver damage and is often used as a biomarker to evaluate liver disease progression. High serum AST levels in HCC indicate poor prognosis. Alanine aminotransferase is another sensitive liver function testing indicator. Alkaline phosphatase is a hydrolytic enzyme associated with the epithelial-mesenchymal transition cell phenotype, which is believed to be the initial stage of HCC microvascular invasion (MVI)[15]. Multiple studies have demonstrated the relationship between the prognosis of HCC after treatment and AST, alanine aminotransferase, and alkaline phosphatase levels[15-18].

Metabolic indicators

Gamma-glutamyl transferase, a key enzyme in glutathione metabolism, not only indicates hepatobiliary damage (especially cholestasis)[19] but also independently predicts OS in primary HCC, while elevated lactate dehydrogenase, a lactate-pyruvate interconversion enzyme linked to tumor metabolic reprogramming, predicts poor prognosis[20].

Inflammatory indicators

There is an association between clinical C-reactive protein levels and parameters of human HCC growth and aggressiveness[21]. Utsumi et al[22] suggests that a low lymphocyte-to-C-reactive protein ratio, accessible, objective, and non-invasive, be an independent risk factor for recurrence. The neutrophil-to-lymphocyte ratio reliably predicts HCC immunotherapy efficacy, serving as a prognostic biomarker by identifying responders and avoiding ineffective treatment[23,24]. Platelets affect thrombosis, inflammation, liver regeneration, and angiogenesis regulation, with platelet count independently predicting HCC recurrence[18] and a higher platelet-to-lymphocyte ratio was associated with a worse OS and RFS[25].

Clinically, interleukin (IL)-2 and IL-6 regulates cell proliferation and survival, which is associated with liver function, tumor traits, and poor prognosis. Wang et al[26] analyzed IL-2 and IL-6 associations in HCC patients undergoing transarterial chemoembolization. Moreover, it’s reported by Peng et al[27] that as transforming growth factor-β (TGF-β) stimulates HCC pathogenesis and metastasis by creating a favorable microenvironment for cancer growth, promoting epithelial-mesenchymal transition, and enhancing fibrogenesis, high TGF-β1 expression can predict a worse OS.

Nutritional and immune-related parameters

The prognostic nutritional index and controlling nutritional status are composite indicators of nutritional immune status with significant value in predicting HCC prognosis. Patients with high prognostic nutritional index or controlling nutritional status scores exhibit significantly worse OS and RFS[28-30].

IMMUNE CELL SUBSETS

Immune system dysregulations, including changes in the number or function of immune cells, cytokine levels, and expression of inhibitory receptors or ligands, are closely associated with HCC (Table 2). Tumor-associated immune cells can be involved, including macrophages, T lymphocytes, cytokines, and relevant interactions.

Table 2 Immune status and biomarkers of immune cells.

Immune components	Immune status
DCs	Decreased antigen presentation, decreased quantity, impaired function
Macrophages	Antigen presentation disorder, activates Th2 immune response, promotes Tregs
MDSC	Inhibits by free radicals, arginase, and TGF-β
Neutrophils	Positively correlated with the stage of cancer
NK cells	The number and the cytolytic activity decrease. Higher total CD56+ NK cells have a good prognosis
T lymphocytes	Decreased, inhibitory receptor expression increased (decreased CD3+ T cells, CD4+ T cells and CD4+/CD8+ T cell ratio. Increased CD8+ CD28- T and CD4+ CD25+ T cells)
Tfh	Decreased CXCR5CD4+, associated with tumor progression
Treg	Increases T cells, inhibits T cells proliferation, IFN-γ secretion, and NK cells response. Poor prognostic markers
Th17 cells	Increased, unspecified effect, and associated with disease progression. Elevated Th17 and Th1 numbers may promote progression as prognostic markers
CIK	Decreased, weakened anti-cancer effect
NK T cells	Dual action, promoting Th2 cytokines. Correlated with invasion and metastasis

DCs: Dendritic cells; MDSC: Myeloid-derived suppressor cells; NK: Natural killer cell; Tfh: T follicular helper cell; Treg: Regulatory T cell; Th: T helper cell; CIK: Cytokine-induced killer cell; TGF-β: Transforming growth factor β; CXCR: C-X-C chemokine receptor; IFN-γ: Interferon-γ.

Dendritic cells

Decreased expression of human leukocyte antigen class I molecules can result in reduced T cells immunity levels, resulting in dendritic cell (DC) presentation failure against HCC-associated antigens[31]. Especially patients infected with hepatitis B and C viruses exhibit varying degrees of peripheral DC defective volume and function. The inability of activated DC to infiltrate tumor tissue results in impaired tumor-specific lymphocyte tending movement and recruitment[32]. CD14+ cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) DC present in peripheral blood lymphocytes and tumor tissues express the immunosuppressive molecule CTLA-4 and programmed cell death protein 1 (PD-1)[33]. High levels of anti-inflammatory cytokines, IL-10, and indoleamine 2,3-dioxygenase inhibition can promote tumor progression and immune escape by attenuating CD4+ T cell immune responses.

Macrophages

Macrophages differentiate into M2 phenotype with reduced expression of immunomodulatory cytokines and antigen presentation. Additionally, they usually express C-C motif chemokine ligand (CCL) 17, CCL22, CCL24 and chemokines, low-level pro-inflammatory cytokines, and reactive oxygen species. This alternative phenotype of macrophages is further involved in activating T helper 2 cell (Th2) immune response, thereby promoting the recruitment and development of regulatory T-cells (Treg). Chronic inflammation has been reported to be primarily associated with higher levels of macrophage colony-stimulating factor and macrophage infiltration and associated with progression and intrahepatic metastases[34,35].

Myeloid-derived suppressor cells

Myeloid-derived suppressor cell is a subset of inflammatory monocytes that can be promoted by upregulating the expression of various factors and regulating T cell responses, including the production of free radicals and TGF-β and the regulation of arginase activity-induced Treg. An increased frequency of myeloid-derived suppressor cell in the peripheral circulation and tumor setting has been reported[36].

Neutrophils

Neutrophils can predict the survival rate. It is inversely proportional to the density of the invasion, and the number of invasions is positively correlated with the cancer stage. Kuang et al[37] reported that chemokines (C-X-C motif chemokine ligand 8) produced by epithelial cells under the influence of Th17 cells can increase the neutrophil population in the peritumoral stromal cell population. Neutrophils produce matrix metalloproteinase (MMP)-9 in tumor tissue, promoting angiogenesis and tumor growth.

Natural killer cells

They are subdivided into two subsets based on cell membrane expression of CD56 and CD16, differing in tissue distribution and immune effects[38]. CD56- dim natural killer (NK) subset is more cytotoxic than the other; however, they can secrete chemokines and pro-inflammatory cytokines for immune modulation in response to stimuli surrounding cells. Activating receptors and inhibitory receptors regulate the activation. Normally, inhibitory receptor action predominates, inhibiting cytotoxic activity. The absence of major histocompatibility complex-I in target and tumor cells provides conditions for NK cells killing by activating NK receptors. Human hepatic resident NK cells are highly cytotoxic in the hepatic sinusoids. The interaction between them and DC induces the expansion of Treg.

Patients with HCC exhibit a reduction in the frequency and absolute number of peripheral NK cells, especially in the CD56- dim NK cell subset. Studies have indicated that tumors infiltrate NK cells less frequently than tumor-adjacent human hepatic resident NK cells. Higher total CD56+ tumor-infiltrating NK cells are associated with better prediction of survival[31]. According to Chu et al[39], the frequency of specific NK cell subsets, including CD11b-/CD27- NK cells associated with delayed HCC progression, increases. Activation of the NKG2D receptor enhances cytotoxicity; however, this potent anti-tumor cell function can be more effective in the early stages and reduced once the tumor progresses. Furthermore, poor clinical prognosis is associated with the expression of PD-1 and CD96 on tumor-infiltrating NK cells.

B lymphocytes

Plasma cells in tumors, especially IgG+ ones, can significantly promote monocyte polarization to inhibitory M2b macrophages. The network of functional interactions between monocytes/macrophages, T and B cells is an important factor in tumor immune tolerance[39]. Imbalance in B-cell subsets is observed in HCC. B cells participate in anti-tumor immune responses through various mechanisms, including antibody production, antigen presentation, and cytokine secretion. Their surface markers, including CD19 and CD20, exhibit altered expression patterns. The proportion of regulatory B cell increase, and they secrete immunosuppressive cytokines, including IL-10 and TGF-β, which suppress the function of CD8+ T cells and NK cells while promoting Treg expansion[40]. This contributes to an immunosuppressive tumor microenvironment, correlating with progression and poor prognosis.

T lymphocytes

The CD3+ and CD4+ T cell frequency, and CD4+/CD8+ T cell ratio were significantly lower than those in healthy individuals, while the CD8+ T cell values were higher. An increased frequency of CD8+ CD28- T and CD4+ CD25+ T cells suggests an increased frequency of suppressor T cells in the peripheral blood[41]. Patients with a higher number of CD4+ T lymphocytes in adjacent tissues always have a better prognosis, while those with a large number of CD8+ T lymphocytes in cancer nest tissues have a better prognosis[42]. Besides, it has been hypothesized that the localization of CD103+ cytotoxic T cells in tumors can determine whether they are anti-cancer. The proportion of precancerous CD103+ cytotoxic T cells is an independent factor in predicting the recurrence, while the proportion of intratumoral ones is an independent factor for good prognosis.

Treg

Treg expresses CD25 and Foxp3, which are closely associated with carcinogenesis. Under normal physiological conditions, native Treg limits autoimmune response. Its frequency increases, and the number is correlated with severity. Patients with advanced HCC exhibit higher intrahepatic CD8+ Foxp3+ Treg than initial patients, suggesting it can be associated with immune evasion. Foxp3+ Treg accumulates highly in tumors, which in turn inhibits DC function, T cell proliferation, and the ability to secrete interferon-γ, promotes tumor progression, and can be used as a poor prognostic marker. Additionally, Treg binds to TGF-β and can inhibit NK cells responses. Based on the animal model of Wakiyama et al[43], it was suggested that the imbalance of Treg to CD8+ T cells can promote HCC progression.

Th17 cells

IL-17 plays a dual role in tumor immunology. Its frequency was significantly higher than that of non-tumor tissues and was positively correlated with microvessel density. It can promote tumor growth by promoting anti-tumor cytotoxic T-cell responses or angiogenesis in peripheral endothelial cells and fibroblasts[33,44]. Compared to non-tumor region, the increase of Th17 and Th1 cells in tumor region correlates with overall disease-free survival. Elevated values can promote tumor progression and serve as a prognostic marker.

NK T cells

Typical ones recognize the non-polymorphic molecule CD1d and present their exogenous lipid antigens, promoting anti-tumor responses by activating effector cells. The decrease is inversely correlated with invasion and metastasis. NK T cells can be a correlation indicator and inhibit tumor progression[32,44]. Table 2 summarizes the functions of various cells above within the tumor immune microenvironment, including changes in their abundance and prognostic implications. It highlights the complexity of the body’s immune status and provides key information on major cell types based on available research.

Cytokines and chemokines

Due to the limited length and the complexity of cytokines and chemokines, we have chosen to summarize their changes and mechanisms of action in liver cancer patients directly in Table 3. Some of these have been mentioned in the previous sections. Again, this represents a selection of what we consider representative findings within the scope of our research. The cytokine environment in the liver of metastatic HCC is biased toward the Th2 profile, with increased levels of anti-inflammatory cytokines and a decrease in pro-inflammatory cytokines (Table 3). Modifications in several pertinent cytokines and their effects on HCC treatment and prognosis are displayed in that.

Table 3 Changes and impacts of cytokines and chemokines.

Immunomodulatory factor	Quantitative changes and functional status
Th1 cytokines	The tumor microenvironment decreases and induces CD8+ T cells (IL-2 is significantly reduced, which is directly related to prognosis while the levels of TGF-β and IL-6 are increased. Others such as IFN-γ, IL-8, IL-15 and IL-18 have been implicated in invasion and metastasis)
Th2 cytokines	Elevated, associated with tumor progression (showing a shift from Th1 to Th2)
Pro-inflammatory cytokines	Involved in pathogenesis and tumor escape
Anti-inflammatory cytokines	Ascending, associated with progression (IL-6 is associated with poor prognosis. IL-37 expression levels were correlated with tumor burden and survival improvement)
Chemokine ligand-receptor axis	Relevant to tumor progression and metastasis (CCR6-CCL20 axis and CXCL12-CXCR4 axis expression were higher, which was positively correlated with clinical stage. Increased CXCR3 expression was correlated with tumor burden and cancer stage)

IL: Interleukin; TGF-β: Transforming growth factor β; IFN-γ: Interferon-γ; Th: T helper cell; CXCR: C-X-C chemokine receptor; CXCL: C-X-C motif chemokine ligand; CCL: C-C motif chemokine ligand; CCR: C-C chemokine receptor.

IL-2 can promote the proliferation and differentiation of T lymphocytes. IL-6 can directly stimulate stem cells’ proliferation and collagen secretion. Its serum level can effectively indicate the prognosis of HCC diagnosis. TGF-β can inhibit the growth of peritumoral tissue cells and contribute to the growth of tumor cells and is the most effective immunosuppressive factor, which can proliferate and differentiate lymphocytes, inhibit cytotoxicity, antagonize cytokines, and promote immune escape and tumor formation[31].

Clinical practice of immune cells

In clinical practice, the immunological components described above can be translated into actionable and measurable indicators through pathological biopsy, immunological assays, and peripheral blood tests. For instance, pathological analysis of tumor tissue can reveal the status of tumor-infiltrating lymphocytes, including T cells and NK cells. A state of high tumor-infiltrating lymphocyte infiltration suggests a more robust anti-tumor immune response, which is associated with better treatment efficacy and improved prognosis. This information directly informs clinical decisions regarding post-treatment monitoring schedules (e.g., frequency and timing of follow-ups) and guides subsequent therapeutic strategies. For example, it can help determine whether to pursue combination therapies, employ immune checkpoint inhibitors, target therapies, or consider adjuvant radiotherapy/chemotherapy.

Conversely, the presence of immunosuppressive cells like Tregs and M2-type macrophages within the tumor microenvironment indicates a state of immune evasion, potential tumor progression, and treatment resistance. Identifying these cells serves as a critical prognostic biomarker for more aggressive disease. In summary, the core task for clinical medicine is to convert this complex immunological knowledge into detectable substances and biomarkers. By fully understanding their clinical significance and integrating these findings with results from other disciplines (e.g., imaging, serology), clinicians can comprehensively assess a patient’s status, dynamically monitor and predict treatment response, and accurately evaluate long-term prognosis.

IMAGING FEATURES

By integrating multiple imaging examinations, multimodal imaging features provide a clear depiction of the lesion’s morphology, size, location, and anatomical relationship with surrounding tissues. Compared to single imaging modalities, it increases diagnostic accuracy and offers better prognosis prediction for patients with HCC. The relationship between specific imaging indicators and the prognosis of patients with HCC is summarized (Table 4).

Table 4 Relationships between imaging features and prognosis of different imaging examinations in hepatocellular carcinoma.

Imaging examination	Imaging features		Prognosis
CT	Tumor number	≤ 3	Good
		> 3	Poor
	Maximum diameter of the tumor	≤ 5 cm	Good
		> 5 cm	Poor
	Tumor envelope	Exist	Good
		Absence	Poor
	Irregular edge reinforcement	Absence	Good
		Exist	Poor
	Distribution of tumors across lobes	Absence	Good
		Exist	Poor
	Envelop invasion (arch-string ratio)	≤ 1	Good
		> 1	Poor
	Diaphragm invasion	Absence	Good
		Exist	Poor
	Cirrhosis	Absence	Good
		Exist	Poor
	The proportion of tumor convex	< 25%	Good
		≥ 25%	Poor
	Portal vein occlusion	Absence	Good
		Exist	Poor
MRI	Peritumoral enhancement	Absence	Good
		Exist	Poor
	Intratumoral necrosis	Absence	Good
		Exist	Poor
	Internal aneurysm	Absence	Good
		Exist	Poor
	Tumor edge	Smooth	Good
		Rag	Poor
	Multifocality	Absence	Good
		Exist	Poor
	Low density ring	Exist	Good
		Absence	Poor
Ultrasound	Tumor enhancement starts for a long time	Absence	Good
		Exist	Poor
	The arterial phase was highly enhanced	Absence	Good
		Exist	Poor
	The portal phase was low and enhanced	Absence	Good
		Exist	Poor
	The delay period is low and enhanced	Absence	Good
		Exist	Poor
DSA	The arrival time is reduced	No	Good
		Yes	Poor
	The peak time rises	No	Good
		Yes	Poor
	Increased filling rate	No	Good
		Yes	Poor
	The curve width is elevated	No	Good
		Yes	Poor
	The mean passage time was increased	No	Good
		Yes	Poor
Radiomics	First-order statistics	Different phases based on CT or MRI
	Texture features
	Wavelet decompositions

CT: Computed tomography; MRI: Magnetic resonance imaging; DSA: Digital subtraction angiography.

Computed tomography

The study by van der Pol et al[45] revealed margin-like arterial-phase enhancement and tumor capsule in liver-enhanced computed tomography (CT) scans, indicating aggressive HCC. The number of tumors has proved to be a valuable prognostic parameter for HCC. According to Lv et al[46], various CT imaging features are closely associated with the occurrence of spontaneous rupture and bleeding in patients with HCC. Based on the above data and Table 4, CT is a more widely used and comprehensive imaging technology in liver cancer evaluation. Clinically, CT is widely used for staging, tumor assessment, and detecting extrahepatic spread.

Magnetic resonance imaging

MVI is an important indicator for evaluating tumors’ metastatic potential and is highly significant for patient prognosis. Research by van der Pol et al[45] demonstrates that magnetic resonance imaging (MRI) reveals distinct imaging features in patients with different MVI grades, including tumor-enhancing peritumoral enhancement, intratumoral necrosis, intratumoral fat, intratumoral arteries, tumor margins, multifocal lesions, and low-density ring patterns, all of which provide significant diagnostic value for MVI grading. Combined with Table 4, MRI provides more distinctive liver cancer imaging, showing greater sensitivity than CT in detecting the malignancy and metastatic potential of small lesions.

Ultrasound

MVI is an important indicator for evaluating tumors’ metastatic potential and is highly significant for patient prognosis. Research by van der Pol et al[45] demonstrates that MRI provides significant diagnostic value for MVI grading in patients with primary HCC. Patients with different MVI grades exhibit distinct qualitative MRI features, including peritumoral enhancement, intratumoral necrosis, intratumoral fat, intratumoral arteries, tumor margins, multifocal patterns, and low-density rings. Combined with Table 4, MRI can show more characteristic images of liver cancer, and its sensitivity is higher than that of CT. In clinical practice, MRI can improve the evaluation of the malignancy and metastatic potential of small lesions.

Digital subtraction angiography

The digital subtraction angiography perfusion imaging technique can directly reflect the spatial distribution of hemodynamics and is important for diagnosing HCC[47]. According to the research of Jiang et al[48] and Becker et al[49]. in advanced HCC, the tumor vascular trophic tissue is larger, longer, and tortuous, and digital subtraction angiography arrival time decreases. In contrast, peak time, filling rate, curve width, and average passage time can be used to evaluate the prognosis of patients with HCC.

Radiomics

Radiomics is a breakthrough in the non-invasive prediction of MVI, early recurrence after hepatectomy, and prognosis after locoregional or systemic therapies[50]. Radiomics also offers advanced tools for tumor microenvironment characterization and prognosis assessment[49,50]. Based on CT or MRI in different phases, radiomics features were extracted as first-order statistics, texture features, and wavelet decompositions. Combined with radiomics features and artificial intelligence models, we can improve accuracy in predicting biological characteristics and prognosis in HCC, such as MVI and tumor recurrence[51-55].

TUMOR MARKERS AND IMMUNOHISTOCHEMICAL MARKERS

The impacts of immunohistochemical markers on the immune microenvironment and prognosis in HCC are summarized (Table 5).

Table 5 Impact on immune microenvironment and prognosis of immunohistochemical markers in hepatocellular carcinoma.

Category	Marker	Impact on immune microenvironment	Impact on prognosis	Clinical decision-guiding value
Immunosuppressive	PD-L1/PD-1	Induces T-cell exhaustion via PD-1/PD-L1	High expression: Improved objective response rate and disease control rate to PD-1/PD-L1 inhibitors in advanced HCC	Predictive: Supports the selection of PD-1/PD-L1 inhibitor therapy
	CTLA-4	Blocks CD28- B7 costimulation; expands Tregs	High expression: Increased Treg infiltration and poorer OS	Emerging target: Informs the potential for combination immunotherapy strategies
	TIM-3	Triggers T-cell apoptosis; polarizes M2 macrophages	Co-expression with PD-1: Reduced OS	Mechanistic insight: Supports the development of dual-targeting approaches
Proliferation/invasion	Ki-67	Drives cell-cycle dysregulation	High expression (> 30%): Increased tumor recurrence after liver transplantation	Prognostic: May necessitate more intensive post-transplant surveillance
	MMP-2/9	Degrades ECM; recruits MDSCs	High expression: Promotes HCC invasion and metastasis	Prognostic: Indicates high risk of metastasis, warranting comprehensive staging and follow-up
Angiogenic	DCP (PIVKA-II)	Activates VEGF pathway	Post-treatment decline: Predicts improved clinical outcomes	Monitoring: Serves as a surrogate biomarker for monitoring treatment efficacy in AFP-negative HCC
	GP73	Promotes angiogenesis via mTOR-VEGFA	High expression: Poor prognosis in HCC	Prognostic: Identifies patients with aggressive disease for more aggressive management
Metabolic dysregulation	AFP	Suppresses DC maturation and contributes to immunosuppression	Baseline AFP < 400 ng/mL or an early AFP response (≥ 75% decline): Predicts improved OS	Predictive and monitoring: Key for patient selection and early efficacy assessment in atezolizumab + bevacizumab therapy
	AFP-L3	Evades variant associated with immune evasion	AFP-L3% > 10%: Poorer OS and higher early recurrence rates	Risk stratification: Identifies high-risk patients for adjuvant therapy or closer monitoring
Tumor suppressor dysfunction	P53	Mutant recruits MDSCs; secretes immunosuppressive cytokines	Drives HCC progression and immunosuppression: Predicts poor prognosis	Prognostic: Suggests high-risk biology, informing the need for adjuvant therapy or enrollment in clinical trials
Immune infiltration	CD3+ T-cells	Mediates antitumor immunity (TCR-MHC)	High intratumoral density: Favorable prognostic factor for improved RFS	Prognostic: May support de-escalation of aggressive therapy in early-stage disease
Emerging targets	LAG-3	Binds MHC-II; synergizes with PD-1 to exhaust T-cells	High expression: Links to inferior RFS and OS	Emerging target: Informs the rationale for combining anti-PD-1 with anti-LAG-3 therapy

PD-L1: Programmed death-ligand 1; PD-1: Programmed cell death protein 1; CTLA-4: Cytotoxic T-lymphocyte-associated protein 4; TIM-3: T-cell immunoglobulin and mucin-domain containing-3; MMP-2/9: Matrix metalloproteinase-2/9; DCP: Des-gamma-carboxy prothrombin; PIVKA-II: Protein induced by vitamin K absence/antagonist-II; Ki-67: Kiel-67; GP73: Golgi protein 73; AFP: Alpha-fetoprotein; AFP-L3: Alpha-fetoprotein L3 fraction; P53: Tumor protein p53; LAG-3: Lymphocyte-activation gene 3; Tregs: Regulatory T cells; ECM: Extracellular matrix; MDSCs: Myeloid-derived suppressor cells; VEGF: Vascular endothelial growth factor; mTOR: Mammalian target of rapamycin; VEGFA: Vascular endothelial growth factor A; DC: Dendritic cell; TCR: T-cell receptor; MHC: Major histocompatibility complex; MHC-II: Major histocompatibility complex class II; OS: Overall survival; HCC: Hepatocellular carcinoma; RFS: Recurrence-free survival.

Alpha-fetoprotein

Alpha-fetoprotein (AFP), a fetal liver glycoprotein, reappears in HCC and is used for surveillance and diagnosis. Marrero et al[56] reported that a level > 400 ng/mL is highly specific for the disease, though sensitivity is limited in early HCC. Zhu et al[57] demonstrated that beyond diagnosis, baseline AFP level and early AFP response are robustly predictive of outcomes with atezolizumab plus bevacizumab therapy in advanced HCC. Patients with baseline AFP < 400 ng/mL or those achieving a ≥ 75% decline in AFP at week 6 demonstrated significantly longer OS and progression-free survival, establishing AFP as a key surrogate biomarker for treatment efficacy[57].

L3 fraction of AFP

L3 fraction of AFP (AFP-L3), a fucosylated variant of AFP, demonstrates high specificity for HCC. Liu et al[58] found that in a large-scale clinical analysis, an AFP-L3% > 10% showed a specificity of 93.06% for HCC diagnosis, with sensitivity varying significantly by tumor size from 37.5% (< 3 cm) to 87.1% (≥ 5 cm). Elevated AFP-L3% is an independent risk factor for poor OS and early recurrence.

Des-gamma-carboxy prothrombin

Des-gamma-carboxy prothrombin is an abnormal prothrombin that demonstrates high specificity as a diagnostic biomarker for HCC, particularly in AFP-negative cases, reported by Liu et al[59] and Chen et al[60] showed that in patients with non-AFP-secreting HCC, des-gamma-carboxy prothrombin serves as a surrogate biomarker for therapeutic response, where early post-treatment decline correlates with improved clinical outcomes.

Golgi protein 73

Liu et al[61] demonstrated that Golgi protein 73 promotes angiogenesis in HCC by activating the mammalian target of rapamycin-vascular endothelial growth factor A signaling pathway. This pro-angiogenic effect contributes to tumor progression. Elevated serum Golgi protein 73 levels are associated with poorer prognosis in HCC patients.

P53 protein

TP53 mutations in HCC cause loss of tumor suppression and acquisition of oncogenic gain-of-function, promoting genomic instability, invasion, and metastasis[62]. Research by Makino et al[63] in liver models shows P53 activation can non-autonomously drive carcinogenesis by fostering an immunosuppressive microenvironment and promoting vascular invasion. P53 aberrations are therefore strongly linked to aggressive disease and poor prognosis.

PD-1/PD-ligand 1/PD-ligand 2

The PD-1/PD-ligand 1 (PD-L1) interaction drives CD8+ T-cell exhaustion[64]. A meta-analysis by Yang et al[65] confirmed that high PD-L1 expression predicts better response to PD-1/PD-L1 inhibitors in advanced HCC. Separately, the work of Yearley et al[66] high PD-L2 expression is associated with primary resistance to anti-PD-1 monotherapy, highlighting the need for dual-targeting strategies.

Kiel-67

Kiel-67 is a nuclear marker of cellular proliferation. Zhang et al[67] reported that expression > 30% predicts higher recurrence risk in HCC patients after liver transplantation, underscoring the need for combined biomarker assessment.

MMPs

MMP-2 and MMP-9 facilitate HCC invasion and metastasis by degrading the extracellular matrix. Tadros et al[68] demonstrates that the tumor suppressor miR-1-3p exerts its anti-invasive effects, at least in part, by directly targeting and downregulating the expression of both MMP-2 and MMP-9, highlighting their crucial role in HCC progression.

Immune checkpoint molecules

In HCC, Dai et al[69] CTLA-4 overexpression is associated with increased Treg infiltration and poorer OS. T-cell immunoglobulin and mucin-domain containing-3 frequently co-expresses with PD1, defining a T-cell exhaustion phenotype that correlates with reduced survival. High CD3+ T-cell density predicts improved RFS. Additionally, lymphocyte-activation gene 3 is upregulated in HCC and its high expression is linked to inferior recurrence-free and OS.

CLINICAL APPLICATIONS

Immunohistochemical markers are correlated with the prognosis of HCC (Figure 2). Blood and biochemical indicators assess the immune status of patients, while immune cell subset detection analyzes dynamic tumor microenvironment changes. Combining these methods aids in formulating individualized treatment plans and improving immunotherapy efficacy prediction. Imaging provides crucial information, such as tumor size, morphology, and location, enabling dynamic monitoring of tumor changes and treatment response. It can also observe immune cell infiltration and tumor angiogenesis, which are closely linked to immunotherapy effectiveness. These dynamic indicators help evaluate treatment outcomes. A comprehensive analysis of blood, biochemical indicators, immune cell subsets, imaging, and immunohistochemical markers significantly enhances prognosis accuracy and clinical application. It helps identify potential immunotherapy beneficiaries, monitor treatment response, and optimize treatment plans, offering a solid foundation for clinical decision-making.

Figure 2 Evaluation of immunohistochemical markers in prognostic risk stratification for hepatocellular carcinoma. By analyzing the expression patterns of key immune checkpoints and biomarkers in the tumor microenvironment, the prognostic risks of HCC were classified into three categories: High risk, intermediate risk, and protective factors. PD-L1: Programmed death-ligand 1; PD-1: Programmed cell death protein 1; TIM-3: T-cell immunoglobulin and mucin-domain containing-3; MMP: Matrix metalloproteinase; DCP: Des-gamma-carboxy prothrombin; GP73: Golgi protein 73; AFP: Alpha-fetoprotein; AFP-L3: Alpha-fetoprotein L3 fraction; LAG-3: Lymphocyte-activation gene 3.

FUTURE PROSPECTS

Multimodal phenomics analysis integrates various data types and methods to assess disease states, providing more accurate prognosis and guiding treatment decisions. This approach holds great promise for patients with liver cancer, helping to monitor disease progression and select appropriate therapies (Figure 3). The research methods outlined below offer a comprehensive framework for advancing this field.

Figure 3 Future prospects of study design and statistic method for prognostic prediction of hepatocellular carcinoma patients based on analysis of multimodal clinical parameters. Based on the multimodal clinical parameters, follow-up studies will center on four key areas: Observation studies, clinical trials, prognosis analyses, and clinical decision tools. Diverse statistical methods, covering linear and nonlinear models as well as machine learning and deep learning models, can be applied for achieving comprehensive and in-depth research results.

Given the rise in obesity, alcohol dependence, and aging populations, observational and controlled trials will be conducted. Observational trials will analyze patient data to identify relationships between clinical characteristics and prognosis, while controlled trials will evaluate the impact of specific interventions. These studies will record demographics, clinical features, and treatment details, with long-term follow-up to gather prognostic data. Statistical methods will establish prognostic models to inform treatment decisions.

Routine blood tests will be analyzed to examine the relationship between traditional liver function markers and HCC prognosis. Machine learning algorithms, such as decision trees and support vector machines, will identify new prognostic indicators. Prognosis prediction models will be developed and validated to enhance clinical decision-making. Bayesian networks will analyze the effects of immune cells, including NK and CD8+ T cells, and their cytokines on HCC progression. These models will clarify the causal relationships between immune cells and cytokines, while random forest algorithms will identify factors related to liver cancer recurrence, aiding in immunotherapy strategies.

Multicenter prospective studies with paired pathological specimens and clinical data will assess the prognostic value of immunohistochemical markers. Longitudinal changes in immunohistochemical profiles will be evaluated to predict treatment responses and survival outcomes. This will lead to the development of multimodal prognostic models integrating radiomic and biomarker data to support HCC management and therapy targeting.

Multimodal approaches have great potential, but their practical use faces challenges, especially in low-income regions. Key issues include limited access to clinical parameters like immune cell subsets and immunochemical markers due to a lack of advanced infrastructure and expertise. Many high-cost biomarker tests are difficult to implement in such settings. For example, detecting immune cell subsets and analyzing immunochemical markers require specialized labs and personnel, which may not be available in resource-limited areas. The complexity of high-dimensional data also demands sophisticated technology, often beyond the reach of low-resource settings. To address these issues, clinicians and researchers should consider simpler, cost-effective alternatives like basic blood tests, imaging, or models based on clinical data for disease prediction and personalized treatment. While these methods may lack the precision of comprehensive multimodal analyses, they can still provide valuable insights.

CONCLUSION

This minireview synthesizes evidence on HCC prognostic factors spanning clinical, hematologic, immunologic, imaging, and immunohistochemical domains. Integrating these parameters via multimodal phenomic analysis enhances prognosis prediction, particularly for immunotherapy outcomes. Elucidating the roles of immune cells and immunohistochemical markers deepens our understanding of this emerging modality and supports personalized treatment planning. Furthermore, this synthesis guides future research in developing robust prognostic models using diverse statistical and computational approaches.