Liver Cancer Immunity

Hepatocellular carcinoma (HCC) is the most common liver cancer, accounting for over 80%–90% of primary liver cancers. HCC is also a leading cause of cancer-specific death, estimated to be the fourth most common cause of cancer-related mortality worldwide.⁽¹⁾ Cholangiocarcinoma (CCA) is the second-most common primary hepatic malignancy after HCC. Although the overall incidence of CCA has increased over recent decades, the 5-year overall survival (OS) remains less than 10%.⁽²⁾ Potentially curative surgical resection or liver transplantation are options for the small subset of patients with early-stage disease. However, most patients with primary liver cancer present with advanced stage disease not amenable to surgical options. Systemic therapies for advanced stage liver cancer have limited efficacy. Hence, there is an essential need for the development of effective medical therapies.

Immuno-oncology has transformed cancer treatment over the past decade. Immune checkpoints are essential in the maintenance of self-tolerance under physiologic conditions.⁽³⁾ Tumors co-opt the antitumor immune response by activating immune checkpoints such as programmed death-1 (PD-1) and its ligand PD-L1, as well as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4). Immune checkpoint inhibition (ICI), using monoclonal antibodies targeting PD-1/PD-L1 and CTLA-4, unleashes preexisting immunity, especially effector CD8⁺ T cells.⁽³⁾ ICI therapies have had substantial benefit with durable responses for a subset of patients. However, most patients do not respond to ICI monotherapy. The tumor immune microenvironment (TIME) affects response to immunotherapies. TIMEs are broadly classified into T cell infiltrated-excluded, infiltrated-inflamed, or infiltrated with tertiary lymphoid structures.⁽⁴⁾ Hence, an enhanced understanding of the immunobiology of TIME will be essential in the development of next-generation immunotherapies.⁽⁴⁾

Immune Ecosystem of Human Liver Cancer

Under physiologic conditions, the liver has a complex microenvironment with constant exposure to gut-derived antigens from dietary and microbial products.⁽⁵⁾ Accordingly, the liver has intrinsic immune tolerogenicity that enables suppression of inappropriate inflammatory responses. Hence, the liver is an immune modulating organ that is poised to respond to deleterious stimuli while maintaining immuno-tolerance.⁽⁶⁾ This closely regulated hepatic immune tolerogenic network is deranged in chronic inflammatory liver disease, facilitating liver tumor development. The TIME of HCC and CCA is complex with diverse populations of innate and adaptive immune cells that affect cancer immune evasion, response to immunotherapy, and patient survival. Recent studies using single-cell analysis have begun to uncover the dynamics of immune cell subsets in tumor ecosystems.

ADAPTIVE AND INNATE LYMPHOCYTES IN LIVER CANCER

Antitumor Lymphocytes

Cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells are antitumor immune cells that play an integral role in cancer immune surveillance and eradication of tumor cells. CD8⁺ T lymphocytes are the primary effector tumor infiltrating lymphocyte (TIL) subset in liver cancer.⁽⁶⁾ Moreover, an increase in CD8⁺ TILs is associated with improved overall patient survival in both HCC and CCA.^(7,8) NK cells are innate lymphocytes with potent cytolytic function that can eradicate tumor cells without prior sensitization.⁽⁹⁾ A decrease in NK cell number or impairment of function occurs in a variety of malignancies including liver cancer. Several factors contribute to NK cell dysfunction in HCC, including up-regulation of the inhibitory receptor NKG2A and increased production of immunosuppressive cytokines such as transforming growth factor β (TGFβ) and IL-10.^(10,11) Additionally, an increased infiltration of CD11b⁻CD27⁻ cells, an immature and inactive phenotype of NK cells with poor cytolytic activity, occurs in HCC. Infiltration of this subset is associated with human HCC progression.⁽¹²⁾ In contrast, high expression of the activating NK cell receptor, natural killer group 2D, correlates with improved disease-free and overall patient survival in CCA.⁽¹³⁾ Preclinical studies using NK cell–based therapies such as infusion of ex vivo expanded human NK cells into CCA xenograft mice have demonstrated an antitumor effect in CCA.⁽¹⁴⁾

Dysregulation of lipid metabolism in nonalcoholic fatty liver disease (NAFLD) alters the adaptive immune response and promotes hepatocarcinogenesis.⁽¹⁵⁾ In preclinical models of NAFLD, lipid dysregulation causes selective loss of CD4⁺ T lymphocytes but not CD8⁺ T lymphocytes. Moreover, fewer CD4⁺ T lymphocytes are found in patients with alcohol-associated and nonalcoholic steatohepatitis than in patients with viral hepatitis, implying that impairment of the adaptive immune response is essential in the progression from NAFLD to hepatocarcinogenesis.⁽¹⁵⁾

Immunosuppressive Lymphocytes

Accumulation of regulatory T cells (Tregs), lymphocytes with a highly immunosuppressive nature that suppress CTLs, occurs in HCC. Increased tumor infiltration of Tregs is associated with CD8⁺ T cell dysfunction, HCC invasiveness as well as progression, and poor patient outcomes.^(16,17) Tregs are also associated with poor recurrence-free survival in CCA.⁽¹⁸⁾ Moreover, Tregs can attenuate the effectiveness of potential antitumor immunotherapies. In a mouse model of HCC, IL-12 therapy resulted in tumor stabilization or regression in 40% of animals. Activation of immunosuppressive mechanisms including increased abundance of Tregs was noted in nonresponder mice.⁽¹⁹⁾ TGFβ has been implicated in promoting Treg production and differentiation with consequent CD8⁺ T cell repression.⁽²⁰⁾ Accordingly, TGFβ inhibition using a specific inhibitor, SM-16, decreased Treg infiltration with resultant HCC tumor regression.⁽²⁰⁾

Dynamics of Adaptive and Innate Lymphocytes in Liver Cancer

Single-cell transcriptomic analysis of tumor-infiltrating lymphocytes has become a powerful tool to dissect the role of these cells in the highly complex TIME of various malignancies. Single-cell RNA sequencing of 5,063 single T cells from 6 patients with HCC identified 11 T cell subsets including five CD8⁺ subsets and six CD4⁺ subsets.⁽²¹⁾ Tregs and exhausted CD8⁺ T cells were preferentially enriched in HCC tumors. A signature gene, layilin (LAYN), was up-regulated on activated CD8⁺ T cells as well as Tregs, and was associated with more repressive and stable Tregs.⁽²¹⁾ Further functional analysis demonstrated a regulatory function of LAYN with inhibition of interferon-gamma (IFN-γ) production and consequent repression of CD8⁺ T-cell function.⁽²¹⁾

Intratumoral heterogeneity is linked to poor patient outcomes, particularly in patients with tumors that are refractory to molecularly targeted chemotherapeutics. Single-cell RNA sequencing analysis of tumors from 19 patients with primary HCC and intrahepatic CCA found an association between tumor transcriptomic diversity and OS in liver cancer.⁽²²⁾ T-cell dysfunction as indicated by lower cytolytic activity was noted in tumors with higher heterogeneity, suggesting that assessment of the tumor ecosystem may help guide response to immune therapy. Increased expression of tumor-derived vascular endothelial growth factor (VEGF) was related to TIME polarization including T-cell dysfunction, providing a mechanistic rationale for the combination of anti-VEGF and ICI in liver cancer.

IMMUNOSUPPRESSIVE MYELOID CELLS IN LIVER CANCER

Macrophages

Tumors can co-opt myeloid cells in the TIME to modulate immune evasion and promote cancer growth.⁽²³⁾ Hence, macrophages and dendritic cells (DCs), which can have antitumor functions, may be skewed by tumors to an immunosuppressive, pro-tumor phenotype (Fig. 1). The presence of immunosuppressive tumor-associated myeloid cells, such as tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs), influences patient survival. Accordingly, the presence of TAMs is linked to poor outcomes in patients with CCA.⁽¹⁸⁾ An extensive immunogenomic analysis of 10,000 tumors across 33 diverse cancer types revealed six immune subtypes of cancer based on differences in macrophage or lymphocyte signatures.⁽²⁴⁾ HCCs (n = 6) were classified under the lymphocyte-depleted subtype with a prominent macrophage signature characterized by a Th1-suppressed and high M2 macrophage response. A high TAM state correlates with poor patient outcomes in HCC based on the Cancer Genome Atlas (TCGA) HCC (LIHC) analysis.⁽²⁵⁾

FIG. 1.

The immune ecosystem of human liver cancer. A tumor-suppressive immune ecosystem is characterized by the presence of immune-stimulatory cell types such as CD8⁺ T lymphocytes, NK cells, and DCs as well as increased IFN-γ and granzyme B production. A tumor-progressive immune ecosystem is characterized by the presence of immunosuppressive MDSC-like macrophages, TAM-like macrophages, LAMP3⁺ DCs, LAYN⁺ regulatory T cells (Tregs), and dysfunctional LAYN⁺ CD8 T cells.

Transcriptome profiling of 75,000 individual CD45⁺ cells from 16 patients with HCC revealed the dynamic nature of various immune cells in HCC.⁽²⁵⁾ Two distinct states of tumor-enriched macrophages were identified: MDSC-like macrophages and TAM-like macrophages. MDSC-like cells had a high expression of S100A family genes FCN1 and VCAN, and low expression of HLA-related genes. In contrast, the gene signature found on TAM-like macrophages resembled a previously described signature in lung cancer TAMs with high expression of APOE, C1QA, C1QB, and TREM2. Moreover, TAM-like cells in HCC had high expression of two additional genes: SLC40A1, which encodes ferroportin, and GPNMB, which encodes type 1 membrane glycoprotein.⁽²⁵⁾ Based on the TCGA-LIHC analysis, both GPNMB and SLC40A1 were linked to poor patient prognosis. Functional assessment demonstrated an inflammatory role of GPNMB and SLC40A1 in TAM-like cells in the HCC TIME.

A few studies have explored the mechanistic basis of TAM-mediated liver cancer progression. Hepatocyte mitochondrial dysfunction and oxidative stress foster CCA carcinogenesis through activation of tumor necrosis factor (TNF)–producing Kupffer cells (KCs), resident hepatic macrophages.⁽²⁶⁾ KC-derived TNF lead to JNK-mediated cholangiocyte proliferation and oncogenic transformation, whereas KC depletion reduced premalignant CCA lesions. Canonical Wnt signaling drives cell proliferation, and is activated in liver cancer. In CCA, Wnt signaling is activated by TAMs with consequent tumor progression.⁽²⁷⁾ Accordingly, TAM depletion or inhibition of Wnt signaling reduced CCA proliferation and augmented apoptosis, with resultant tumor regression. There is also evidence that CCAs may commandeer macrophages to foster a tumor-supportive immune niche. Cellular spheroids generated from CCA cells educated macrophages to a TAM phenotype with invasive properties. These in vitro educated TAMs had a similar molecular phenotype to macrophages isolated from human resected CCA specimens.⁽²⁸⁾ These studies provide insight into the mechanistic underpinnings of TAM-mediated liver cancer progression, although further work is necessary to elucidate the role of TAMs and their interactions with other elements of the liver cancer TIME.

Dendritic Cells

DCs are antigen-presenting cells that activate the adaptive immune response by migrating to tumor-draining lymph nodes after acquiring tumor antigens and activating T cells. DCs are broadly categorized as conventional DCs (cDCs), which are highly phagocytic antigen-presenting cells or plasmacytoid DCs, which are not phagocytic and require activation.⁽²⁹⁾ Flow cytometry and immunohistochemical-based studies have demonstrated a correlation between mature cDCs and CD4⁺/CD8⁺ T-cell infiltration in CCA, and a decrease in the number of TNFα-producing cDCs in both HCC and CCA tumors.^(30,31) Based on single-cell analysis, three subsets of DCs enriched in HCC are cDC1, cDC2, and LAMP3⁺ DCs; the latter is a more mature form of DCs that migrate from tumors to lymph nodes.⁽²⁵⁾ LAMP3⁺ DC signature is associated with exhausted CD8⁺ T cell and Treg signature, suggesting that these DCs may be linked to T-cell dysfunction.⁽²⁵⁾ Accordingly, although cDCs with maturation features are typically associated with an enhanced CD8⁺ T-cell activation, these results signify the complex and dynamic nature of the TIME and the multiple mechanisms by which tumors can co-opt the TIME to evade immune surveillance.

Myeloid-Derived Suppressor Cells

MDSCs are immature myeloid cells that are pathologically activated in the setting of chronic inflammation. MDSCs have potent immunosuppressive properties and inhibit CTLs as well as NK cells through production of arginase, inducible nitric oxide synthase, indoleamine 2,3-dioxygenase, reactive oxygen species, TGFβ, and IL-10.⁽³²⁾ The frequency of MDSCs is significantly increased in human HCC and CCA, and their accumulation in the peripheral blood of patients with HCC is associated with decreased overall survival.^(33,34) MDSCs exert their immunosuppressive effects in liver cancer through induction of Tregs as well as attenuation of T-cell and NK-cell function.^(33,35,36) In preclinical tumors models, phosphodiesterase 5 (PDE5) inhibition diminishes tumor-immunosuppressive mechanisms through inhibition of MDSC function.⁽³⁷⁾ Adoptive cell transfer of cytokine-induced killer cells (CIKs) in mice bearing subcutaneous HCC tumors increased infiltration of MDSCs, reducing the antitumor effect of CIK immunotherapy.⁽³⁸⁾ Treatment with a PDE5 inhibitor enhanced antitumor efficacy of CIK immunotherapy by inhibiting MDSC accumulation and function.⁽³⁸⁾

DYSBIOSIS OF GUT MICROBIOTA AND LIVER CARCINOGENESIS

Under physiologic conditions, the liver is constantly exposed to the gut microbiome through the portal circulation. Consequently, chronic liver disease is associated with alteration in the gut microbiota or dysbiosis as well as increased translocation of intestinal bacteria. Impairment of the bile acid metabolism–microbiota crosstalk promotes inflammation and can contribute to the development of liver cancer.⁽³⁹⁾ Accumulation of gut-derived endotoxin or lipopolysaccharide (LPS) occurs in animal models of carcinogen-induced hepatocarcinogenesis.⁽⁴⁰⁾ Accordingly, antibiotic mediated LPS reduction or ablation of its receptor, toll-like receptor 4 (TLR4), had an antitumor effect. Reconstitution of TLR4-expressing myeloid cells in TLR4-deficient mice restored carcinogen-induced hepatic inflammation and proliferation. Similarly, TLR4 and the intestinal microbiota fostered HCC progression, whereas gut sterilization significantly reduced HCC tumor burden, particularly in the later stages of hepatocarcinogenesis (Fig. 2).⁽⁴¹⁾

FIG. 2.

Dysbiosis of gut microbiota and liver carcinogenesis. Intestinal microbiota modify bile acids (BA) and shift the balance between primary BA and secondary BA. Primary BA can regulate chemokine levels and hence influence accumulation of anti-tumor natural killer T cells (NKT). Intestinal microbiota also foster HCC progression via enhanced lipopolysaccharide (LPS)-toll-like receptor 4 (TLR4) signaling, whereas antibiotics or TLR antagonists reduce HCC tumor burden. Abbreviations: BA, bile acid; CXCL16, chemokine (C-X-C motif) ligand 16; CXCR6, cysteine-X-cysteine receptor 6; and LSEC, liver sinusoidal endothelial cell.

The commensal microbiome influences T-cell response and immune checkpoint inhibitor efficacy in melanoma.⁽⁴²⁾ Alteration of commensal gut bacteria induced an antitumor effect in a mouse model of primary liver cancer through a selective increase in hepatic CXCR6⁺ natural killer T (NKT) cells.⁽⁴³⁾ Vancomycin-mediated elimination of gram positive bacteria increased the conversion of primary bile acids to secondary bile acids. This increased expression of C-X-C motif chemokine ligand 16 on liver sinusoidal endothelial cells with consequent augmentation of antitumor immunity through NKT cell recruitment (Fig. 2). The finding that alteration of the commensal gut bacteria has a liver-selective antitumor effect indicates that therapies modulating the intestinal microbiome may be a potential treatment option for liver cancer.

Immunotherapy in Liver Cancer

TARGETING IMMUNE CHECKPOINTS IN HCC

Single-Agent Activity of Antibodies Targeting CTLA-4

The anti-CTLA-4 antibody, tremelimumab, was evaluated in a single-arm phase 2 study of 21 patients with advanced HCC, Child-Pugh A or B, and hepatitis C.⁽⁴⁴⁾ Grade 3 aspartate aminotransferase and alanine aminotransferase elevations were noted in 45% and 25% of patients, respectively, but were most often noted to occur early after the first dose and not associated with a parallel decline in liver function. The objective response rate (ORR) was 17% in the 17 patients who were evaluable for response, and another 58.8% of patients had stable disease as their best response. In an intention-to-treat analysis including all 21 patients, the median time to progression was 6.48 months (95% confidence interval [CI] 3.95–9.14) and median OS was 8.2 months (95% CI 4.64–21.34).

Single-Agent Activity of Antibodies Targeting PD-1

Nivolumab, a fully human immunoglobulin G4 monoclonal anti-PD-1 antibody, was first tested for its safety and efficacy in a phase 1/2 study of over 250 patients with advanced HCC, CheckMate 040.⁽⁴⁵⁾ The ORR per RECIST 1.1 criteria was 20%; responses occurred in all cohorts, independent of etiology or previous exposure to sorafenib. A favorable safety profile was confirmed in the expansion cohorts with grade 3 and 4 treatment-related adverse events (AEs) at 19%, consisting mostly of transient laboratory abnormalities. Nivolumab was subsequently granted accelerated approval by the Food and Drug Administration (FDA) for the treatment of patients with advanced HCC who had previously been treated with sorafenib in September 2017. The safety and tolerability of single-agent anti-PD-1 therapy noted with nivolumab was duplicated in Keynote 224, a phase 2 study of pembrolizumab in patients with advanced HCC; this study showed an ORR of 17% (95% CI 11–26) in 104 patients.⁽⁴⁶⁾ This study used a detailed and centralized review of hepatic events and reported a 3% rate of hepatic immune-mediated events.⁽⁴⁶⁾ Camrelizumab, another anti-PD-1 therapy, was assessed in a multicenter, open-label, randomized phase 2 trial in China (n = 303).⁽⁴⁷⁾ The ORR with camrelizumab was 14.7% (95% CI 10.3–20.2), and grade 3 or 4 treatment-related AEs occurred in 22% of the 217 patients who received camrelizumab. Several other single-agent ICI targeting the PD-1/PD-L1 axis, including atezolizumab (NCT04157985), durvalumab (NCT03847428), toripalimab (NCT03949231) and tislelizumab (NCT03412773), are currently under investigation in phase 3 trials of HCC.

Nivolumab and pembrolizumab have also been evaluated in phase 3 trials. Keynote 240 was a randomized phase 3 study of pembrolizumab versus best supportive care in patients who failed sorafenib, with co-primary endpoints of progression-free survival (PFS) and OS.⁽⁴⁸⁾ Despite a numerically superior median OS for pembrolizumab versus placebo (13.9 [11.6–16] months vs. 10.6 [8.3–13.5] months; P = 0.0238), the trial did not reach statistical significance, as the prerequired P value to reach significance was 0.0174. In CheckMate 459, nivolumab was compared with sorafenib in systemic therapy–naïve patients with advanced HCC.⁽⁴⁹⁾ The median OS was 16.4 (13.9–18.4) months versus 14.7 months (11.9–17.2), respectively, with a hazard ratio (HR) of 0.85 (0.72–1.02). When considering subsequent therapies, 20% of patients on the sorafenib arm received a checkpoint inhibitor in second line, and another 11% received investigational agents, some of which were immune-oncology agents. Grade 3–4 treatment– related AEs were noted in 22% of patients on the nivolumab arm and 49% of patients on the sorafenib arm. There was clinically meaningful difference in quality-of-life outcomes favoring nivolumab based on the FACT-hep disease-specific questionnaire. In summary, despite the fact that single-agent anti PD-1 agents manifested consistent ORRs between 14% and 20% with durable responses in phase 1/2 trials, the phase 3 studies of both nivolumab in first line and pembrolizumab in second line failed to show statistically significant improvements in OS. The rapidly evolving landscape of available systemic therapies for HCC being used in a sequential manner and some statistical design limitations may have affected the outcomes of these trials. Nonetheless, despite the evidence of clinical benefit in some patients, it is plausible that the single-agent activity was not sufficient to show significant improvements in median OS in an unselected population. As a result, the development of biomarkers to improved patient selection and the pursuit of combination therapies present possible solutions to harness the full potential of checkpoint inhibition in HCC.

Emerging Biomarkers for Single-Agent Anti-PD-1 Antibodies

In patients with advanced HCC treated with single-agent anti-PD-1 therapy, several potential biomarkers may be associated with improved survival and response. In CheckMate 040, patients with advanced HCC received nivolumab monotherapy, regardless of PD-L1 status. Although response was noted in both PD-L1 positive and negative patients, the response rate in patients with increased tumor cell PD-L1 expression was significantly higher (complete response [CR] + partial response [PR] versus stable disease [SD]; P = 0.00009).⁽⁵⁰⁾ Moreover, CD3 expression was also significantly associated with response (CR + PR vs. SD; P = 0.05). In the subset of patients with evaluable tumor samples (n = 37), several inflammatory signatures such as the IFN-γ-related mRNA profile and T-cell exhaustion signature had a significant correlation with improved response and OS with anti-PD-1 therapy.^(50,51) In contrast, Wnt/β-catenin mutations such as activating catenin beta-1 (CTNNB1) or inactivating AXIN1 mutations are associated with poor outcomes in patients with HCC treated with anti-PD-1.⁽⁵²⁾ In summary, this continues to be an area of active investigation; it is possible that a composite of multiple biomarkers may be useful in improving patient selection.

COMBINATORIAL APPROACHES WITH IMMUNE CHECKPOINT INHIBITORS

Targeting PD-1/PD-L1 Plus VEGF Axis

Anti-VEGF therapy decreases the activity of MDSCs and Tregs, enhances PD-L1 expression, and increases CTL infiltration.⁽⁵³⁾ Accordingly, single-agent VEGF inhibition has shown modest activity in HCC.⁽⁵⁴⁾ Moreover, preclinical studies have demonstrated a synergistic antitumor effect with combinatorial blockade of PD-1 and the VEGF axis.⁽⁵⁵⁾ Imbrave150 (NCT03434379) is a landmark phase 3 clinical trial comparing the PD-L1 inhibitor atezolizumab plus bevacizumab (n = 336), a VEGF inhibitor, with sorafenib (n = 165) in treatment-naïve patients with unresectable HCC.⁽⁵⁶⁾ Trial inclusion criteria limited eligibility to patients who were Child-Pugh A. Primary analysis data demonstrated that with a median follow-up of 8.6 months, there was a 42% reduction in risk of death with the combination group, with an OS HR of 0.58 (95% CI: 0.42–0.79; P = 0.0006). Median PFS was 6.8 in combination versus 4.3 months in sorafenib (HR 0.59; 95% CI: 0.47–0.76; P < 0.0001). Grade 3–4 AEs were reported in 57% of patients receiving the combination and 55% of patients receiving sorafenib. Although these are potentially paradigm-shifting results, as this is the first study to show superiority of a first-line systemic therapy over sorafenib, mature data and longer follow-up are needed. Moreover, safety in a real-world population will also need to be established, and further studies will be necessary to determine whether this combination is safe in patients with more advanced liver disease who have portal hypertension and an increased bleeding risk.

Lenvatinib, a multikinase inhibitor that inhibits VEGF receptor (VEGFR) 1–3, is FDA-approved as first-line therapy in patients with unresectable HCC based on the multicenter, randomized, open-label, noninferiority REFLECT trial.⁽⁵⁷⁾ The combination of pembrolizumab and lenvatinib was associated with an ORR of 36.7%, with a median PFS of 9.7 months in unresectable HCC.⁽⁵⁸⁾ Grade 3–4 treatment-related AEs were noted in 73% of patients, with 4 (13%) patients experiencing fatal AEs. A phase 3 study of lenvatinib plus pembrolizumab versus lenvatinib alone in patients with advanced unresectable HCC is currently ongoing (LEAP-002; NCT03713593) (Table 1).

TABLE 1.

Ongoing Clinical Trials of Immunotherapy Targeting the Tumor Microenvironment in Patient With Liver Cancer

Intervention	Target	Trial Type	Trial Description
Immune checkpoint Inhibition plus tyrosine kinase inhibition
HCC
Atezolizumab + bevacizumab	PD-L1; VEGF	Phase 3, randomized, open label	Advanced or metastatic, treatment-naïve HCC; Child-Pugh A (480 pts)	NCT03434379
Durvalumab + bevacizumab	PD-L1; VEGF	Phase 2, randomized, open label, multicenter	Advanced HCC (433 pts)	NCT02519348
HLX10 + HLX04	PD-1; VEGF	Phase 2, single arm, open label, multicenter	Advanced HCC with progression or intolerant toxicity after standard treatment; BCLC stage B/C (150 pts)	NCT03973112
Sintilimab + IBI305	PD-1; VEGF	Phase 1b, singe arm, open label	Advanced HCC; BCLC stage B/C; Child-Pugh score ≤ 7 (45 pts)	NCT04072679
		Phase 2/3, randomized, open label, multicenter	First-line treatment of advanced HCC; BCLC stage B/C; Child-Pugh score ≤ 7 (566 pts)	NCT03794440
Nivolumab + sorafenib	PD-1; multiple kinases	Phase 3, randomized, open label, multicenter	Advanced, treatment-naïve HCC; Child-Pugh A (1,723 pts)	NCT02576509
		Phase 2, nonrandomized, multicenter	Advanced, treatment-naïve HCC (40 pts)	NCT03439891
Nivolumab + lenvatinib	PD-1; multiple kinases	Phase 2, single arm, open label, multicenter	Advanced HCC; Child-Pugh ≤ 6 (50 pts)	NCT03841201
		Phase 1b, nonrandomized, open label	Advanced HCC; Child-Pugh A (30 pts)	NCT03418922
Nivolumab + cabozantinib	PD-L1; VEGFR2/c-MET	Phase 1/2, randomized, open label	Advanced HCC with or without chronic viral hepatitis; Child-Pugh B (1,097 pts)	NCT01658878
Nivolumab + regorafenib	PD-1; VEGFR2/TIE2	Phase l/2a, single arm, open label	Advanced HCC with progression after sorafenib (69 pts)	NCT04170556
		Phase 2, single arm, open label	Chemotherapy-naïve, unresectable or metastatic HCC; Child-Pugh A (42 pts)	NCT04310709
Pembrolizumab + sorafenib	PD-1; multiple kinases	Phase 1b/2, single arm, open label	Advanced or metastatic liver cancer; Child-Pugh A (27 pts)	NCT03211416
Pembrolizumab + lenvatinib	PD-1; multiple kinases	Phase 2, single arm, nonrandomized	Advanced hepatobiliary malignancies including HCC and BTC (50 pts)	NCT03895970
		Phase 1b, open label	Advanced HCC; BCLC stage B/C (150 pts)	NCT03006926
		Phase 3, randomized controlled, multicenter	Advanced, treatment-naïve HCC; BCLC stage B/C (750 pts)	NCT03713593
Pembrolizumab + regorafenib	PD-1; VEGFR2/TIE2	Phase 1b, nonrandomized, open label	Advanced, treatment-naïve HCC; BCLC stage B/C (57 pts)	NCT03347292
Camrelizumab + apatinib	PD-1; VEGFR2	Phase 2, single arm, open label	Advanced-stage HCC; BCLC stage B/C (40 pts);	NCT04014101
		phase 2, single arm, open label	unresectable HCC (30 pts)	NCT03793725
		Phase 3, randomized, open label, multicenter	Advanced, treatment-naïve HCC; BCLC stage B/C (510 pts)	NCT03764293
		Phase 2 , single arm, open label	Advanced HCC; Child-Pugh A (190 pts)	NCT03463876
Sintilimab + lenvatinib	PD-1; multiple kinases	Phase 2, single arm, open label	Local advanced HCC; BCLC stage B/C; Child-Pugh score ≤ 7 (56 pts)	NCT04042805
Sintilimab + anlotinib	PD-1; multiple kinases	Phase 2, single arm, open label	Advanced HCC; BCLC stage B/C; Child-Pugh A (20 pts)	NCT04052152
Toripalimab + axitinib	PD-1; multiple kinases	Phase 2, single arm, open label, single center	Advanced hepatobiliary malignancies including HCC and CCA (60 pts)	NCT04010071
Toripalimab + sorafenib	PD-1; multiple kinases	Phase 1/2, single arm, open label	Unresectable HCC with portal vein tumor thrombus; Child-Pugh score ≤ 7 (39 pts)	NCT04069949
Tislelizumab + regorafenib	PD-1; VEGFR2/TIE2	Phase 2, randomized, open label	Advanced HCC (125 pts)	NCT04183088
AK105 + anlotinib	PD-1; multiple kinases	Phase 1b/2, open label, multicenter	Unresectable HCC; BCLC stage B/C; Child-Pugh score ≤ 7 (30 pts)	NCT04172571
Durvalumab + tivozanib	PD-L1; VEGFR1–3	Phase 1b/2, open label	Advanced, treatment-naïve HCC; Child-Pugh A (42 pts)	NCT03970616
Durvalumab + cabozantinib	PD-L1; VEGFR2/c-MET	Phase 1b, single arm, open label	Advanced gastroesophageal cancer including HCC (30 pts)	NCT03539822
Atezoiizumab + cabozantinib	PD-L1; VEGFR2/c-MET	Phase 3, randomized controlled	Advanced, treatment-naïve HCC; BCLC stage B/C (740 pts)	NCT03755791
CCA
Pembrolizumab + lenvatinib	PD-1; multiple kinases	Phase 2, single arm, open label	Advanced, refractory, primary liver cancer or BTC (50 pts)	NCT03895970
		Phase 2, single arm, open label	Advanced refractory solid tumors including BTC; Child-Pugh score 5–6 (180 pts)	NCT03797326
Pembrolizumab + ramucirumab	PD-1; VEGFR2	Phase 1, single arm, open label	Advanced refractory, biopsiable cancers including BTC (155 pts)	NCT02443324
Camrelizumab + apatinib	PD-1; VEGFR2	Phase 2, nonrandomized, open label	Advanced primary liver cancer including BTC; Child-Pugh A or B (152 pts)	NCT03092895
Sintilimab + anlotinib + gemcitabine + cisplatin	PD-1; multiple kinases; chemotherapy	Phase 2a, randomized controlled, multicenter	Unresectable or metastatic BTC (80 pts)	NCT04300959
TQB2450 + anlotinib	PD-L1; multiple kinases	Phase 1b/2, single arm, open label	Advanced, refractory BTC or HCC (60 pts)	NCT03825705
Avelumab + regorafenib	PD-L1; VEGFR2/TIE2	Phase 1/2, nonrandomized, open label	Advanced, refractory digestive tumors, not MMR-deficient (212 pts)	NCT03475953
Immune microenvironment-targeted therapy
HCC
Nivolumab + galunisertib	PD-1; TGF-β inhibitor	Phase 1b/2, nonrandomized, single arm, open label	Advanced refractory solid tumors including HCC (75 pts)	NCT02423343
Nivolumab + BMS-986205	PD-1; IDO1 Inhibitor	Phase 1/2, single arm, open label	Advanced HCC; Child-Pugh A (23 pts)	NCT03695250
Nivolumab + cabiralizumab	PD-1; CSF1R	Phase 2, randomized controlled	Advanced HCC; Child-Pugh score ≤ 7 (74 pts)	NCT04050462
CCA
Pembrolizumab + sargramostim	PD-1; GM-CSF	Phase 2, single arm, open label	Advanced biliary cancers (42 pts)	NCT02703714
Nivolumab + cabiralizumab	PD-1; CSF1R	Phase 2, randomized, open label	Resectable, biopsiable BTC (16 pts)	NCT03768531
Nivolumab + entinostat	PD-1; histone deacety-lase inhibitor	Phase 2, nonrandomized, open label	Advanced, treatment-naïve CCA or PDAC (54 pts)	NCT03250273
ABBV-181 + ABBV-368	PD-1; CD40 agonist	Phase 1, nonrandomized, open label	Advanced solid cancers including CCA (170 pts)	NCT03071757

Note: A www.ClinicalTrials.gov search was performed using the terms “liver cancer,” “hepatocellular carcinoma,” “liver neoplasm” for HCC; “biliary tract cancer,” “cholangiocarcinoma,” “biliary carcinoma,” “bile duct,” or “biliary tract” for CCA; and “hepatobiliary malignant tumors” for both HCC and CCA. The search identified immunotherapy trials with a status of “Recruiting,” “Not yet recruiting,” “Active, not recruiting,” or “Enrolling by invitation,” and trials without the inclusion of a specific liver cancer cohort or without adequate information available were excluded. Search was updated as of March 30, 2020.

Abbreviations: BCLC, Barcelona Clinic Liver Cancer; CD40, cluster of differentiation 40; CSF1R, colony stimulating factor 1 receptor; GM-CSF, granulocyte-macrophage colony-stimulating factor; IDO1, indoleamine 2,3-dioxygenase 1; and PDAC, pancreatic ductal adenocarcinoma.

Regorafenib, a multikinase inhibitor with activity against kinases involved in angiogenesis (VEGFR 1–3) and oncogenesis (RET, c-RAF/RAF-1), is FDA-approved as a second-line agent for patients with advanced HCC who have failed sorafenib treatment.⁽⁵⁹⁾ Cabozantinib is another tyrosine kinase inhibitor (TKI) that is approved as a second-line and third-line agent for advanced HCC.⁽⁶⁰⁾ Combinatorial regimens of these TKIs plus ICI are currently under investigation in several ongoing clinical trials (Table 1).

Although these combinatorial regimens have significant antitumor potential, their efficacy may be limited by AEs. Unlike bevacizumab, which specifically targets VEGF, TKIs targeting the VEGF axis are multikinase inhibitors with a greater potential for off-target effects and higher rate of AEs. A meta-analysis of bevacizumab versus VEGFR TKIs in patients with metastatic colorectal cancer demonstrated higher rates of AEs group (diarrhea, fatigue, thrombocytopenia, neutropenia, and hypertension) with VEGFR TKIs compared with bevacizumab.⁽⁶¹⁾ We eagerly await the results of ongoing clinical trials in advanced HCC to determine the efficacy as well as safety of these combinatorial regimens.

Targeting PD-1/PD-L1 Plus CTLA-4

The safety and efficacy of combination ipilimumab, an anti-CTLA-4 antibody, and nivolumab was assessed in sorafenib-treated patients with advanced HCC in a CheckMate 040 cohort.⁽⁶²⁾ Based on an ORR of 32% and median OS of 23 months, the FDA granted accelerated approval to combination ipilimumab 3 mg/Kg and nivolumab 1 mg/Kg in March 2020 for treatment of advanced HCC previously treated with sorafenib. This combination did have a relatively high rate of grade 3–4 treatment-related AEs at 53%, and 57.1% of patients with immune-mediated events required steroids. Several ongoing clinical trials are investigating the safety and efficacy of dual immune checkpoint inhibition in human liver cancer (Table 2).

TABLE 2.

Ongoing Clinical Trials of Dual Immune Checkpoint Inhibition in Patients With Liver Cancer

Intervention	Target	Trial Type	Trial Description	ClinicalTrials.gov Identifier
Dual immune checkpoint inhibition
HCC
Nivolumab + ipilimumab	PD-1; CTLA-4	Phase 3, randomized, multicenter	Advanced, treatment-naïve HCC; Child-Pugh score 5 or 6 (1,084 pts)	NCT04039607
		Phase 1/2, randomized, open label	Advanced HCC; Child-Pugh B (1,097 pts)	NCT01658878
Durvalumab + tremelimumab	PD-L1; CTLA-4	Phase 3, randomized, open label, multicenter	Treatment-naïve, unresectable HCC; BCLC stage B/C (1,310 pts)	NCT03298451
		Phase 2, randomized, open label, multicenter	Advanced HCC (433 pts)	NCT02519348
		Phase 2, two arms, open label	Intermediate-stage HCC (30 pts)	NCT03638141
Nivolumab + relatlimab	PD-1; LAG-3	Phase 1/2a, randomized, open label	Advanced, treatment-naïve solid tumors including HCC (1,500 pts)	NCT01968109
TSR-042 + TSR-022	PD-1; TIM-3	Phase 2, single arm, open label	Advanced or metastatic HCC; BCLC stage B/C (42 pts)	NCT03680508
CCA
Nivolumab + ipilimumab	PD-L1; CTLA-4	Phase 2, single arm, open label	Advanced, refractory solid tumors including BTC (707 pts)	NCT02834013
		Phase 2, randomized, open label	Unresectable, treatment-naïve BTC; Child-Pugh A (64 pts)	NCT03101566
Durvalumab + tremelimumab	PD-L1; CTLA-4	Phase 1, nonrandomized, open label	Advanced, refractory, biopsiable solid tumors including BTC (269 pts)	NCT01938612

Note: A www.ClinicalTrials.gov search was performed using the terms “liver cancer,” “hepatocellular carcinoma,” “liver neoplasm” for HCC; “biliary tract cancer,” “cholangiocarcinoma,” “biliary carcinoma,” “bile duct,” or “biliary tract” for CCA; and “hepatobiliary malignant tumors” for both HCC and CCA. The search identified immunotherapy trials with a status of “Recruiting,” “Not yet recruiting,” “Active, not recruiting,” or “Enrolling by invitation,” and trials without the inclusion of a specific liver cancer cohort or without adequate information available were excluded. Search was updated as of March 30, 2020.

Abbreviations: LAG-3, lymphocyte-activation gene 3; TIM-3, T-cell immunoglobulin mucin-3.

IMMUNOTHERAPY BEYOND CHECKPOINT INHIBITORS

Cell-based immunotherapies such as adoptive cell transfer with TILs or CIKs have potent antitumor effects with little cytotoxicity to normal cells. The efficacy and safety of activated CIKs was assessed in the adjuvant setting in 230 patients with HCC who had undergone surgical resection or radiofrequency ablation (RFA) or percutaneous ethanol injection.⁽⁶³⁾ Median PFS was 44 months in the immunotherapy group versus 30 months in the control group; OS was not reached in either group, but OS was longer in the CIK treatment group (HR: 0.21; 95% CI: 0.06–0.75; P = 0.008). Although overall AEs were significantly higher in the CIK group (62% vs. 41%; P = 0.002), the serious AE rate was similar (7.8% vs. 3.5%; P = 0.15).

Despite these encouraging results, isolating tumor-specific T cells from patients with cancer remains a challenge. Gene-modified T-cell therapy can overcome this limitation by engineering T cells targeting specific tumor antigens. These approaches include tumor antigen–specific T-cell receptor and chimeric antigen receptor technology, and are currently being evaluated in patients with solid tumors, including liver cancer (Table 3).

TABLE 3.

Ongoing Clinical Trials of Cell-Based Therapy in Patients With Liver Cancer

Intervention	Trial Type	Trial Description	ClinicalTrials.gov Identifier
Cell-based therapy
HCC
NK cells	Phase 1/2, nonrandomized	Advanced HCC; Child-Pugh A/B (200 pts)	NCT04162158
iNKT cells + CD8 T cells	Phase 1/2, single-arm, open label	Advanced solid tumor including HCC (40 pts)	NCT03093688
iNKT cells + TACE	Phase 2, randomized, open label	Advanced HCC; BCLC stage C (144 pts)	NCT04011033
CIKs	Phase 1, single arm, open label	Advanced solid tumor including HCC (24 pts)	NCT04282044
	Phase 2, randomized, single center	Advanced liver cancer; Child-Pugh score ≤ 7 (80 pts)	NCT03146637
CIKs + TACE	Phase 2, randomized, open label, multicenter	Intermediate stage HCC undergoing TACE; Child-Pugh B (78 pts)	NCT02856815
CIKs + RFA	Phase 3, randomized .open label	Recurrent HCC (210 pts)	NCT02678013
CIKs + CD3-MUC-1 + Cryotherapy	phase 2, randomized, single center	Advanced liver cancer, BCLC stage C; Child-Pugh score ≤ 7 (90 pts)	NCT03484962
	Phase 1, single arm, open label	Advanced GPC3-positive HCC; Child-Pugh score < 7 (30 pts)	NCT03198546
Anti-GPC3 CAR-T	Phase 1, single arm, open label	Advanced GPC3-positive HCC; Child-Pugh score < 7 (20 pts)	NCT04121273
	Phase 1, single arm, open label	Advanced GPC3-positive HCC; BCLC stage B/C (15 pts)	NCT03884751
	Phase 1, sequential assignment, open label	Advanced GPC3-positive HCC; BCLC stage B/C (36 pts)	NCT03980288
	Phase 1, single arm, open label	Advanced GPC3-positive HCC; Child-Pugh score < 7(14 pts)	NCT02905188
Anti-EpCAM CAR-T	Phase 1, single arm, open label	EpCAM positive solid tumor including HCC (60 pts)	NCT03013712
Anti-EGFR vlll CAR-T/TCR-T	Phase 1/2, single arm, open label, multicenter	EGFR positive solid tumor including HCC (50 pts)	NCT03941626
Anti-CD147-targeted CAR-T	Phase 1, single arm, open label	Advanced HCC; Child-Pugh score ≤ 7 (34 pts)	NCT03993743
Anti-NKG2D-based CAR T	Phase 1, single arm, open label	NKG2D positive relapsed/refractory solid tumor including HCC (10 pts)	NCT04270461
Autologous AFP-specific TCR T cells	Phase 1, single arm, open label	AFP positive HCC; Child-Pugh score ≤ 6 (24 pts)	NCT03132792
	Phase 1, single arm, open label	AFP positive, unresectable HCC; BCLC stage B/C; Child-Pugh score ≤ 6 (9 pts)	NCT03971747
NY-ESO-1-specific TCR T cells	Phase 1, nonrandomized, open label	Advanced solid tumor including HCC (22 pts)	NCT02869217
	Phase 1, single arm, open label	NY-ESO-1 expressing advanced cancer including HCC (43 pts)	NCT01967823
CCA
Autologous CIKs + RFA	Phase 2/3, nonrandomized, single blind	Unresected cholangiocarcinoma without extrahepatic metastasis (50 pts)	NCT02482454
Autologous CD8 T cells + pembrolizumab	Phase 1, single arm, open label	Advanced GI malignancies including CCA (40 pts)	NCT02757391
Autologous Tern cells + RFA/chemotherapy	Phase 2, randomized, open label	ICC after radical resection (20 pts)	NCT03820310
Autologous transfer TIL	Phase 2, single arm, open label	Unresectable, refractory BTC (59 pts)	NCT03801083
Autologous MUC-1 CAR T	Phase 1/2, single arm, open label	MUC-1 positive ICC (9 pts)	NCT03633773
MSLN-specific TCR T cells	Phase 1/2, single arm, open label	Advanced mesothelin-positive cancer including CCA (70 pts)	NCT03907852

Note: A www.ClinicalTrials.gov search was performed using the terms “liver cancer,” “hepatocellular carcinoma,” and “liver neoplasm” for HCC; “biliary tract cancer,” “cholangiocarcinoma,” “biliary carcinoma,” “bile duct,” or “biliary tract” for CCA; and “hepatobiliary malignant tumors” for both HCC and CCA. The search identified immunotherapy trials with a status of “Recruiting,” “Not yet recruiting,” “Active, not recruiting,” or “Enrolling by invitation,” and trials without the inclusion of a specific liver cancer cohort or without adequate information available were excluded. Search was updated as of March 30, 2020.

Abbreviations: AFP, alpha-fetoprotein; CAR-T, chimeric antigen receptor T cells; CD3, cluster of differentiation 3; CD8, cluster of differentiation 8; CD147, cluster of differentiation 147; EGFR, epidermal growth factor receptor; EpCAM, epithelial cell adhesion molecule; GPC3, glypican 3; HBV, hepatitis virus B; iNKT, invariant natural killer T cells; GMUC-1, mucin 1; NKG2D, killer cell lectin like receptor K1; NY-ESO-1, New York esophageal squamous cell carcinoma 1; TACE, transarterial chemoembolization; Tcm, csentral memory T; TCR, T cell receptor; TIL, tumor infiltrating lymphocytes.

IMMUNOTHERAPY IN EARLIER STAGES OF DISEASE

Immunotherapy in the neoadjuvant setting when a higher endogenous tumor antigen burden is present in the primary tumor can augment T-cell priming, and has the potential to eliminate micrometastasis, which would be the source of postsurgical tumor recurrence.⁽⁶⁴⁾ Several ongoing advanced phase trials assessing the safety and efficacy of ICI in the adjuvant/neoadjuvant setting are underway in advanced liver cancer (Table 4).

TABLE 4.

Ongoing Clinical Trials of Neoadjuvant/Adjuvant Immunotherapy in Patients With Liver Cancer

Intervention	Target	Trial Type	Trial Description	ClinicalTrials.gov Identifier
Neoadjuvant Immunotherapy
HCC
Pembrolizumab	PD-1	Phase 1, single arm, open label	Adjuvant or neoadjuvant treatment of resectable HCC (45 pts)	NCT04224480
		Phase 1, single arm, open label	Adjuvant or neoadjuvant therapy with curative treatment such as resection or RFA; Child-Pugh A (50 pts)	NCT03337841
Cemiplimab	PD-1	Phase 2, randomized, multi-cohort	Resectable HCC (94 pts)	NCT03916627
Nivolumab + ipilimumab	PD-1; CTLA-4	Phase 2, single arm, open label	HCC with potential for curative surgical resection; Child-Pugh A (40 pts)	NCT03510871
		Phase 2, randomized, open-label	HCC with potential for curative surgical resection (45 pts)	NCT03222076
		Phase 2, single-arm, open-label	Resectable HCC; Child-Pugh A (32 pts)	NCT03682276
Nivolumab + BMS-813160/BMS-986253	PD-1; CCR2/5-inhibitor or IL-8	Phase 2, randomized, open label	Resectable NSCLC/HCC (50 pts)	NCT04123379
Camrelizumab + apatinib	PD-1; VEGFR2	Phase 1/2, single arm, open label, multicenter	Downstaging/bridging of HCC before liver transplant; Child-Pugh A or B (120 pts)	NCT04035876
		Phase 2, single arm, open label	Perioperative treatment of HCC (20 pts)	NCT04297202
Sintilimab + TACE	PD-1; transarterial chemoembolization	Phase 2, single arm, open label	Early and intermediate stage HCC not suitable for primary resection; BCLC stage A/B (61 pts)	NCT04174781
CCA
Nivolumab + cabiralizumab	PD-1; CSF1R	Phase 2, randomized, open label	Resectable BTC (16 pts)	NCT03768531
Adjuvant Immunotherapy
HCC
Nivolumab	PD-1	Phase 2, single arm, open label, multicenter	Advanced HCC treated by electroporation (NIVOLEP) with curative intent; Child-Pugh A (50 pts)	NCT03630640
		Phase 3, randomized, double blinded	HCC at high risk of recurrence after curative hepatic resection or ablation; Child-Pugh score 5 or 6 (530 pts)	NCT03383458
Pembrolizumab	PD-1	Phase 3, randomized, double blinded, two-arm	HCC after surgical resection or local ablation; Child-Pugh A (950 pts)	NCT03867084
Toripalimab	PD-1	Phase 2/3, randomized-controlled, double-blind	Local advanced HCC after curative hepatic resection; Child-Pugh A (402 pts)	NCT03859128
Nivolumab + cabozantinib	PD-1; VEGFR2/c-MET	Phase 1, single arm, open label	Local advanced HCC after resection (15 pts)	NCT03299946
Nivolumab + PI 101	PD-1; ropeginterferon alfa-2b	Phase 1/2, randomized, open label	After curative surgery of HBV-related HCC (72 pts)	NCT04233840
Durvalumab + bevacizumab	PD-L1; VEGF	Phase 3, randomized-controlled, multicenter	HCC at high risk of recurrence after curative hepatic resection or ablation; Child-Pugh score of 5 or 6 (888 pts)	NCT03847428
Atezolizumab + bevacizumab	PD-L1; VEGF	Phase 3, randomized, open label, multicenter	HCC at high risk of recurrence after curative hepatic resection or ablation; Child-Pugh A (662 pts)	NCT04102098
HBV-specific TCR T cells	HBV antigen	Phase 1, single-arm, open label	Hepatitis B virus (HBV) related HCC after liver transplantation (10 pts)	NCT02686372
		Phase 1, single-arm, open label	Recurrent, HBV-related HCC after liver transplantation (10 pts)	NCT02719782
		Phase 1, single arm, open label, single center	HBV related HCC post hepatectomy or radiofrequency ablation; BCLC stage C (10 pts)	NCT03899415
		Phase 1/2, open label, single arm, multicenter	Recurrent, HBV-related HCC after liver transplantation (72 pts)	NCT03634683

Note: A www.ClinicalTrials.gov search was performed using the terms “liver cancer,” “hepatocellular carcinoma,” and “liver neoplasm” for HCC; “biliary tract cancer,” “cholangiocarcinoma,” “biliary carcinoma,” “bile duct,” or “biliary tract” for CCA; “hepatobiliary malignant tumors” for both HCC and CCA. The search identified immunotherapy trials with a status of “Recruiting,” “Not yet recruiting,” “Active, not recruiting,” or “Enrolling by invitation,” and trials without the inclusion of a specific liver cancer cohort or without adequate information available were excluded. Search was updated as of March 30, 2020.

Abbreviations: BTC, biliary tract cancer; CCR2/5, C-C chemokine receptor types 2/5; HBV, hepatitis virus B; NSCLC, non-small-cell lung carcinoma.

Locoregional therapy such as RFA, transarterial chemoembolization, radioembolization, or external radiation therapy is an option for treatment of unresectable HCC limited to the liver. The combination of locoregional therapy with ICI has garnered interest based on the premise that locoregional modalities such as RFA may mediate immunogenic cell death and release of antigens that promote a peripheral immune response.⁽⁶⁵⁾ This peripheral immune stimulation can theoretically be augmented by ICI. In a cohort of 32 patients with HCC, the anti-CTLA-4 antibody tremelimumab was administered in patients who underwent RFA. PR was noted in 5 of 19 evaluable patients, and increased intratumoral CD8⁺ T-cell accumulation was noted in patients with response.⁽⁶⁵⁾ The combination of ICI and locoregional therapy is currently under active investigation in several clinical trials of liver cancer (Table 5).

TABLE 5.

Ongoing Clinical Trials of Immune Checkpoint Inhibition in Combination With Locoregional Therapy in Liver Cancer

Intervention	Trial Type	Trial description	ClinicalTrials.gov Identifier
Immune checkpoint Inhibition plus locoregional therapy
HCC
Nivolumab (anti-PD-1) + Y90	Phase 2, nonrandomized, open label, single center	Advanced HCC; Child-Pugh A (40 pts)	NCT03033446
	Phase 1/2, one arm, open label	Advanced HCC; Child-Pugh A or B (35 pts)	NCT02837029
Nivolumab (anti-PD-1) + SIRT	Phase 2, single arm, open label, multicenter	HCC candidates for locoregional therapy (40 pts)	NCT03380130
	Phase 2/3, randomized, two arm	Intermediate stage HCC; Child-Pugh B or C (522 pts)	NCT04268888
Nivolumab (anti-PD-1) + TACE	Phase 1, nonrandomized, open label	Advanced HCC; Child-Pugh A (14 pts)	NCT03143270
	Phase 2, single arm, open label	Intermediate stage HCC; Child-Pugh A (49 pts)	NCT03572582
Pembrolizumab (anti-PD-1) + RFA/MWA	Phase 2, one arm, open label	Early-stage HCC; Child-Pugh score ≤6 (30 pts)	NCT03753659
Pembrolizumab (anti-PD-1) + SBRT	Phase 2, single arm, open label	Advanced HCC after treatment with sorafenib; Child-Pugh A (30 pts)	NCT03316872
Pembrolizumab (anti-PD-1) + TACE	Phase 1b, one arm, open label	Unresectable HCC; Child-Pugh score < 7 (26 pts)	NCT03397654
Sintilimab (anti-PD-1) + TACE ± MWA	Phase 1, single arm, open label	Unresectable HCC; BCLC stage B/C (45 pts)	NCT04220944
Toriplimab (anti-PD-1) + RFA/MWA	Phase 1/2, randomized, open label	Unresectable HCC; Child-Pugh A or B (120 pts)	NCT03864211
Durvalumab (anti-PD-L1) + TACE	Phase 3, randomized controlled, multicenter	Locoregional HCC; Child-Pugh A (600 pts)	NCT03778957
Durvalumab (anti-PD-L1) + tremelimumab (anti-CTLA-4) + TAC/RFA	Phase 2, nonrandomized, open label	Unresectable, refractory HCC or BTC; Child-Pugh A/B (90 pts)	NCT02821754
	Phase 2, single arm, open label	Unresectable HCC or BTC; Child-Pugh A (70 pts)	NCT03482102
	Phase 2, nonrandomized, open label
		Intermediate stage of HCC (30 pts)	NCT03638141
CCA
Pembrolizumab (anti-PD-1) + SBRT	Phase 2, randomized, open label	Unresectable, treatment-naïve ICC, eligible for radiotherapy; Child-Pugh A (184 pts)	NCT03898895
Durvalumab (anti-PD-L1) + tremelimumab (anti-CTLA-4) + TACE/RFA/cryoablation	Phase 2, nonrandomized, open label	Unresectable, refractory HCC or BTC (90 pts)	NCT02821754
	Phase 2, single arm, open label	Unresectable HCC or BTC (70 pts)	NCT03482102

Note: A www.ClinicalTrials.gov search was performed using the terms “liver cancer,” “hepatocellular carcinoma,” and “liver neoplasm” for HCC; “biliary tract cancer,” “cholangiocarcinoma,” “biliary carcinoma,” “bile duct,” or “biliary tract” for CCA; and “hepatobiliary malignant tumors” for both HCC and CCA. The search identified immunotherapy trials with a status of “Recruiting,” “Not yet recruiting,” “Active, not recruiting,” or “Enrolling by invitation,” and trials without the inclusion of a specific liver cancer cohort or without adequate information available were excluded. Search was updated as of March 30, 2020.

Abbreviations: MWA, microwave ablation; TACE, transarterial chemoembolization; SBRT, stereotactic body radiotherapy.

IMMUNE CHECKPOINT INHIBITORS IN BILIARY CANCERS

The efficacy of ICI monotherapy in biliary tract cancers (BTCs) has been disappointing thus far based on interim results from the KEYNOTE-158 trial, an ongoing phase 2, single-arm, open-label trial of pembrolizumab in advanced cancers, including a cohort of BTC. Based on the interim analysis, the ORR was 5.8% in 104 patients with BTC and the duration of response ranged from 6.2 to over 15 months.⁽⁶⁶⁾ The ORR was slightly higher in a cohort of patients with BTC with PD-L1-positive tumors. However, there was no significant difference in the OS (7.2 vs. 9.6 months). Limited data indicate the potential benefit of ICI monotherapy in patients with BTC with microsatellite instability due to mismatch repair deficiency (MSI-high/MMR-deficient). In a prospective trial of 86 patients with advanced MSI-high/MMR-deficient tumors, the disease control rate in patients with BTC (n = 4) was 100%; 1 patient had complete response (CR) and 3 had stable disease.⁽⁶⁷⁾ Based on its efficacy in this cohort of patients, pembrolizumab was granted accelerated FDA approval in MSI-high/MMR-deficient tumors including BTC.

In a phase 2 study of patients with advanced BTC, nivolumab resulted in PR in 10 (22%) and SD in 17 (37.8%) of 45 response-evaluable patients.⁽⁶⁸⁾ All patients who responded had microsatellite stable tumors. In a phase 1 clinical trial of the anti-PD-L1 antibody durvalumab with (n = 65) or without tremelimumab (n = 42) in patients with advanced BTC, the combination had PR in 7 patients compared with only 2 patients in the durvalumab monotherapy group.⁽⁶⁹⁾ Grade 3 or 4 AEs were noted in 23% of patients in the combination arm, and one death related to drug-induced liver injury was reported.

Although single-agent ICI in HCC has shown a trend toward improvement in survival and quality of life, based on the currently available data, the efficacy of single-agent ICI in BTC has been subpar. CCAs are desmoplastic tumors with an abundant tumor-immune microenvironment characterized by immunosuppressive TAMs and MDSCs.⁽⁷⁰⁾ An effective approach will likely use a combination of immunotherapies targeting the innate and adaptive immune response in CCA. Combination immunotherapy strategies using ICI are currently under investigation in multiple ongoing clinical trials (Tables 1–5).

Future Perspectives

Recent efforts have begun to uncover the complex hepatic immune network in chronic inflammatory liver diseases as well as liver tumors. Further studies using single cell–based approaches, such as single-cell transcriptomics and multiplexed imaging, are required to illustrate the complex cellular and phenotypic heterogeneity of the liver cancer ecosystem. The information gleaned from these studies will be imperative in the development of successful treatment approaches. The intricate liver cancer TIME with diverse immune cell subsets and extensive tumor-immune cell crosstalk also underscores the need for combination therapeutic approaches targeting these components. Accordingly, the first successful phase 3 trial of systemic therapy in HCC demonstrating survival benefit compared with sorafenib harnessed immune checkpoint inhibition with VEGF inhibition. More studies are needed to identify effective combination therapies for patients with advanced-stage liver cancer. Additionally, immunotherapy holds significant potential in earlier stages of disease, and several trials investigating the effectiveness of immune-based approaches in the neoadjuvant and adjuvant setting are ongoing. Finally, development of biomarkers that can effectively predict immunotherapy response will be essential in identifying the optimal therapeutic target and selecting the appropriate therapy for each individual patient.

Abbreviations:

AE

adverse event

AFP

alpha-fetoprotein

BCLC

Barcelona Clinic Liver Cancer

CCA

cholangiocarcinoma

cDC

conventional DCs

CI

confidence interval

CIK

cytokine-induced killer cell

CR

complete response

CTLA-4

cytotoxic T-lymphocyte-associated protein 4

CTL

cytotoxic T lymphocyte

DCs

dendritic cells

FDA

Food and Drug Administration

HCC

hepatocellular carcinoma

HR

hazard ratio

ICI

immune checkpoint inhibition

KC

Kupffer cell

LAYN

layilin

MDSC

myeloid-derived suppressor cell

MSI

microsatellite instability

NK

natural killer

ORR

objective response rate

OS

overall survival

PD-1

programmed death 1

PD-L1

programmed death ligand 1

PFS

progression-free survival

PR

partial response

RFA

radiofrequency ablation

SD

stable disease

TAM

tumor-associated macrophage

TGFβ

transforming growth factor β

TIL

tumor-infiltrating lymphocyte

TIME

tumor immune microenvironment

TKI

tyrosine kinase inhibitor

TLR4

toll-like receptor 4

TNF

tumor necrosis factor

Treg

regulatory T cells

VEGF

vascular endothelial growth factor

VEGFR

vascular endothelial growth factor receptor