BTCs exhibit distinct genomic alterations that orchestrate oncogenic signalling and immune evasion. For example, FGFR2 fusions (≈15%) constitutively activate the MAPK/STAT3 pathways, upregulating PD-L1 via STAT3 binding to its promoter and recruiting immunosuppressive TAMs through CCL2 secretion, thereby establishing an immune-cold microenvironment (12). Similarly, IDH1/2 mutations (≈20%) drive the accumulation of (R)-2-hydroxyglutarate, which inhibits TET2-mediated demethylation and T-cell GAPDH activity, leading to impaired HLA-I antigen presentation and reduced IFN-γ production (13). In contrast, TP53 loss (35.5%) activates NF-κB-dependent IL-8 secretion, recruiting MDSCs that deplete arginine via ARG1 overexpression and deposit extracellular matrix (ECM) barriers to block T-cell infiltration (14–16). Moreover, KRAS mutations (≈27%) induce IL-8-mediated NETosis to physically trap T cells and compete for glutamine, suppressing mTOR-dependent T-cell function (17, 18). In GBC, HER2 amplification (27.2%) transfers HER2 to dendritic cells via tumour-derived exosomes, inhibiting DC maturation and antigen presentation (19), whereas SMAD4 inactivation in eCCA (11.3%) hyperactivates TGF-β signalling to directly promote Treg differentiation and collagen deposition by CAFs (20, 21). Notably, BRCA-deficient tumours (3.6%) presented elevated TMB (10.0 vs. 6.0 mut/Mb; P<0.001) and microsatellite instability (MSI-H: 17.9%), suggesting increased susceptibility to immune checkpoint blockade (22, 23). Overall, these driver mutations define molecular subtypes and offer therapeutic targets. Genomic heterogeneity stems from clonal evolution during chronic inflammation. In PSC-associated BTC specifically, TP53/KRAS mutations synergize with bile acid metabolic aberrations. This synergy drives malignancy (24, 25). Single-cell analyses revealed that ErbB pathway mutations in GBC promote tumour progression via T-cell exhaustion (26). Cellular origins further diversify ICC molecular profiles: Hepatocyte-derived iCCA frequently exhibits TERT promoter mutations, whereas cholangiocyte-derived tumours harbour BAP1 deletions (27–29). Epigenetic dysregulation (e.g., RBM10 splicing factor mutations) and homologous recombination defects (e.g., BRCA germline mutations) contribute to genomic instability (30, 31). In addition, spatial multiomics approaches, including single-cell sequencing and spatial transcriptomics, have revealed the significant intratumor heterogeneity present in BTCs. These technologies provide deeper insights into the complex cellular composition of tumours and their microenvironments, shedding light on how these heterogeneous regions influence treatment resistance (32, 33). For example, intratumor variation in immune cell infiltration and metabolic reprogramming has been linked to the development of resistance to immunotherapies and targeted therapies (33, 34). These findings underscore the importance of considering the spatial organization and functional diversity of tumours when therapeutic strategies are designed, as localized subpopulations within the tumour may exhibit differential responses to treatment (35, 36) ( Figure 1 ).

In the development and progression of BTC, abnormal DNA methylation, histone modifications, and noncoding RNA regulation form a complex epigenetic network. Abnormal DNA methylation, through the collaborative action of UHRF1/DNMT1 [by recruiting the HELLS chromatin remodelling complex to facilitate the recognition of hemimethylated CpG (37)] and methylation recognition mechanisms mediated by MBD2, drives chemotherapy resistance (38) and presents genome-wide differential methylation features across anatomical subtypes (iCCA, eCCA, and GBC) (39). In clinical practice, methylation markers in bile [such as early cholangiocarcinoma (CCA) markers in PSC patients (40)], differentially methylated regions (DMRs/DHMRs) in plasma circulating free DNA (cfDNA) (41, 42), and the F12 gene CpG site (43) can serve as liquid biopsy tools, improving diagnostic accuracy for malignant biliary strictures. Targeted therapies for IDH1 mutation-associated methylation abnormalities (44) and DNMT3A overexpression [persistently present in the progression from cholelithiasis to gallbladder cancer (45, 46)] have shown potential for personalized treatment (47, 48). With respect to histone modifications, abnormal expression of the histone acetyltransferase KAT2B (49) and methyltransferase G9a [which promotes BTC invasion through the Hippo pathway LATS2/YAP regulation (50, 51)] drives tumour progression, whereas the heterochromatin protein HP1α regulates iCCA proliferation by interfering with the interferon pathway (52). Among noncoding RNAs, downregulation of circUGP2 (53) and abnormal expression of circACTN4 and cPKM [which promote chemotherapy resistance through the PKM2/β-catenin axis (38)] influence prognosis. LINC00511 (54) and miR-27a-3p [targeting the FOXO1/PI3K/AKT pathway (55, 56)] regulate tumour stem cell properties, whereas exosomal circRNAs [such as bile-derived CCA-circ1 (57)] and circNFIB [through miR-412-3p/PIK3R3 inhibition of metastasis (58)] can serve as molecular subtype biomarkers. These mechanisms, which are mediated by RNA splicing via RBM10 (59) and signalling pathways such as the PI3K/AKT pathway (60), lay the foundation for prognosis evaluation and targeted therapy development in BTC ( Figure 2 ).

In iCCA, TAMs play a pivotal role in driving immunosuppression by upregulating PD-L1, which promotes T-cell exhaustion (61). This process is further enhanced by abnormally activated regulatory T cells (Tregs), which decrease antitumour immunity by suppressing effector T-cell function and the secretion of proinflammatory cytokines, such as IL-10 and TGF-β (62). The cross-talk between TAMs and Tregs creates a mutually reinforcing immunosuppressive loop in which TAMs secrete factors such as IL-6 and GM-CSF, which can promote Treg expansion and activation (63–65). IL-6, produced by both tumour and stromal cells, not only enhances cancer stem cell (CSC) proliferation but is also a key mediator of systemic inflammation, which is correlated with poor prognosis in CCA patients (66, 67). IDH1-mutant tumours exacerbate immune suppression through metabolic reprogramming, increasing the production of metabolites such as 2-HG, which inhibits the function of TET2 and reduces T-cell activation, contributing to resistance against ICIs (68–70). Furthermore, VEGF-C-driven lymphangiogenesis contributes to an immunosuppressive microenvironment by facilitating lymph node metastasis and further suppressing immune responses through the recruitment of immunosuppressive cells (71, 72).

Additionally, cancer-associated fibroblasts (CAFs), derived from TGF-β1-activated hepatic stellate cells (HSCs) that express α-SMA, play a critical role in remodelling the TME. CAFs secrete extracellular matrix (ECM) components that create mechanical barriers to promote invasiveness and chemoresistance in iCCA (73–75). CAFs not only increase tumour progression through the secretion of protumour factors, such as VEGF and IL-6, but also induce vasculogenic mimicry (VM), which accelerates metastasis in gallbladder cancer (67, 76, 77). Heterogeneous CAF subpopulations interact with myeloid-derived suppressor cells (MDSCs) and inhibit T-cell function by secreting immunosuppressive cytokines, such as IL-10 and TGF-β. This interaction is facilitated by the Notch1 signalling pathway, which not only amplifies tumour malignancy but also helps CAFs recruit MDSCs to the TME (78, 79). CAFs can transfer oncogenic molecules, such as miRNAs and proteins, via exosomes, further contributing to resistance in genomically distinct subtypes, including ERBB2-amplified tumours (80–82). These interactions between CAFs, TAMs, Tregs, and MDSCs create a complex immune-suppressive network that supports immune evasion and tumour progression in cholangiocarcinoma. These findings underscore the importance of targeting the functional diversity of these immune cells using combinatorial strategies, such as Notch inhibitors or immunotherapies, to overcome resistance and improve therapeutic outcomes (65, 83) ( Figure 3 ).

Multiomics remodelling has revealed profound metabolic heterogeneity in ICCA, including enhanced glycolysis, lipid dysregulation, and mitochondrial adaptation (84–86). Tumour cells prioritize endogenous glycogen over glucose as the primary glycolytic carbon source under hypoxia, with elevated glycogen phosphorylase activity fuelling tumour progression (84). Lipid metabolism alterations, such as increased synthesis and oxidation, and PGC-1α-mediated mitochondrial reprogramming sustain proliferation and represent therapeutic vulnerabilities (87). IDH1 mutations generate (R)-2-hydroxyglutarate, which inhibits α-ketoglutarate-dependent enzymes to induce epigenetic dysregulation and immune evasion (68). Cancer stem cells (CSCs) rely on mitochondrial metabolism, whereas TAMs exacerbate immunosuppression via glycolytic shifts (86, 87). These metabolic adaptations correlate with anatomical subtypes (e.g., ERBB2-amplified iCCA vs. PIK3CA-mutant extrahepatic tumours) and influence chemotherapy/targeted therapy efficacy (88–90). Multiomics analyses further revealed metabolic heterogeneity across subtypes: gallbladder cancer shows lipid-driven hypoxia adaptation, whereas extrahepatic tumours exhibit PI3KCAH1047R-driven transformation (90–92). Metabolic crosstalk within the TME involves CAFs that secrete lactate to promote immunosuppression and TAMs that enhance glycolysis to suppress immunity (53, 93, 94). Exosome-mediated signalling coordinates angiogenesis and fibrosis, whereas metabolic–immune network dysregulation (e.g., GPR109A pathway anomalies) underpins poor prognosis and therapy resistance (95–97). IDH1-mutant tumours resist ICIs via epigenetic–immune crosstalk, whereas CPS1-deficient tumours disrupt the urea cycle to alter the pH of the TME (98–100). Targeting metabolic nodes, such as glycolytic enzymes, mitochondrial pathways, or FGFR2/IDH1-related aberrations, may reverse immunosuppression and increase treatment sensitivity (18, 53, 94, 101). Overall, the bidirectional feedback between metabolic reprogramming and TME remodelling drives cholangiocarcinoma progression, necessitating precision strategies that integrate metabolic subtyping with immune and targeted therapies (86, 102, 103) ( Figure 4 ).

ICIs, particularly PD-1/PD-L1 inhibitors and CTLA-4 inhibitors, are among the most successful immunotherapy strategies. PD-1/PD-L1 inhibitors work by disrupting immune suppression between tumour cells and immune cells, promoting the activation of T cells and enhancing antitumour immune responses (104–106). PD-1 inhibitors, such as nivolumab and pembrolizumab, have entered clinical trials for the treatment of BTC and have shown some efficacy (107, 108). This finding indicates the limitations of monotherapy in BTC and highlights the immune suppressive characteristics of the tumour microenvironment in this cancer.

The combination of immune checkpoint inhibitors with chemotherapy has been increasingly investigated in BTC, with preliminary results suggesting that this combination can significantly improve clinical outcomes. For example, the combination of pembrolizumab with gemcitabine/cisplatin has improved progression-free survival (PFS) and overall survival (OS) in advanced BTC patients compared with chemotherapy alone (3). This combination approach harnesses the cytotoxic effects of chemotherapy while activating the immune response through immune checkpoint inhibitors, providing a more effective treatment option for BTC patients (109). Combining immunotherapy with targeted therapy is another promising strategy currently being explored for BTC. The combination of PD-1 inhibitors with FGFR2 inhibitors or IDH1 inhibitors aims to enhance immune responses and overcome tumour resistance mechanisms. For example, the combination of PD-1 inhibitors with FGFR2 inhibitors has shown potential in clinical trials, with some patients experiencing delayed disease progression (88, 110–113). These combination therapies aim to target both immune escape pathways and molecular drivers of tumour growth, offering a multipronged approach to improve BTC treatment outcomes.

These combination strategies offer several advantages in BTC treatment. First, they increase treatment efficacy by simultaneously targeting immune evasion pathways and molecular drivers of tumour growth, which improves clinical outcomes. Mechanistic studies support these clinical findings: Mechanistically, targeting TAMs and their PD-L1 upregulation can effectively alleviate immune suppression and promote T-cell-mediated tumour killing (114). Additionally, combining chemotherapy with immune checkpoint inhibitors synergistically amplifies the therapeutic response and improves PFS and OS (115). In addition, the combination of targeted therapies with immunotherapy helps overcome tumour resistance mechanisms, providing a more comprehensive treatment approach. Targeted drugs, such as FGFR2 inhibitors and IDH1 inhibitors, block tumour cell proliferation and signalling pathways, thus increasing the effectiveness of immunotherapy and restoring the ability of the immune system to recognize tumours (116, 117). This multifaceted strategy aims to address the complexity of BTC and may increase the sustainability of clinical responses.

However, combination therapy also faces several challenges. The combined use of chemotherapy and immune checkpoint inhibitors may increase toxicity and immune-related adverse events (irAEs), which may limit the tolerability of these treatments in some patient populations (115). Additionally, tumour heterogeneity in BTC is a significant challenge because the molecular characteristics of a patient may dictate the response to combination treatment. Mechanistic research suggests that molecular features, such as FGFR2 and IDH1 mutations, may influence immune evasion and resistance mechanisms (118, 119). Further research on the relationship between tumour molecular characteristics and treatment response is crucial to optimize combination therapies. Finally, reliable biomarkers to identify patients who will benefit the most from combination therapies are lacking, and the discovery of such biomarkers remains a significant challenge. Mechanistic studies have pointed to the potential of identifying biomarkers, such as key molecules in the tumour microenvironment, which may help optimize patient selection. Ongoing research is needed to identify suitable biomarkers for patient selection and refine these combination treatment strategies. Despite these challenges, combination therapy remains a promising avenue for improving BTC treatment outcomes (120), and further clinical trials and mechanistic studies are necessary to optimize these approaches.

Critically, TAMs are the predominant immune cells in the BTC TME. They play pivotal roles in promoting both tumour progression and immune suppression. TAMs secrete cytokines, such as IL-6 and TNF-α, that inhibit T-cell activation while promoting the proliferation of cancer stem cells. Moreover, TAMs upregulate immune checkpoints, such as PD-L1, which further suppresses CD8+ T-cell activity, thereby facilitating immune escape and promoting tumour growth (87, 121, 122). Tregs are another crucial cell type in the BTC TME that significantly contribute to immune suppression. Tregs secrete immunosuppressive cytokines, such as TGF-β and IL-10, to directly inhibit the function of effector T cells (e.g., CD8+ T cells). This action suppresses antitumour immunity. The accumulation of Tregs in the BTC TME is associated with poorer prognosis, indicating their role in immune evasion and resistance to treatment (123–125). MDSCs constitute another class of immune-suppressive cells that accumulate in the TME under the influence of tumour-secreted factors. MDSCs suppress T-cell activation and promote tumour progression by secreting immunosuppressive molecules, such as arginase-1 (Arg1) and inducible nitric oxide synthase (iNOS) (126). In BTC, the presence of MDSCs is correlated with immune suppression and disease progression (121, 127). CAFs are a key cell type in the TME and contribute to tumour progression through the secretion of ECM components (such as collagen and fibronectin) and protumourigenic factors (such as VEGF and IL-6). CAFs remodel the ECM, which can create physical barriers that impede immune cell infiltration. In addition, CAFs promote immune evasion by secreting lactate and other metabolic products, further suppressing immune responses (128).

One approach to reshaping the TME involves targeting immune-suppressive cells, such as TAMs, Tregs, and MDSCs. TAMs can be targeted using antibodies against CD40 or CSF1R, which reduce the number of TAMs and restore immune responses (129, 130). Targeting Tregs using anti-CTLA-4 antibodies or anti-CCR4 antibodies, which can deplete Tregs and promote effector T-cell function, has been proposed as another strategy to reshape the TME (131). Additionally, targeting MDSCs with anti-GM-CSF antibodies can reduce their accumulation and reverse immune suppression in the TME (130). The dense ECM in the TME often acts as a physical barrier that prevents immune cells from infiltrating tumour tissues. Strategies to target ECM components are being developed to increase immune cell infiltration. For example, the use of matrix metalloproteinase (MMP) inhibitors has been shown to degrade the ECM barrier, promoting immune cell entry and enhancing the effects of immunotherapy (132). Furthermore, targeting ECM components secreted by CAFs can help alleviate mechanical barriers in the TME and improve immune responses (133, 134). The metabolic reprogramming of both tumour and immune cells in the TME plays a crucial role in immune evasion. The acidic and hypoxic conditions in the TME, which are caused by altered metabolic pathways, suppress immune cell function. Targeting these metabolic pathways in tumour and immune cells, such as glycolysis and lipid metabolism, has been proposed as a strategy to restore immune function and increase the efficacy of immunotherapy (135, 136). For example, inhibiting lactate metabolism in tumour cells can reverse immune suppression and enhance the antitumour effects of immune cells (137).

Combining immunotherapy with strategies to reshape the TME has shown promising results in preclinical models. Targeting both immune-suppressive pathways in the TME and enhancing immune responses has synergistic effects (138). For example, combining PD-1 inhibitors with CAF-targeting therapies has resulted in increased antitumour effects in preclinical studies (139). This approach not only improves immune cell function but also suppresses tumour growth, providing a more effective treatment strategy for BTC.

FGFR2 gene fusions or amplifications are common in iCCA, with approximately 10–15% of iCCA patients harbouring FGFR2 alterations. FGFR2 inhibitors, such as pemigatinib and futibatinib, have shown significant efficacy in FGFR2 fusion-positive patients (93, 140). Combining FGFR2 inhibitors with PD-1 inhibitors (e.g., nivolumab) may enhance the effects of immunotherapy. This potential benefit arises because FGFR2 inhibitors inhibit tumour cell proliferation. Moreover, they enhance immune system responses (141). IDH1 mutations occur in a subset of BTC patients, particularly in iCCA patients. IDH1 inhibitors (e.g., ivosidenib) restore normal metabolic pathways, inhibiting tumour cell growth. In early clinical trials, the combination of IDH1 inhibitors with immunotherapy has shown potential for enhancing immune responses (142). HER2 overexpression or amplification is observed in some gallbladder cancers and iCCA. HER2-targeted therapies (e.g., trastuzumab) in combination with immunotherapy (e.g., PD-1 inhibitors) may enhance antitumour immune responses and overcome immune evasion (55, 143).

The combination of immune checkpoint inhibitors, such as nivolumab or pembrolizumab, with chemotherapy (e.g., gemcitabine and cisplatin) has shown superior efficacy in BTC treatment (3). Compared with chemotherapy alone, the combination of PD-1 inhibitors with chemotherapy significantly improves PFS and OS. Additionally, immunotherapy helps overcome chemotherapy resistance, enhancing immune responses against tumours (144). Chemotherapy induces tumour cell death, leading to the exposure of tumour antigens, which enhances immune system recognition of the tumour (145). Moreover, chemotherapy increases immune cell infiltration into the tumour, creating a more favourable environment for immunotherapy. The combination of chemotherapy and immune checkpoint inhibitors not only enhances immune responses but also provides complementary mechanisms to combat tumour growth, improving therapeutic efficacy (105, 107).

Tumour cells enhance glycolysis and lipid metabolism to promote growth while suppressing immune cell function. Inhibiting tumour cell glycolysis or lipid biosynthesis restores immune cell antitumour functions, increasing the efficacy of immunotherapy. The combination of metabolic inhibitors with immunotherapy has demonstrated significant synergistic effects in preclinical models, particularly in BTC models (3, 146). Metabolic reprogramming in the TME creates an acidic and hypoxic environment that suppresses immune cell activity. Targeting these metabolic pathways and restoring normal metabolic states may reverse the immune-suppressive environment in the TME, thereby enhancing the effects of immunotherapy (147). For example, the use of glucose metabolism inhibitors or lactate removal agents has been shown to effectively restore antitumour responses of the immune system and improve immunotherapy response rates (148).

BTC tumour cells evade immune surveillance through several mechanisms. These processes primarily involve 1) the upregulation of immune checkpoints (such as PD-L1 and CTLA-4), 2) the accumulation of TAMs and regulatory Tregs, and 3) the reprogramming of tumour metabolic pathways. These mechanisms limit the immune response and contribute to tumour immune escape, thereby exacerbating resistance (157). For example, high PD-L1 expression in BTCs directly inhibits CD8+ T-cell-mediated antitumour effects (158). TAMs and Tregs, through the secretion of immunosuppressive cytokines such as IL-10 and TGF-β, further weaken immune responses (159). Metabolic reprogramming in the TME is another important factor leading to immune evasion. BTC cells promote glycolysis, lipid synthesis, and lactate secretion, creating an acidic, hypoxic, and nutrient-deprived microenvironment. This environment not only provides a growth advantage to tumour cells but also suppresses immune cell functions. For example, the accumulation of lactate in the TME impairs T-cell efficacy and suppresses immune responses (160).

Although ICIs have demonstrated efficacy in various cancers, immune-related adverse events (irAEs) associated with these treatments have become a significant clinical challenge. irAEs refer to autoimmune reactions triggered by the activation of the immune system, which can result in damage to various organs, including the skin, liver, lungs, and endocrine system (161, 162). In BTC patients, the incidence of irAEs due to ICIs is relatively high, and their management requires careful consideration of the patient’s immune status and comorbidities. This management is especially challenging in patients with advanced BTC or those with multiorgan metastasis, where irAEs complicate treatment. Therefore, balancing efficacy with the management of irAEs remains a major clinical challenge in BTC immunotherapy.

The management of irAEs in BTC patients involves a range of strategies aimed at controlling immune overactivation while maintaining the antitumour efficacy of ICIs. First-line treatment typically involves systemic corticosteroids, with additional immunosuppressive agents such as mycophenolate mofetil or infliximab for severe cases (163–165). Temporary interruption of ICIs may be recommended for mild irAEs, with reintroduction upon symptom resolution. Preventive measures focus on pretreatment screening to identify high-risk patients, particularly those with a history of autoimmune disorders. Low-dose corticosteroids before treatment and personalized monitoring using biomarkers predictive of immune activation are areas of active research. Although these strategies can reduce the severity of irAEs, further clinical validation is needed before they can be widely implemented.

Balancing efficacy with safety is a critical challenge in BTC immunotherapy, especially given the potential for severe irAEs. To optimize outcomes, clinicians must tightly monitor patients by assessing organ function and tracking symptoms regularly. An individualized approach is necessary, wherein the decision to continue or pause ICI therapy is based on the patient’s overall health, the severity of irAEs, and potential survival benefits (166). Furthermore, combining ICIs with other therapies, such as chemotherapy or targeted treatments, may help mitigate the immune activation seen with monotherapy, potentially reducing the risk of irAEs while retaining treatment efficacy (167, 168). Ongoing research into biomarkers to predict irAEs and methods to prevent their occurrence will be essential for improving the safety profile of ICIs in BTC and enhancing the clinical management of these patients (169).

Currently, the response rate of BTC patients to immunotherapy remains low, making patient selection crucial. However, accurately identifying patients who are most likely to benefit from immunotherapy is difficult due to the lack of effective predictive biomarkers. While some biomarkers, such as PD-L1 expression, microsatellite instability (MSI), and the TMB, have been shown to have predictive value in other cancers, their utility in BTC remains unvalidated (170, 171). Biomarker identification for immunotherapy response in BTC remains in its early stages. Emerging technologies, such as liquid biopsy, ctDNA, and genomic analysis, offer the potential for earlier and more accurate assessment of treatment efficacy and resistance. Furthermore, research suggests that the interaction between immunotherapy and the TME may affect the treatment response, highlighting the potential of TME components as predictive biomarkers (172).

Although reshaping the TME to increase immunotherapy efficacy has become a significant research focus, translating this strategy into clinical practice still faces multiple challenges. First, the complexity and heterogeneity of the TME hinder the complete elimination of immunosuppressive components (172, 173). Additionally, TME-reshaping drugs may interact with immunotherapies, potentially causing side effects. For example, targeting CAFs (cancer-associated fibroblasts) may lead to adaptive changes in tumour cells that result in therapy resistance (174). Therefore, precisely targeting immune-suppressive components within the TME while minimizing side effects and maximizing the effectiveness of immunotherapy remains a key focus for future research.

Although numerous clinical trials related to immunotherapy are underway, some limitations still remain in the design of clinical trials for BTC. First, most clinical trials have focused on late-stage BTC patients, often overlooking patients with early-stage or locally advanced disease. Second, the heterogeneity of trial results is significant, due in part to individual patient differences, tumour heterogeneity, and variations in immune responses. Therefore, larger, more rigorously designed clinical trials are needed to validate the true effects and safety of immunotherapy in BTC (175) Table 2 .

Chimeric antigen receptor T-cell (CAR-T) therapy is an emerging treatment strategy that involves genetically modifying a patient’s own T cells to recognize and attack tumour cells. Although CAR-T-cell therapy has achieved significant success in haematologic cancers, its application in solid tumours, including BTCs, faces challenges. Research indicates that the primary challenge for CAR-T-cell therapy in solid tumours is the ability to effectively penetrate the tumour microenvironment and enhance T-cell infiltration (176, 177). Current research focuses on optimizing CAR-T-cell therapy design, including targeting specific antigens found in BTCs, such as Mucin 1 and CEA, to improve efficacy and reduce side effects (178). Tumour vaccines, particularly mRNA-based vaccines, constitute another emerging strategy in immunotherapy. mRNA vaccines deliver specific tumour antigen genes to the patient’s body, inducing the immune system to recognize and attack tumour cells. mRNA vaccines that target BTCs are currently in the preclinical phase, and these vaccines have the potential to enhance antitumour immune responses (179). For example, mRNA vaccines that target KRAS mutations have shown therapeutic potential in other solid tumours and may be used for BTC treatment in the future (180).

T-cell immunoreceptor with Ig and ITIM domains (TIGIT) and lymphocyte-activation gene 3 (LAG-3) are newly identified immune checkpoint molecules that play critical roles in immune suppression in various cancers. Research indicates that the upregulation of TIGIT and LAG-3 is closely associated with immune evasion and tumour resistance (181–183). Currently, monoclonal antibodies that target TIGIT and LAG-3 are undergoing clinical trials. Combining PD-1/PD-L1 inhibitors with TIGIT or LAG-3 inhibitors may improve treatment responses in BTC patients (184). V-domain Ig suppressor of T-cell activation (VISTA) is another newly discovered immune checkpoint molecule that plays a key role in regulating T-cell function (185). VISTA upregulation in the TME has been shown to play a role in immune suppression, particularly in solid tumours. Antibodies that target VISTA are currently under development and may offer new treatment options for BTC patients when combined with existing immunotherapies (186).

AI and deep learning technologies can analyse patient genomic data, immune phenotypes, and clinical characteristics to identify biomarkers associated with immunotherapy response. These technologies can help clinicians select the patients most likely to benefit from immunotherapy and provide personalized treatment plans (187). AI algorithms can assist in optimizing immunotherapy regimens, including selecting the best immune checkpoint inhibitors, targeted therapies, and combination therapies. By simulating different treatment pathways and predicting treatment responses, AI can provide more precise therapeutic decisions for BTC patients (188). Furthermore, AI can integrate the analysis of the gut microbiome to further enhance personalized treatment strategies (189). Increasing evidence shows that the gut microbiome significantly influences immune therapy responses, with certain microbiome compositions enhancing immune responses and improving the efficacy of immunotherapy (190, 191). AI can analyse the microbiome data of patients to identify those whose microbiome features may affect immune therapy outcomes, thereby optimizing treatment plans (192). By integrating genomic data, immune phenotypes, clinical characteristics, and microbiome data, AI provides more comprehensive decision support for personalized treatment plans, helping to improve treatment efficacy and patient survival (193).