Biliary tract cancer (BTC) is primarily classified into three clinical subtypes based on anatomical location: intrahepatic cholangiocarcinoma (iCCA), extrahepatic cholangiocarcinoma (eCCA), and gallbladder cancer (GBC). Extrahepatic cholangiocarcinoma can be further subdivided into perihilar and distal types, with significant differences in clinical presentation, treatment strategies, and prognosis among subtypes. The causes of biliary tract cancer are complex and varied. Established risk factors include gallstones, primary sclerosing cholangitis, hepatolithiasis, liver fluke infection, chronic viral hepatitis, and cirrhosis. In recent years, metabolic factors such as obesity and non-alcoholic steatohepatitis have been increasingly linked to a higher risk of BTC (1, 2). The global incidence of BTC is rising, particularly for the iCCA subtype. Epidemiological data show that between 1990 and 2019, the number of BTC cases and related deaths increased by 84.8% and 81.8%, respectively. The prognosis for BTC is very poor, with a median overall survival of approximately 9 months and a one-year survival rate of about 51% (3, 4). BTC also shows high molecular and genomic heterogeneity. This heterogeneity not only drives cancer progression but also poses major challenges for precision therapy. Due to its insidious onset, rapid progression, limited treatment options, and dismal prognosis, BTC constitutes a major social and healthcare burden (5).

Surgery is the first-choice curative treatment for BTC. However, the often asymptomatic nature of early-stage disease results in the majority of patients being diagnosed at an advanced stage. This substantially limits the applicability of surgical intervention and locoregional therapies, with only approximately 22% of patients ultimately qualifying for resection (3, 6). For patients with unresectable advanced disease, systemic pharmacotherapy is the mainstay. The combination of gemcitabine and cisplatin is established as the standard first-line chemotherapy regimen for BTC (7). However, traditional chemotherapy offers limited benefits. It provides only short-term remission for some patients, with most tumors rapidly developing resistance (8). While targeted therapies based on molecular profiling have emerged as an alternative to overcome chemotherapy resistance, their clinical application is constrained by challenges including limited biomarker detection rates, narrow applicability, and acquired resistance.

In recent years, with in-depth exploration of the tumor immune microenvironment and immunology, immunotherapy has gradually become a major research focus in cancer treatment (9, 10). This strategy enhances anti-tumor immune responses and overcomes immune evasion, demonstrating strong therapeutic potential. Immune checkpoint inhibitors (ICIs), particularly antibodies targeting programmed cell death protein 1 (PD-1) and programmed cell death ligand 1 (PD-L1), are now part of standard first-line regimens for many cancers (11). For BTC, regimens such as cisplatin plus gemcitabine with durvalumab or cisplatin plus gemcitabine with pembrolizumab are recommended as first-line options (12, 13). However, ICI monotherapy shows limited efficacy in BTC, with the majority of patients developing treatment resistance (14). This treatment resistance is closely linked to the immunosuppressive tumor microenvironment (TME) in BTC. The TME consists of tumor cells, immune cells, and stromal cells, and it is the key factor that limits ICI effectiveness and causes resistance (15).

Adoptive cell therapy (ACT) is an emerging tumor immunotherapy strategy centered on utilizing immune effector cells to attack tumors. The standard procedure can be summarized in three key steps: first, collecting immune-reactive cells from the patient’s peripheral blood mononuclear cells or tumor tissue; second, activating and expanding these cells ex vivo; and finally, reinfusing the expanded effector cells into the patient to mediate antitumor immunity (16). ACT has proven highly effective in treating hematological malignancies, primarily based on efficient targeting of specific antigens like cluster of differentiation (CD) 19 and B-cell maturation antigen. Its application is continuously expanding into the realm of solid tumors, where growing preclinical and clinical data are confirming its potential and paving the way for broader use (17–21).

BTC, as a malignancy characterized by both high heterogeneity and strong immunosuppressive properties, has limited responses to conventional therapies, creating an urgent clinical need for new strategies. In this context, various adoptive cell therapy modalities, such as chimeric antigen receptor T cells (CAR-T), tumor-infiltrating lymphocytes (TIL), cytokine-induced killer cells (CIK), natural killer cells (NK), and T-cell receptor-engineered T (TCR-T) cells, are being extensively explored for BTC treatment (Figure 1). This review will detail the research progress of these ACT strategies in BTC through dedicated sections. We will also discuss the key challenges currently faced in the field and share perspectives on potential future directions.

Chimeric antigen receptor T-cell therapy is an immunotherapy strategy that involves genetically engineering a patient’s own T cells to express chimeric antigen receptors targeting specific tumor antigens. It has recently emerged as a promising direction in BTC treatment. The structure of CARs has evolved through generations, but its core consistently comprises three functional domains: an extracellular antigen recognition domain, a transmembrane domain, and an intracellular signaling domain (22). The extracellular domain is typically a single-chain variable fragment (scFv) derived from a monoclonal antibody. This fragment is generated by linking the variable regions of the antibody heavy and light chains with a flexible linker. It is responsible for the specific recognition of epitopes on tumor cell surfaces, thereby forming the molecular basis of CAR targeting. The transmembrane domain, serving as an anchor connecting the extracellular and intracellular parts, is usually derived from molecules like CD3ζ, CD8, or CD28. This domain not only anchors the CAR to the T-cell membrane but also plays a crucial role in signal transduction and receptor stability. The intracellular signaling domain is the core component that activates T-cell functions. The first-generation CARs only contained the primary CD3ζ signaling domain, which often resulted in limited T-cell activation, poor expansion, and insufficient persistence. To overcome these limitations, a key innovation in second-generation CARs was the addition of specific co-stimulatory domains like CD28 or 4-1BB, fused directly to the CD3ζ chain. This pivotal enhancement significantly improved T-cell activation, proliferation, cytokine production, and persistence. Building on this, third-generation CARs further incorporated two distinct co-stimulatory domains, aiming to synergistically amplify anti-tumor responses. However, challenges posed by the solid TME prompted the development of fourth-generation CARs, also known as T cells redirected for universal cytokine killing (TRUCKs). These are engineered based on earlier generations but are equipped with inducible cytokine modules. Upon recognizing tumor antigens, these CAR-T cells trigger direct cytotoxicity alongside localized cytokine release. This helps remodel the immune microenvironment, recruit and activate innate immune cells, and ultimately achieve more potent tumor eradication (23, 24). Currently, fifth-generation CAR-T therapies are still under development and refinement. Their core features include a modular design philosophy employing a “universal receptor + targeting module” split-receptor system; the use of gene editing technologies like clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 (CRISPR/Cas9) or transcription activator-like effector nuclease for precise cell engineering; and the integration of T-cell receptor signaling, co-stimulatory signals, and cytokine signaling pathways for synergistic CAR-T cell activation. Such designs aim to improve the universality, persistence, resistance to exhaustion, and safety of CAR-T cells, opening new potential strategies for solid tumor immunotherapy (25–27). In BTC, CAR-T therapy primarily targets tumor-associated antigens such as Mucin 1 (MUC1), Epidermal Growth Factor Receptor (EGFR), integrin alpha-v beta-6 (αvβ6), CD133, and B7 Homolog 3 (B7-H3) (Table 1).

Tumor-infiltrating lymphocyte therapy is an adoptive cell treatment based on T lymphocytes isolated from the patient’s own tumor tissue. TIL therapy relies on isolating T cells with inherent tumor recognition capability from the TME, expanding them ex vivo, and reinfusing them into the patient to enhance anti-tumor immune responses. Compared to most other adoptive cell therapies, TIL therapy does not require genetic engineering. This may lower the risk of adverse events associated with genetic modification. Simultaneously, it mediates specific tumor cell killing, demonstrating good targeting and therapeutic potential (54, 55). It has shown significant clinical efficacy in solid tumors like melanoma and has recently expanded into research for gastrointestinal malignancies (56–58).

In BTC, TILs demonstrate the potential to target tumor-specific antigens. Evidence shows that TILs can recognize neoantigens derived from the breakpoint region of FGFR2-TDRD1 fusions. Furthermore, CD4⁺ T cells and their T-cell receptors that possess this specific recognition ability can be successfully isolated (59). This finding confirms the immunogenicity of FGFR2 fusions, providing a theoretical basis for utilizing TILs to target oncogenic fusions. In clinical practice, a case report involved a patient with metastatic cholangiocarcinoma. The patient received a TIL infusion, where about 25% of the T cells were reactive to an ERBB2IP mutation. Following this treatment, the patient experienced sustained tumor stabilization and partial regression for over one year. Notably, when the disease later progressed, a second infusion was given. This second infusion contained a highly purified (>95%) population of mutation-specific TILs. The result was again tumor shrinkage, demonstrating the durable anti-tumor activity of TILs in BTC (60). To optimize TIL therapy, a research team established an orthotopic mouse model of CCA to systematically evaluate two expansion strategies: traditional CD3 agonist-based expansion and tumor antigen-driven expansion. Results showed that while both methods effectively expanded TILs, the tumor antigen-driven expansion method demonstrated stronger cytotoxicity in vitro, maintaining high killing activity even at low effector:target ratios. Meanwhile, the persistence of these TILs was also significantly superior to those expanded by the traditional method in vivo (61).

Recent strategies based on modifying peripheral blood-derived T cells offer new directions for TIL therapy. One study proposed a “three-in-one” strategy, creating super circulating TIL-like cells (ScTILs) for treating advanced BTC. In the corresponding clinical trial, ScTILs demonstrated a manageable safety profile with no severe cytokine release syndrome (CRS) or neurotoxicity observed, which was likely attributable to the omission of both lymphodepleting chemotherapy and high-dose IL-2 support. In patients with normal baseline B-cell levels receiving an appropriate dose, the median overall survival reached 18.3 months, superior to existing standard treatment options (62). This strategy provides a new approach to overcome limitations of traditional TIL therapy regarding cell source, production timeline, and in vivo persistence. TILs hold significant value in the integrated diagnosis and treatment of BTC. As key components of the TME, their characteristics provide important insights for efficacy prediction and prognosis assessment (63). Meanwhile, TIL-based adoptive cell therapy is opening a promising innovative treatment pathway for advanced patients.

Natural killer cells are crucial components of the innate immune system, capable of recognizing and killing tumor cells without the need for antigen-specific priming. NK cell activity is controlled by the balance between inhibitory receptors and activating receptors on their surface. Importantly, their function is not restricted by the major histocompatibility complex (MHC). In cancer immunotherapy, NK cells attract considerable attention due to their broad anti-tumor activity and relatively low risk of toxicity. NK cells primarily kill tumor cells and promote anti-tumor immunity by releasing perforin/granzymes, expressing Fas ligand, mediating antibody-dependent cellular cytotoxicity (ADCC), and secreting pro-inflammatory cytokines (64–66). NK cells currently used for adoptive immunotherapy are mainly derived from autologous or allogeneic peripheral blood, umbilical cord blood, or the NK-92 cell line. NK cells derived from induced pluripotent stem cells (iPSC-derived NK cells), leveraging the core advantage of being “off-the-shelf” cell products, offer a novel strategy to address these bottlenecks. Recent research has successfully developed feeder-free monolayer culture systems capable of efficiently inducing high-purity, functional iPSC-derived NK cells (iNK cells). Transcriptomic analysis shows that iNK cells closely resemble peripheral blood-derived NK cells at the gene expression level. In functional assays, iNK cells effectively kill cholangiocarcinoma cell lines, such as KKU-055 and KKU-213A, in both 2D and 3D tumor spheroid models. Their cytotoxic ability is greater than that of the NK-92 cell line, making them a superior cell source for clinical use in cancer immunotherapy (67).

Preclinical studies confirm that ex vivo expanded NK cells can significantly inhibit tumor growth in CCA xenograft models, with multiple infusions not eliciting significant toxicity (68). Comparative studies showed that NK cells exhibited stronger direct killing activity in short-term cytotoxicity assays compared to Vδ2 γδ T cells. Particularly noteworthy, the combination with an anti-EGFR monoclonal antibody significantly enhanced NK cell-mediated ADCC (69). Engineered NK cells demonstrate potential for precise targeted therapy. CAR-NK cells targeting cellular-mesenchymal epithelial transition factor (cMET) showed specific killing activity against cMET-high CCA cell lines (>80% cell death) but no significant effect on cMET-low cells (70). Recently, novel physical regulation methods have further expanded NK cell application prospects. For example, piezoelectric nanoparticles targeting NK cells (αCD56-P@BT/αNK1.1-P@BT) can generate mechanical and electrical signals under ultrasound stimulation, effectively enhancing NK cell migration, tumor infiltration, and killing performance by activating TRP ion channel-mediated calcium influx and cytoskeleton rearrangement. In a mouse model of iCCA, NK cells loaded with these nanoparticles exhibited excellent anti-tumor effects (71). Clinically, Leem et al. conducted a phase I/IIa trial evaluating the safety and efficacy of allogeneic NK cells combined with the PD-1 inhibitor Pembrolizumab in patients with chemotherapy-refractory advanced BTC. Results showed no direct drug-related serious adverse events in the combination therapy. The overall response rate (ORR) in the per-protocol set reached 50.0%, with a disease control rate (DCR) of 62.5% and a median PFS of 4.1 months. These outcomes are significantly better than historical data for Pembrolizumab monotherapy (72). This study provided the first clinical validation of the synergistic anti-tumor potential of NK cells combined with immune checkpoint inhibitors in BTC. In summary, based on optimized NK cell expansion techniques and the development of combination anti-tumor strategies, NK cell therapy shows broad prospects in BTC.

Cytokine-induced killer cells are a heterogeneous group of immune effector cells. Their main component is CD3⁺CD56⁺ NKT-like cells. These cells combine the strong ability to proliferate outside the body from T cells with the non-MHC-restricted killing function of NK cells. These cells can be efficiently induced and massively expanded from human peripheral blood mononuclear cells, bone marrow, or umbilical cord blood-derived mononuclear cells by sequential addition of IFN-γ, anti-CD3 monoclonal antibody, and high-dose recombinant human IL-2 (73). CIK cells primarily induce tumor cell apoptosis via the granzyme-perforin pathway, and their cytotoxicity depends on the NKG2D receptor recognizing stress-related ligands like MICA/B expressed on tumor cell surfaces (74, 75). Beyond direct tumor cell killing, CIK cells can also modulate immune responses by secreting various cytokines (e.g., IL-2, IFN-γ, TNF-α, IL-4) (76).

In severe combined immunodeficient (SCID) mouse human CCA xenograft models, CIK cells demonstrated the ability to specifically infiltrate tumor tissue and effectively inhibit tumor growth. Further research indicated that co-culturing CIK cells with dendritic cells (DCs) could enhance their anti-tumor activity, while using purified CD3⁺CD56⁺ subsets could avoid the immunosuppressive effects induced by tumor RNA-pulsed DCs (77). The research team led by Morisaki conducted two independent studies exploring synergistic killing mechanisms. These studies focused on cytokine-activated killer (CAK) lymphocytes in combination with different drugs. In an in vitro study on CCA, combining CAK cells with cetuximab significantly enhanced the killing of CCA cell lines via ADCC (78). In a model of chemotherapy-resistant metastatic solid tumors, pretreatment with gemcitabine increased the expression of MICA/B on tumor cell surfaces. This upregulation enhanced CAK cell cytotoxicity through the NKG2D receptor (79). Together, these two studies indicate that combining CAK cells with either targeted drugs or chemotherapy agents can synergistically enhance anti-tumor effects through distinct immune mechanisms. Furthermore, studies show that overexpressing the inducible T-cell costimulator (ICOS) in CIK cells significantly improves their proliferation and cytokine secretion, particularly production of IFN-γ. This enhancement increases their killing activity against BTC cells (80, 81). When ICOS binds to its ligand (ICOSL) on tumor cells, it activates the phosphatidylinositol 3-kinase/protein kinase B and extracellular signal-regulated kinase signaling pathways. This activation promotes CIK cell survival and strengthens their anti-apoptotic ability, thereby improving their persistence and cytotoxicity within the tumor microenvironment.

Clinical research data support the safety and preliminary efficacy of CIK/CAK cells in BTC treatment. A study reported that among 5 patients with chemotherapy-resistant metastatic BTC receiving gemcitabine combined with CAK therapy, 2 achieved a complete response and 3 achieved stable disease, with overall survival ranging from 16 to 64 months. Importantly, no severe adverse events related to cell infusion were observed, suggesting a favorable safety profile (79). Furthermore, a case report detailed a patient with iCCA whose disease progressed after bi-specific antibody conjugated CIK cell therapy. This patient subsequently achieved a complete response after receiving a PD-1 inhibitor. In this case, the sequential strategy was well-tolerated with no reported severe adverse events, suggesting that CIK therapy may effectively prime the immune microenvironment for subsequent checkpoint blockade without compounding toxicity (82). These findings suggest that CIK cell therapy can effectively inhibit BTC progression, and its combination with other therapies represents an important future direction for anti-cancer treatment. However, attention must also be given to potential adverse reactions, such as cytokine release syndrome and autoimmune diseases, which may be more pronounced in combination therapies. The safety profile of these combination approaches needs thorough evaluation alongside efficacy assessments.

TCR-T therapy involves genetically reprogramming a patient’s T cells. This process equips them with a new T-cell receptor that can identify specific tumor antigens presented by MHC molecules on cancer cells. Upon recognition, the engineered T cells effectively eliminate the tumor. In contrast to CAR-T therapy, which is confined to targeting surface antigens, TCR-T technology can detect antigens derived from a tumor’s intracellular proteins. This capability significantly expands the range of targetable tumors and positions TCR-T as a highly promising strategy for treating solid tumors (83). Nonetheless, the application of this innovative therapy in BTC is still in early exploration, with early-phase trials including NCT05194735.

Cancer-testis antigens (CTAs) are considered ideal targets for immunotherapy. This is because they are specifically expressed in various cancers, but in normal tissues, their presence is largely restricted to immune-privileged sites such as the testis and placenta (84, 85). Data show that in about 33.7% of iCCA cases, CTAs and HLA class I molecules are co-expressed (86). This co-expression creates a functional basis for TCR-T cells to recognize and attack tumor cells. Studies targeting New York esophageal squamous cell carcinoma 1 (NY-ESO-1) have further strengthened the case for CTA-directed therapy. In both preclinical and clinical settings, NY-ESO-1-specific TCR-T cells strongly suppressed localized and disseminated tumors (87–89). Another promising CTA in iCCA is melanoma-associated antigen (MAGE) A3. Its expression increases gradually during the transition from biliary epithelial dysplasia to invasive carcinoma, indicating it may play a role in cancer development. This dynamic expression also offers a pathological rationale for targeting MAGE-A3 in immunotherapy (86). In TCR-T development, high-affinity TCRs targeting the MAGE-A3:112–120 epitope were successfully screened using an HLA-A*02:01 transgenic mouse model. These TCRs could not only recognize MAGE-A3-derived peptides but also cross-react with homologous antigens like MAGE-A12, MAGE-A2, and MAGE-A6 (90). This cross-reactivity significantly broadens their potential targeting range.

Alpha-fetoprotein (AFP) is a well-established clinical biomarker for hepatocellular carcinoma (HCC). Its specificity and high expression levels in HCC patients make it an ideal target for TCR-based therapy. Given that some iCCA cases also express AFP, this strategy could potentially be extended to AFP-positive iCCA patients. Existing evidence indicates that AFP is not only highly expressed in HCC but also detected in a subset of iCCA tumor cells, and these AFP-positive cells exhibit cancer stem cell properties like self-renewal and high tumorigenicity (91). Multiple studies confirm that TCR-T cells can specifically recognize and kill AFP-positive tumor cells, further supporting the feasibility of this strategy in iCCA (92–94).

In BTC, driver mutations exhibit significant heterogeneity across anatomical subtypes. Genomic analyses indicate that TP53 (26%) and KRAS (18%) are among the most prevalent alterations (95). Notably, KRAS mutations are enriched in extrahepatic cholangiocarcinoma and gallbladder cancer compared to intrahepatic cases. This distinct mutational landscape underscores the particular promise of TCR-T therapies targeting these shared neoantigens for subsets of patients, especially those with extrahepatic disease. Exemplifying this approach, the Phase I/II trial NCT05194735 engineers autologous T cells via a non-viral Sleeping Beauty transposon/transposase system to target specific KRAS and TP53 mutations. Preliminary reports suggest these TCR-T products can be successfully manufactured and administered without dose-limiting toxicities, providing early evidence for the feasibility and safety of targeting intracellular driver mutations through an MHC-restricted mechanism (96).

In summary, antigens such as MAGE-A3, AFP, and KARS have been detected in a subset of biliary tract cancers. Targeting these antigens with TCR-T therapy has already shown promising preclinical and clinical results in other cancer types, providing a strong rationale and a translational foundation for further investigating their use in BTC.

Adoptive cell immunotherapy is a key branch of cancer treatment with broad potential for biliary tract cancers. This field includes several approaches, such as CAR-T, TIL, CIK, NK, and TCR-T therapies, each with its own distinct features. Recent advances in immunology and genetic engineering have accelerated both basic and clinical research in this area. Meanwhile, multiple clinical trials are investigating adoptive cell immunotherapy for biliary tract cancers (Table 2). Despite this progress, the clinical application for BTC remains challenging due to its complex tumor biology and microenvironment. Addressing these hurdles will demand innovative, multi-level solutions that integrate knowledge across different fields.

The TME of BTC is highly heterogeneous and immunosuppressive. Its structure is built upon a dense extracellular matrix and various stromal cells. It is also populated by immunosuppressive cells, including M2 macrophages and Tregs, and filled with factors like TGF-β and IL-10 (97–99). These components create a powerful network that blocks drug delivery and weakens immune cell function. Consequently, this promotes tumor progression and metastasis. Research shows that only a tiny fraction of therapeutic T cells successfully reach the tumor core, and this limited infiltration greatly reduces treatment effectiveness (100).

Beyond the hindrances of the TME, effector cells themselves face significant functional deficiencies and survival challenges. CAR-T cells often fail to achieve long-term persistence in vivo due to insufficient support from key cytokines and rejection by the host immune system (101). In addition, CAR-T cells are prone to exhaustion, especially after prolonged exposure to tumors. This exhausted state is marked by high levels of inhibitory receptors. Studies using 3D tumor models show that CAR-T cells can lose their tumor-killing ability relatively quickly after becoming exhausted (52). TIL and CIK therapies contain mixed populations of cells, and not all of these cells contribute to fighting the tumor. Some may even suppress the activity of other immune cells (102, 103). Meanwhile, NK cells are highly sensitive to the process of cryopreservation and thawing, often suffering a substantial loss in both viability and killing activity. This sensitivity currently restricts their clinical effectiveness (104).

Target-related challenges also pose major obstacles for adoptive cell therapy. Tumor cells can evade immune attack by downregulating HLA molecules or target antigens, causing immunotherapies to lose their target. This is known as antigen escape (105, 106). The high heterogeneity of BTC further complicates treatment. The expression of key targets often varies significantly between patients and even between different tumor sites in the same individual (95, 107, 108). Another critical issue is “on-target, off-tumor” toxicity. If the target antigen is also present on normal tissues, the therapeutic cells may attack healthy cells, causing serious side effects (109). For instance, the cMET antigen is expressed in about 50%-60% of CCA patients, making it a potential target (110). However, because cMET is also found in normal bile ducts and other tissues, targeting it carries a risk of severe damage to these healthy organs, directly threatening patient safety (70).

Each technological pathway also has its specific bottlenecks. Traditional CAR-T cells lack effective safety switches and are prone to significant exhaustion in environments with high PD-L1 expression, which can lead to severe CRS (111, 112). The TIL preparation process is complex, time-consuming, and difficult to standardize. It requires high-dose IL-2 for expansion, and the T cells are prone to exhaustion during this process (113, 114). Furthermore, obtaining sufficient tumor tissue can be a major barrier for advanced patients. For CIK cell therapy, a key drawback is that IL-2, while crucial for inducing CIK cells, also promotes the expansion of Tregs. This can potentially counteract the anti-tumor effect (115, 116). NK cell therapy, on the other hand, faces source limitations. NK cells derived from peripheral or cord blood are limited in number and heterogeneity, and iPSC differentiation technology is not yet fully mature. TCR-T technology is strictly limited by HLA restriction. Its efficacy is dependent on matching specific HLA types, limiting the broad application of TCR-T cells (117, 118).

To address these challenges, researchers have proposed various innovative strategies, primarily focusing on three dimensions: genetic engineering, optimization of combination therapies, and innovations in manufacturing processes. In genetic engineering, modifying the CAR structure through approaches like converting inhibitory signals or blocking multiple inhibitory receptors can significantly enhance CAR-T cell tumor-killing capacity. For example, one approach uses a PD-1–CD28 switch receptor. When this modified CAR-T cell encounters PD-L1, it turns the inhibitory signal into a stimulating one. In experiments, these cells achieved a 70.69% killing rate against PD-L1–high BTC cells and showed reduced exhaustion markers (35). Complementing receptor modification, gene editing via CRISPR/Cas9 enables the precise disruption of inhibitory pathways. One strategy involves targeting multiple immunosuppressive molecules, such as PD-1 and receptors for cytokines like TGF-β, IL-10, and IL-6. This approach has been shown to generate stronger CAR-T cells that demonstrate improved tumor infiltration and increased IFN-γ secretion. In humanized cholangiocarcinoma models, this ultimately leads to significant tumor regression and prolonged survival (119). Separately, armored CAR-T or TCR-T cells are engineered to remodel the immunosuppressive tumor microenvironment. They secrete pro-inflammatory cytokines, which enhance T-cell activity and attract endogenous immune cells. Novel approaches now use endogenous tumor-specific promoters to confine payload delivery to the tumor site. This localized release reduces systemic toxicity and improves the balance between treatment strength and safety (120–122). Concurrently, safety mechanisms such as “off-switches” or suicide genes have been introduced to induce CAR elimination if needed (123, 124). Meanwhile, developing dual-target CAR structures and sequential treatment strategies can effectively counter antigen escape by tumor cells. These approaches expand target recognition to include multiple antigens, enhancing coverage of heterogeneous tumors and reducing the risk of relapse due to antigen loss (39, 125). Beyond expanding the target repertoire, further engineering sophistication is aimed at enhancing the precision of tumor recognition. Although not yet reported in BTC, logic-gated CAR systems represent a highly promising strategy to overcome heterogeneity while minimizing on-target, off-tumor toxicity. These systems require the presence (or absence) of multiple tumor-associated antigens for full T-cell activation, thereby theoretically confining activity to tumor cells expressing the exact antigen combination while sparing normal tissues that may express only one (126). The successful application of such strategies in other solid tumors provides a solid foundation for their exploration given BTC’s complex antigenic profile. Collectively, these next-generation engineering strategies aim to not only enhance the intrinsic fitness of effector cells against exhaustion but also dynamically reshape the immunosuppressive environment and adapt to the heterogeneous landscape of biliary tract cancer.

In combination therapy strategies, the combination of immune checkpoint inhibitors with adoptive cell therapy shows synergistic effects and has received empirical support in treating various malignancies (127–129). In CCA research, the combination of B7-H3 CAR-T cells with PD-1 blockade significantly enhanced tumor clearance in patient-derived tumor spheroid models (52). On the other hand, combining adoptive cell therapy with radiotherapy or chemotherapy can induce a more comprehensive anti-tumor immune response. Studies indicate that low-dose radiotherapy can optimize the TME by inducing iNOS+/M1-type macrophage phenotypic reprogramming while upregulating chemokine secretion, thereby synergistically potentiating the tumor infiltration and anti-tumor activity of adoptive T cells (130, 131). Chemotherapeutic drugs like gemcitabine can induce immunogenic cell death, promoting the exposure and presentation of tumor antigens. Thus, integrating adoptive cell therapy with modalities like checkpoint blockade, radiotherapy, or chemotherapy constitutes a powerful therapeutic approach by leveraging their different mechanisms of action (79).

Innovations in manufacturing processes and product forms have also made significant progress. In the NK cell field, advances in iPSC differentiation technology make the large-scale production of “off-the-shelf” NK cell products possible (Figure 2). Feeder-free monolayer culture systems can yield high-purity, highly cytotoxic iPSC-derived NK cells, whose cytotoxicity is superior to the traditional NK-92 cell line (67, 132). In the T-cell therapy field, the development of ScTILs represents an innovative strategy. Instead of extracting cells from tumor tissue, ScTILs are isolated from a patient’s blood. These PD-1-positive T cells are genetically engineered using lentiviral vectors to express an enhanced receptor that reverses immunosuppressive signals from the tumor microenvironment. A CD19-targeting CAR is also introduced, enabling rapid expansion in the body via B-cell activation (62). Unlike traditional TIL therapy, ScTILs do not require tumor resection, lymphodepleting chemotherapy, or high-dose IL-2. This results in shorter production time, improved safety, and broader accessibility. Clinical studies in advanced BTC have shown significant survival benefits with this approach. Another emerging modality is TCR-like antibodies. These antibodies can specifically recognize peptide-MHC complexes on tumor cells and exert anti-tumor effects through several mechanisms: they can be engineered into CAR-T cells, mediate effector functions like ADCC, or directly trigger tumor cell apoptosis. Unlike TCR-T therapy, TCR-like antibodies do not require complex cell manipulation or patient-specific expansion. They can be used as “off-the-shelf” products, broadening the potential of cellular immunotherapy (133, 134). The introduction of nanotechnology further enhances the efficacy and safety of adoptive cell therapy. Nanocarriers enable targeted delivery of gene editing tools, RNA, or cytokines. This approach improves immune cell activation, function, and persistence in vivo, while helping to modulate the immunosuppressive tumor microenvironment (135–138). Simultaneously, nanotechnology can also serve as a novel regulation method, providing new technical means for precisely controlling effector cell function (71, 139).

Emerging strategies are looking beyond directly targeting cancer cells. Instead, they focus on remodeling the non-tumor components of the TME, such as cancer-associated fibroblasts, macrophages, and dendritic cells, and disrupting their immunosuppressive interaction networks. Cancer-associated fibroblasts that express fibroblast activation protein are particularly important in iCCA. They not only build a dense scar-like barrier but also orchestrate a suppressive network by secreting chemokines. Specifically, the signaling axis involving signal transducer and activator of transcription 3 and C-C motif chemokine ligand 2 actively recruits myeloid-derived suppressor cells, while C-X-C motif chemokine ligand 12 and its receptor contribute to the exclusion of effector T cells (140, 141). Preclinical studies show that CAR-T cells designed to target fibroblast activation protein can successfully remove these supportive cells. This depletion severs the communication link between cancer-associated fibroblasts and myeloid-derived suppressor cells, breaks down the physical barrier and makes tumors more responsive to subsequent treatment with tumor-targeting CAR-T cells (142, 143). Parallel to stromal depletion, exploiting the plasticity of myeloid cells represents another prospective avenue. Tumor-associated macrophages typically maintain an immunosuppressive M2 phenotype through crosstalk with tumor cells via the PD-1/PD-L1 axis (144). Chimeric antigen receptor macrophages are a new tool that can penetrate dense tumors and reverse this polarity. Beyond mediating phagocytosis, CAR-macrophage cells secrete pro-inflammatory cytokines that repolarize the local niche from M2 to M1 (145–147). A phase I trial (NCT04660929) evaluating anti-HER2 CAR-macrophages is currently underway, offering a translational precedent for HER2-positive BTC. Furthermore, effective anti-tumor immunity relies on the interplay between T cells and antigen-presenting cells, which is often disrupted in the TME. Engineered T cells secreting Fms-like tyrosine kinase 3 ligand and lymphotactin have been shown to actively recruit and activate conventional type 1 dendritic cells. This restored interaction between dendritic cells and T cells promotes robust antigen spreading and endogenous T cell clonal expansion, effectively counteracting tumor heterogeneity (148). These multi-pronged strategies are therefore particularly promising for the treatment of BTC. The tumor microenvironment in BTC is typically highly complex, fibrotic, and immunosuppressive. By targeting these critical cellular nodes and their communication links, this integrated approach aims to transform this hostile setting into one that enables sustained anti-tumor immunity.

Future breakthroughs in adoptive cell therapy for BTC will require a smart and combined use of the strategies we have discussed. The next generation of treatments will likely not rely on a single type of engineered cell. Instead, they will form a living drug system that can dynamically adapt to BTC’s complex tumor microenvironment. This approach promises to overcome key challenges like immunosuppression, T-cell exhaustion, and target heterogeneity. Ultimately, it could provide BTC patients with safer and more durable treatment options.

The treatment of biliary tract cancer faces severe challenges due to its high heterogeneity and immunosuppressive microenvironment. Adoptive cell therapies, including CAR-T, TIL, NK, CIK, and TCR-T, offer new hope for advanced patients by activating or engineering immune cells to precisely target tumors. Although showing potential in target selection and clinical research, this field remains constrained by key bottlenecks such as TME suppression, limited effector cell infiltration and functional exhaustion, target heterogeneity, and off-tumor toxicity. Future breakthroughs depend on strategies like genetic engineering and combination with immune checkpoint inhibitors to synergistically overcome these barriers, thereby promoting clinical translation and improving patient outcomes.