Tumor immunotherapy uses the body’s immune cells to eliminate tumor cells, with T cells being the main effector cells. Tumor elimination by T cells begins with the recognition of tumor antigens (8), which are self-molecules altered by gene mutation, protein truncation, protein misfolding, or abnormal post-translational modification (9). Tumor antigens are divided into tumor-associated antigens (TAAs) and tumor-specific antigens (TSAs) (10).TAAs are macromolecules found on the surface of tumor cells and can also be found in normal tissues. Nevertheless, TAAs are highly expressed in tumors, most of which have central tolerance and weak immunogenicity (11). TSAs, or neoantigens, are not expressed in normal somatic cells and do not exhibit central tolerance. TSAs are more specific to tumor cells and have strong immunogenicity. Neoantigens can be presented on the cell surface and subsequently recognized by T cells under the action of major histocompatibility complex (MHC) molecules, causing T cell activation and promoting T cell-mediated attack and removal of tumor cells (12). Neoantigens contribute to tumor-specific immune responses and have been used as targets for novel, precise, and personalized tumor immunotherapy (13).Currently, neoantigen therapy is mainly used in immune checkpoint inhibitors (ICIs) and cellular immunotherapy (adoptive immunotherapy and tumor vaccines).

The tumor microenvironment (TME) refers to the environment required for tumor cell appearance and development and contains tumor cells, extracellular matrix, immune cells, and the cytokines, metabolites, and exosomes they release (14). Different organs have biologically unique microenvironments, and different thyroid cancer types also have greatly differing immunological TMEs (15). Immune cells in the ATC TME mainly consist of tumor-associated macrophages (TAMs), tumor-associated mast cells (MCs), dendritic cells (DCs), natural killer cells (NKs), cytotoxic T cells (CTLs), and regulatory T cells (Tregs) (16). However, some scientists believe that ATC only involves cancer cells, macrophages, and vascular endothelial cells and almost no other cells (17). Due to the low prevalence and low possibility of surgery for ATC, most of the current studies address the role of the phenotypic characteristics of tumor-associated immune cells in the pathogenesis or progression of papillary thyroid carcinoma, whereas there are few studies on ATC.

Thyroid cancer has an extremely wide range of TAM density, with TAM density being the highest in ATC (18). TAMs can directly promote tumor occurrence, development, and metastasis by releasing various inflammatory factors, growth factors, and matrix proteases. TAMs can also indirectly promote tumor progression by mediating tumor angiogenesis and immunosuppression (19, 20). TAMs are mainly divided into M0, M1, and M2 types (21). M0 TAMS are usually dormant, while the proinflammatory M1 TAMs mainly produce cytokines such as tumor necrosis factor-α (TNF-α) and interleukin-1 (IL-1), which inhibit and kill tumor cells. The anti-inflammatory M2 TAMs produce factors such as IL-10 and IL-13, which promote tumor formation and development (22). In ATC, TAMs account for >50% of nuclear cells, and the M2 type is predominant (17). M2-like TAMs accelerate ATC cell metastasis by activating the insulin receptor (IR)-A/insulin-like growth factor 1 receptor (IGF1R)-mediated PI3K–AKT–mTOR pathway and upregulating IGF-1/IGF-2 (23). It is also believed that only ATC contains a branched TAM network to promote ATC invasion through metabolic and nutritional functions (17).

The ATC microenvironment has significantly increased tumor-infiltrating lymphocytes (TILs) as compared with normal thyroid tissues, and most of them are CD8+T cells (24). However, only one investigator demonstrated that CD8+ T cells enhance their killing effect on ATC cells by secreting granzyme, TNF-α, and interferon-γ (IFN-γ) (25). In patients with ATC, NKs kill tumor cells by inducing apoptosis via the release of cytolytic granules (26). MCs are the first immune cells recruited during inflammation and are associated with ATC aggressiveness (27). ATC has a higher number of MCs than other forms of thyroid cancer. By producing IL-6, TNF-α, and CXC chemokine ligand 8 (CXCL8)/IL-8, MCs aid epithelial–mesenchymal transition in ATC cells (28). Galectin-9 aids MC support of ATC cell adhesion, angiogenesis, metastasis, and tumor immune escape (29). Most of the body’s professional antigen-presenting cells are termed DCs, yet the amount of DCs in ATC is much lower than that of other thyroid cancers (30, 31).

Soluble mediators such as cytokines and chemokines are mainly released by tumor-infiltrating immune cells and can also be released by cancer cells. ATC cells secrete CXCL9 and CXCL10, which are chemotactic to T cells (24). IL-4 and IL-10 upregulate the anti-apoptotic proteins B-cell lymphoma-2 (Bcl-2) and Bcl-xL to promote ATC cell progression and chemotherapy resistance (32). CXC chemokine receptor 4 (CXCR4) is involved in the local invasion and distant metastasis of ATC cells mediated by stromal cell-derived factor (33). The transcription factor CREB3L1 activates the extracellular matrix signaling pathway, maintains the fibroblast-like characteristics of ATC cells, reshapes the tumor interstitial microenvironment, and drives ATC malignant transformation (34). ATC is also linked to elevated levels of other cytokines, including IL-6, IL-1, granulocyte colony-stimulating factor, macrophage colony-stimulating factor, CXCL8, and CXCR4 (35, 36). One obvious characteristic of ATC is the high percentage of immunosuppressive cytokines (24). By altering the immunological microenvironment, immunotherapy can reinstate the tumor-killing capacity of anti-tumor immune cells (37).

Single-cell RNA sequencing (scRNA-seq) is a novel technology that enables comprehensive analysis of the cell composition and transcriptional phenotype of malignant cells and surrounding immune cells by examining the transcriptome information of individual cells (38). In one study, scholars used scRNA-seq to distinguish tumor cells from normal cells based on differences in copy number (39). They estimated that the average prediction accuracy for identifying tumor cells was 97%. Lu sequenced the single-cell RNA of ATC patients and classified the cell types (40). They divided the cells into eight main types and found that in patients with ATC, the number of endothelial cells decreased significantly, while the number of myeloid cells increased significantly. Furthermore, they also discovered that ATC cells exhibited overexpression of mesenchymal and glial genes. Compared with papillary thyroid carcinoma, the number of M2 macrophages in ATC significantly increased, while the number of M1 macrophages decreased. In ATC tumors, the total depleted cells account for more than 50% of the T cell population. It is also found that TRα1, as a transcription factor, plays a role in inhibiting tumor growth through various signaling pathways. ScRNA-seq revealed that the induction of PAX8 by TRα1 transformed the cell landscape of ATC from one state to another through a transcription program (41). ScRNA-Seq not only deepens our understanding of ATC cell composition and heterogeneity but also offers new insights for personalized therapy and the development of molecular markers.

Immune cells have cell surface receptors known as ICIs, which control T lymphocyte activation and effector activities (42). The most well-characterized ICIs are programmed cell death protein 1 (PD-1) and cytotoxic T lymphocyte antigen 4 (CTLA-4). The other ICIs include lymphocyte activation gene 3, T-cell immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain, T-cell immunoglobulin domain and mucin domain 3 (43). A trial of the PD-1 inhibitor spartalizumab in patients with advanced/metastatic ATC reported an overall response rate of 19%, including a complete response of 7% and a partial response of 12%. PD-L1-positive individuals had a higher response rate (29%) than PD-L1-negative patients (0%) (44). An objective response rate of 16% and a 1-year survival rate of 38% were noted for ATC treated with pembrolizumab or nivolumab (45). However, only one patient in a trial using stereotactic body radiation therapy and the CTLA-4 inhibitor tremelimumab or the PD-L1 inhibitor durvalumab for metastatic ATC lived for more than a year (46).Furthermore, additional research reported the efficacy of pembrolizumab and spartalizumab for patients with ATC, which demonstrated the anti-tumor action of ICIs in ATC management (47, 48).

Before starting immunotherapy, most of the patients in these trials receiving ICIs underwent surgery, radiation, chemotherapy, and even targeted therapy. Six patients with metastatic ATC who had not responded to radiation, chemotherapy, or radioiodine therapy were included in a trial evaluating the combination of lenvatinib and pembrolizumab. The results demonstrated that 66% of the patients had a full response, 16% had stable illness, and 16% had progressing disease. The median progression-free survival for all patients was 16.5 months. Half of the ATC patients were still receiving therapy at the time of data cutoff, with treatment durations of 1–40 months (49). Following surgery, a 67-year-old patient with ATC was treated with the PD-1 inhibitor sintilimab in conjunction with the antiangiogenic drug anlotinib. The patient demonstrated notable tumor shrinkage and an 18.3-month maintained remission (50). After receiving combination therapy consisting of the anti-PD-1 antibody camrelizumab and the multitargeted kinase inhibitor famitinib, a patient with locally advanced unresectable ATC underwent postoperative radiotherapy and complete surgical resection after computed tomography demonstrated a partial lesion response after three cycles. The patient demonstrated a very good quality of life about 24 months after diagnosis (51). Another study reported that three patients with ATC died within six months despite treatment with surgery, radiation, and chemotherapy in addition to pembrolizumab (52). Based on the aforementioned trials, it is believed that the treatment effect might also be connected to the duration of treatment overlap. Hence, more research is needed to fully understand the synergistic mechanism and tolerance of combined treatments. Currently, there are several reports on ongoing ATC studies involving PD-1/PD-L1 inhibitor (NCT05453799, and NCT05119296), or PD-1 combined with other treatment (NCT04171622, NCT04675710, NCT05696548, NCT04238624, NCT05659186, NCT03181100, NCT04400474, and NCT04579757). Please refer to Tables 1 , 2 for a summary of trials.

Adverse effects are unavoidable when using ICIs. According to NCT02404441, grade 1 or 2 diarrhea, pruritus, exhaustion, and fever were the most frequent treatment-related adverse effects in patients with advanced/metastatic ATC receiving spartalizumab (44). The most common grade 2–4 adverse events were hypertension, fatigue, weight loss/anorexia, oral mucositis, diarrhea, joint/muscle pain, and hand-foot syndrome in six patients with metastatic ATC who were treated with combined lenvatinib and pembrolizumab after all other forms of surgery, chemotherapy, or radiotherapy had failed. However, the lenvatinib-induced adverse effects necessitated treatment stopping in two patients (49). One patient receiving pembrolizumab treatment developed severe grade 4 colitis (bloody stools, severe dehydration, and hypotension), necessitating extensive fluid resuscitation and systemic steroids (53). Despite the ICI adverse effects reported in several studies, most of them can be alleviated with symptomatic therapy. Therefore, ICIs have emerged as a new tool in ATC treatment.

Adoptive immunotherapy involves injecting immune cells with anti-tumor activity to either directly kill or activate the body’s immune response to kill tumor cells. It includes TILs, T cell receptor engineered T cells (TCR-T), chimeric antigen receptor T cell therapy (CAR-T), and NK therapy. TIL therapy was effective in treating numerous solid tumors, including melanoma, breast, cervical, head and neck, stomach, liver, esophagus, and lung cancers (54). The effect of autologous TILs (LT-145 or LN-145-S1) on patients with ATC and some other cancers is being studied in NCT03449108. As an alternative to TIL therapy, T cells can be isolated from peripheral blood and targeted to tumor cells by transferring synthetic TCR or CAR using gene therapy techniques (55). TCR-T involves transfecting TCR α- and β-chain genes into T cells, which can recognize TSAs. This results in structural alterations in the TCR antigen-binding region of the T cells, enabling them to specifically recognize the corresponding tumor antigens. T lymphocytes expressing specific TCRs can recognize human leukocyte antigen (HLA)–peptide complexes on the surface of tumor cells. Antigenic stimulation signals are transmitted by phosphorylating the intracellular region of the immune receptor tyrosine activation motifs, which activates the T cell immune response (56). Patients with colorectal cancer, synovial sarcoma, and metastatic melanoma demonstrated a notable clinical response to TCR-T. CAR overcomes some of the limitations of TCR. However, the success of CAR-T therapy has been concentrated in hematologic tumors, whereas only partial progression has been observed in solid tumors to date (57, 58). Ongoing studies are evaluating the safety and tolerability of AIC2 CAR-T cells in patients with relapsed/refractory poorly differentiated thyroid cancer and ATC (NCT04420754). Animal studies have demonstrated that adoptive cell therapy based on NKs significantly inhibited metastatic growth in ATC lung metastasis model mice (59). An ongoing study (NCT05194709) is evaluating the efficacy, safety, and pharmacokinetics of anti-5T4 CAR-NK cells in patients with advanced solid tumors, and the NCT03415100 study is evaluating the safety and feasibility of CAR-NK cell therapy targeting NKG2D ligands in patients with metastatic solid tumors.

TIL-treated patients with advanced cutaneous melanoma experienced many adverse events, such as thrombocytopenia, fever, chills, neutropenia, and tachycardia (60). In colorectal cancer, TCR-T causes severe transient inflammatory colitis (61). The most common adverse effects of CAR-T in other tumors are cytokine release syndrome and immune effector cell-associated neurotoxicity syndrome (62). Grade 3 cytokine release syndrome also occurred in two patients in a phase I study of CD33 CAR-NK cells in patients with relapsed or refractory acute myeloid leukemia (63). However, the adverse effects of adoptive immunotherapy in ATC are unclear and require further clarification.

Tumor vaccines can stimulate the body to generate active specific immunity against tumor cells or cells or molecules in the TME that promote tumor growth. ATC typically has a high mutation burden, and most of the neoantigens in ATC can be used for tumor vaccines. Thus, tumor vaccines might be more effective in ATC (64). The current ATC vaccines mainly include DC vaccines and oncolytic virus (OV) vaccines. Triiodothyronine can enhance DC ability to stimulate cytotoxic T-cell responses, thereby enhancing anti-tumor responses (65). DC vaccines have been successfully used to treat medullary thyroid carcinoma, and the treatment of thyroid cancer might be more effective than that for other tumors (66). OVs use natural or genetically modified viruses to specifically infect and lyse cancer cells but do not harm normal cells. The anti-cancer activities of OVs are derived from multimodal cancer-killing mechanisms. First, OVs infect and replicate in cancer cells, inducing tumor cell lysis and releasing infectious viral progeny that spreads to the surrounding tumor cells. Second, OV-mediated oncolysis of tumor cells initiates the release of tumor-associated antigens, cellular danger-associated molecular pattern signals, and cytokines that promote the maturation of antigen-presenting cells and activate antigen-specific CD4 and CD8 T cell responses. Third, OVs can specifically infect and destroy tumor vascular endothelial cells and stromal cells and destroy tumor blood vessels by promoting endostatin and angiostatin production (67). Furthermore, OVs induced cell death and tumor regression in cultured ATC cells and mouse models and modulated the ATC microenvironment to shift the M2-type TAMs to the proinflammatory M1 phenotype (68, 69). Although adverse events such as fever, injection site reactions, headache, and vomiting occurred in patients with glioma and melanoma treated with the OV vaccine, they were all tolerable (70). Nevertheless, patients with ATC have not received a tumor vaccine. It is expected that the development of DC And OV vaccines will become a means of treating ATC ( Figure 1 ).