Thyroid tissue is highly sensitive to radiation in the early stage of development, so ionising radiation exposure is a clear pathogenic factor of thyroid cancer, especially prominent in children (23). Initial reports linking pediatric exposure to thyroid carcinoma appeared in the 1950s: it was reported that 10 of the 28 thyroid cancer patients aged 18 and under had received X-ray treatment for thymus hyperplasia within 4–16 weeks after birth (24). This relationship was corroborated following the Chernobyl meltdown in 1986, when widespread radioactive contamination drove a surge in thyroid carcinoma diagnoses among children and adolescents residing in the vicinity (25). Additionally, diagnostic procedures such as dental X-rays and CT scans, as well as radiotherapy to the head, neck, or chest, are also considered to increase the risk of thyroid cancer (26, 27).

Ionising radiation triggers oxidative stress in biological systems through direct and indirect means, and affects local and systemic processes through a variety of mechanisms, including accelerating ageing, causing genetic instability and mutation, causing cell membrane damage and cell death, changing enzyme function and cell metabolism, damaging mitochondrial function and initiating the cancer process (28, 29). Further research shows that ionising radiation may trigger RET/PTC gene rearranment (30). In thyroid papillary carcinoma, the RET proto-ocarcinoma gene can produce a variety of chimeric oncogenes, which are collectively known as RET/PTC. RET/PTC rearraising is a key molecular event, and its frequency has been confirmed to increase with the increase of radiation dose among atomic bomb survivors (31). Therefore, reducing ionising radiation exposure, especially among children and adolescents, may play a key role in the prevention of thyroid cancer.

Accumulating evidence identifies BRAF activating mutations and RET/PTC rearrangements as central drivers in the pathogenesis and progression of papillary thyroid carcinoma (Figure 1). Thyroid cell division and maturation are governed by BRAF, a specific serine/threonine kinase. The mutation of this gene will drive the occurrence and development of tumours (32, 33). Among them, the T1799A point mutation of the BRAF gene will produce a BRAF-V600E variant, causing the kinase to be in a state of continuous activation (34, 35). Approximately 45% of papillary thyroid carcinoma cases harbor the BRAF-V600E mutation, and the mutation is closely related to adverse clinical pathological characteristics, including invasive histological types, higher risk of recurrence and reduced reactivity to radioactive iodine treatment (36, 37). RET/PTC rearrangements, among the earliest genetic events in thyroid tumorigenesis, are exclusively identified in PTC and occur in 20–40% of sporadic adult tumors (30, 38).

TSH is a well-known stimulator of thyroid cells. Excess TSH continuously stimulates follicular cells, leading to their hypertrophy and proliferation, and ultimately undergoing malignant transformation through stages of nodules, focal proliferation, and adenomas (39). Multiple studies have recognized serum TSH as an independent marker for diagnosing thyroid malignancies, with higher TSH concentrations correlating with increased cancer risk and greater tumor aggressiveness in well-differentiated thyroid cancer (40, 41). Despite remaining within standard limits, this factor is linked to increased cancer susceptibility and accelerated tumor growth (42, 43).

There is a significant gender difference in thyroid cancer: the rate of women suffering from the disease is 2.9 times that of men (44). The carcinogenic effect of oestrogen is mainly realised by its binding to the receptors ER-α and ER-β, and these complexes then interact with oestrogen-reactive elements, thus activating multiple cell survival pathways (45). Therefore, the endoplasmic reticulum has become a key cell and transcription regulation hub in the proliferation mechanism of endocrine-related malignant tumours. Studies show that oestrogen can promote the proliferation of human thyroid cancer cells by activating the filament activated protein kinase (MAPK) pathway (46). Estrogen may also stimulate transfer phenotypes, including adhesion and migration, by inhibiting the components of the calcium mucin-cyclic protein system (47).

Iodine intake is closely related to the frequency of detection of thyroid cancer. Data shows that in iodine-deficient areas, the proportion of invasive ATC and FTC is higher (48). In the normal physiological state, the proliferation rate of thyroid follicular cells is slow. However, when iodine intake is insufficient, elevated serum thyroid-stimulating hormone (TSH) will strongly stimulate its proliferation, increasing its growth rate by 5 to 30 times, resulting in significant follicular cell proliferation and glandular enlargement (49). When thyroid cells are in a state of rapid proliferation, their sensitivity to mutagenic factors such as radiation, chemical carcinogens and oxidative stress may be significantly enhanced, making it easier to accumulate additional genetic mutations.

On the contrary, it is also important to note that excessive intake of iodine can be harmful. Excessive iodine consumption exacerbates the risk of dysfunction, particularly in subjects with a history of thyroid pathology or antecedent iodine deprivation (50). Thyroid dysfunction associated with iodine excess is most often clinically insignificant and reversible, whereas iodine-induced hyperthyroidism may, in susceptible individuals, reach a life-threatening severity. Evidence has indicated that excessive iodine consumption may be associated with thyroid carcinogenesis, potentially mediated by enhanced oxidative damage to DNA (51). Therefore, more long-term large-scale case-control study data is needed to accurately estimate iodine intake in order to evaluate the potential relationship with iodine deficiency and excessive iodine intake.

Lipid metabolism in tumors is primarily characterized by significant reprogramming of fatty acid (FA) metabolism. Disruption of the balance between FA synthesis and degradation can profoundly affect key hallmarks of tumor biology, including, but not limited to, cancer cell migration, proliferation, apoptosis, and local invasion (52). Dysregulation of fatty acid metabolism contributes to thyroid cancer development. Upregulation of stearoyl-CoA desaturase 1 (SCD1) is evident in both ATC and well-differentiated subtypes, serving a critical function in maintaining ATC cell viability and growth (53). Acetyl-CoA carboxylases (ACCs) are rate-limiting enzymes in the cytoplasm that regulate fatty acid metabolism. Research indicates that circPCNXL2 is overexpressed in PTC tissues and cells, promoting cell proliferation by inhibiting ACC1 phosphorylation and enhancing fatty acid synthesis (54). Therefore, given that the high progression of thyroid cancer is closely related to lipid metabolism disorders, intervention in these metabolic pathways has become a highly potential treatment direction.

In the tumour microenvironment (TME) of thyroid cancer, macrophages, as the main interstitial component, have a double regulating effect on the course of the disease: they can both inhibit and promote the development of cancer (55, 56). T In the TME of thyroid cancer, tumour-assiated macrophages (TAMs) usually differentiate into two phenotypes with very different functions: M1 and M2. In the early stage of cancer, M1 macrophages play the role of clearing cancer cells and activating immune responses; while M2 tends to promote tumour growth and spread. Current research generally believes that the functional imbalance of the TME is an important driving force for the progression and metastasis of thyroid cancer (57). Although the body’s immune system can effectively remove early tumour cells, the TME will eventually help cancer cells escape immune attacks by inhibiting the killing ability of immune-effecting cells (58–60). The infiltration density of M2 macrophages is significantly positively correlated with the malignant progression of thyroid cancer, which makes the cell group a highly potential key treatment target (61, 62).

The iron transport system in the cell precisely regulates the steady-state balance of iron. If this system fails, it may lead to iron overload, which in turn becomes a key factor in iron poisoning (74). Physiologically, iron serves as a fundamental trace element, present primarily in Fe²⁺ or ferric (Fe³⁺) states (75). Following intestinal uptake or the breakdown of erythrocytes, the plasma ferroxidase ceruloplasmin oxidizes Fe²⁺ species into the Fe³⁺ form (76). Following this, iron combines with transferrin to form a complex, which is taken up by cells via transferrin receptor 1 (63). In the endosome, metal reductase STEAP3 reduces iron ions to Fe²⁺, and then transports them to the cytoplasm by divalent metal ion transporter 1 (77). After entering the cytoplasm, iron can enter the dynamically unstable iron pool or be stored in ferritin - ferritin is a polymeric assembly structure composed of light chains and heavy chains (78, 79). Surplus cellular iron is eliminated via Ferroportin, the efflux channel encoded by SLC40A1, which facilitates the translocation and subsequent oxidation of iron ions (80).

The precise absorption, storage and release regulation of iron is the key to maintaining the iron homeosis of cells (81). Once this dynamic balance is broken, it will cause intracellular iron overload. Excess iron ions will catalyse lipid peroxidation through Fenton reaction and Haber-Weiss cascades, thus irreversibly triggering ferroptosis (77, 79). The latest research shows that disorders of iron metabolism can lead to excessive accumulation of iron elements in cells, which has been proven to be a key driver of inducing ferroptosis (82, 83). Elucidating the molecular controls governing iron metabolism is essential for defining the fundamental pathways of ferroptosis (84).

The characteristic feature of ferroptosis is the fatal accumulation of lipid peroxide. Its toxicity mainly comes from the oxidation of polyunsaturated fatty acids (PUFAs) (85). PUFAs, especially arachidonic acid and prostaglandin E2, are the main targets of lipid peroxidation reactions. The oxidation process can be initiated by specific enzyme catalysis or spontaneous non-enzymatic chemical reactions (86, 87). In the enzyme catalytic pathway, arachidonic acid is first activated into AA-CoA under the action of ACSL4, and then AE-PE is esterified by LPCAT3 and phosphatidylethanolamine. The complex undergoes an oxidation reaction under the catalysis of lipoxygenase or cytochrome P450 oxidase, and finally produces lipid peroxide (88, 89). The Fenton reaction promotes the non-enzymatic lipid oxidation reaction by producing highly active hydroxyl radicals (90). The accumulation of these toxic substances will change the biophysical properties of cell membranes and cause oxidative damage to intracellular biological macromolecules and organelles, eventually causing iron-dependent cell death.

Inhibiting the activity of ACSL4 or LPCAT3 can reduce the accumulation of phospholipid peroxide and make cells resistant to ferroptosis (91). In contrast, monounsaturated fatty acids (MUFAs), such as oleic acid, limit the integration of polyunsaturated chains into membrane lipids through competitive assimilation. This competitive displacement limits the substrate available for oxidation, effectively dampening cellular susceptibility to ferroptosis (92). Dissecting the functional interplay among ACSL4, LPCAT3, LOXs, and POR is fundamental to decoding the etiology of ferroptosis and pinpointing new strategies for intervention.

Amino acid metabolism exerts substantial control over ferroptosis, primarily by governing the System Xc^-–glutathione–GPX4 signaling axis (93). The Xc^- system, composed of the SLC7A11 and SLC3A2 subunits, serves as a cysteine/glutamate antiporter. Its primary physiological role is to facilitate the export of intracellular glutamate in exchange for the import of extracellular cystine (16). Following internalization, cystine undergoes reduction to form cysteine, thereby supplying the rate-limiting substrate for glutathione biosynthesis. Subsequently, the sequential catalytic action of glutamate-cysteine ligase and glutathione synthase assembles glutamate, cysteine, and glycine into glutathione, the obligatory cofactor for GPX4 (94). As a key phospholipid peroxidase, GPX4 uses glutathione to reduce toxic phospholipid peroxides into harmless alcoholic substances, thus protecting the cell membrane structure and effectively inhibiting the occurrence of ferroptosis (66).

When using substances such as elastin, sorafenib, sulfasalazine, or high concentration of glutamic acid to block the Xc− function of the system, it will interfere with the uptake of cysteine, reduce the level of glutathione in cells, inhibit GPX4 activity, lead to the accumulation of phospholipid peroxide and cause ferroptosis (94, 95). Similarly, direct inhibition of GPX4 (such as the use of RSL3) will also destroy the redox homeostas and promote the occurrence of fatal lipid peroxidation (96). In addition to this classical pathway, other metabolic pathways also participate in the regulatory network of ferroptosis, including the transsulphur pathway of converting methionine into cysteine, and the regulatory role of glutamine catabolism. These mechanisms together reveal the multi-level regulation of amino acid metabolism to ferroptosis sensitivity in normal cells and malignant cells (97, 98).

Several pathways suppress lipid peroxidation through distinct mechanisms, primarily the ferroptosis inhibitory protein 1(FSP1)–CoQ10, mitochondrial DHODH–CoQ10, and GPX4 systems, and their inactivation directly contributes to the occurrence of ferroptosis. FSP1, located at the plasma membrane, reduces CoQ10 to its active form using NAD(P)H, scavenging lipid peroxyl radicals independently of the classical GSH–GPX4 pathway (99). Evidence suggests that FSP1 functions in tandem with GPX4 to orchestrate the control of ferroptosis within retinal pigment epithelial tissues (100, 101). In mitochondria, besides GPX4, DHODH also reduces CoQ10, preventing ferroptosis (102). Together, GPX4 (cytoplasm and mitochondria), FSP1 (plasma membrane), and DHODH (mitochondria) form three complementary antioxidant defense systems (103).

It is worth noting that the functions of DHODH and GPX4 in mitochondria can compensate for each other, but cytoplasmic GPX4 or plasma membrane FSP1 cannot replace the mitochondrial protective function of DHODH (104, 105). On the contrary, the impaired function of DHODH can promote ferroptosis. Type I interferon, for instance, enhances manganese-induced ferroptosis via enzyme inhibition, whereas DHODH silencing can restrain the expansion of cervical cancer cells and promote ferroptotic cell death.

Beyond these canonical pathways, the cellular antioxidant repertoire includes several alternative defense systems. For example, the GCH1-mediated tetrahydrobiotic (BH4) synthesis pathway can not only directly remove lipid peroxides, but also promote endogenous coenzyme Q10 synthesis through membrane remodelling, thus indirectly inhibiting ferroptosis. Coenzyme Q10 is an endogenous lipid with redox activity. Its side chain is composed of ten isopentane units, which is crucial to its function of maintaining cell homeostasis (100). Moreover, depletion of CoQ10H₂ directly induces ferroptosis (106). Collectively, these studies demonstrate a close association between antioxidant systems and the regulation of ferroptosis.

As one of the most important cancer suppressor genes, p53 regulates many key life activities such as cell cycle blockage, ageing, apoptosis and ferroptosis, thus effectively inhibiting the occurrence and development of malignant tumours (107, 108). Under oxidative stress conditions, p53 will participate in the regulation of cell death, and its effect will converge in ferroptosis and apoptosis pathways. For example, in lung cancer cell experiments, it was found that when the p53 gene is silenced, most of the cells can still survive under reactive oxygen exposure, with a survival rate of more than 90%; on the contrary, when reactive oxygen activates p53, the cell death rate will rise to about 40%, and this death effect can be affected by the ferroptosis inhibitor Fer- 1 Significantly prevent. This shows that the activation of p53 will guide cells towards ferroptosis under certain conditions (109).

P53 regulates ferroptosis through a variety of molecular mechanisms. By suppressing SLC7A11, p53 reduces cystine import through system Xc^-, weakens GPX4 activity, and promotes ROS accumulation, ultimately inducing ferroptosis (107, 110). Moreover, as described by Ou et al. (2016), p53 can facilitate ferroptotic cell death through the SAT1–ALOX15 axis (111). In this study, the induction of SAT1 did not affect the expression level of GPX4, and the overexpression of SLC7A11 could not save cell death, indicating that this pathway was independent of GPX4 and system XC ^-. Further experiments showed that SAT1 depended on ALOX15 (an enzyme catalyzing arachidonic acid peroxidation) to trigger ferroptosis, which could be effectively blocked by an ALOX15 inhibitor.

However, some studies also pointed out that p53 may inhibit ferroptosis under certain conditions. Tarangelo et al. (2018) found that stable expression of wild-type p53 decreases cellular susceptibility to ferroptosis and diminishes system Xc^- activity (112). Based on studies of CDKN1A (p21), it has been proposed that p53, via the p21 pathway, can restrict ferroptotic cell death in tumors. This highlights the dual role of p53 in ferroptosis: both negative regulation by inducing p21 and positive regulation by inhibiting the system Xc−. The regulatory effect of p53 on ferroptosis is highly situationally dependent, and it is dynamically regulated by the cell microenvironment and the dominant signalling pathway, so its precise mechanism still needs to be continuously explored (113).

The voltage-dependent anion channel (VDAC), residing in the mitochondrial outer membrane (MOM), mediates both the exchange of ions and metabolites and the generation of ROS (114). Abnormal function of the VDAC channel contributes to the execution of ferroptotic cell death. Blocking VDAC oligomer formation mitigates the accretion of both intracellular iron and reactive oxygen species, thereby conferring cytoprotection against ferroptosis via the curtailment of lipid oxidation (115). Especially in the mouse model, inhibitors such as vbit-12 and fer-1 showed protective effects. In addition, Yagoda et al. (2007) demonstrated that the direct association of erastin with VDAC induces mitochondrial dysfunction and subsequent oxidative burst, thereby initiating the ferroptotic process (116). Specific isoforms, particularly VDAC3, have been implicated in the iron-dependent cell death pathways activated by erastin and RSL5 (117), although the specific mechanism is not completely clear (118).

Beyond the VDAC axis, the regulation of ferroptosis encompasses several secondary mechanisms, including sulphur metabolism, heme oxygenase-1 activity, transferrin routing, and the peroxidative cascade of polyunsaturated fatty acids. Iron ions potentiate the accrual of lipid reactive oxygen species, and this combination of factors collectively drives the onset of cellular demise (119). In essence, VDAC is prerequisite to the convergence of mitochondrial metabolism and oxidative stress; however, its precise role in the ferroptosis pathway still awaits thorough elucidation.

Within the TME, diverse cell populations modulate ferroptotic susceptibility, functioning to either suppress or drive the pathway. For instance, CD8+ T cells sensitize tumors to ferroptosis through the secretion of interferon-γ (IFNγ) (120, 121). In terms of mechanism, IFNγ simultaneously downregulates SLC7A11 to limit the intake of cysteine, and upregulates ACSL4 to promote the biosynthesis of polyunsaturated phospholipids rich in ararachidic acid. This process weakens the antioxidant ability and promotes lipid peroxidation at the same time, eventually leading to the death of ferroptotic cell death. Therefore, the simultaneous existence of IFNγ and arachidonic acid in the TME has become a key determinant of ferroptosis in cancer cells (120, 121).

In addition to CD8+ T cells, neutrophils also promote ferroptosis by transferring their particles containing medullary peroxidase to tumour cells (120). In contrast, cancer-related fibroblasts (CAFs), which have a profound impact on both congenital and adaptive immunity, tend to inhibit ferroptosis (122). This function of CAFs can be achieved not only by increasing the expression of long-stranded non-coding RNA DLEU1, but also by releasing miR-522 (123). In gastrointestinal tumors, anoctamin 1 inhibits ferroptosis and elevates TGF-β release, which attracts CAFs and weakens CD8⁺ T cell–driven antitumor immunity, thereby promoting immunotherapy resistance (124). These findings highlight the reciprocal interactions among tumor cells, fibroblasts, and immune cells in dictating ferroptotic outcomes.

Cytokine signalling further increases this complexity. TGF-β1 will activate the signalling pathway that depends on SMAD, thus promoting the ferroptosis of cancer cells; while interleukin-1β (IL-1β) enhances the production of NADPH and stabilises iron-sulphur clusters by acetylation of niacinamide nucleotide transhydrogenase, thus inhibiting ferroptosis (125, 126).

In addition to cytokines, metabolic signals in the TME also play a decisive role. This environment is usually characterised by high accumulation of lactic acid and limited glucose availability (127, 128). Under the condition of glucose deficiency, AMPK will phosphorylate acetyl coenzyme A carboxylase to maintain resistance to ferroptosis (129, 130). At the same time, excessive lactic acid not only inhibits immune function, but also enhances the resistance of tumour cells to ferroptosis by stimulating the synthesis of monounsaturated phospholipids or further acidifying the TME (131, 132). Interestingly, acidosis is not universally inhibits ferroptosis; cancer cells in an acidic environment are particularly vulnerable to ferroptosis induced by n-3 and n-6 polyunsaturated fatty acids (133).

In a word, the regulation of ferroptosis in the tumour is an interaction between cytokines derived from immunity and matrix and the metabolic status of TME changes, depending on the specific situation. Future treatments will need to use the properties that promote ferroptosis while offsetting its resistance mechanism, so as to shift the balance to ferroptosis in favour of cancer cells.

Macrophages, neutrophils, and natural killer (NK) cells are the principal constituents of the innate immune system, forming the frontline immunological barrier against tumor development (134). In addition to directly killing malignant cells, they also coordinate and enhance adaptive immune responses. However, their function is strictly regulated by the ferroptosis and the biologically active metabolites it produces (135). In the TME, tumour-related macrophages (TAMs) mainly exhibit immunosuppressive M2 phenotypes, which prompts extensive research to focus on reducing the number of M2 type TAMs or reprogramming them into anti-tumour M1 states (136). Recent findings demonstrate that M1 TAMs exhibit enhanced resilience to ferroptosis compared to M2 TAMs, suggesting that susceptibility to this cell death modality is dictated by the macrophage functional state (137).

The differential vulnerability of tumor-associated macrophages (TAMs) to ferroptosis is attributable, in part, to their distinct metabolic profiles and molecular identity. The different susceptibility of M1 and M2 TAMs to ferroptosis can be partly attributed to the differences in their metabolism and molecular properties. These free radicals can effectively remove lipid free radicals produced by the 15-lipooxygenase (15-LOX) pathway, thus protecting M1 type TAMs from ferroptosis (137). In addition, M1 macrophages showed an increase in the level of ferritin heavy chain 1 - which contributes to ferritin assembly and intracellular iron storage - at the same time, the level of iron-output transporter iron transporter is reduced (138). These changes work together to reduce the unstable iron pools in M1 TAMs, thus enhancing their inherent resistance to ferroptosis.

Given their heightened vulnerability, M2 TAMs are more susceptible to ferroptosis than M1 subtypes. This suggests a therapeutic strategy for the selective elimination of immunosuppressive M2 macrophages while sparing the M1 population. Such approaches not only purge the immunosuppressive population but also potentially re-engineer the surviving macrophages toward an anti-tumor M1 phenotype, consequently boosting the therapeutic power of cancer immunotherapy (139). Supporting this concept, many studies have shown that iron nanoparticles, a powerful ferroptosis inducer, can drive the transformation of M2 TAMs into M1-like cells, ultimately enhancing their anti-tumour function (140, 141).

In addition to the method of nanoparticle-mediated, the signalling pathway also has a key impact on ferroptosis and macrophage polarisation. For example, receptor tyrosine kinase TYRO3 has become a key regulatory factor in the TME (142). TYRO3 fosters an immunosuppressive microenvironment through a dual mechanism: counteracting ferroptosis in cancer cells and shifting the M1/M2 macrophage ratio towards the M2 subtype. Conversely, pharmacological blockade of TYRO3 simultaneously compels tumor cells into ferroptosis, shifts macrophage polarization toward the anti-tumor M1 state, and potentiates the therapeutic response to PD-1 checkpoint inhibition (142).

However, emerging evidence indicates that the relationship between tumour-associated macrophages (TAMs) and ferroptosis is more complex and subtle than previously recognised. In some contexts, ferroptosis promotes the accumulation of TAMs and drives their polarisation toward an immunosuppressive M2 phenotype, whereas blocking ferroptosis has been shown to hinder this M2 polarisation. This context-dependent effect has been observed in glioblastoma models, highlighting the dual and situational roles of ferroptosis in regulating TAM behaviour (143, 144).

In short, these studies highlight a complex and situational relationship between ferroptosis and macrophage polarisation, revealing the seemingly contradictory effect of ferroptosis against tumour immunity.

In the TME, medullary inhibitory cells (MDSCs) are the main cell groups that actively inhibit immune response (145). The significant ferroptosis resistance of these cells is mainly due to the inhibitory effect of ASAH2 on the p53–HO-1 signalling axis (146). Targeting ASAH2 by drug or genetic means can trigger ferroptosis in MDSCs, which then enhances the infiltration of CD8+ T cells and strengthens the anti-tumour immune response (146).

Attention has also been paid to the prospective advantages of combining methods that promote ferroptosis with strategies targeting MDSCs. In the mouse model of liver cancer, the deletion of GPX4 increased the infiltration of CD8+ T cells. However, it also increases the expression of PD-L1 on tumour cells and promotes the collection of polymorphic nuclear MDSCs, ultimately limiting the tumour inhibition effect (147). Importantly, the combined use of GPX4 inhibitors and the strategy of inhibiting the recruitment of MDSCs significantly improved the therapeutic effect of anti-PD-1 and prolonged the survival of mice with GPX4 wild tumours (147).

Recent evidence shows that polymorphic nuclear medullary inhibitory cells (PMN-MDSCs) play a role that depends on specific situations and sometimes seems contradictory in the regulation of ferroptosis (148). In the TME, PMN-MDSCs themselves are prone to ferroptosis, partly due to hypoxia leading to the downward regulation of GPX4 expression. Paradoxically, inducing ferroptosis of such cells may enhance their immunosuppressive activity, which is likely to be achieved by releasing phospholipid oxide that can inhibit the reaction of T cells. Therefore, inhibiting ferroptosis rather than inducing ferroptosis may more effectively inhibit tumour growth by reducing the immunosuppressive effect mediated by PMN-MDSCs; while the use of ferroptosis inducers for treatment may inadvertently accelerate the progression of the tumour (148).

It is worth noting that some observations differ from most of the findings obtained in the immunosound mouse model. Although most studies show that ferroptosis, which triggers immunosuppressive cells, can enhance the body’s anti-tumour immune response, and systemic administration of ferroptosis inducers in immune-healthy mice can usually inhibit tumour progression, several significant exceptions have been reported. For example, those neutrophils that infiltrate tumours and exhibit immunosuppressive characteristics similar to PMN-MDSCs show strong resistance to ferroptosis (149). The observed ferroptosis resistance is related to the up-regulation of acetic acid decarboxylase 1 (ACOD1), which can promote the synthesis of lycolic acid and trigger the protection mechanism mediated by NRF2 against ferroptosis. On the contrary, the deletion of ACOD1 will reduce the recruitment of neutrophils, enhance the anti-tumour immune response, and enhance the effect of immunotherapy intervention. The root cause of these differences may be due to differences in experimental design, differences in animal models, or the intrinsic heterogeneity of PMN-MDSCs and tumour-related neutrophils themselves. This highlights the urgent need for in-depth research on how ferroptosis regulates the TME and thus affects the anti-tumour immune response.

NK cells are another vital part of the intrinsic immune system (150). In the TME, the oxidative stress caused by lipid peroxidation is related to NK cell disfunction. Activating the NRF2 pathway can restore the function and anti-tumour efficacy of NK cells, while blocking ferroptosis helps NK cells survive in the tumour (148, 151). In general, these observations show that NK cells are particularly susceptible to ferroptosis in the TME.

In short, these results highlight the complex and situationally dependent effects of ferroptosis on different inherent immune cell groups in the TME. Immunosuppressive M2 tumour-related macrophages and tumour-infiltrated neutrophils seem to be particularly sensitive, and natural killer cells also show significant susceptibility. On the contrary, medullary inhibitory cells (MDSCs) show different degrees of resistance or susceptibility according to the specific conditions of the TME. A more comprehensive understanding of how ferroptosis interacts with innate immune cells in the TME is crucial to the rational development and improvement of cancer immunotherapy.

Adaptive immune cells, especially B cells and T lymphocytes, will undergo functional changes in the TME. These changes enable them to further regulate the development of tumours (152, 153). The mechanism is rooted in the iron-deficient cancer cell secretion group, in which the components (including lipid peroxide) will destroy the function of effective immune cells and weaken their ability to identify tumour antigens. Similar to inherent immune cells, the adaptive immune cell group itself is inherently susceptible to ferroptosis, but this sensitivity also depends on the specific biological background.

B cells are very important for anti-tumour immunity, mainly by producing antibodies and regulating T cell responses (154). It is worth noting that congenital B-like cells, including B1 cells and marginal B cells, have particularly active lipid metabolism and rely on GPX4 to maintain their antibody production and overall function. Therefore, when GPX4 is missing, these cells are highly sensitive to ferroptosis (155). Although the data shows that targeting GPX4 may become a feasible strategy for specific B cell malignant tumours, comprehensive research is still needed to fully understand the extent to which ferroptosis affects the anti-tumour immune function of B cell dependence.

Many studies have shown that under various experimental conditions, T cells show considerable resistance to ferroptosis, which means that ferroptosis in cancer cells can be specifically induced without compromising the anti-tumour response mediated by T cells. Preclinical data demonstrate that genetic ablation of Slc7a11 in murine models suppresses tumor expansion without compromising T cell proliferation or function. Consequently, this intervention has been shown to potentiate the therapeutic efficacy of immune checkpoint blockade (156). Similarly, ferroptosis induced by cysteine deprivation or supplementation of arachidonic acid supports (and even enhances) the proliferation and effect function of T cells in the TME while limiting tumour growth. On the contrary, inhibiting ACSL4-driven ferroptosis will weaken the anti-tumour activity mediated by T cells (120, 121).

Inhibiting GPX4 by chemical inhibition or gene knockout can promote an increase in the number of CD4+ and CD8+ T cell groups in the three-negative breast cancer mouse model (139). In addition, it is reported that a variety of nanoparticle-mediated ferroptosis induction methods can also increase the infiltration of T cells into tumours (157, 158). In short, these research results show that in many cases, inducing ferroptosis does not impair the viability of T cells, but will enhance the anti-tumour immune response mediated by T cells (159, 160).

Contrary to previous reports that T cells can generally resist ferroptosis, some studies have shown that inhibiting GPX4 can make T cells very vulnerable. In the model of specific knockout of GPX4 in T cells, these cells will quickly accumulate lipid peroxide in their membranes and cause ferroptosis (161). Notably, CD8+ T lymphocytes exhibit greater vulnerability to ferroptosis induced by GPX4 inhibition compared to malignant cells (162). Elevated CD36 expression in tumor-infiltrating CD8+ T cells, which facilitates fatty acid uptake, exacerbates lipid peroxidation and ferroptosis, thereby diminishing their overall anti-tumor efficacy. It is noteworthy that the function of CD8+ T cells can be saved by CD36 gene knockout or drug-induced ferroptosis inhibition (163, 164).

The GPX4-mediated ferroptosis protection is crucial for CD4+ T cells. Regulating T cells (Treg cells), which are famous for their inherent immunosuppressive activity and inhibition of anti-tumour immune response, show significant resistance to ferroptosis, which is likely due to their high GPX4 levels. Selective deletion of GPX4 in Treg cells will cause ferroptosis, which will lead to enhanced anti-tumour immunity (165, 166). Similarly, the survival and function of CD4+ T cell subgroups that promote anti-tumour immunity - follicular-assisted T cells (T_FH cells) also depend on GPX4 to maintain, although the specific pathways of controlling the ferroptosis of T_FH cells in the TME have not been fully clarified (167).

In general, these research results show that triggering ferroptosis by missing SLC7A11 or GPX4 can differentially affect the function of different T cell subgroups. T cells exhibit a marked dependence on GPX4 rather than SLC7A11. This reliance likely reflects the low abundance and minimal functional contribution of SLC7A11 within this lineage (168, 169). By inhibiting SLC7A11 or supplementing arachidonic acid-triggered ferroptosis, the anti-tumour response driven by CD8+ T cells can be maintained and in some cases enhanced. In contrast, inhibiting GPX4 produces different results according to different T cell subgroups: triggering ferroptosis in immunosuppressive Treg cells will enhance the anti-tumour immune response, while inhibiting GPX4 in CD8+ or T_FH cells will impair their anti-tumour function.

Together, these observations at the level of adaptive immunity illustrate that the immunological consequences of ferroptosis depend not only on the pathway being targeted, but also on the specific immune cell subsets involved.

Extending beyond adaptive immune cells, ferroptosis within the tumor microenvironment affects tumor cells as well as multiple innate immune populations, resulting in divergent immunological outcomes. Ferroptosis in tumor cells is frequently accompanied by lipid peroxidation and the release of stress-associated signals, which can support antigen presentation and promote immune cell recruitment under certain conditions (120, 139, 157, 170, 171). In contrast, when ferroptotic stress directly impacts immune effector cells, including macrophages, neutrophils, or natural killer cells, their viability and functional state may be altered in ways that either relieve or reinforce local immunosuppression (137, 143, 148, 149).

Accumulating evidence indicates that these opposing effects are shaped by the cellular identity of the affected population and the local metabolic environment. Immunosuppressive myeloid cells, such as M2-polarized macrophages, often display higher sensitivity to ferroptosis than their pro-inflammatory counterparts, suggesting a potential opportunity to selectively remodel the immune microenvironment (137, 140, 141, 172). However, ferroptosis has also been reported to promote the recruitment or functional polarization of myeloid cells toward immunosuppressive states in specific tumor contexts, underscoring that its immunological impact is not uniformly beneficial (143, 144).

Taken together, ferroptosis should be regarded as a context-dependent regulator of antitumor immunity within the tumor microenvironment rather than a uniformly immunostimulatory or immunosuppressive process. Its biological and therapeutic effects are determined by the cell types involved, the intensity and duration of ferroptotic stress, and the surrounding metabolic and inflammatory conditions (148, 149). In thyroid cancer, where long-term disease control is often required, these considerations highlight the importance of precisely modulating ferroptosis to suppress tumor growth while preserving protective immune function.

As a unique intracellular selenium protein, GPX4 uses glutathione as a reducing agent to prevent lipid peroxide from accumulating in cells, which is crucial to maintaining cell redox balance and preventing oxidative damage (173). Unlike other GPX family members, GPX4 is relatively low in dependence on glutathione and can directly remove lipid peroxides on the membrane, which makes it the core regulatory factor of ferroptosis (174). In thyroid cancer tissue, the expression of GPX4 is significantly increased, and its overexpression is closely related to the enhancement of tumour invasion and the poor prognosis of patients (175). Clinical research confirms that the overall survival rate of patients with high expression of GPX4 is significantly lower than that of patients with baseline expression level (176).

Mechanistically, high GPX4 expression safeguards thyroid carcinoma from both ferroptosis and metastatic dissemination. Conversely, reducing or ablating GPX4 triggers the accumulation of ROS, depletes glutathione (GSH) reserves, and precipitates ferroptotic cell demise, thereby diminishing cellular viability (175). Whether it is genetic intervention or the use of pharmacological inhibitors such as RSL3 and diaryl ether compounds, it can effectively target GPX4, thus inducing ferroptosis, impairing mitochondrial function, inhibiting mTOR pathways, enhancing DNA damage and reducing the migration of thyroid cancer cells (176). The sensitivity of thyroid cancer cells to GPX4 inhibition is heterogeneous, which mainly depends on their genetic background - especially the presence or absence of BRAF, RAS, PIK3CA gene mutations and TERT promoter mutations. This susceptibility of mutation dependence has been confirmed in experiments in traditional single-layer culture and three-dimensional sphere systems (177). In short, these evidences emphasise that GPX4 is a core determinant of ferroptosis control and the development of thyroid cancer, and its inhibition is positioned as a potential treatment development direction in the clinical environment.

As a component of the Xc− reverse transporter, the transmembrane protein SLC7A11 is responsible for the exchange of cysteine and glutamic acid, and is a key inhibitor of ferroptosis (178, 179). Its expression is significantly increased in thyroid cancer, promotes tumour proliferation by inhibiting ferroptosis, and is associated with poor prognosis (180). Diverse molecular mediators influence ferroptosis in thyroid malignancies through the specific modulation of SLC7A11. For example, cyclic RNAs (circRNAs) such as circ-0067934 and circ-0008274 can enhance the expression of SLC7A11, thus promoting malignant progression; meanwhile, ETS variant transcription factor 4 (ETV4) has been proven to upregulate SLC7A11 in vitro and in vivo, which in turn promotes the proliferation and migration of cancer cells (181, 182). On the contrary, tumour inhibitory genes and epigenetic regulatory factors enhance the sensitivity of cells to ferroptosis by inhibiting the activity of SLC7A11. Tumour inhibitor p53 effectively limits the progression of thyroid cancer by inhibiting the transcription of SLC7A11 and promoting the occurrence of ferroptosis (183).

Likewise, BAP1 limits glutathione levels by suppressing SLC7A11 expression. This inhibition facilitates lipid peroxide buildup, thereby precipitating ferroptosis (184). In the epigenetic mechanism, m6A RNA methylation has become a key determinant of thyroid tumour biology. In thyroid papillary carcinoma, the reduced activity of demethylase ALKBH5 and FTO will change cell homeostasis, which is conducive to malignant growth. Experimental reactivation of these enzymes can offset this effect by reprogramming the expression of SLC7A11 and restoring the susceptibility to ferroptosis of cells, thus inhibiting tumour expansion (180). In terms of mechanism, ALKBH5 promotes ferroptosis through the ALKBH5–TIAM1–Nrf2/HO-1 axis, while FTO reduces the expression of SLC7A11 through the regulation of m6A dependence, thus increasing the levels of ROS and Fe²+ (20). In general, these research findings highlight that SLC7A11, as the core hub of ferroptosis regulation and an attractive therapeutic target for thyroid cancer, is subject to multiple regulation of transcriptional, post-transcriptional and epigenetic mechanisms, providing multiple ways for clinical intervention.

The ferritin autophagy process is a key link in cells to release iron, increase unstable iron pools and therefore be more sensitive to ferroptosis, in which NCOA4 acts as a cargo receptor (185). Therapeutic activation of this pathway has been demonstrated to suppress thyroid tumor progression. For instance, vitamin C induces ferroptosis by triggering ferritinophagy. In a parallel mechanism, SIRT6 potentiates NCOA4-dependent ferritinophagy, a process that expands the intracellular labile iron pool and renders cells vulnerable to ferroptosis (186, 187). In vivo research evidence further supports this mechanism. The use of ferroptosis-inducing agent sulfopyridine can effectively inhibit tumour growth driven by SIRT6 upregulation (187). These findings show that the SIRT6–NCOA4–ferrin autophagy axis is the core regulatory node of ferroptosis, and also provides a promising pathway for the treatment intervention of thyroid cancer.

In oncology, non-coding RNAs act as fundamental regulators of gene expression and cellular signaling networks. The most extensively researched subclasses include microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs) (188, 189). These molecules involve almost all aspects of tumour biology, including pathogenesis, diagnostic evaluation, prognostic judgement and treatment response (190, 191). The capacity of non-coding RNAs to modulate ferroptosis is increasingly acknowledged, establishing these molecules as central mediators and promising therapeutic targets in thyroid cancer progression. Among these, Long noncoding RNAs (lncRNAs) are transcripts defined as being longer than 200 nucleotides. Through interaction with DNA, RNA or proteins, they play a regulatory role at multiple levels such as chromatin dynamics, transcription program, RNA processing, stability, translation and competitive endogenous RNA (ceRNA) networks (192, 193).

In thyroid papillary carcinoma, the high expression of long-chain non-coding RNA CERS6-AS1 and LASP1 was associated with adverse clinical outcomes, while miR-497-5p showed the opposite expression pattern (194). In terms of mechanism, inhibiting CERS6-AS1 will reduce the vitality of cancer cells and enhance their susceptibility to ferroptosis by regulating the miR-497-5p/LASP1 axis (194). In contrast, circular RNAs (circRNAs) are characterized by a covalently closed, single-stranded architecture that confers exceptional stability and tissue-specific distribution. These cyclic RNA molecules can play a variety of functions such as transcription regulatory factors, miRNA sponges or protein scaffolds in the TME (195, 196). CircKIF4A has been recorded in a variety of malignant tumours, including thyroid cancer. It regulates GPX4 activity by adsorbing miR-1231, thus inhibiting the occurrence of ferroptosis and promoting tumour proliferation and migration (197, 198). Similarly, circ_0067934 adsorbs miR-545-3p through sponge action, thus driving the overexpression of SLC7A11, thus enhancing tumour cell proliferation and protecting cells from ferroptosis (199).

In short, these studies together reveal a precise regulatory network driven by non-coding RNA, which coordinates the process of ferroptosis in thyroid cancer, and also highlights the broad prospect of non-coding RNA as a potentially feasible target for diagnostic markers and therapeutic interventions.

In addition to intrinsic molecular regulators, increasing attention has been directed toward nanomedicine-based strategies as modulators of ferroptosis, thereby providing new perspectives on thyroid cancer progression. Accumulating evidence indicates that engineered nanoparticles can interfere with intracellular iron homeostasis, promote reactive oxygen species (ROS) generation, and intensify lipid peroxidation cascades, all of which converge on the execution of ferroptotic cell death (200). Such nano-enabled amplification of oxidative stress is particularly relevant to ferroptosis, given its dependence on iron-catalyzed lipid peroxide accumulation and redox imbalance.

Beyond directly perturbing redox and iron metabolism, nanomedicine platforms also offer unique advantages in the controlled delivery of ferroptosis-related cues. By improving pharmacokinetics, tumor accumulation, and intracellular release of ferroptosis-inducing agents, nanoformulations may enhance tumor selectivity while reducing systemic exposure and off-target toxicity (200, 201). In this context, nanocarrier-based systems have been proposed as effective tools to potentiate oxidative stress–dependent, non-apoptotic cell death pathways in cancer cells, particularly in tumors exhibiting resistance to conventional therapies.

Although direct evidence in thyroid cancer remains limited, these advances collectively suggest that nanomedicine-assisted regulation of ferroptosis represents a promising adjunct strategy to influence tumor behavior. Further investigation using thyroid cancer–specific models will be required to clarify how nano-enabled ferroptosis modulation may be integrated into disease progression studies and future therapeutic frameworks.

Despite these promising features, several major translational challenges remain. First, systemic safety is a central concern, because ferroptosis induction is inherently linked to disruption of redox buffering and iron–lipid homeostasis, which are also essential for normal tissue integrity. Reviews in the nanomedicine–ferroptosis field have emphasized that the same nano-enabled amplification of oxidative stress that is therapeutically desirable in tumors may, if not sufficiently confined, translate into dose-limiting toxicities, particularly in metabolically active organs that are prone to oxidative injury (e.g., liver and kidney) or highly sensitive to lipid peroxidation–driven damage (e.g., nervous system) (202, 203). Second, achieving truly tumor-selective ferroptosis remains nontrivial. Even when nanoformulations improve pharmacokinetics, heterogeneous vascular permeability, stromal barriers, and variable cellular uptake can lead to inconsistent intratumoral deposition. Under such conditions, incomplete tumor-specific accumulation may increase systemic exposure and enable off-target lipid peroxidation and ferroptosis-like injury in normal tissues. This concern is particularly relevant for ferroptosis-triggering nanomaterials whose physicochemical properties (e.g., surface chemistry and catalytic activity) can strongly influence ferroptotic potency and toxicity risk, highlighting the need for rigorous structure–activity and safety profiling before clinical translation (203, 204). Third, the field still lacks reliable, non-invasive biomarkers to monitor ferroptosis engagement in patients, which complicates both efficacy assessment and real-time safety surveillance during early-phase trials. Current ferroptosis readouts largely rely on tissue-based molecular or biochemical measurements, limiting their feasibility for longitudinal monitoring. Recent work has begun to explore imaging-oriented solutions, including system Xc^-–linked PET tracer approaches for tracking ferroptosis-related target engagement in vivo, as well as ferroptosis-associated MRI strategies that can sensitively detect ferroptosis-linked organ injuries (205, 206). However, these methods remain in early development and require further validation, standardization, and clinical correlation before they can serve as routine trial endpoints.

Taken together, these considerations suggest that while nanomedicine may enhance delivery and efficacy of ferroptosis-based interventions, clinical feasibility will likely depend on simultaneously improving tumor selectivity, defining a safety window that minimizes systemic and off-target ferroptotic damage, and developing practical non-invasive monitoring tools to guide dosing and patient selection.

Thyroid cancer includes distinct subtypes, such as PTC, FTC, ATC, and MTC. Current studies on ferroptosis in thyroid cancer are unevenly distributed across these subtypes, and direct comparative analyses remain limited. Most mechanistic investigations of ferroptosis-related regulators, including GPX4, SLC7A11, and iron metabolism–associated pathways, have focused on PTC, whereas data from FTC, ATC, and MTC are relatively scarce (207–209).

With respect to iron metabolism, transferrin receptor 1 (TFRC/CD71), a key mediator of cellular iron uptake, has been examined in thyroid cancer with inconsistent results (207–210). Some studies have reported increased TFRC expression in thyroid cancer tissues compared with normal thyroid tissue (208, 210), whereas others have shown TFRC downregulation or no significant difference in PTC (207, 209). These findings indicate that TFRC expression in thyroid cancer remains controversial, and whether such differences are related to tumor stage or subtype has not been systematically evaluated. At the cellular level, ATC cell lines such as 8505C have been reported to exhibit relatively low CD71 expression and increased tolerance to iron overload and ferroptosis-related stress (187). In contrast, FTC-derived cells, including FTC-133, show higher sensitivity to ferroptosis induction under certain experimental conditions (187).

Differences among subtypes have also been observed in ferroptosis-related regulatory pathways. In ATC models, overexpression of SIRT6 has been shown to upregulate NCOA4 expression, increase intracellular Fe²⁺ levels, and suppress tumor cell growth (187), indicating that ferritinophagy-associated iron release participates in ferroptosis regulation in aggressive thyroid cancer. By comparison, in PTC models, overexpression of the fat mass and obesity-associated protein (FTO) has been reported to downregulate SLC7A11, induce ferroptosis, and inhibit tumor formation (180). These findings indicate that ferroptosis-related regulators involved in different thyroid cancer subtypes are not identical.

However, current evidence remains limited and fragmented, making it difficult to draw definitive conclusions regarding subtype-specific differences in ferroptosis regulation in thyroid cancer. Further clarification of ferroptosis heterogeneity will require more refined analyses integrating tumor subtype, iron metabolism, and ferroptosis regulation.

Standard therapeutic management for thyroid cancer comprises surgical extirpation, succeeded by radioiodine ablation and suppression of TSH levels (12). For papillary thyroid cancer in the thyroid gland, especially tumours with a diameter of 1 to 4 cm, the best surgical scope has always been controversial. Current protocols from the American Thyroid Association (ATA) and the National Comprehensive Cancer Network (NCCN) advocate for risk-adapted surgical strategies. While total thyroidectomy is mandated for patients with high-risk features, thyroid lobectomy remains a viable alternative for specific low-risk cohorts (231, 232). Although early studies favoured total thyroidectomy based on the belief that it could improve the survival rate, recent evidence shows that, after taking into account risk factors, lobectomy and total thyroidectomy are comparable in terms of recurrence and survival outcomes (233, 234). Lobectomy has the advantages of thyroidectomy, including fewer complications, hypoparathyroidism and reduced risk of laryngeal nerve damage, and the possibility of maintaining normal thyroid function, although it has limitations in postoperative monitoring and recurrence detection (235, 236). Therefore, surgical decisions ought to be guided by the patient’s individual factors, including multifocal tumour, contralateral nodules, anaesthesia risk, quality of life considerations, and the expected degree of dependence on thyroid hormone treatment.

In differentiated thyroid cancer, the application of radioactive iodine has evolved into a risk stratification strategy, which is divided into residual tissue ablation, auxiliary intervention and targeted treatment of residual or recurrent diseases according to its use (231). According to the current guidelines of the ATA, radioiodine treatment is mainly recommended for patients with medium and high-risk characteristics, while those who are classified as low-risk and in good postoperative condition usually benefit little from it (231). Current literature bolsters the recommendation to employ lower radioactive iodine doses (e.g., 30 mCi) to mitigate systemic toxicity and enhance patient well-being (237, 238). Therefore, in the past two decades, the application of radioiodine in limited thyroid cancer has decreased (239). Efficacy monitoring largely depends on serum thyroglobulin (Tg) levels and imaging evaluation; however, for patients with anti-thyroglobulin antibodies in the body or patients after partial thyroidectomy, the reliability of this monitoring is reduced, which highlights the replacement biomarkers and personalised follow-up the demand for visiting strategies (12).

Thyroid hormone suppression therapy aims to mitigate the mitogenic effect of TSH on thyroid cancer cells; consequently, its clinical deployment has evolved toward increased selectivity and individualization. Early studies supported its universal use, but later evidence showed that the benefits were limited for low-risk or long-term remission patients, and its main advantage was in patients with medium and high risk or disease activity (240, 241). Concerns about side effects such as cardiovascular complications and osteoporosis, as well as emerging and even questioning data on its efficacy in invasive cancer, further narrowed its indications (242, 243). Future prospective studies may need to combine tumour genotypes to more accurately define which patients can really benefit from thyroid-stimulating hormone inhibitory therapy.

Although ferroptosis has shown potential in thyroid cancer treatment, several obstacles limit its clinical translation. First, selective delivery of ferroptosis inducers to thyroid cancer tissue and/or metastatic lesions remains a major challenge. Thyroid cancer frequently involves cervical lymph nodes and may metastasize to distant organs such as the lung and bone. Nanomedicine-based approaches have been explored to improve drug delivery efficiency and reduce off-target exposure (200); however, current evidence is largely derived from in vitro or animal studies, and their clinical feasibility in thyroid cancer has yet to be established. Second, systemic toxicity associated with ferroptosis induction cannot be overlooked (244). Ferroptosis is not restricted to malignant cells, and normal tissues such as the liver and kidney are also sensitive to iron imbalance and lipid peroxidation. Non-selective induction of ferroptosis may therefore cause off-target tissue injury, a concern that becomes more relevant in the setting of long-term treatment or combination therapy. Third, reliable and clinically applicable methods for monitoring ferroptosis in humans are still lacking. Current evaluation of ferroptosis mainly relies on the expression of key regulatory molecules in tumor tissues (such as GPX4, SLC7A11, and FSP1), as well as indicators related to lipid peroxidation and iron metabolism. While these approaches are useful for mechanistic studies and tissue-level assessment, they are largely based on tissue samples or experimental models and do not meet clinical requirements for non-invasive, repeatable, and dynamic monitoring. In addition, lipid peroxidation and iron metabolic status in humans are highly dynamic and influenced by inflammation, metabolic conditions, and concomitant treatments, which further limits their specificity and stability as clinical indicators. Consequently, translating current molecular and tissue-based assessment strategies into tools suitable for patient follow-up and treatment response evaluation remains a significant barrier to the clinical application of ferroptosis-based approaches in thyroid cancer.

Although progress has been made in surgical resection, radioiodine treatment and thyroid-stimulating hormone inhibition therapy, a considerable number of thyroid cancer patients (nearly 20%) eventually experience disease recurrence or have treatment resistance, which significantly affects the long-term survival outcome (12, 13). These limitations highlight the urgent need to go beyond conventional programs and develop innovative treatment strategies (245). Experimental evidence shows that regulating ferroptosis-related factors such as GPX4, FSP1 and GCH1 can overcome the drug resistance mechanism and inhibit tumour progression (246). Although this strategy of targeting ferroptosis has been widely studied in highly invasive malignant tumours, its application in thyroid cancer is still insufficient, so the research attention has been relatively limited. Given the persistent difficulties in managing refractory and recurrent malignancy, the ferroptosis pathway offers a viable therapeutic target for enhancing outcomes in thyroid cancer patients.

Accumulating data indicate that specific natural products and synthetic small molecules can suppress thyroid cancer progression by triggering ferroptosis. Among many natural compounds, curcumin is a bioactive polyphenol extracted from the rhizome of turmeric, which is traditionally used as a pigment and flavouring agent. It has received more and more attention because it shows broad-spectrum anti-cancer activity in malignant tumours of the mammary gland, liver, lung and stomach (247, 248). In terms of mechanism, curcumin can induce reactive oxygen (ROS) surge and accelerate lipid peroxidation. These effects are driven by simultaneously upregulating heme oxygenase-1 (HO-1) and inhibiting GPX4 activity (249, 250). In thyroid cancer, these coordinated molecular changes not only enhance oxidative stress, but also promote the interaction between HO-1 and GPX4, eventually pushing the redox balance to ferroptosis (251). By using this iron-promoting death mechanism, curcurmin can effectively hinder the survival and progression of tumour cells, which highlights its potential as a natural compound associated with the treatment of thyroid cancer.

Vitamin C is not only an important nutrient to promote iron absorption, but also can effectively regulate the process of ferroptosis in thyroid cancer (252). In addition to known effects such as immune regulation, vitamin C promotes ferritin to release free iron through autophagy, and these iron ions then produce a large amount of reactive oxygen through the Fenton reaction. The continuous accumulation of reactive oxygen continues to promote the lipid peroxidation reaction, and finally induces ferroptosis (253). Active oxygen continues to accumulate, continuously promoting lipid peroxidation reaction, and finally inducing ferroptosis (19). In undifferentiated thyroid carcinoma models, vitamin C administration markedly curtails tumor cell proliferation by instigating the ferroptosis pathway. In short, these research findings highlight that vitamin C, as a promising preparation, can use ferroptosis to therapeutically inhibit invasive thyroid cancer (19).

Neferine, a bisbenzylisoquinoline alkaloid isolated from the seed embryo of Nelumbo nucifera, possesses diverse medicinal properties, including anticancer, anti-inflammatory, antioxidant, and cardioprotective activities (254, 255). In thyroid cancer, studies have reported that neferine can induce ferroptosis through multiple mechanisms (256). Specifically, neferine increases intracellular iron levels and reactive oxygen species while downregulating key ferroptosis regulators, such as SLC7A11 and GPX4, thereby disrupting cellular redox homeostasis. In addition, neferine promotes apoptosis in thyroid cancer cell lines, including IHH-4 and CAL-62, and suppresses the Nrf2/HO-1/NQO1 antioxidant signaling pathway, further sensitizing tumor cells to ferroptosis. These effects have also been validated in animal models, in which neferine treatment effectively promotes tumor cell death and inhibits thyroid cancer progression in vivo (256). In short, these findings indicate that neferine is a promising natural compound targeting ferroptosis for thyroid cancer treatment.

In addition to natural products, synthetic derivatives also show hope; for example, the dihydroisoxazole derivative DE16 inhibits mitochondrial polarisation and down-regulates GPX4, imitates the role of classical de-iron inducers, and exerts a strong anti-tumour effect in thyroid cancer (257). It is noteworthy that even clinically useable targeted drugs, such as anlotinib, are associated with the regulation of ferroptosis.

Anlotinib is a multi-target small molecule tyrosine kinase inhibitor, which has shown strong anti-angiogenic activity in thyroid cancer. Recent studies have extended its mechanism of action beyond angiogenesis, revealing its key role in regulating ferroptosis. Anlotinib administration downregulates key ferroptosis-associated proteins, including transferrin, ferritin light chain (FTL), and GPX4. Conversely, it amplifies oxidative stress by elevating ROS and lipid peroxidation (258, 259). Markers for apoptosis, pyroptosis, and necroptosis remained unaltered, identifying ferroptosis as the primary mechanism mediating this anti-tumor response. The study of the mechanism further shows that anlotinib can induce protective autophagy, which can fight the ferroptosis. On the contrary, blocking autophagy will enhance the ferroptosis effect of anlotinib and enhance its tumour inhibition effect in in vitro and in vivo models (259).

The BRAFV600E mutation is one of the most common carcinogens in thyroid papillary carcinoma, but the treatment of tumours carrying this mutation is increasingly challenged by the development of treatment resistance (260). In order to meet this challenge, recent studies have explored strategies that transcend the inhibition of classical kinase. It is worth noting that a compound based on diaryl ether was designed to utilise the ferroptosis pathway in thyroid cancer cells (257). The molecule induces lipid peroxidation-mediated cell death, rather than traditional cell apoptosis, by inhibiting the expression of GPX4 and destroying the polarisation stability of mitochondria. Through the coordinated modulation of redox homeostasis and mitochondrial function, these compounds demonstrate potent anti-tumor efficacy. These findings position ferroptosis induction as a viable therapeutic avenue for BRAF V600E-mutant thyroid cancer.

Overall, although accumulating preclinical studies suggest that ferroptosis may represent a potential therapeutic vulnerability across multiple thyroid cancer subtypes, the current body of evidence remains largely confined to mechanistic exploration, and its overall translational maturity is clearly insufficient (see Table 1). Most available studies rely heavily on in vitro culture systems and simplified xenograft models, which are typically derived from a limited number of thyroid cancer cell lines and therefore fail to adequately capture the pronounced heterogeneity observed in patients, particularly with respect to metabolic states, tumor–stroma interactions, and immune microenvironmental features. Consequently, whether ferroptosis sensitivity observed under experimental conditions can be extrapolated to more differentiated tumors or clinically complex disease contexts remains unsupported by direct evidence.

Drug exposure levels and in vivo feasibility constitute especially prominent barriers in the translational development of ferroptosis-based strategies. Several naturally derived compounds, including curcumin (251), vitamin C (19), shikonin (261), and Tenacissoside H (262), have been reported to induce or potentiate ferroptosis-associated phenotypes in specific experimental models; however, these effects are often achieved at relatively high treatment concentrations, the systemic attainability and safety of which remain insufficiently validated in vivo. In parallel, small-molecule GPX4 inhibitors such as RSL-3, ML-162, and ML-210 are widely employed as mechanistic tools, yet their pharmacokinetic properties, tissue distribution, and potential toxicities in thyroid cancer–relevant animal models have not been systematically characterized, substantially limiting their prospects for clinical advancement (177).

Beyond single-agent approaches, several emerging intervention paradigms further underscore the gap between conceptual innovation and translational evidence. For instance, studies centered on deubiquitination regulatory axes have implicated the USP8–GPR34 signaling pathway in the establishment of ferroptosis tolerance, with the USP8 inhibitor DUB-IN-3 demonstrating anti-tumor activity in preclinical settings (263). Nevertheless, such findings remain restricted to a small number of experimental systems, and their applicability across different thyroid cancer subtypes, as well as their long-term therapeutic effects, requires further validation. Similarly, combination strategies integrating ferroptosis induction with conventional chemotherapy—such as the co-administration of isobavachalcone and doxorubicin—are currently supported primarily by preliminary preclinical observations, without systematic evaluation of dosing sequence, cumulative toxicity risk, or impact on tumor recurrence control (264).

In addition, nanomedicine-based strategies offer novel technical avenues for ferroptosis intervention but simultaneously introduce additional layers of translational complexity (265). Ultrasound-responsive nanoplatforms exemplified by FCIPL are designed to enhance intratumoral drug penetration and amplify ferroptotic effects within a theranostic framework. However, their therapeutic performance is highly dependent on external physical parameters, including ultrasound energy, tissue penetration depth, and spatial targeting precision, all of which may differ substantially between experimental models and human thyroid tumors. At present, studies exploring such platforms in thyroid cancer remain limited, and their clinical operability and reproducibility have yet to be rigorously assessed.

Taken together, although ferroptosis-centered strategies demonstrate measurable anti-tumor potential at the preclinical stage, their translation into clinical application is constrained by multiple challenges, including limited model representativeness, insufficient pharmacokinetic and safety data, and increased technical complexity. Future investigations should move beyond isolated mechanistic validation toward research paradigms that more closely reflect clinical reality, thereby clarifying the true therapeutic value of ferroptosis across distinct thyroid cancer subtypes.