The immune system plays a pivotal role in the detection and elimination of malignant cells, a process known as tumor surveillance. This complex network involves various immune cell types, including T cells, B cells, natural killer (NK) cells, dendritic cells, and macrophages (16). These components work in concert to identify and destroy cells that exhibit aberrant characteristics, such as mutations or altered antigen expression (17). Through ongoing surveillance, the immune system can often target nascent tumors before they develop into clinically detectable disease.

Central to the immune response are immune checkpoints, which are regulatory molecules that modulate T cell activity (18). These checkpoints maintain homeostasis within the immune system, preventing excessive responses that could lead to autoimmunity ( Figure 1 ) (19). Key checkpoints include PD-1, programmed cell death ligand-1 (PD-L1), and cytotoxic T lymphocyte-associated protein 4 (CTLA-4) (20). The PD-1 pathway, in particular, represents a critical checkpoint in T cell regulation. PD-1 is expressed on the surface of activated T cells, Treg cells and macrophages, while its ligands, PD-L1 and PD-L2, are often overexpressed on tumor cells and in the tumor microenvironment (TME) (21). When PD-1/PD-2 binds to PD-L1 on tumor cells, it sends an inhibitory signal that dampens T cell activation and proliferation, allowing tumors to evade immune detection (22).

The checkpoint molecule CTLA-4 competes with the co-stimulatory receptor CD28 for binding to B7–1 and B7–2 on antigen-presenting cells (APCs) (23). This negative regulation occurs early in T cell activation and serves to modulate immune responses, thereby preventing autoimmunity (24). ICIs enhance T cell activation by blocking these pathways (25). TME plays a critical role in shaping the immune response to tumors. This complex ecosystem encompasses not only tumor cells but also various immune cells, stromal components, and the extracellular matrix (26). TME can influence both tumor susceptibility to immune attack and the capacity of immune cells to mount effective responses (27).

Within the TME, several immunosuppressive mechanisms are at play, including recruitment of Tregs, secretion of immunosuppressive cytokines (e.g., IL-10, TGF-β), and the presence of myeloid-derived suppressor cells (MDSCs) (28). Furthermore, metabolic changes within the TME can further inhibit T cell function. Tumors often exhibit altered metabolic pathways that lead to the accumulation of adenosine, a metabolite known to suppress T cell activation (29, 30). Targeting these immunosuppressive elements within the TME represents a promising strategy to enhance the efficacy of checkpoint blockade therapies (31).

The immune system utilizes checkpoints to regulate responses, preventing autoimmunity while enabling the detection of malignant cells (32). ICIs disrupt these regulatory signals, unleashing a potent anti-tumor immune response (9). T cell exhaustion, characterized by a progressive loss of effector functions, often results from chronic antigen exposure, such as in tumors (33). Exhausted T cells exhibit elevated levels of inhibitory receptors, including PD-1, CTLA-4, and TIM3, along with reduced proliferation and cytokine production (34).

Emerging evidence indicates that ICIs can restore the functionality of exhausted T cells, enabling them to regain their immune activity (10, 35, 36). However, the clinical efficacy of these agents can be highly heterogeneous. While some patients experience improved survival outcomes, others may not respond or exhibit subsequent relapses following initial benefits (37–39). This variability underscores the need for a deeper understanding of the mechanisms driving both responses and resistance to ICIs.

Current research focuses on identifying biomarkers capable of predicting responses, regulating the TME, and exploring combination strategies (40–42). Combining ICIs with other modalities—such as chemotherapy, targeted therapy, or novel agents that modulate the TME (43–45)—has shown promise in preclinical and clinical studies (46–49). Integrating AI and machine learning to analyze large datasets offers a powerful approach for predicting patient responses based on genetic and epigenetic profiles, including biomarkers like TMB and MSI (50). Furthermore, novel therapies that modulate microbiomes are being investigated for their potential to enhance ICI efficacy by influencing the immune landscape (51). By addressing the multifaceted mechanisms of immune evasion and optimizing treatment regimens, it may be possible to improve outcomes for a broader range of patients (52). In conclusion, immune checkpoint therapy has revolutionized the approach to cancer treatment by leveraging the body’s immune defense mechanisms. Understanding the intricacies of the immune system, checkpoint regulation, and tumor biology are fundamental for further advancements in this field. As research continues to unravel the complexities of immune evasion and response, there is hope for optimizing ICI therapies and improving the prognosis for patients with cancer.

The advent of immunotherapy has profoundly transformed cancer treatment, especially with the development of ICIs that capitalize on the immune system’s capacity to recognize and eliminate malignant cells by targeting immune checkpoint pathways exploit by tumor to evade immune detection and destruction (53). A nuanced understanding of key immune checkpoints, particularly the PD-1/PD-L1 pathway and CTLA-4, is essential to appreciate how ICIs enhance anti-tumor immunity. In this review, we will explore the mechanisms of these critical checkpoints, their role in tumor biology, and the significant clinical impacts of ICIs in cancer therapy.

The FDA has approved numerous ICIs for lung cancer and other malignancies, demonstrating significant improvements in overall survival for select patients (90). ICIs have fundamentally transformed cancer treatment paradigms, leading to several approval milestones for various tumor types, including melanoma, lung cancer, urothelial carcinoma, and head and neck squamous cell carcinoma (91–95). These ICIs have demonstrated consistent clinical benefits, including improved overall survival, prolonged progression-free survival, and enhanced quality of life for many patients, particularly those with advanced disease who previously had limited treatment options (96).

The clinical impact of ICIs extends beyond traditional treatment modalities. In cancers like NSCLC, where outcomes were previously heavily dependent on chemotherapy, the introduction of PD-1/PD-L1 inhibitors has dramatically shifted prognosis (97). Studies have shown that patients treated with these ICIs experience significant survival benefits, often with a lower incidence of severe toxicities compared to conventional chemotherapy (98). These findings establish immune checkpoint therapy not only as a viable option for patients but also as a cornerstone of modern oncological treatment.

However, research and clinical experience indicate that the efficacy of ICIs is limited to 30-40% of patients. Some patients may experience initial benefits, but over time, the therapeutic effect may wane, leading to disease progression. This phenomenon is largely due to the emergence of drug resistance, which limits the broader application of ICIs.

Primary resistance to ICIs is characterized by a lack of therapeutic response in patients with NSCLC ( Figure 3 ) (110). This complex phenomenon is influenced by the tumor’s molecular and immunological features as well as the composition of the TME. Understanding the factors that contribute to primary resistance is essential for developing strategies to enhance the efficacy of ICIs.

The TME plays a critical role in shaping the immune response and determining the effectiveness of ICI therapies (111). Various immunosuppressive cells, such as MDSCs and Tregs, can hinder T cell activation and inhibit anti-tumor immunity (112, 113). MDSCs, a heterogeneous population of myeloid cells that expand in cancer, impede T cell responses by producing immunosuppressive factors, such as arginase, nitric oxide, and reactive oxygen species (ROS) (114). These factors can induce T cell apoptosis, promote T cell anergy, and inhibit the activity of effector T cells. Accumulation of MDSCs in the TME is often associated with poor prognosis and may contribute to primary resistance to ICIs (115).

Tregs are another critical component of the immunosuppressive TME (116). While they maintain peripheral tolerance and prevent overt immune responses. Tregs protect tumors from immune attack by suppressing effector T cell responses characterized by high expression of PD-L1 or CTLA4. Elevated levels of Tregs in the TME signify an unfavorable prognosis and can hinder T cell activation in responses to ICIs (117). The immunosuppressive TME is further shaped by the secretion of cytokines. For instance, the secretion of transforming growth factor-beta (TGF-β) and interleukin-10 (IL-10) can lead to immunosuppressive conditions, thereby reducing T cell function and promoting primary resistance to ICIs (118).

The genetic and molecular characteristics of tumors significantly impact their immunogenicity and response to immune checkpoint therapy. Tumors with low mutation burdens and specific genetic alterations typically exhibit reduced neoantigenicity, impairing immune recognition (119). TMB, defined as the total number of mutations per megabase of DNA within a tumor genome (120), is associated with immunogenicity. High TMB tumors are generally more likely to produce neoantigens that can trigger an immune response (121). Conversely, tumors with low TMB, such as those harboring specific driver mutations or exhibiting chromosomal stability, tend to produce fewer neoantigens, resulting in reduced immune recognition and often leading to poor responses to ICIs (120, 122).

Importantly, recent studies have shown that some patients with low PD-L1 expression and TMB still demonstrate significant responses to ICIs (123). This observation suggests the need for exploring alternative predictive models beyond TMB and PD-L1. The T-cell inflamed gene signature (TIS), characterized by the presence of immune genes indicative of T cell activation, has emerged as a promising alternative marker (124). Higher TIS scores are associated with better responses to therapy, underscoring the importance of a robust intertumoral T cell presence that can potentially override low PD-L1 and TMB levels (125). Tumors may also employ various strategies to evade immune detection, significantly contributing to primary resistance. For instance, downregulation of major histocompatibility complex (MHC) molecules on tumor cells severely limits the presentation of tumor antigens to T cells, impairing T cell-mediated recognition and targeting by the immune system (126). Alongside TME factors and genetic profiles, tumors may utilize alternative mechanisms, such as reduced expression of co-stimulatory signaling molecules and secretion of inhibitory cytokines, to avoid immune detection (127).

Recent research has revealed novel mechanisms of immune escape that further illuminate additional layers of complexity in primary resistance. For instance, studies have shown that tumors utilize the cGAS-STING pathway to subvert immune responses (128). Specifically, adenylate succinate lyase (ADSL) is implicated in the suppression of cGAS-STING signaling, which is crucial for activating innate immune responses (129). ADSL interferes with the synthesis of cyclic GMP-AMP (cGAMP), a critical downstream mediator of the cGAS-STING pathway (130). By downregulating this pathway, tumors can evade detection by the innate immune system, facilitating immune evasion and primary resistance. Targeting ADSL to restore STING pathway activation represents a promising therapeutic strategy that could enhance immune surveillance and improve patient responses to ICIs (129).

Additionally, the role of N-acetyltransferase 8-like (NAT8L) and its metabolite N-acetylaspartic acid (NAA) has been identified in regulating protein acetylation, influencing tumor cell metabolism, and facilitating immune evasion (131). Elevated levels of NAT8L in the TME have been linked to reduced immune activity, thus contributing to therapeutic resistance (132). By targeting NAT8L with specific inhibitors in combination with ICIs, it may be possible to restore acetylation patterns conducive to improved anti-tumor immunity and overcome resistance (133, 134).

Furthermore, tumors may also lose the expression of specific targetable antigens normally recognized by the immune system (133). This antigenic variability can result from selective pressure exerted by immune responses, allowing the emergence of subclonal populations that do not express prominent tumor antigens, resulting in a lack of recognition by T cells (135). Additionally, tumor cells can downregulate or lose MHC class I expression, impeding T cell recognition and activation (136), thereby granting a survival advantage against immune-mediated destruction (137). The upregulation of soluble PD-L1 in the TME can also inhibit CD28-mediated co-stimulation, further complicating T cell activation and contributing to primary resistance (138, 139). In summary, understanding the mechanisms behind ADSL-mediated cGAS-STING suppression and the role of NAT8L/NAA in immune escape offers new avenues for therapeutic intervention. Targeting these pathways may enhance the efficacy of ICIs and improve outcomes for patients with primary resistance to treatment.

Acquired resistance develops in patients who initially respond to ICI therapy but subsequently exhibit a decline in therapeutic efficacy. This phenomenon underscores the dynamic interplay between the immune system and tumor evolution, revealing how tumors adapt to immune pressures over time (140). One prominent mechanisms of acquired resistance involve alteration in the tumor’s antigenic profile (141, 142). Similar to primary resistance, acquired resistance can occur when tumors lose expression of antigens targeted by ICIs ( Figure 4 ) (143). Tumors that initially express high levels of antigens targeted by ICIs may evolve to downregulate or eliminate these antigens in response to immune pressure, preventing T cell re-recognition (144) ( Figure 4 ).

Analysis of tumor biopsies has shown that acquired resistance may stem from defects in interferon-gamma (IFN-γ) signaling pathways. For example, acquisition of JAK mutations represents an early step in drug resistance development, inducing resistance to the anti-proliferative effect of IFN-γ and leading to immune resistance and tumor recurrence (145). In-depth studies have unearthed additional mechanisms, such as deficiencies in antigen presentation (146), depletion of neoantigens (147), and tumor-mediated immunosuppression (148), all contributing to acquired resistance to ICIs and subsequent tumor progression.

Tumor heterogeneity facilitates the emergence of subclones that express different antigens profiles, making them less recognizable to T cells, which contribute to treatment failure. This reinforces the need for ongoing monitoring of tumor antigen profiles during treatment (149). The dynamic nature of the TME contributes to acquired resistance through alterations in immune cell composition over time (150). Changes in myeloid and lymphoid cells within the TME can lead to increased immunosuppressive activity, diminishing the effectiveness of previously effective ICI therapies. For example, while activated effector T cells may initially infiltrate the tumor (151), their numbers may dwindle over time due to the recruitment of immunosuppressive cells, such as MDSCs and Tregs, further diminishing the anti-tumor immune response (152).

Chronic stimulation of T cells in the TME can lead to T cell exhaustion, characterized by the upregulation of inhibitory receptors such as PD-1, CTLA-4, and LAG-3, as well as reduced cytokine production (153). This state of dysfunction complicates future therapeutic interventions aimed at reinvigorating T cell responses. Tumors can activate alternative signaling pathways in response to immunotherapy, affording them survival advantages despite immune pressure. For instance, the aberrant activation of the JAK/STAT pathway can enhance the expression of anti-apoptotic proteins and promote tumor cell survival (16). Changes in signaling within the TME, influenced by both tumor and immune cell interactions, can foster an environment that promotes tumor growth while inhibiting effective immune surveillance (154). Mechanisms such as activation of the NF-κB pathway can further promote survival and proliferation of tumor cells, leading to persistent and ongoing growth despite the presence of ICIs (155).

Tumors can activate alternate signaling pathways (e.g., the JAK/STAT pathway) to promote survival despite immune pressure (32). To maximize the efficacy of ICIs, it is vital to consider non-TME factors such as host genetics and comorbidities in understanding resistance mechanisms. These elements can influence treatment outcomes significantly and require further exploration. Furthermore, comprehensive strategies for combination therapies should delve into the optimal sequencing and timing of ICIs, including the comparison of neoadjuvant versus adjuvant administrations. By integrating knowledge of both primary and acquired resistance mechanisms and defining patient stratification based on genetic and tumor characteristics, future therapeutic strategies can be developed to overcome resistance barriers and improve patient outcomes in ICI therapy.

As ICIs have transformed cancer treatment, the management of irAEs has become an essential aspect of care (156). These adverse effects can range from mild to severe and may involve various organ systems, including the skin, gastrointestinal tract, liver, and endocrine glands. Effective toxicity management strategies include proactive monitoring and early intervention. For instance, corticosteroids are commonly employed to manage severe irAEs such as pneumonitis and colitis. Immune modulation techniques, including the use of ICIs in conjunction with other therapies, may enhance the overall safety profile of treatment regimens (157).

Combination therapies often pose additional risks for irAEs; therefore, prioritizing patient education about potential side effects and providing clear guidelines for reporting symptoms are critical (158). This proactive approach ensures timely interventions and minimizes the impact of irAEs on patients’ quality of life. In conclusion, a comprehensive understanding of resistance mechanisms, including the molecular pathways responsible for primary and acquired resistance and the management of irAEs, is essential for optimizing the therapeutic use of ICIs. By addressing these challenges, we can develop future therapeutic strategies aimed at overcoming resistance and improving patient outcomes in NSCLC and beyond.

AI and machine learning are increasingly being utilized to predict patient responses to ICIs. By analyzing large datasets, these technologies can identify patterns and biomarkers, such as TMB and MSI, which correlate with patient responsiveness to therapy. For example, AI algorithms can assess genomic and transcriptomic profiles of tumors to determine the likelihood of therapeutic success with ICIs, facilitating tailored treatment options for patients (161). Integrating these AI-driven insights into clinical practice may significantly enhance the selectivity and efficacy of ICI therapies (50).

Emerging evidence also highlights the role of gut microbiota in modulating the effectiveness of ICI treatment and influencing irAEs (162). Microbiome-modulating therapies are being explored as a potential avenue to enhance ICI efficacy (163). The gut microbiota produces various metabolites that can influence immune responses, with short-chain fatty acids (SCFAs), such as butyrate showing promise for their anti-inflammatory properties (164). Research indicates that butyrate-based treatments may enhance ICI outcomes while mitigating side effects. Studies have also identified specific pathways, like the PD-L2-RGMB axis, by which the gut microbiota could promote responses to PD-1 blockade, suggesting an innovative strategy to improve outcomes for patients who do not respond to conventional immunotherapy (165).

Combining ICIs with other therapeutic modalities has garnered significant attention, effectively enhancing anti-tumor immune responses.

Understanding the TME is crucial for devising strategies to improve the efficacy of ICIs. The presence of various immunosuppressive cells and factors within the TME can significantly hinder immune responses against tumors. Innovative strategies targeting these immunosuppressive elements may enhance ICI effectiveness.

The era of precision medicine emphasizes tailoring cancer treatment based on individual patient characteristics, particularly through advancements in genomics and proteomics (215). This personalized approach is critical for developing strategies for ICIs therapies. Identifying predictive biomarkers is essential for optimizing patient selection and improving treatment efficacy. Biomarker-driven approaches allow clinicians to ascertain which patients are most likely to benefit from ICIs, thus minimizing unnecessary toxicity in non-responders (216). Among the most promising predictive biomarkers are TMB and MSI. High TMB is indicative of a greater number of mutations, often leading to the production of more neoantigens, which enhances the likelihood of a robust immune response to ICIs (12). MSI, a condition characterized by the accumulation of insertion or deletion mutations in microsatellite regions of DNA, has also shown predictive value in assessing response to PD-1 blockade (217).

Advancements in genomic, transcriptomic, and proteomic technologies have facilitated comprehensive tumor profiling, enabling the identification of various additional biomarkers. Novel markers, such as immune cell infiltration patterns, particularly tumor-infiltrating lymphocytes (TILs), and specific genetic mutations (e.g., KRAS and STK11), are being actively explored for their potential to predict resistance or response to immunotherapies (218). These insights are fundamental in refining treatment approaches and improving patient outcomes.

Additionally, liquid biopsy techniques that analyze circulating tumor DNA (ctDNA) offer a non-invasive method for real-time monitoring of tumor dynamics and treatment responses (219). This approach allows for continuous evaluation of tumor characteristics, enabling timely adjustments to personalized treatment plans based on evolving biomarker profiles.

Epigenetic modifications can profoundly influence gene expression and contribute to immune evasion. Novel approaches that target these epigenetic alterations hold the potential to enhance sensitivity to ICIs.

The combination of different immunotherapeutic agents is another promising avenue for enhancing anti-tumor responses.

AI and machine learning (ML) are playing an increasingly pivotal role in cancer treatment, especially in predicting patient responses to ICIs (248). These technologies are set to revolutionize clinical decision-making by analyzing multi-omics data, including genomics, transcriptomics, epigenomics, radiomics, and clinical data, to inform predictions regarding ICI responses, acquired resistance, and the likelihood of irAEs.

By harnessing extensive datasets, AI can provide valuable insights into complex biomarkers like TMB and MSI, guiding clinicians in making informed therapeutic decisions tailored to specific characteristics of patient populations (249). Furthermore, AI-driven methodologies allow for dynamic patient monitoring through advanced techniques, such as CT imaging and radiomics, alongside digital pathology and ctDNA assessments. This holistic integration ensures timely identification of changes in tumor biology, facilitating rapid adjustments to therapy (250).

AI’s potential to optimize personalized treatment options holds vast promise, enabling clinicians to select appropriate interventions—whether single-agent therapies or bespoke combination strategies tailored for specific subsets of patients. Nonetheless, challenges such as data standardization, model interpretability, and the need for prospective validation must be addressed to ensure the reliability and applicability of AI-driven insights in clinical contexts. Large-scale projects utilizing real-world data and biobanks are set to propel the field forward by refining predictive models and validating AI applications across diverse populations, ultimately enhancing our understanding of how to tailor immunotherapy based on specific patient characteristics (251).

As our understanding of resistance mechanisms to ICIs, numerous clinical trials are currently underway to investigate combinatorial treatment approaches to overcome these challenges. These trials are critical for evaluating the efficacy and safety of new therapeutic strategies, informing best practices, and optimizing patient management (141).

One promising area of investigation includes combining ICIs with other treatment modalities, such as targeted therapies, chemotherapy, and radiotherapy. Recent clinical trials focusing on PD-1 inhibitors combined with novel agents targeting specific tumor mutations have shown early signs of efficacy (252), emphasizing the importance of personalized medicine. Such combinations leverage distinct mechanisms of action, potentially eliciting robust immune responses while counteracting resistance pathways. However, it is essential to critically assess the feasibility and toxicity of these combinations. For instance, while synergistic effects may be anticipated, the increased complexity of treatment regimens may lead to heightened toxicity profiles, necessitating careful monitoring and management of adverse effects.

Moreover, certain strategies have demonstrated negative data, highlighting the complexities involved in translating preclinical promise into therapeutic benefit. Clinical failures of cytokine therapies illustrate the challenges of achieving meaningful outcomes in a clinical setting (3). Recognizing these limitations enriches our understanding of the current treatment frameworks and informs future research directions by emphasizing the need for robust Phase III data to establish efficacy and safety.

Additionally, novel agents such as bispecific T-cell engagers (BiTEs) and CAR-T therapies are being explored to enhance antitumor activity by more effectively redirecting the immune system against cancer cells (253). However, these promising approaches also raise concerns regarding their feasibility and potential toxicities. The predominance of Phase I/II trial data for these modalities often limits our understanding of long-term outcomes and safety in diverse patient populations. Ongoing trials will be pivotal in revealing how these combinations influence various TMEs and the overall immune landscape, which may ultimately reshape treatment paradigms.

Furthermore, patient stratification is becoming increasingly central to clinical trials as researchers seek to identify specific populations that may benefit most from combination therapies. Biomarkers such as TMB, PD-L1 expression, and T-cell receptor (TCR) diversity will be critical for distinguishing responders from non-responders, thereby enhancing clinical outcomes while reducing unnecessary toxicities associated with ineffective therapies (243).

Translating innovative therapies from research to clinical practice requires robust regulatory frameworks. The rapid pace of advancements in immune checkpoint therapy and combination strategies challenge regulatory bodies tasked with ensuring patient safety while fostering an environment conducive to innovation (254). Regulatory frameworks must adapt quickly to allow timely approval of novel agents and combination therapies that demonstrate promising safety and efficacy profiles.

In evaluating therapies comprising multiple modalities, regulators will need to navigate the complexities of risk-benefit analyses, considering each treatment’s distinct safety profiles and potential toxicities. Cost-effectiveness analysis will be paramount, as the integration of novel therapies can place significant financial burdens on healthcare systems (255). Determining the economic implications of these combinations is essential for assessing their feasibility for widespread adoption within standard clinical practices, especially in publicly funded healthcare systems.

Consequently, enhancing public education regarding these innovative therapies is vital for fostering informed discussions and empowering patients. Initiatives that clearly articulate the potential benefits and risks of new treatments will help cultivate engaged patient communities and ultimately lead to better health outcomes. Navigating these regulatory and societal considerations is essential for successfully implementing immune checkpoint therapies (256).

As cancer research increasingly focuses on patient-centric approaches, involving patients in research discussions has grown critical. Patients provide invaluable insights and experiences that can guide investigators in designing studies that are not only scientifically rigorous but also resonate with real-world needs (257). Incorporating patient perspectives early in the research process can inform the development of clinical trial endpoints that matter most to patients, such as quality of life, symptom management, and long-term survivorship (258). These discussions can direct future investigations to focus on outcomes that align with patient priorities, potentially enhancing participant recruitment and retention rates.

Furthermore, involving patients in study design and execution fosters a sense of ownership and empowerment, encouraging a stronger commitment to research initiatives. Patient advocacy organizations can play instrumental roles in bridging the gap between researchers and the patient community, promoting dialogue that drives innovation and ensures research efforts address pressing challenges within the cancer landscape.

Incorporating patient-reported outcomes (PROs) into clinical trials will be essential for capturing the real-world experiences of patients undergoing immune checkpoint therapy (259). Understanding the psychological, emotional, and social implications of treatment enables a comprehensive understanding of therapeutic efficacy (260). Moreover, PROs can illuminate how combination therapies impact patient well-being, providing invaluable data to inform clinical practice (261). As cancer research continues to evolve, embracing patient involvement as a core component of study design will be paramount. This approach enhances research relevance and fosters the development of therapies that authentically meet the needs of the patients they aim to serve.