Although immunotherapy for cancer has been intensely studied, significant and durable antitumor response with limited severe treatment-related toxicity has only been observed since FDA’s approval of immune checkpoint inhibitors (ICIs) that target specific immune checkpoints (1–3). Over the past decade, ICIs have quickly become the primary systemic treatment option for advanced melanoma and non-small cell lung cancer (NSCLC) with expanding indications for solid tumors arising from other anatomical sites (2–5). ICIs that are currently in clinical use mostly target the PD-1/PD-L1(PD-(L)1) immune checkpoint at the site of peripheral tumor (6). This strategy of targeting the primary mechanism of escape from cancer immunosurveillance at the peripheral tumor sites has led to the restoration of adaptive anti-tumor immunity in an overall immunosuppressive local tumor micro-environment (TME) without causing a significant systemic response in normal tissues (1, 7). Thus, leading to long lasting anti-tumor response and a low incidence of severe toxicities in a number of patients (4, 5). Based on previous studies on the treatment of solid tumors with anti-PD-(L)1 inhibitors, the tumor immune micro-environment (TIME) of responders to ICIs has been characterized in contrast to non-responders (8–13). An overall inflamed TIME in which IFN-γ driven tumor expression of PD-L1, and tumor infiltration by functional CD8⁺ T cells have been consistently identified in responders (8, 12, 13). On the contrary, a paucity of tumor infiltration by functional CD8⁺ T cells has been clearly demonstrated in non-responders. Tumor progression after an initial response may also occur after acquiring additional mechanisms of immune evasion, such as loss-of-function mutations or other genomic alterations, which can subsequently induce an overall non-inflamed TIME (14–17). Currently, the response rate to ICI remains low in non-lymphoid solid tumors, such as NSCLC (5, 17). This makes more effective treatment strategies to overcome ICI resistance urgently needed. With increased understanding of the mechanisms of antitumor immunity and its impairment, many strategies to enhance antitumor immunity and overcome immunosuppression within the TIME, subsequently converting a non-inflamed or “cold” TIME into an inflamed one have been proposed (18–20). As one major local treatment strategy for cancer, radiotherapy (RT) has been shown to have immunostimulatory properties, which can be further exploited to serve this purpose (21–25). In this review, the impairments leading to suboptimal antitumor immunity and poor response to ICIs, RT’s immunogenic properties, and the rationales for combining RT with immunomodulatory agents to remodel the TIME in order to restore or augment anti-tumor immunity will be discussed.

Immune surveillance of cancer has long been known to exist, which allows the immune system to eradicate malignant lesions as they arise in the human body (7, 26, 27). To initiate anti-tumor immunity for tumor elimination ( Figure 1 ), a coordinated activation of both innate and adaptive immunity is required (26–29). In the process, tumor antigens and the detection of “danger signals” from the tumor by the innate immune sensors, trigger the recruitment and activation of antigen presentation cells (APCs) (28–33). Subsequently, tumor antigen cross-presentation by APCs that travel to the tumor draining lymph node (tdLN) leads to the priming and activation of T cells (33). Co-stimulation signals, which may be further tuned by co-inhibitory signals, are required for T cells to be fully activated; while sub-optimal activation only leads to T cell anergy (34, 35). Fully activated T cells are able to express a unique set of “homing” chemokine receptors that are accompanied by increased expression of related chemokines in the TME. Thus, allowing for the homing of activated T cells to the tumor and its infiltration (33). One well known example is the CXC receptor 3(CXCR3), and its ligands CXCL9, CXCL10, and CXCL11, which interactions in tumor infiltration may also involve other chemokines, such as CXCL5 (33, 36, 37). At the site of tumor, selectin ligands and other adhesion molecules are also required for the binding of blood vessels and extravasation by T cells (33, 36). After tumor infiltration, cytotoxic CD8⁺ T cells will recognize tumor cells through recognizing tumor antigens presented by the major histo-compatibility complex class I (MHC-I) molecules and induce T cell mediated cytotoxicity (33).

Immune escape by tumor cells eventually develops. T cell activation within the TIME will stimulate the expression of PD-1 and release of IFN-γ by T cells (2). IFN-γ then induces PD-L1 expression by tumor cells as well as other cells within the TIME, resulting in the inhibition of the local immune response through PD-1/PD-L1 interaction (2, 38, 39). The process of PD-1/PD-L1 mediated effector T cell exhaustion is also called adaptive immune resistance. It serves as a dominant feedback mechanism to maintain peripheral tolerance, and is a common mechanism of immune escape by various cancers (2, 6, 40). Also, it makes re-invigoration of tumor infiltrating cytotoxic T lymphocytes (CTLs) within the TIME through targeting the PD-1 axis a very effective therapeutic strategy for cancer (1–6, 8, 41). An inflamed TIME that is characterized by IFN-γ induced PD-L1 expression by mostly tumor cells, and the presence of functional tumor infiltrating lymphocytes (TILs) is often associated with such re-invigoration, which is only observed in a limited number of patients with cancers such as NSCLC (2, 8, 41). This implies the presence of other impairment(s) within the steps of anti-tumor immunity generation.

Various mechanisms of immune escape by cancer leading to a paucity of T cell infiltration or a lack of activated T cells within the TME have been characterized in recent years (17, 33, 42–44). Major impairments in the generation of anti-tumor immunity, include the paucity of tumor neoantigens, impaired antigen presentation, lack of T cell priming and activation, poor tumor infiltration by activated T cells, poor tumor cell recognition/impaired IFN-γ signaling, and alterations in the composition and properties of immune cells within the peripheral TIME ( Figure 2 ).

Mutations arise and accumulate during carcinogenesis and cancer progression, leading to the generation of tumor neoantigens (45, 46). Such neoantigens induce T cell specific reactivity, especially in patients who respond to immune checkpoint blockade (47–49). A high tumor mutation burden (TMB) consistently correlated with a higher incidence of durable clinical response (DCR) and improved survival after treatment with anti-CTLA-4 or anti-PD-L1 ICIs for cancers which are known to harbor a higher level of somatic mutations (50–54). This observation is due to the higher likelihood of a high neoantigen load in the presence of a high TMB (55). As shown in a cohort of melanoma patients, DCR to ICIs is not guaranteed by the presence of a high TMB (50). Instead, specific mutation associated neo-epitopes are required to elicit a cytotoxic anti-tumor T cell response in long-term responders to ICIs, and their occurrence is low even in the presence of a high TMB. This is due to immunoediting or the loss of tumor clones which harbored these mutant neoantigens over time (50, 52, 55). Subsequently, immune ignorance develops with poor TIL infiltration of the tumor. As shown in NSCLC, the repression of clonal neoantigen expression is ongoing in an inflamed TIME through various mechanisms, such as DNA copy number loss, suppression of gene transcription, epigenetic, as well as post-translational modification (56). Loss of immunogenic tumor mutation associated neoantigens has also been identified as one mechanism of acquired resistance to ICIs in NSCLC patients who underwent treatment with anti-PD-1 or combined anti-PD-1 and anti-CTLA-4 ICIs (57). Additional strategies to induce or maintain an adequate immunogenic mutant neoantigen repertoire in cancer patients are needed. However, high TMB and increased PD-L1 expression together represent key characteristics of an immunogenic tumor that has developed PD-1 axis mediated adaptive immune resistance, which may be used to select patients for treatment with anti-PD-1 or anti-CTLA-4 ICIs (50, 52).

As evidenced by a lack of DCR to ICIs in the presence of very high TMB, additional mechanisms of immune escape leading to a paucity of TILs within the TIME exist (50, 52). The antigen cross-presentation process, a key step in the generation of adaptive anti-tumor immunity, can be impaired in multiple ways, leading to a non-inflamed TIME. Among them, the repression of type I conventional dendritic cell (cDC₁) trafficking into the TME due to tumor-intrinsic activation of the WNT/β-catenin signaling pathway has been well studied (58–61). A paucity of CD103⁺ DCs was observed when WNT/β-catenin was expressed in melanoma, which was due to the higher expression of transcriptional repressor AFT3. As a result, AFT3 inhibited the expression of CCL4, which is the key chemokine for guiding cDC₁ migration into the tumor (58). Activated b-catenin signaling due to mutations or somatic copy number alteration is frequently observed in non-inflamed tumors (59). As shown in NSCLC, b-catenin expression is associated with shorter overall survival (OS) and significantly less tumor infiltrating CD8⁺ T cells, especially in the presence of a high TMB (60, 61). The trafficking of DCs into the tdLN can also be suppressed by tumor intrinsic alterations, such as increased expression of liver X receptor-α (LXR-α) ligands that would bind to LXR-α on DCs to inhibit CCR7 expression, which is required for DC migration to the tdLN (62, 63).

Other than the disruption of DC trafficking, activation and maturation of DCs can also be impaired. The maturation of DCs is induced through the detection of danger associated molecular patterns (DAMPs), which are mostly composed of intracellular proteins, and pieces of cytosolic tumor DNA from dying tumor cells by innate immune sensors (29, 64, 65). This leads to increased IFN-β signaling, which induces downstream stimulation of antigen presentation, DC maturation and trafficking to the tdLN. The innate immunity induced by DAMP or cytosolic DNA may be attenuated in tumors with increased cyclooxygenase-driven prostaglandin E2 (PGE₂) production, leading to decreased DC recruitment and maturation (66, 67). On the other hand, decreased innate immunity may also result from the neutralization or inhibition of DAMP signaling, which inhibits subsequent DC activation (68, 69).

DC recruitment into the tumor, and differentiation are also mediated through NK cells, which produce cDC₁ chemo-attractants CCL5, CXCL1, and survival stimulating cytokine FLT3LG. However, NK cells’ viability and chemokine production may be suppressed by tumor PGE₂ overexpression, which demonstrates how the tumor may exert its influence on the TIME (70, 71). Other intrinsic tumor factors may also suppress DC maturation, such as vascular endothelial growth factor (VEGF); or induce a tolerogenic DC phenotype (DCreg), such as interleukin (IL)-6, transforming growth factor (TGF)β (72). At last, cross presentation by DCs can also be impaired due to the accumulation of oxidized lipids as a result of oxidatively truncated lipids’ binding the heat shock protein (hsp) 70, which prevents the translocation of peptide-MHC complex to the cell surface (73, 74). Activation of CD8⁺ T cells also requires additional stimulatory signals from the APCs, including co-stimulatory signals and APC-released cytokines (64, 75). Peripheral tolerance can develop with suboptimal co-stimulation, which may be alleviated by agonist antibodies to the co-stimulation checkpoints, activating cytokines, or inhibitors to inhibitory checkpoints, such as CTLA-4 (33).

Chemo-taxis of effector T cells to the tumor depends on the expression of CXCR3 on T cells and the presence of its ligands, CXCL9, CXCL10, and CXCL11 within the TIME (76). Within the tumor, Batf3-dependent cDC₁s that express CXCL9 and CXCL10 are required for effector T cell chemo-taxis (77). A lack of tumor infiltration by T cells is observed in their absence. On the contrary, their presence is associated with a state of T cell exhaustion, and response to ICIs (78). More recently, macrophage-derived CXCL9 and CXCL10 were also found to be required for CD8⁺ T cell infiltration into the TIME, and any response to ICIs (79). Chemokines can be modified by post-translational modifications, which have been associated with impaired T cell chemo-taxis to the tumor (33, 80).

In addition, T cell extravasation and tumor infiltration can be suppressed by endothelial cells and stromal cells. CTLs must adhere to the endothelial lining for extravasation, and this may be inhibited by the vascular endothelial growth factor (VEGF) mediated down-regulation of adhesion molecules, or the induction of CTL apoptosis through Fas ligand (FasL) overexpression that are induced by VEGF, IL-10, or PGE₂ within the TME (81, 82).

The tumor stroma has largely been known to be immunosuppressive, which can form dense areas of fibrosis in advanced cancer, leading to a state of poor T cell infiltration into the tumor (83). Increased collagen density has been associated with decreased CD8⁺ T cell infiltration into the tumor (84). Similarly, cancer associated fibroblasts (CAFs) have been shown to induce an immunologically excluded phenotype mediated by increased TGFβ signaling within fibroblasts (85). Other mechanisms of CD8⁺ T cell inhibition by CAFs include CXCL12 mediated CTL exclusion; PD-L1/L2, FasL upregulation; generation of adenosine; IL-6 mediated DC conversion to DCreg; and blunting of TCR signal transduction (86). In addition, CAFs can secret immunosuppressive factors and stimulate the recruitment of suppressive immune cells into the TIME (87). Tumor stroma along with the ill-formed tumor vasculature represent areas of great therapeutic potential for the generation of a more immunogenic TIME (83, 87, 88).

Reduced tumor neoantigen recognition may result from decreased MHC-I expression caused by the presence of genomic alterations of MHC-I or any component of the antigen processing machinery (APM), such as β2-microglobulin (B2M) (72, 89). Alternatively, it may result from the reduced MHC-I up-regulation due to impaired IFN-γ signaling caused by JAK1/2 mutations (14). Complete loss of the B2M gene causes complete loss of MHC-I expression (14, 15). Genomic alterations of B2M and JAK1/2 have been associated with acquired resistance to ICIs along with MHC-I loss of homozygosity (LOH) (14, 15, 90, 91). Other mechanisms leading to decreased MHC-I expression also exist, which include genomic alterations of the transactivator of MHC-I related genes, NLCR5 (NOD-like receptor family, caspase recruitment domain containing 5); the expression of embryonic transcription factor, DUX4; epigenetic silencing; and post-translational re-direction to autophagy (72, 92, 93).

At last, alterations of the immune cell composition within the TIME can lead to suppression of anti-tumor immunity. One such alteration is the increased presence of terminally exhausted CD8⁺ T cells. Initially observed after chronic antigen exposure, T cell exhaustion is characterized by increased PD-1 expression its co-expression with additional inhibitory checkpoint receptors, and hierarchical loss of cytolytic function in CD8⁺ T cells (94–97). TILs appear to be at different stages of exhaustion with effector functions preserved in progenitor exhausted CD8⁺ T cells, but lost in terminally exhausted CD8⁺ T cells (95–97). Progenitor exhausted T cells appear to have greater chromatin accessibility; higher expression of stimulatory cytokines, co-stimulatory checkpoints, and survival/memory molecules; while terminally exhausted T cells had more accessibility to and expression of co-inhibitory receptors, effector molecules, and transcription factors associated with exhaustion (98). As shown in NSCLC, increased degree of CD8⁺ T cell exhaustion is correlated with increased PD-1 expression and the co-expression of additional co-inhibitory receptors, such as TIM-3, TIGIT, and CTLA-4 at later stages of cancer progression (99, 100). PD-1 appears to be the primary mediator of CD8⁺ T cell exhaustion. However, anti-tumor response to anti-PD-1 therapy alone is poor in tumors infiltrated by terminally exhausted TIL’s, which mostly reside in the TIME (97–100). On the contrary, combined blockade to PD-1 and other co-inhibitory checkpoints was shown to restore effector function in terminally exhausted TILs, and generate significant anti-tumor activity when PD-1 and additional co-inhibitory checkpoints are co-expressed (97, 100).

The immune cell composition may also be sculpted to have an increased presence of suppressive immune cells within the TIME. The inhibitory roles of regulatory T cells (Tregs), and myeloid derived suppressor cells (MDSCs) have been well characterized, while tumor associated macrophages (TAMs) and tumor associated neutrophils (TANs) can develop pro-tumor or anti-tumor properties (101–106). Tregs, Forkhead box P3 (FOXP3) expressing CD4⁺ T cells, are recruited into the TIME through up-regulation of CCR4, CCR5, CCR8, and CCR10 mediated chemotaxis in inflamed tumors by both tumor and immune cells; and CCR4 mediated chemotaxis due to increased CCL22 expression in EGFR mutant non-inflamed tumors, exert their inhibitory effects on CTLs through IL-2 depletion, binding and capturing of co-stimulatory signals on dendritic cells, and the production of immunosuppressive factors, such as TGFβ, IL-10, FGL₂, VEGF, and granzymes (101, 102). MDSCs are derived from polymorphonuclear (PMN) or monocytic myeloid cells in the presence of cancer induced inflammation (103, 104). They secrete a series of immunosuppressive mediators, such as IL-10, PGE₂, TGFβ, free radicals; inducing M₂ macrophages and Tregs, while suppressing CTL adhesion, TCR expression, activation, and survival (105, 106). TAM and TAN may polarize into immunostimulatory/anti-tumor, or immunosuppressive/pro-tumor types by stimulatory signals, such as IFN-γ, or suppressive signals, such as, TGFβ (105–107). While type 1(M₁) macrophages have potent antigen presentation and phagocytotic properties, type 2 (M₂) macrophages are immunosuppressive for which they express inhibitory immune checkpoints, such as PD-L1; and secrete a series of suppressive factors, such as TGFβ, IL-10, IDO, and arginase 1 (Arg1), which stimulate Treg function, inhibit DC maturation, and suppress CTL function (105, 106). Similar to TAMs, TANs can secrete a series of immunosuppressive mediators and express a series of inhibitory immune checkpoints, once induced by suppressors, such as TGFβ (107).

Various mechanisms of immune escape by cancer may lead to poor response to ICIs, which shed light on the development of additional therapeutic strategies to induce an inflamed TIME. As a major local treatment modality for cancer, RT has been shown to have immunogenic properties (21). Ablative doses of radiation can cause immunogenic cell death (ICD) through the stimulation of antigen presentation, leading to increased maturation and recruitment of effector CD8⁺ T cells ( Figure 3 ) (108–112). Its immunostimulatory properties may lead to the restoration of an overall inflamed TIME and enhanced antitumor immunity. Furthermore, such immunogenic antitumor effects may be significantly augmented when RT is combined with immunomodulatory agents, such as ICIs (21, 22).

Immunostimulatory properties of radiation allows it to act synergistically with immunomodulatory agents to induce a CD8⁺ T cell mediated immunogenic antitumor response that is stronger than from either alone ( Figure 3 ). Strong local and systemic antitumor response (response outside of the irradiated field, which is also known as the “abscopal effect”) were elicited when RT is combined with immunomodulatory agents stimulating the proliferation and/or activation of DCs, Th₁ CD4⁺ T cells, and CD8⁺ T cells (189–192). This was due to IFN-I mediated DC infiltration, CD8⁺ cross priming, and subsequent effector CD8⁺ T cell infiltration into the TIME (191, 192). Such antitumor response was consistently observed when RT was combined with an anti-CTLA-4 ICI in vivo (193–196). The intensity of an abscopal response generated from such combinations are dose fractionation and treatment sequence dependent. As shown by Dewan et al, an abscopal effect was most prominent after combining an anti-CTLA-4 antibody with 8 Gy x 3 fractions vs. 6 Gy x 5 fractions, or 20 Gy x 1 fraction with SF irradiation failing to induce any such effect (194). The three-fraction regimen was further shown to have stronger distant effect than the five-fraction regimen. When compared with concurrent delivery of RT and an anti-CTLA-4 ICI, ICI delivery after RT was shown to have reduced distant therapeutic effect. Increased antitumor immunity with the combined treatment is associated with increased CTLs, CD8/CD4 ratio, reduced Tregs within the TIME, and increased CD8⁺ T cell clonality (195, 196).

Unlike anti-CTLA-4 ICIs, which suppress the inhibition of CD8⁺ T cell activation within the more proximal site of tdLNs, anti-PD-(L)1 ICIs are more recognized for their ability to overcome PD-(L)1 mediated inhibition of CTL function within the peripheral TIME (6). PD-L1 expression is upregulated in DCs and tumor cells after irradiation, while acute decrease in PD-1 expression in CD8⁺ T cells within the TIME is observed after RT with an ablative dose (197, 198). Significant improvement in local and abscopal antitumor responses was observed after combined treatment with RT and an anti-PD-L1 ICI (197–199). Increases in response were more prominent with hypofractionated schedule delivering ablative doses, which results in increased CD8⁺ T cell infiltration & function; and reduced intratumoral MDSCs resulting from decreased trafficking and CD8⁺ T cell mediated direct killing requiring TNFα (197, 199). After ablative doses, latent increase in PD-1 expression by tumor-infiltrating T cells and elevated PD-L1 expression by tumor cells at both primary and secondary tumor sites were observed (200, 201). This was accompanied by decreased intratumoral Tregs.

Anti-PD-1 ICIs combined with RT induced significant increases in tumor-antigen specific and memory CD8⁺ T cells, as well as CD8⁺/Treg ratio within the peripheral TIME, subsequently leading to significantly amplified local and abscopal antitumor responses (201–204). The increased CD8⁺/Treg ratio mainly resulted from a reduction of Tregs within the peripheral TIME after the combined treatment, which is accompanied by increased CD8⁺ T cell clonality at both primary and secondary tumor sites (202, 203). Contrary to the reduction of intratumoral MDSCs observed after ablative doses, increased intratumoral accumulation of MDSCs has been observed after combined treatment with conventionally fractionated radiation and an anti-PD-1 antibody (203). Overall, pre-clinical data supports the use of hypofractionated RT schedules in combined RT and anti-PD-(L)1 treatment strategies.

The benefit of RT and anti-PD-(L)1 combinations also depends on features of the tumor residing tissue and the tumor itself. For example, LD irradiation with 4 Gy x 3 fractions may induce significant CD8⁺ T cell tumor infiltration and increased CD8⁺ T cell function when combined with an anti-PD-1 antibody (205). This combination led to significant antitumor response in a KRAS^G12D mutant orthotopic mouse model of lung adenocarcinoma, which depended on the presence of lung tissue residing club cells that express synaptosome-associated protein 23 (Scgb1a1). This is due to club cell secretory proteins’ immunostimulatory roles, such as suppression of myeloid cells (205). Alternatively, STK11/LKB1 mutations, which frequently co-occurs with KRAS-mutations in lung adenocarcinomas, were associated with poor synergy between RT and anti-PD-1 in vivo (200). This may result from increased CTL exhaustion in the presence of increased CD8⁺ T cell infiltration and CD8⁺/Treg ratio after the combined treatment. These examples demonstrate the influences of various tumor-related features on the TIME, which may dictate tumor response to combined RT and ICI treatment. Additional therapeutic strategies are needed to further augment the efficacy of RT and ICI combinations in these situations.

The PD-1/PD-L1 axis is a main mediator of CD8+ T cell exhaustion with in the TIME of resistant melanomas after combined treatment with RT and an anti-CTLA-4 ICI (206). In resistant tumors, increased tumor expression of PD-L1 and the proportion of PD-1⁺ EOMES⁺ CD8⁺ T cells that do not express markers for activation, Ki67 and GzmB were found. Activated CD8⁺ T cells and CD8⁺/Treg ratio markedly increased within the TIME after an anti-PD-L1 ICI was added to the RT and anti-CTLA-4 ICI combination. Dual CTLA-4 and PD-(L)1 blockade combined with radiation led to a CR rate of 80% and significant prolongation of survival in vivo. Therefore, significant antitumor immunity may be generated by combining RT and the targeting of multiple stimulatory and/or inhibitory checkpoints. This strategy may be further explored in the context of combining RT with dual immune checkpoint blockade (ICB) (207–212).

Further improvement of antitumor response was observed when an OX40 agonist, an anti-TIM-3, or an anti-TIGIT antibody was added to the RT and anti-PD-(L)1 antibody combination (210–212). Significant increase in CD8⁺ T cell and CD103⁺ DC infiltration, as well as decreased CD8+ T cell exhaustion were observed with the addition of an OX40 agonist (210). This led to further improvement in local and abscopal responses, leading to an survival advantage over radiation combined with an anti-PD-1 antibody alone. Similar survival advantage was observed with the addition of an anti-TIM-3, or an anti-TIGIT antibody (211, 212). The added benefit from an anti-TIGIT antibody was dependent on radiation dose fractionation. TIGIT expression by CD8⁺ T cells was increased by 8 Gy x 3 fractions, but decreased by 2 Gy x 18 fractions (212). Other strategies to further enhance immunostimulatory effects of RT combined with ICIs may include addition of other immunostimulatory agents, such as stimulatory cytokines (e.g., IL-2, IL-12) and activators of innate immunity sensor (213–218). Alternatively, additional therapeutic advantage may be gained by adding agents reducing the level of immunosuppression within the TIME to RT and ICI combinations, such as antibodies against TGFβ, VEGF, or PI3K; M₂ TAM or MDSC reducing agents; inhibitors of suppressive metabolite production, such as CD73 antibodies; and inhibitors to other checkpoints, such as CD47 antibodies (199, 209, 219–224).

The TIME is sculpted by the tumor’s intrinsic features and those of the various immune cells, stromal cells, endothelial cells within the TME. Various mechanisms of failure to generate anti-tumor immunity exist, resulting in a noninflamed or “cold” TIME which is associated with primary or acquired resistance to ICIs and other immunomodulatory agents. Radiotherapy’s immunostimulatory effects are offset by its immunosuppressive effects. Its immunostimulatory effects may be augmented when combined with ICIs and other immunomodulators in a manner that is both dose/fractionation and sequence dependent. This strategy may stimulate the generation of antitumor immunity, and/or reduce immunosuppression caused by suppressive mediators and cells within the TIME. Therefore, combining RT with immunotherapy may help overcome resistance to ICIs and other forms immunotherapy, and enhance established treatments for cancer with promising early clinical evidence emerging. In addition, strategies to enhance the efficacy of such combinations warrant further exploration.

AC and NN contributed to conception and design of the study. AC wrote the first draft of the manuscript. All authors contributed to the article and approved the submitted version.