Immunotherapy resistance in colorectal cancer: therapeutic strategies and biomarker-guided approaches Authors: Mahdi Jafarlou, Vahid Jafarlou, Samaneh Nouri, Nazila Alizadeh, Zahra Azimzadeh Tabrizi, Mahsa Iravani, Farid Ghorbaninezhad, Deniz Shakiba, Behzad Baradaran Published: 2026 ABSTRACT AbstractOvercoming immunotherapy resistance remains a major challenge in colorectal cancer (CRC), particularly because most tumors are microsatellite stable/proficient mismatch repair (MSS/pMMR) and respond poorly to immune checkpoint inhibitors (ICIs). In contrast, deficient mismatch repair/high microsatellite instability (dMMR/MSI-H) tumors show greater sensitivity to PD-1/PD-L1 and CTLA-4 blockade due to high tumor mutational burden, neoantigen generation, and immune-inflamed tumor microenvironments. However, both primary and acquired resistance continue to limit durable clinical benefit. Resistance arises from tumor-intrinsic mechanisms, including impaired antigen presentation, MHC-I downregulation, oncogenic signaling, and chromatin remodeling defects, as well as tumor-extrinsic mechanisms such as immune suppression, stromal exclusion, inhibitory cytokines, metabolic stress, and gut microbiome dysbiosis. This review distinguishes itself by linking specific resistance barriers to therapeutic strategies rather than discussing immunotherapy approaches in isolation. We summarize emerging strategies, including rational ICI combinations, targeted therapies, adoptive cell therapies, CAR-T cells, BiTEs, cancer vaccines, oncolytic viruses, microbiome-modulating approaches, and cytokine/adjuvant-based therapies. We also highlight clinically relevant biomarkers, including MSI/MMR status, tumor mutational burden, immune contexture, Immunoscore, circulating tumor DNA, microbiome profiles, and spatial or AI-assisted multi-marker models. This review summarizes emerging strategies to enhance therapeutic efficacy in CRC and maps each modality to the tumor-intrinsic or tumor-extrinsic resistance barriers it is most likely to overcome.Keywords: Colorectal cancer, Immunotherapy resistance, Immune checkpoint inhibitors, MSI-H, MSS, Tumor microenvironment Introduction Global burden and clinical challenges of colorectal cancer (CRC) CRC remains one of the most common malignancies worldwide and accounts for approximately 10% of all cancer cases. It is also the second leading cause of cancer-related death globally, highlighting its persistent clinical burden despite advances in screening and treatment [1]. In the United States, CRC remains a major cause of cancer mortality, with 158,850 new cases and 55,230 deaths expected in 2026 [2]. Although screening has reduced incidence and mortality in older adults, the increasing burden of early-onset CRC underscores the need for more effective prevention, detection, and treatment strategies [3]. Immunotherapy in CRC: achievements and current limitations CRC is typically managed through conventional therapies such as surgical procedures, radiation therapy, and chemotherapy. Standard treatment for metastatic CRC (mCRC) still relies largely on chemotherapy-based regimens, which are limited by therapeutic resistance, dose-limiting toxicity, and a treatment plateau [4, 5]. Immunotherapy has transformed CRC treatment by restoring or enhancing antitumor immune responses through immune checkpoint inhibition, adoptive cell transfer, cancer vaccines, and cytokine-based approaches. Over the past decade, immune checkpoint inhibitors (ICIs) targeting programmed cell death protein 1 (PD-1), its ligand PD-L1, and cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) have demonstrated durable clinical. This success in CRC is largely based on the identification of dMMR/MSI-H tumors, which are particularly sensitive to ICIs such as pembrolizumab and have supported FDA approval in this setting [6]. This sensitivity is driven by the hypermutated nature of dMMR/MSI-H tumors, which leads to a high tumor mutational burden (TMB) and the production of many neoantigens [7]. This creates a highly immunogenic tumor microenvironment (TME) enriched with tumor-infiltrating lymphocytes (TILs), which can be reactivated by blocking inhibitory checkpoints such as PD-1, PD-L1, and CTLA-4 [7]. However, this benefit is largely restricted to the minority of patients with dMMR/MSI-H tumors. In contrast, most CRCs are pMMR/MSS or MSI-L and are generally characterized by lower neoantigen burden, reduced T-cell infiltration, weaker inflammatory signaling, and an immunologically cold or excluded TME. As a result, these tumors show limited responsiveness to current ICI monotherapy, making primary resistance in MSS/pMMR CRC one of the central challenges in the field [8]. Furthermore, even within the responsive dMMR/MSI-H population, acquired resistance can emerge while on continuous treatment after an initial period of control [9]. Mechanisms of resistance to immunotherapy in CRC The efficacy of immunotherapy is often constrained by immunosuppressive mechanisms within the tumor microenvironment, including regulatory immune cells, inhibitory cytokines, metabolic limitations, and immune checkpoint signaling. Collectively, these factors promote immune evasion and represent a major barrier to effective antitumor responses, particularly in solid tumors such as CRC [10]. Mechanisms of immunotherapy resistance in CRC are multifactorial and deeply implicated with the biology of the tumor and TME. Resistance can be mainly categorized into tumor-intrinsic and tumor-extrinsic factors [11]. Tumor-intrinsic resistance refers to cancer cell-autonomous mechanisms that impair immune recognition or reduce susceptibility to immune-mediated killing. These include defective antigen presentation caused by downregulation or loss of major histocompatibility complex class I (MHC-I), mutations affecting interferon-γ signaling pathways, activation of oncogenic signaling cascades such as Wnt/β-catenin and MAPK pathways, and epigenetic alterations including chromatin remodeling defects within the SWI/SNF complex [12–14]. Collectively, these alterations diminish tumor immunogenicity, reduce T-cell priming, and impair cytotoxic lymphocyte recognition of malignant cells. In contrast, tumor-extrinsic resistance is mediated by the surrounding TME, which actively suppresses antitumor immunity. This immunosuppressive milieu is characterized by infiltration of regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), cancer-associated fibroblasts (CAFs), and M2-polarized tumor-associated macrophages (TAMs), all of which inhibit effector T-cell activity and promote immune evasion [10, 15]. In addition, suppressive cytokines and soluble mediators such as TGF-β, VEGF, IL-10, and indoleamine 2,3-dioxygenase (IDO) further reinforce immune dysfunction by impairing T-cell trafficking, promoting T-cell exhaustion, and limiting antigen-presenting cell maturation [11, 16]. Metabolic competition within the TME, including depletion of glucose, glutamine, arginine, and tryptophan, also restricts the proliferation and function of tumor-infiltrating lymphocytes [10, 17]. Other extrinsic factors associated with immunotherapy resistance include gut microbiomes, which can impact tumor immunity by promoting cancer through pro-tumor microbiota, while probiotics may improve immunotherapy [11]. Gut microbiota has also emerged as an important host-related determinant of immunotherapy response in CRC. The intestinal microbiome can influence antitumor immunity by regulating antigen presentation, dendritic-cell (DC) maturation, T-cell priming, cytokine production, epithelial barrier integrity, and microbial metabolites [18]. Dysbiosis may promote immune resistance through chronic inflammation, immunosuppressive myeloid expansion, altered tryptophan and short-chain fatty acid metabolism, and impaired cytotoxic T-cell activity [19]. Conversely, favorable microbial communities may enhance immune checkpoint blockade responses by supporting interferon-related signaling and systemic antitumor immunity. These findings provide the rationale for microbiome-modulating strategies, including fecal microbiota transplantation, probiotics, prebiotics, dietary interventions, and engineered bacterial platforms, which are discussed in later Sect. [20]. Together, these intrinsic and extrinsic mechanisms establish a highly immune-resistant ecosystem that limits the efficacy of current immunotherapeutic strategies in CRC. Emerging strategies to overcome immunotherapy resistance Given the multifaceted nature of immune resistance in CRC, considerable efforts have been directed toward the development of novel therapeutic strategies aimed at enhancing immunotherapy efficacy. Current research is exploring strategies beyond ICI monotherapy. Combinational approaches incorporating chemotherapy, targeted agents, radiotherapy, and innovative modalities such as adoptive cell therapies (including chimeric antigen receptor T-cell (CAR-T) therapy and bispecific T-cell engagers (BiTEs)), cancer vaccines, oncolytic viruses (OVs) therapy, and microbiome-modulating therapies, all aimed at boosting immune activation and reversing resistance [21–25]. Furthermore, neoadjuvant immunotherapy regimens are being explored to leverage systemic immunity before surgery, especially in MSI-H CRC, potentially improving long-term outcomes [26]. Predictive biomarkers and future perspectives The development of reliable predictive biomarkers is critical for optimizing patient selection and guiding personalized immunotherapeutic strategies in CRC. Beyond MSI status, ongoing research is exploring additional biomarkers related to tumor genetics, epigenetic markers, immune infiltration patterns, and host-related factors [27]. A comprehensive understanding of these parameters is essential for identifying patients most likely to benefit from immunotherapy and for guiding rational combination strategies. However, many existing reviews discuss immunotherapeutic modalities and biomarkers separately, without clearly linking specific resistance mechanisms to therapeutic selection and biomarker-guided stratification. The key knowledge gap addressed in this review is how tumor-intrinsic alterations, tumor-extrinsic immune suppression, microbiome-related factors, and emerging biomarkers can be integrated into a practical framework for overcoming immunotherapy resistance in CRC. Therefore, this review integrates current and emerging immunotherapeutic strategies, resistance mechanisms, and biomarker development into a practical framework for rational therapeutic selection and personalized CRC immunotherapy. CRC molecular subtypes Given the pronounced high heterogeneity of CRC at both the genetic and molecular levels, a detailed understanding of its molecular subtypes is essential for interpreting therapeutic responses and optimizing immunotherapy strategies. Various classification systems, including the Jass, Ogino system, Colon Cancer Subtypes (CCS), Sadanandam CRC Assigner System, Colon Cancer Molecular Subtype, and CRC Intrinsic Subtypes, have been designed to categorize CRC [28–33]. An international consortium developed the Consensus Molecular Subtypes (CMS) model to address inconsistencies in CRC subtype classifications. CRC is classified into four CMSs [34]. Beyond its molecular stratification value, the CMS classification provides a clinically meaningful framework not only for informing prognosis but also for predicting response to immunotherapy in CRC. Among the four subtypes, CMS1 (immune MSI), representing about 14% of CRCs, is usually found on the right side of the colon [34]. They harbor a median overall survival (OS) of just 15.9 months in metastatic cases [35]. These tumors are characterized by hypermutated tumors with MSI-H/dMMR [34], which results in a very high TMB rate, exceeding 20 mutations per megabase, in over 80% of CMS1 tumors [36]. They often have BRAF mutations and show a CpG island methylator phenotype (CIMP-H) in addition to decreased signaling through the WNT pathway [34, 37]. The immune-rich tumor microenvironment TME, with high expression of cytokines (CXCL9, CXCL10, IL-15, IFN-γ), tertiary lymphoid structures (TLS), and upregulated HLA class I/II genes, makes CMS1 tumors highly responsive to immunotherapy, such as anti-PD-1/PD-L1 antibodies [36, 38, 39]. Nonetheless, immune evasion via PD-1, PD-L1, CTLA-4, and LAG3 expression can limit efficacy, suggesting potential benefit from combination immunotherapies [40]. Therefore, CMS1 tumors represent the most immunotherapy-responsive CMS subgroup, although co-expression of inhibitory checkpoints may justify combination checkpoint blockade in selected cases. CMS2 (Canonical), accounting for approximately 37% of all cases, is left-sided colon cancer characterized by epithelial differentiation, Wnt and MYC pathway activation, APC loss, KRAS mutations, and TP53 inactivation [34, 35, 37]. Although CMS2 tumors often show a relatively better prognosis, they typically display an immune-cold phenotype with low immune-cell infiltration and limited inflammatory signaling [35]. CMS2 tumors exhibit an immune-desert TME with minimal immune infiltration, resulting in poor responsiveness to ICIs [40]. Therefore, therapeutic strategies for CMS2 should prioritize immune-priming approaches, such as anti-EGFR therapies, such as cetuximab, radiotherapy, cytokine/adjuvant-based therapy, or other combinations designed to convert an immune-desert TME into a more inflamed state [35, 41–43]. CMS3 (Metabolic) accounts for approximately 13% of all CRCs with intermediate survival outcomes, with a median OS of around 18.6 months in metastatic settings and is mainly defined by metabolic dysregulation, altered glutaminolysis and lipid metabolism, and frequent KRAS mutations [34, 35]. It represents moderate activation of WNT and MYC, with PIK3CA mutation and IGFBP3 overexpression, without BRAF mutations [37]. This subtype is distinguished by frequent KRAS mutations (found in 68% of cases), which drive constitutive activation of the MAPK pathway, contributing to resistance to immunotherapies [34]. Targeted metabolic therapies, KRAS inhibitors, or combination regimens with ICIs may improve responses in CMS3 tumors with metabolic dysregulation. CMS4 (mesenchymal), accounting for approximately 23% of CRCs, is associated with aggressive behavior, advanced disease, and poor prognosis [34]. Unlike CMS2, CMS4 is not simply immune-cold; it often contains abundant lymphocyte and macrophage infiltration but remains functionally immunosuppressed due to stromal activation, CAF enrichment, TGF-β/CXCL12 signaling, and angiogenic factors such as VEGF [39, 40]. This immune-excluded and stromal-rich phenotype limits the efficacy of ICIs alone. Accordingly, CMS4 tumors may benefit more from combination strategies targeting stromal suppression, angiogenesis, and immune exclusion, including chemotherapy plus anti-angiogenic therapy such as bevacizumab [44]. In practical terms, immune cells may be present but spatially excluded, exhausted, or suppressed by the stromal compartment, limiting the efficacy of ICI monotherapy. Therefore, CMS4 tumors may benefit from combination strategies targeting stromal suppression, angiogenesis, TGF-β signaling, or immune exclusion, including anti-angiogenic or anti-TGF-β-based approaches. Overall, CMS classification is most useful for immunotherapy when interpreted through immune responsiveness rather than molecular features alone: CMS1 is generally immune-hot and ICI-sensitive, CMS2 and CMS3 are predominantly immune-cold, and CMS4 is immune-infiltrated but stromal-excluded and immunosuppressed. Another molecular subtyping system of CRC is the coordinate immune response cluster (CIRC) typing, which stratifies CRC patients into four groups, including groups A, B, C, and D, derived from the clustered expression of a gene set [45]. Group A patients exhibit high expression of CIRC genes, characterized by MSI-H, POLE mutations, and upregulation of immune checkpoint genes, including CTLA-4, PD-L1, LAG-3, and TIM3, which correlate with a high mutational load and immune infiltration, suggesting that they may benefit from checkpoint blockade (ICB) therapy. In contrast, patients with lower CIRC expression have frequent mutations in KRAS, BRAF, NRAS, TP53, PIK3CA, and PTEN. Group D patients with RAS mutations exhibit resistance to ICIs. Collectively, CMS and CIRC classifications predict ICI responsiveness by identifying immune-inflamed subtypes (e.g., CMS1 and CIRC-A) enriched for MSI-H/dMMR tumors with pre-existing antitumor immunity [34, 45]. However, these frameworks show limited predictive value for immunotherapeutic approaches beyond ICIs, including adoptive cell therapies, bispecific antibodies, cancer vaccines, and OVs, as response to emerging strategies depends on additional determinants such as antigen expression, stromal architecture, immune cell trafficking, and metabolic features of the TME [46]. Moreover, intratumoral heterogeneity and temporal dynamics of immune states further constrain the predictive accuracy of static transcriptomic classifications [47]. Consequently, precision immunotherapy in CRC will likely require integrative subtyping models that incorporate molecular classification with immune phenotyping, spatial transcriptomics, and biomarker-based TME assessment to optimize patient selection for both ICIs and next-generation combinatorial immunotherapies. Furthermore, recent spatial multi-omics studies have refined CRC immune classification beyond conventional MSI/MSS and CMS subtypes. In mCRC patients receiving immunotherapy, spatial immune mapping showed that C1QC⁺ resident tissue macrophages were enriched in responders regardless of MSI or MSS status. Their co-localization with CD4⁺ T cells, supported by MHC-II–mediated antigen presentation, was associated with better response, whereas CAFs disrupted this interaction in non-responders [48]. These findings suggest that immune-hot and immune-cold states are not defined only by immune-cell abundance, but also by spatial immune organization, antigen-presenting niches, and stromal barriers. Therefore, future CRC stratification should integrate molecular subtype, spatial multi-omics, and immune–stromal architecture to improve immunotherapy selection. Overall, these subtype-specific immune features indicate that CMS classification is most useful when interpreted through the lens of immune responsiveness: CMS1 is generally immune-hot, CMS2 and CMS3 are predominantly immune-cold, and CMS4 is immune-infiltrated but stromal-excluded and immunosuppressed. These immune phenotypes provide the rationale for the therapeutic strategies discussed in the next section, as immune-hot tumors may benefit from checkpoint blockade, whereas immune-cold or immune-excluded tumors often require immune priming, stromal remodeling, or rational combination approaches. Current immunotherapeutic modalities and adjunctive strategies in CRC Building on the molecular and immune phenotypes described above, this section summarizes current and emerging immunotherapeutic strategies for overcoming subtype-specific resistance barriers in CRC. Over the last decade, immunotherapy has gained prominence as an alternative to surgery, chemotherapy, and radiation for the treatment of different cancers, revolutionizing oncology and becoming an established therapy for particular subgroups of CRC [47]. Immune checkpoints are a category of molecules present on the surface of immune cells that regulate the degree of immunological activation [49]. They inhibit autoimmune disorders and initiate immune-mediated attacks on healthy cells. In cancers, immune checkpoints, including PD-1 and CTLA-4, are inappropriately activated, leading to a weaker tumor immune response [41, 50]. ICB treatment has shown remarkable therapeutic benefits on a broad range of tumor types due to its capacity to disrupt the interaction between immune checkpoints and their receptors [51, 52]. ICIs, especially PD-1 and CTLA-4 inhibition, have shown the most efficacy in treating MSI-H/dMMR cancers, resulting in high response rates and long-lasting disease control [53, 54]. The approval of pembrolizumab and nivolumab, either individually or together, has revolutionized the standard of therapy for this group of patients [55]. Conversely, most MSS/pMMR exhibit resistance to ICIs owing to a low tumor mutational load, limited tumor-infiltrating lymphocytes, and immunosuppressive cytokine networks that provide a cold TME [56]. Addressing these obstacles has emerged as a significant research priority, promoting the creation of innovative strategies for improving immune activation. These include adoptive cell transfer (e.g., TILs, TCR-engineered cells), CAR-T therapy, antibody-based approaches such as EGFR blockade and BiTEs, therapeutic cancer vaccines, OVs therapy, microbiome modulation, and cytokine/adjuvant immunotherapy. Numerous techniques are being examined alongside with ICIs or traditional chemotherapy/radiotherapy to enhance antigen presentation, T-cell trafficking, and effector function synergistically [11]. Importantly, biomarker maturity is not uniform across these approaches. It is well established for ICIs, where MSI-H/dMMR and high TMB enrich for response, and for anti-EGFR therapy, where RAS/BRAF status guides treatment selection. By contrast, for ACT, CAR-T, BiTEs, vaccines, OVs, microbiome-modulating therapies, and cytokine/adjuvant strategies, biomarker use remains largely investigational and will be highlighted only where supporting data are available. This section evaluates each therapy modality, outlining its mechanism, clinical evidence, and contribution to overcoming immune resistance (Fig. 1). Strategies to overcome therapeutic resistance in CRC. Schematic overview of emerging and established immuno-oncology approaches that can restore anti-tumor immunity in CRC. Wedges illustrate: (i) Cytokine therapy administration or induction of IL-12, IL-18, and TNF-α to boost effector responses. (ii) ICI therapy—blockade of PD-1/PD-L1 and other checkpoints (e.g., TIM-3) to enhance T-cell activation. (iii) Adoptive cell therapy expansion of tumor-infiltrating lymphocytes (TILs) or infusion of CAR-T cells. (iv) EGFR-targeted antibody agents such as cetuximab or panitumumab to inhibit EGF/EGFR signaling and suppress growth and metastasis. (v) Bispecific T-cell engagers (BiTEs) antibodies that link T cells to tumor antigens to promote tumor-cell lysis;. (vi) Oncolytic virus therapy engineered viruses (e.g., adenovirus) sometimes armed with IL-12 or tumor-antigen genes to increase anti-tumor activity. (vii) Microbiome therapy manipulation of the gut microbiota (e.g., FMT) to potentiate systemic anti-tumor immune responses. Abbreviations: Ab: antibody, BiTE: bispecific T-cell engager, CAR: chimeric antigen receptor, CEA: carcinoembryonic antigen, CRC: colorectal cancer, EGF: epidermal growth factor, EGFR: epidermal growth factor receptor, FMT: fecal microbiota transplantation, ICI: immune checkpoint inhibitor, IL: interleukin, MHC: major histocompatibility complex, PD-1: programmed cell death protein-1, PD-L1: programmed death-ligand 1, TIL: tumor-infiltrating lymphocyte, TIM-3: T-cell immunoglobulin and mucin-domain containing-3, TNF-α: tumor necrosis factor-alpha Biomarker-guided patient selection is discussed when clinically established. For emerging modalities, biomarker evidence is presented as exploratory rather than definitive. To directly address the mechanistic heterogeneity of resistance, the following subsections also indicate which tumor-intrinsic and tumor-extrinsic resistance barriers are most plausibly targeted by each therapeutic modality. Given the marked difference in immunotherapy responsiveness between MSI-H/dMMR and MSS/pMMR CRC, Table 1 provides a comparative overview of the current evidence, therapeutic rationale, and clinical relevance of major immunotherapy-based strategies across these two settings. Table 1Comparative therapeutic rationale and current evidence for immunotherapy-based strategies in MSI-H/dMMR versus MSS/pMMR colorectal cancerTherapeutic strategyMSI-H/dMMR CRCMSS/pMMR CRCClinical implicationRefPD-1/PD-L1 blockadeHigh clinical benefit; established standard in advanced diseaseLimited activity as monotherapyMost effective in MSI-H/dMMR;MSS tumors require combination approaches [56–58]Dual checkpoint blockadeImproves immune activation and may deepen responsesInvestigational; modest benefit in selected combinationsMore established in MSI-H/dMMR;requires biomarkers in MSS disease [57, 59, 60]Chemotherapy + ICIMay enhance antigen release but not always requiredMixed results; limited efficacy in unselected MSS CRCMore relevant for immune priming inMSS tumors [40, 61, 62]Radiotherapy + ICIMay further enhance local and systemic immune activationPromising strategy to convert cold tumors into inflamed tumorsParticularly relevant for immune-coldor immune-excluded MSS CRC [63, 64]Targeted therapy + ICIUseful in molecularly selected subgroupsPotential benefit in selected settings, such as BRAF- or MEK-targeted combinationsRequires mutation-guided patientselection [65, 66]ACT/CAR-T/BiTEsPotentially useful but not dependent only on MSI statusPromising for MHC-defective or antigen-positive tumorsRequires antigen expression andcontrol of TME barriers [67–70]Cancer vaccinesStronger rationale due to high neoantigen burdenRequires personalized neoantigens and combination with immune-priming agentsMSI-H is more naturally vaccine-suitable;MSS needs combination strategies [71, 72]Oncolytic virusesMay amplify pre-existing immune responsesPromising for immune-cold tumors by inducing antigen release and inflammationEspecially relevant for MSS CRCsensitization to ICIs [73–75]Microbiome modulationMay support ICI responsivenessInvestigational adjunct to improve immune activationNeeds validation, safety monitoring,and biomarker-guided use [22, 76]Cytokine/adjuvant strategiesMay enhance effector immune responsesMay help overcome weak priming and poor interferon signalingMost useful as part of rationalcombinations [60, 77] Comparative therapeutic rationale and current evidence for immunotherapy-based strategies in MSI-H/dMMR versus MSS/pMMR colorectal cancer Most effective in MSI-H/dMMR; MSS tumors require combination approaches More established in MSI-H/dMMR; requires biomarkers in MSS disease More relevant for immune priming in MSS tumors Particularly relevant for immune-cold or immune-excluded MSS CRC Requires mutation-guided patient selection Requires antigen expression and control of TME barriers MSI-H is more naturally vaccine-suitable; MSS needs combination strategies Especially relevant for MSS CRC sensitization to ICIs Needs validation, safety monitoring, and biomarker-guided use Most useful as part of rational combinations ICIs The development of ICIs, such as monoclonal antibodies (mAbs) targeting PD-1, PD-L1, and CTLA-4, has fundamentally transformed the treatment paradigm and prognosis of various solid tumors [57, 58]. Nonetheless, the clinical efficacy of this novel treatment strategy is markedly diverse across various tumor types, perhaps due to the significant heterogeneity in immunophenotypes [78]. In CRC, immune checkpoint inhibition has become an established therapeutic option for patients with dMMR/MSI-H tumors, particularly in the metastatic setting [59]. Indeed, ICIs have transformed the treatment of MSI H/dMMR CRC, with PD-1 inhibition being established as a first-line standard [79]. The phase III KEYNOTE-177 trial established pembrolizumab monotherapy as the standard first-line treatment for patients with MSI-H/dMMR mCRC, demonstrating significantly improved progression-free survival compared with conventional chemotherapy. At more than five years of follow-up, pembrolizumab achieved a median overall survival of 77.5 months with durable responses, further confirming its long-term clinical benefit in this population [80] Another checkpoint is CTLA-4, which is mostly regulated by FOXP3 on Tregs, resulting in an immunosuppressive TME [81]. Synergistic therapeutic effect has been shown by the combination inhibition of PD-1 and CTLA-4, leading to approval by the FDA for circumstances such as metastatic CRC with the MMR/MSI-H phenotype [82]. Recent clinical studies are evaluating the effectiveness of PD-1 and CTLA-4 inhibition in combination with supplementary therapies, such FOLFOX plus bevacizumab (NCT05627635) or panitumumab (NCT03442569) [60, 83]. LAG-3 is an inhibitory receptor expressed on both CD4+ and CD8+ T lymphocytes upon antigen stimulation. It is consistently expressed on FOXP3+ Tregs, leading to increased levels of the immunoregulatory cytokines IL-10 and TGF-β, which inhibit tumor-specific responses from T-cells [67]. Preclinical findings suggest that simultaneous inhibition of LAG-3 and PD-1 may rejuvenate T-cell functionality and overcome anergy. A phase I clinical study revealed that the combination of the anti-LAG-3 antibody favezelimab and the anti-PD-1 antibody pembrolizumab in patients with advanced MSS mCRC had notable antitumor efficacy, with an OS of 8.3 months and a median progression-free survival of 2.1 months [84]. Phase II research (NCT03642067) is now being conducted to evaluate the safety and effectiveness of nivolumab when combined with relatlimab in patients with advanced MSS CRC [85]. Adoptive cell therapy (ACT) Among several immunotherapy approaches, ACT has emerged as one of the most promising therapies. ACT is the administration of ex vivo-expanded and/or genetically engineered immune cells, including T cells or natural killer (NK) cells, that are precisely directed to tumor locations to improve immunity against tumors [68]. ACT for CRC includes various strategies such as CAR-T cells, TCR-engineered T cells (TCR T), tumor-infiltrating lymphocytes (TILs), and NK-based adoptive methods, with the objective of improving anti-tumor immunity through the administration of potent immune cells into patients [86]. CAR-T treatment is pivotal in ACT; however, its effectiveness in CRC is limited by antigen heterogeneity, an immunosuppressive TME, and on-target off-tumor damage [87]. Recent reviews outline combinatorial tactics such as CRISPR/Cas9 edits and checkpoint inhibitor co-administration to overcome these challenges [88]. Novel methods such as TCR T and CAR NK enhance therapeutic capabilities by targeting intracellular antigens and providing innate cytotoxicity, respectively [89]. Preclinical CRC models indicate that the combination of CAR-T therapy with PD-1 inhibition improves T-cell infiltration, persistence, and anti-tumor activity [90]. Advanced platforms, such as Vδ1+ Delta One T (DOT) cells, exhibit promise cytotoxicity against both MSI and MSS CRC in organoid and xenograft models; however the presence of inhibitory receptors could restrict performance [87, 91]. Furthermore, cytokine-induced killer (CIK) cell treatment, in combination with chemotherapy or surgery, has shown enhanced overall and progression-free survival in retrospective CRC populations, with carcinoembryonic antigen (CEA) levels suggested as a predictive biomarker for responses [92]. At present, ACT selection in CRC is driven mainly by target-antigen expression and tumor immune context rather than by a single validated predictive biomarker. CAR-T therapy as a specialized ACT platform Building upon the broader principles of ACT, CAR‑T cell approaches introduce genetic engineering to enhance tumor specificity and antitumor potency. Born in 1989 [93] as a technique to guide the specificity of T cells in an MHC-unrestricted manner, CAR-T therapy has been recently employed to boost the immune system’s response to cancer cells when the usual response has failed [69]. CAR-T treatment marks an important milestone in ACT, delivering tailored and robust immune activation [94]. Despite its revolutionary success in hematologic malignancies, transferring CAR-T cells to CRC has been challenging [95]. Preclinical and clinical findings highlight barriers such as antigen heterogeneity, poor T-cell trafficking, immunosuppressive microenvironments, and on-target off-tumor toxicity, limiting effectiveness in solid tumors like CRC [96]. Early-phase studies have targeted many CRC-associated antigens as CAR targets, including CEA, HER2, GUCY2C, TAG 72, EpCAM, and CD133 [97]. Accordingly, current biomarker development for CRC CAR-T remains largely antigen-based, with candidate tumors selected according to surface-antigen expression and accessibility rather than a single validated genomic marker. Therapeutic responses were either negligible or nonexistent for the majority of targets; however, others, like CEA targeted CAR-T cells, showed tumor cell persistence and tolerability [98]. Nevertheless, various solutions aim to overcome this limitation: One approach includes CRISPR-based gene editing to increase localized production of cytokines such as IL-12/IL-2 within the tumor, thereby boosting effectiveness while minimizing systemic toxicity. Preclinical CRC models revealed near 100% eradication in solid tumor systems [99]. Emerging treatments include modified microorganisms that regulate the TME to promote CAR-T trafficking and activity in gastrointestinal malignancies. These distinctive combinatorial techniques hold promise for unlocking CAR-T’s full potential in CRC [21, 100]. Antibody-based therapies Despite the promise of cell-based therapies, antibody-based approaches offer a more scalable and clinically accessible strategy for modulating tumor–immune interactions. Antibody-based therapeutics are fundamental to targeted therapy in CRC, using the highly specific nature of mAbs to inhibit oncogenic signaling and activate immune effector mechanisms [101]. In mCRC, anti-EGFR antibodies, including cetuximab and panitumumab, are widely used for RAS wild-type tumors, since they inhibit ligand-induced EGFR activation, downstream MAPK signaling, and enhance antibody-dependent cellular cytotoxicity (ADCC). Their effectiveness depends on biomarker-driven selection, while activating RAS or BRAF mutations result in primary resistance [102]. In addition to conventional mAbs, novel forms including BiTEs have arisen. These molecules concurrently attach to a TAA and CD3 on T cells, controlling polyclonal T-cell cytotoxicity toward cancer cells irrespective of MHC status [103, 104]. BiTE constructs targeting CEA and EpCAM are undergoing clinical assessment in CRC, with preliminary signals of efficacy in extensively pretreated patients [105]. Next-generation antibody methods are being developed to address intrinsic and acquired resistance caused by EGFR ectodomain mutations, the activation of bypass pathways (such as HER2 and MET), or downstream abnormalities in KRAS/NRAS [106]. This includes antibody-drug conjugates (ADCs) that transport cytotoxins directly to neoplastic cells, glycoengineered mAbs with improved FcγR affinity for enhanced ADCC, and multispecific antibodies that simultaneously target numerous epitopes or receptors [107]. EGFR-targeted antibodies The EGFR is highly expressed in a minority of CRC and is crucial for promoting tumor proliferation, survival, and metastasis via the activation of the RAS–RAF–MEK–ERK signaling pathway [108]. Cetuximab, a chimeric IgG1 mAb, and panitumumab, an entirely human IgG2 antibody, have received approval for the treatment of RAS wild-type mCRC [109]. These drugs exert their anticancer effects by competitively reducing ligand binding to EGFR, hence inhibiting downstream mitogenic signaling, and by activating immune effector cells to generate ADCC [102]. In addition to mediating ADCC, cetuximab can enhance antitumor immunity by activating NK cells through Fcγ receptor engagement, thereby promoting cytokine release, dendritic cell maturation, and subsequent priming of tumor-specific CD8+ T cells [110, 111]. These immunomodulatory effects provide a rationale for combining EGFR-targeted antibodies with other immunotherapeutic strategies. The addition of EGFR antibodies to chemotherapy (FOLFIRI or FOLFOX) dramatically increases progression-free and OS in patients with RAS wild-type malignancies, according to clinical studies like CRYSTAL, PRIME, and FIRE-3. Crucially, biomarker-driven selection is essential since patients with activating mutations in KRAS, NRAS, or BRAF develop resistance and do not benefit from EGFR-targeted treatment [112]. Notwithstanding early responses, acquired resistance often emerges within 12–18 months, powered by EGFR ectodomain mutations (e.g., S492R), secondary RAS mutations, or the activation of alternative signaling pathways, including HER2 and MET amplification [108]. Ongoing researches emphasize combination therapies, next-generation EGFR inhibitors, and dual-targeting antibodies for overcoming these resistance mechanisms [70]. BiTEs BiTEs are engineered antibodies that concurrently bind a tumor-associated antigen (TAA) on cancer cells and the CD3ε segment of the T-cell receptor complex, therefore diverting polyclonal T cells to lyse tumor cells without reliance on peptide-MHC recognition [113]. This strategy is especially promising in MSS CRC, where inadequate antigen presentation restricts immune identification. In CRC, CEA is one of the most thoroughly examined targets for BiTE development [23]. CEA-BiTEs, including cibisatamab (CEA-TCB), have shown significant T-cell activation, cytokine release, and tumor regression in preclinical models, while early-phase studies indicate disease stability and random tumor reduction in severely pretreated patients [114]. BiTEs that target EpCAM are also being developed, taking advantage of the protein’s high expression in epithelial malignancies [115]. Primary concerns are on-target off-tumor toxicity, notably gastrointestinal inflammation resulting from CEA/EpCAM expression in normal mucosa, and the potential for cytokine release syndrome (CRS) [101]. Strategies for addressing these challenges include step-up dosage regimens, half-life-extended formulations, and tumor-conditional activation by protease-cleavable linkers [115]. Current phase I/II studies are assessing BiTEs alongside PD-1 inhibition or chemotherapy to improve T-cell infiltration and prevent immune evasion in MSS CRC [71]. Cancer vaccines Cancer vaccines aim to provoke or enhance anticancer immune responses by improving tumor antigen presentation and facilitating the activation of tumor-specific T cells [116]. In CRC, many vaccination platforms have been explored, including peptide-based, whole-cell, DC-based, and nucleic acid (DNA/RNA) vaccines [117]. Early-phase studies have shown the immunogenicity of peptide-based vaccinations that target TAAs, including guanylyl cyclase C (GUCY2C), mucin-1 (MUC1), and CEA [22]. Nonetheless, therapeutic effectiveness in advanced CRC is limited owing to tumor-induced immune suppression and reduced T-cell infiltration in MSS disease [118]. DC vaccines, loaded with TAA or neoantigens, have shown encouraging immune activation, particularly when used in combination with ICIs to suppress inhibitory pathways [72]. Recent advancements focus on individualized neoantigen vaccines tailored to specific tumor mutational patterns. In MSI-H/dMMR CRC, the elevated tumor mutational load facilitates the detection of immunogenic neoepitopes, thereby promoting robust CD8+ T-cell responses [119]. Therefore, MSI-H/dMMR and other hypermutated contexts currently represent the most plausible biomarker-enriched settings for personalized vaccine development, whereas MSS disease will likely require individualized neoantigen or combination-based strategies. Early clinical findings indicate that these vaccines may elicit long-lasting T-cell immunity and might improve the effectiveness of PD-1 inhibition [120]. Preclinical and early-phase studies have shown that personalized neoantigen mRNA vaccines may elicit robust CD8⁺ T-cell and CD4⁺ helper responses against CRC-specific antigens, indicating that they are actively being developed for the prevention of CRC [73, 121]. Despite these advancements, challenges persist, including antigen heterogeneity, immune evasion, and the immunosuppressive TME. Investigations are underway into rational combination tactics, including ICIs, cytokine adjuvants, or microbiome modification to enhance vaccination effectiveness in CRC [122, 123]. OVs therapy OVs are genetically engineered or naturally occurring viruses that selectively replicate within tumor cells, inducing direct lysis and releasing TAAs that can stimulate systemic immunity [124]. In CRC, OVs are being studied both as monotherapy and in combination with ICIs or standard treatments, with particular emphasis on microsatellite-stable (MSS) tumors that are poorly responsive to ICIs [125]. Adenovirus-based vectors are among the most extensively investigated in CRC [74]. Preclinical studies using adenoviruses armed with cytokines such as IL-12 and a modified IL-18 have demonstrated enhanced CD8⁺ T-cell and NK-cell infiltration, improved antigen presentation, and prolonged survival. Similarly, constructs incorporating GM-CSF or IL-15 have shown potential to reshape the TME toward a more immunogenic state, thereby sensitizing tumors to ICIs [24, 88]. Other OV platforms are also advancing. Reovirus, when combined with PD-1 blockade in MSS CRC models, improved antigen presentation, expanded CD8⁺ T cells, and reduced immunosuppressive macrophages [126]. Vaccinia virus derivatives are being explored as delivery vehicles and immune activators. Among RNA virus approaches, vesicular stomatitis virus (VSV-IL-2) has shown the ability to boost antigen-specific CD8⁺ responses, while Coxsackievirus A21 (CVA21) inhibited tumor growth and modulated cytokines such as IFN-γ, IL-10, and TGF-β in CRC mouse models [127, 128]. A major limitation of OV therapy remains efficient intratumoral delivery, especially for metastatic lesions. Investigational strategies include nanoparticle encapsulation and mesenchymal stem cell carriers, while hepatic artery infusion is still considered theoretical in CRC [129]. Clinically, the vaccinia-derived OV PexaVec has reached phase I/II testing (NCT03206073) in combination with durvalumab and tremelimumab for MSS CRC. The regimen was feasible and biologically active, though responses were modest [130]. These data, together with translational evidence showing OV-induced upregulation of T-cell activation genes in CRC responders, underscore the potential of OVs to convert cold tumors into immunologically hot lesions and to broaden immunotherapy efficacy in CRC [131]. Additional mechanistic evidence supports the rationale for combining OVs with immune checkpoint blockade [132, 133]. Li and Sun emphasized that OVs can shift therapeutic emphasis from direct viral oncolysis toward coordinated activation of host antitumor immunity, thereby helping convert immunologically cold tumors into hot tumors and improving responsiveness to ICIs [132]. In addition, a recent mechanistic study showed that m1A inhibition can enhance oncolytic herpes simplex virus-elicited antitumor immunity by downregulating MYC/PD-L1 signaling and preventing OV-induced immune escape [133]. Although this evidence was generated outside CRC, it highlights a relevant mechanism by which OV-induced immune activation may be strengthened through modulation of checkpoint-related resistance pathways [133]. Microbiome-modulating therapies The gut microbiome significantly influences CRC biology and responsiveness to immunotherapy through antigen presentation, cytokine signaling, and microbial metabolites. Dysbiosis characterized by enrichment of Fusobacterium nucleatum has been linked to CRC progression and chemoresistance; a recent mechanistic study showed that F. nucleatum suppresses chemotherapy-induced pyroptosis and correlates with higher postoperative recurrence, highlighting it as a therapeutic target [134]. Emerging reviews emphasize that microbiota modulation may regulate the efficacy of ICIs across solid tumors, including CRC, through strategies such as fecal microbiota transplantation (FMT), probiotics, prebiotics, dietary interventions, or engineered strains [135]. This approach is particularly relevant for MSS/pMMR CRC, where ICIs have limited activity. A 2025 focused review outlined how microbiota modification combined with PD-1/PD-L1 blockade could improve antitumor immunity, citing early clinical findings and molecular rationale for patient stratification [25]. Supporting this, another study summarized eight preclinical studies and five clinical trials testing FMT or experimental microbial interventions with ICIs in CRC; reported benefits included tumor growth control and enhanced immune activation, although confirmatory phase III data remain pending [136]. Preclinical studies further show that FMT plus anti-PD-1 improves survival and reshapes microbial composition in CRC mouse models, favoring T-cell responses [137]. Specific commensals are also under investigation: Akkermansia muciniphila has demonstrated immunostimulatory properties, with proposed mechanisms including barrier enhancement, improved antigen presentation, and DC priming, while translational research is identifying biomarkers of response [42]. In synthetic biology, engineered E. coli Nissle 1917 expressing CRC neoantigens induced robust antitumor immunity and tumor regression in mice, illustrating a programmable tool to convert cold tumors into inflamed, T-cell–infiltrated states [138, 139]. Overall, microbiome-modulating therapies are evolving from associative observations to interventional strategies. Current evidence supports their role as adjuncts, particularly in MSS CRC, while ongoing studies aim to define optimal modalities, safety, and biomarker-guided patient selection [25]. Although microbiome-directed strategies may enhance immunotherapy responsiveness in CRC, their clinical interpretation requires caution. Most reported microbial signatures remain associative rather than proven causal determinants and may be influenced by diet, antibiotic exposure, geography, tumor stage, prior therapy, sampling time, and sequencing methods. Therefore, candidate taxa or microbial metabolites should be validated in mechanistic studies and prospective trials. Moreover, FMT, probiotics, dietary modulation, and engineered bacterial platforms still face translational barriers, including donor selection, pathogen transmission, batch variability, engraftment stability, long-term safety, and protocol standardization. Thus, microbiome modulation should currently be considered an investigational adjunct rather than an established CRC immunotherapy strategy [140, 141]. Cytokine and adjuvant immunotherapies Cytokine-based immunotherapies are being redefined to improve efficacy and reduce toxicity in CRC. Early cytokines such as high-dose IL-2 and IFN-α were limited by systemic adverse effects and insufficient durability [77]. Current approaches focus on engineered cytokines that selectively activate effector T and NK cells while avoiding Treg expansion. A publication in Nature Reviews Immunology described IL-2 variants biased toward IL-2Rβγ and IL-15 superagonists, both designed to enhance cytotoxic lymphocyte persistence and showing synergy with PD-1 blockade in preclinical CRC models [142]. Locoregional cytokine delivery provides another innovation. Immunocytokines, such as CD45-directed constructs, have been shown to safely induce systemic antitumor immunity when administered locally, reshaping the TME with reduced systemic toxicity [77]. Similarly, Cheng et al. demonstrated in 2025 that an oncolytic adenovirus encoding IL-12 and a modified IL-18 significantly increased CD8⁺ T-cell infiltration and prolonged survival in CRC mouse models, highlighting the potential of cytokine-armed vectors to remodel the tumor milieu [143]. Adjuvant immunotherapies are advancing in parallel. Toll-like receptor (TLR) agonists, including CpG oligodeoxynucleotides, have been tested with ICIs in gastrointestinal cancers, improving DC activation and interferon signaling [144]. STING agonists are also under active investigation. In CT26 CRC models, a branched glycopolymer nanoassembly delivering the STING agonist diABZI enhanced PD-L1 blockade by activating type I interferon signaling and reprogramming the TME toward a state enriched in effector T-cell activity, thereby improving responsiveness to immunotherapy [145]. Together, these advances highlight cytokine engineering and adjuvant strategies as promising tools to overcome immunotherapy resistance in CRC. Matching therapeutic strategies to specific resistance barriers Because immune resistance in CRC arises from both tumor-intrinsic and tumor-extrinsic mechanisms, each therapeutic strategy is expected to overcome only a defined subset of resistance barriers rather than all mechanisms simultaneously. ICIs primarily counteract adaptive immune suppression mediated by inhibitory receptors such as PD-1, PD-L1, CTLA-4, and LAG-3, and are most effective when tumors retain antigen presentation capacity and contain pre-existing tumor-reactive lymphocytes, as observed in many MSI-H/dMMR tumors [11, 59, 79, 80]. However, ICIs alone are unlikely to fully overcome irreversible defects in antigen presentation, such as severe MHC-I or B2M loss, or profound immune exclusion driven by stromal and cytokine-mediated barriers. Adoptive cell therapies can partially bypass inadequate endogenous T-cell priming by supplying expanded or engineered effector cells. TIL- and TCR-based approaches are most relevant when tumor antigens are present and antigen presentation remains functional, whereas CAR-T cells and CAR-NK cells can recognize surface antigens in an MHC-independent manner and may therefore help address resistance associated with impaired MHC-I–restricted recognition [86, 89, 91]. Nevertheless, their efficacy in CRC remains limited by antigen heterogeneity, poor tumor trafficking, stromal exclusion, immunosuppressive cytokines, and metabolic stress within the TME [98]. Armored or gene-edited CAR platforms, cytokine-secreting CAR cells, and combinations with checkpoint blockade or microbiome engineering are being developed to overcome these extrinsic barriers [98]. Antibody-based therapies address a different spectrum of resistance mechanisms. EGFR-targeted antibodies are most useful for tumors driven by EGFR signaling and can enhance immune effector mechanisms through antibody-dependent cellular cytotoxicity, but their benefit is restricted by RAS/BRAF mutations and acquired bypass signaling through EGFR ectodomain mutations, HER2, or MET activation [108]. In contrast, BiTEs redirect polyclonal T cells toward tumor-associated antigens such as CEA or EpCAM and can mediate tumor-cell killing independently of classical peptide-MHC recognition, making them particularly relevant for tumors with defective antigen presentation but preserved surface-antigen expression [115]. Cancer vaccines primarily aim to overcome insufficient tumor antigen priming by increasing the breadth and durability of tumor-specific T-cell responses. Neoantigen-based vaccines are most rational in MSI-H/dMMR, POLE/POLD1-mutant, or otherwise hypermutated tumors with a higher probability of immunogenic epitopes, whereas MSS tumors often require combination with ICIs, adjuvants, cytokines, or microbiome-modulating strategies to overcome poor T-cell infiltration and immune suppression [116]. May address cold or poorly inflamed tumors by inducing immunogenic tumor-cell lysis, releasing tumor antigens, activating innate immune pathways, and promoting CD8 + T-cell and NK-cell infiltration; therefore, they are particularly relevant for resistance associated with poor antigen release, weak antigen presentation, and low immune infiltration [65]. Microbiome-modulating therapies target host- and microenvironment-related resistance, including dysbiosis, altered cytokine signaling, impaired antigen presentation, and unfavorable microbial metabolites. Strategies such as FMT, probiotics, prebiotics, dietary modulation, and engineered bacteria may enhance systemic and intratumoral immunity, especially as adjuncts to ICIs in MSS/pMMR CRC [136]. Cytokine and adjuvant immunotherapies are designed to overcome insufficient immune activation, impaired DC maturation, weak interferon signaling, and reduced effector T-cell or NK-cell persistence. IL-2/IL-15-based agents, IL-12/IL-18-armed vectors, TLR agonists, and STING agonists can reprogram the TME toward a more inflamed state, although their ability to overcome severe stromal exclusion, nutrient depletion, or irreversible tumor-intrinsic genetic defects remains incomplete and likely requires rational combinations [143, 146]. Overall, these therapeutic classes should be viewed as complementary rather than interchangeable. MHC-I downregulation may be more effectively addressed by MHC-independent approaches such as CAR-based therapies and BiTEs; immune exhaustion and checkpoint-mediated suppression by ICIs and next-generation checkpoint combinations; defective priming by vaccines, OVs, and adjuvants; immunosuppressive cytokine and stromal barriers by cytokine/adjuvant strategies, anti-angiogenic or anti-TGF-β-based combinations; metabolic and microbiome-associated resistance by microbiome and metabolic interventions; and oncogenic pathway-driven resistance by biomarker-selected targeted therapies combined with immunotherapy. This framework supports a resistance-mechanism-guided approach to combination therapy in CRC. Biomarkers of response and resistance to CRC immunotherapy The reduced efficacy of Immunotherapy in CRC patients led to the subsequent investigation of identifying biomarkers that predict response or resistance to immunotherapy in CRC, which is critical for patient stratification and improving therapeutic outcomes. The following sections summarize the most established biomarkers, including microsatellite instability, tumor mutational burden, PD‑L1 expression, and features of the TME, as well as emerging biomarkers that may further refine patient selection in the future. MSI, dMMR, and TMB Genetic and molecular biomarkers, including MSI, dMMR, and TMB, have emerged as significant and reliable predictors in this field. Among the most validated predictors are MSI and dMMR. Tumors classified as MSI-H contribute to robust immune activation, which correlates strongly with benefit from immune ICIs, prompting regulatory approval of pembrolizumab for MSI-H/dMMR mCRC [6]. However, the majority of CRCs are MSS, which typically exhibit primary resistance to ICIs, underscoring the necessity for additional biomarkers. TMB has emerged as a significant predictive marker, with a higher rate linked with tumor neoantigen load and improved immunotherapy responses across various cancers [147]. In CRC, MSI-H tumors are often TMB-high (TMB-H); however, subsets of MSS tumors with elevated TMB also demonstrate responsiveness to ICIs [148]. High TMB does not necessarily guarantee a positive reaction to immunotherapy; the quality of mutations tends to be more important than their quantity, and evaluating tumor neoantigen burden (TNB) may enhance the assessment of TMB. Rousseau et al. demonstrated that in Stage IV colorectal patients, high TMB is not useful in predicting response to ICIs [149]. Oncogenic mutations: KRAS and BRAF Mutations in oncogenes such as KRAS and BRAF have recently been shown to impact immunomodulation [150]. BRAF mutation is considered a predictive factor as it has been identified as a negative prognostic marker in CRC, promoting the development of TME (e.g., increased TILs, PD-L1 expression) [151]. However, a meta-analysis comparing the overall response rate (ORR) of immunotherapy in MSI-H CRC patients found no significant differences in ORR between those with BRAF mutations and those with wild-type [152], which indicates a poor predictive value of BRAF mutation in ICI response in MSI-H CRC. BRAF mutation may suggest a potential marker for combination strategies (e.g., BRAF inhibitor + ICI) [153]. A recent phase 2 study evaluating the combination of encorafenib, cetuximab, and nivolumab in patients with MSS, BRAF-V600E-mutated mCRC indicated that this combination could induce a transient MSI-H phenotype and potentially enhance response to anti-PD-1 antibodies [153]. KRAS mutation is associated with a more immunosuppressive TME through increased Tregs and reduced activated CD4+ memory T cells [154]; however, it’s not a strong independent predictor of immunotherapy response. Some studies suggest that KRAS-mutant tumors may upregulate PD-L1 expression, but this does not consistently translate to better ICI outcomes [155]. However, Co-mutation with TP53 may enhance immunogenicity and potentially improve response to anti-PD-1 therapy [156]. PD-1/PD-L1 axis as predictive biomarkers TME markers, including PD-1/PD-L1 Expression, TILs, and IFN-γ, can be associated with response to immunotherapy in CRC. A meta-analysis has shown that PD-L1 expression can act as a biomarker for an unfavorable prognosis and the adverse clinicopathological characteristics [157]. In contrast, PD-L1 expression on tumor cells has shown varied outcomes and has failed as a prognostic marker due to a lack of standardization [158], demonstrating the fact that PD-L1 expression is not a reliable biomarker. Several issues, including Tumor stage, the sampling location, and timing, are significant factors in the results of PD-L1 staining [159]. Additionally, the absence of a standardized evaluation and scoring method limits its clinical significance. TILs and immune contexture The TILs, characterized by higher infiltration of CD8+ T cells, CD4+ T cells, Th1 cells, and IFN-γ upregulation, can also serve as a marker to predict the efficacy of immunotherapy [63]. A higher percentage of CD39+ cells in CD8+ TILs may indicate more effective PD-1/PD-L1 signaling, suggesting CD39+ as a potential marker [160]. In addition, TME immune types (TMITs) are classified into four groups based on PD-L1 expression and TILs, highlighting the significance of PD-L1 and TIL recruitment [161]. PD-L1-positive TILs are categorized as TMIT type I, which benefits from ICIs therapy [161]. TMIT stratification can help distinguish between cold and hot tumors, improving immunotherapy strategies. IFN-γ signaling and immune resistance mechanisms IFN-γ signaling within the TME also plays a complex role. Clinically, sufficient IFN-γ has been linked to positive responses to immunotherapy; however, both insufficient and excessive IFN-γ (by upregulating immune inhibitory molecules such as PD-L1, IDO1, and Arg1) signaling can lead to immune resistance [146, 162]. Enhancing IFN-γ signaling through combinations such as DNA methyltransferase inhibitors plus ICIs, or modulating IFN-γ receptor stability, may overcome resistance in CRC [163, 164]. Emerging and non-invasive biomarkers Additional emerging biomarkers include epigenetic and transcriptomic markers, microbiome profiles, circulating biomarkers such as ctDNA and blood cytokines, and clinical factors. These biomarkers are summarized in Tables 1 and 2; Fig. 2. Beyond single biomarkers, multi-marker scoring systems and AI-driven models are increasingly being explored to improve prediction of immunotherapy response in CRC [165]. The Immunoscore, based on the density and spatial distribution of CD3⁺ and CD8⁺ T cells in the tumor center and invasive margin, provides a standardized measure of antitumor immune contexture and may complement MSI/MMR and TMB status [166]. In parallel, artificial intelligence and machine-learning approaches can integrate histopathology, radiomics, genomics, transcriptomics, immune-cell infiltration, and liquid biopsy data to generate composite predictive models. These integrative tools may help identify immune-hot, immune-excluded, and immune-cold tumors more accurately than single markers alone [167]. However, clinical implementation still requires external validation, assay standardization, and prospective evaluation in CRC immunotherapy trials. The identification of reliable predictive biomarkers remains essential for optimizing immunotherapy outcomes in CRC. While MSI‑H/dMMR status currently represents the most clinically validated biomarker, additional factors such as TMB, immune infiltration patterns, and microbiome composition are increasingly recognized as contributors to treatment response. Continued investigation of these biomarkers may improve patient stratification and help overcome resistance to immunotherapeutic strategies. Overall, biomarker development in CRC is moving from single-marker assessment toward integrated models that combine MSI/MMR status, TMB, immune contexture, spatial organization, circulating biomarkers, and microbiome profiles. Table 2Positive biomarkers for CRC immunotherapyMarkerAssociation with the immune response Ref Genetic alterations MSI-H/DmmrHigh neoantigen load, inflamed TME, High TMB, Established predictive biomarker for ICIs.[6, 7]High TMBAssociated with neoantigen load and improved response to ICIs.[147, 148]CMS1Immune-activated with High inflammatory infiltrates and high MSI-H.[34]CIRC Group AUpregulation of immune checkpoint genes expression (CTLA-4, PD-L1, LAG-3, TIM3)[45]POLE MutationLinked to upregulated PD-L1; predictive of ICIs response.[168]POLD1 MutationInvolves DNA repair and is associated with MSI.[169]TP53/KRAS co-mutationAssociated with higher PD-L1 expression and T cell infiltration; linked to benefit from PD-1 inhibitors.[156]DNA damage response (DDR) gene mutationsEnhanced immune responses to tumor cells, Diversification of immune receptors[170]CMTM6 expressionPredicts the response rate to PD-1/PD-L1 inhibitors[171]High NF1 expressionRestrains overactive RAS function, correlates with enhanced effectiveness of EGFRi.[172]PIK3CA MutationAssociated with increased TMB, CD8 + T cell infiltration, and PD-L1 expression.[173]RNF43 MutationBetter antitumoral response to BRAFi and/or EGFRi[174] TME biomarkers Low NLRAssociated with a Better response to immunotherapy, including CAR-T therapy[175]TILsCorrelate with better responses to ICIs.[176]IFN-γ levels/signaturesSufficient IFN-γ promotes antigen presentation and T cell activity, leading to positive immunotherapy responses, while both low and high IFN-γ signaling can cause immune resistance.[162, 164, 177]High ISReflects cytotoxic T cell and memory T cell infiltration in the central and invasive margins (IMs) and correlates with better immunotherapy outcome[178]ILC3 infiltrationPromotes a more immune-hot TME in CRC, enhancing immunotherapy responses and supporting microbiota maintenance for effective antitumor immunity.[179] Epigenetic and transcriptomic biomarkers MGMT protein deficiencyEnhances adaptive immune reactions against tumors[180]immune-associated miRNA prognostic signature (IAMIPS)Patients classified as low-risk by IAMIPS have an immune-hot phenotype, TMB-H, and increased expression of immune checkpoints, associated with a better response to ICIs.[181] Circulating Biomarkers (Liquid Biopsy) ctDNA levelEarly changes in ctDNA levels during ICI therapy were strongly associated with long-term outcomes.[182]Th1 (CD126-CD4+) cellsUsing flow cytometry, a higher Th1 rate was observed in the blood samples of anti-PD-1 responding patients compared with non-responders.[183]Circulating arginineThere is a correlation between ARG levels and clinical ICI activity[184] Gut microbiome Fusobacterium nucleatumIncreased levels are associated with improved response to PD-L1 blockade, possibly via STING pathway activation and immune cell recruitment.[185]Bifidobacterium sppEnhancing anti-PD-L1 therapy response[186]Bacteroides fragilisPromotes the efficacy of anti-PD-L1 therapy.[187]Akkermansia muciniphilaEnhances the effectiveness of PD-1 blockade in non-responders by bringing CCR9 + CXCR3+CD4 + T cells[188]Lactobacillus caseiEnhances CD8 + TIL infiltration and cytotoxic cytokine secretion, boosting the efficacy of anti-PD-1 therapies.[189]MSI-H: Microsatellite Instability-High, dMMR: Deficient Mismatch Repair, TMB: Tumor Mutational Burden, ICIs: Immune Checkpoint Inhibitors, CMS1: Consensus Molecular Subtype 1, CIRC: Colorectal Cancer Intrinsic Subtype Classification, POLE: DNA Polymerase Epsilon, POLD1: DNA Polymerase Delta 1, TILs: Tumor-Infiltrating Lymphocytes, DDR: DNA Damage Response, Th1: T Helper Type 1, CMTM6: CKLF Like MARVEL Transmembrane Domain Containing 6, NF1: Neurofibromin 1, EGFRi: Epidermal Growth Factor Receptor Inhibitors, BRAFi: BRAF Inhibitors, CAR T: Chimeric Antigen Receptor T cell therapy, IFN-γ: Interferon Gamma, IS: Immunoscore, ILC3: Innate Lymphoid Cells type 3, MGMT: O6-Methylguanine-DNA Methyltransferase, IAMIPS: Immune-Associated MicroRNA Prognostic Signature, ctDNA: Circulating Tumor DNA, ARG: Arginine, PD-L1: Programmed Death-Ligand 1, PD-1: Programmed Death Protein 1, STING: Stimulator of Interferon Genes, CCR9: C-C Chemokine Receptor Type 9, CXCR3: C-X-C Chemokine Receptor Type 3, CD8+: Cluster of Differentiation 8 positive (cytotoxic T cells), CD4+: Cluster of Differentiation 4 positive (helper T cells) Positive biomarkers for CRC immunotherapy MSI-H: Microsatellite Instability-High, dMMR: Deficient Mismatch Repair, TMB: Tumor Mutational Burden, ICIs: Immune Checkpoint Inhibitors, CMS1: Consensus Molecular Subtype 1, CIRC: Colorectal Cancer Intrinsic Subtype Classification, POLE: DNA Polymerase Epsilon, POLD1: DNA Polymerase Delta 1, TILs: Tumor-Infiltrating Lymphocytes, DDR: DNA Damage Response, Th1: T Helper Type 1, CMTM6: CKLF Like MARVEL Transmembrane Domain Containing 6, NF1: Neurofibromin 1, EGFRi: Epidermal Growth Factor Receptor Inhibitors, BRAFi: BRAF Inhibitors, CAR T: Chimeric Antigen Receptor T cell therapy, IFN-γ: Interferon Gamma, IS: Immunoscore, ILC3: Innate Lymphoid Cells type 3, MGMT: O6-Methylguanine-DNA Methyltransferase, IAMIPS: Immune-Associated MicroRNA Prognostic Signature, ctDNA: Circulating Tumor DNA, ARG: Arginine, PD-L1: Programmed Death-Ligand 1, PD-1: Programmed Death Protein 1, STING: Stimulator of Interferon Genes, CCR9: C-C Chemokine Receptor Type 9, CXCR3: C-X-C Chemokine Receptor Type 3, CD8+: Cluster of Differentiation 8 positive (cytotoxic T cells), CD4+: Cluster of Differentiation 4 positive (helper T cells) Factors shaping response and resistance to CRC immunotherapy. Central schema summarizing four major determinant domains: immune checkpoint molecules, genetic and molecular biomarkers, tumor microenvironment, and tumor mutational burden (TMB), and how they modulate outcomes to immune checkpoint inhibitor (ICI) therapy in colorectal cancer (CRC). Top panels illustrate favorable contexts: (1) effective PD-1/PD-L1 blockade restoring T-cell activation; (2) MSI-H/dMMR tumors generating abundant neoantigens and CD8⁺ T-cell infiltration, promoting tumor cell death. Right panels depict resistance mechanisms: (3) impaired antigen presentation (↓MHC-I), MSS status, and oncogenic WNT/β-catenin and TGF-β signaling driving immune exclusion and proliferation; (4) an immunosuppressive TME enriched for M2 macrophages and arginase-1 with high tumor PD-L1, limiting TIL recruitment and fostering resistance. Left/bottom panels show additional modifiers: (5) persistent PD-1 signaling causing T-cell exhaustion; (6) low TMB together with IDO and TGF-β activity favoring proliferation; (7) high TMB/oncogenic mutations (e.g., KRAS, BRAF) increasing neoantigen load and sensitivity to ICIs; (8) a reprogrammed TME with sufficient IFN-γ, decreased tumor PD-L1, and abundant Th1/CTL-type TILs overcoming resistance. Green arrows denote factors/contexts associated with improved response; red arrows indicate drivers of resistance. Abbreviations: BRAF: B-Raf proto-oncogene, CD8+: CD8-positive T cell, CRC: colorectal cancer, CTL: cytotoxic T lymphocyte, dMMR: deficient mismatch repair, IDO: indoleamine-2,3-dioxygenase, ICI: immune checkpoint inhibitor, IFN-γ: interferon-gamma, KRAS: Kirsten rat sarcoma viral oncogene homolog, MHC-I: major histocompatibility complex class I, M2: M2 macrophage (alternatively activated), MMR: mismatch repair, MSI-H: microsatellite instability–high, MSS: microsatellite stable, PD-1: programmed cell death protein-1, PD-L1: programmed death-ligand 1, TGF-β: transforming growth factor-beta, Th1: T helper type 1 cell, TILs: tumor-infiltrating lymphocytes, TMB: tumor mutational burden, TME: tumor microenvironment, WNT/β-catenin: WNT signaling pathway involving β-catenin Table 3Negative biomarkers for CRC immunotherapyMarkerAssociation with immune responseRefGenetic alterationsMSS/pMMRLow neoantigen load, cold tumor microenvironment[190]CMS2Immune desert or cold tumor, classical type[34]CMS3Immune desert or cold tumor, classical type[34]CMS4Immunosuppressive phenotype[34]MDM4 AmplificationAssociated with immune evasion and poor response to therapy[191]EGFR mutationBiomarkers for resistance to anti-EGFR[192]RAS mutationBiomarker of resistance to anti-EGFR[192]ALK MutationOften associated with microsatellite stable (MSS) status, they exhibit resistance not only to anti-EGFR therapies but also show limited benefit from immunotherapy, possibly due to an immunosuppressive tumor microenvironment[193]MET RearrangementAssociated with tumor metastasis and immune escape mechanisms in CRC by inducing acidic microenvironments, generally linked to a more aggressive phenotype and resistance to immunotherapy[194]STK11/LKB1 expressionLow STK11 expression is linked with worse overall survival, and without direct correlation with lymphocyte infiltration, it contributes to an immune-cold microenvironment and decreased ICB efficacy in colorectal adenocarcinoma, especially in patients with KRAS mutations[195]PTEN LossPotential predictor of resistance to cetuximab[196]B2M MUTATIONProbable deficient response to ICIs[197]JAK1/2 MUTATIONFails to respond to anti-PD-1[198]Nf1 MUTATIONResults in resistance.predictor of response to anti-EGFR targeting antibodies for RAS and BRAF V600 wild-type tumors[199]Epigenetic and transcriptomic biomarkersDNMT3A AlterationInfluences epigenetic regulation DAB2IP transcription and consequently activates the MEK/ERKnd PI3K/Akt signaling cascades [200]Circulating biomarkers (Liquid Biopsy)DKK1 (High Serum Expression)Inhibits the antitumor immune reaction through the GSK3β/E2F1/T-bet axis in CD8 + T cells, predicting poor tumor response to PD-1 blockade in dMMR/MSI CRCs. [201]IL-17Promotes resistance to ICIs by upregulating PD-L1 expression through pathways like p65/NRF1/miR-15b-5p, contributing to immune escape [202]LAGE3 (High Expression)Correlates with poor immune infiltration in CRC, and an immunosuppressive TME that is less responsive ICIs [203]High Circulating LDH LevelsContributes to an immunosuppressive TME and poorer cancer immunotherapy outcomes via increased tumor glycolysis, hypoxia, and burden. [204]APRILNegatively relates to immune cell infiltration and expression of CD163 in CRC, suggesting a less favorable TME for effective immunotherapy. [205]BAFFNegatively related to immune cell infiltration and expression of CD163 in CRC, suggesting a less favorable TME for effective immunotherapy. [205]IL-8Negatively related to immune cell infiltration and expression of CD163 in CRC, suggesting a less favorable TME for effective immunotherapy [205]MMP2Negatively related to immune cell infiltration and expression of CD163 in CRC, suggesting a less favorable TME for effective immunotherapy [205]TME biomarkersImmune ExhaustionInvolves CD4 + and CD8 + CD3+ T cells, B Lymphocytes, and Conventional NK Cells and correlates with resistance to immunotherapies [206, 207]CAFsAssociated with immunosuppression and ICIs resistance [208]Increased Circulating Tsens (Senescent CD4 + T cells)Exhibits impaired proliferation and function, leading to poor anti-tumor response and lower immunotherapy effectiveness [209]TGF-β Signaling ActivationLinked to resistance to ICIs like anti-PD-1/PD-L1 [210]VEGF-AInfluences T cell exhaustion by inducing the expression of the transcription factor TOX in T cells in anti-PD-1-resistant MSS CRC [211]Clinical factorsMalignant Ascites and Peritoneal MetastasesPatients with dMMR/MSI-H CRC with peritoneal metastases and ascites should be considered as a peculiar subgroup with highly unfavorable outcomes to current ICI-based therapies [212]Antibiotic UseAdministration of broad-spectrum antibiotics harms ICI treatment outcome by inducing intestinal dysbiosis, altering the normal gut microbiota [213]Microsatellite Stable, pMMR: Proficient Mismatch Repair, CMS: Consensus Molecular Subtype, MDM2/4: Mouse Double Minute 2/4 (gene amplification), EGFR: Epidermal Growth Factor Receptor, RAS: Rat Sarcoma (gene family), ALK: Anaplastic Lymphoma Kinase, MET: MET Proto-Oncogene, Receptor Tyrosine Kinase, STK11/LKB1: Serine/Threonine Kinase 11/Liver Kinase B1, PTEN: Phosphatase and Tensin Homolog, B2M: Beta-2-Microglobulin, JAK1: Janus Kinase 1, Nf1: Neurofibromin 1, DNMT3A: DNA Methyltransferase 3 Alpha, DKK1: Dickkopf WNT Signaling Pathway Inhibitor 1, IL-17: Interleukin 17, LAGE3: Leukocyte Antigen 3, LDH: Lactate Dehydrogenase, APRIL: A Proliferation-Inducing Ligand, BAFF: B-cell Activating Factor, IL-8: Interleukin 8, MMP2: Matrix Metallopeptidase 2, TME: Tumor Microenvironment, CAFs: Cancer-Associated Fibroblasts, Tsens: Senescent T cells, TGF-β: Transforming Growth Factor Beta, VEGF-A: Vascular Endothelial Growth Factor A, ICI: Immune Checkpoint Inhibitor, ICIs: Immune Checkpoint Inhibitors, TMB-H: Tumor Mutational Burden-High, MSI-H: Microsatellite Instability-High, TOX: Thymocyte selection-associated high mobility group box protein Negative biomarkers for CRC immunotherapy Results in resistance. predictor of response to anti-EGFR targeting antibodies for RAS and BRAF V600 wild-type tumors Microsatellite Stable, pMMR: Proficient Mismatch Repair, CMS: Consensus Molecular Subtype, MDM2/4: Mouse Double Minute 2/4 (gene amplification), EGFR: Epidermal Growth Factor Receptor, RAS: Rat Sarcoma (gene family), ALK: Anaplastic Lymphoma Kinase, MET: MET Proto-Oncogene, Receptor Tyrosine Kinase, STK11/LKB1: Serine/Threonine Kinase 11/Liver Kinase B1, PTEN: Phosphatase and Tensin Homolog, B2M: Beta-2-Microglobulin, JAK1: Janus Kinase 1, Nf1: Neurofibromin 1, DNMT3A: DNA Methyltransferase 3 Alpha, DKK1: Dickkopf WNT Signaling Pathway Inhibitor 1, IL-17: Interleukin 17, LAGE3: Leukocyte Antigen 3, LDH: Lactate Dehydrogenase, APRIL: A Proliferation-Inducing Ligand, BAFF: B-cell Activating Factor, IL-8: Interleukin 8, MMP2: Matrix Metallopeptidase 2, TME: Tumor Microenvironment, CAFs: Cancer-Associated Fibroblasts, Tsens: Senescent T cells, TGF-β: Transforming Growth Factor Beta, VEGF-A: Vascular Endothelial Growth Factor A, ICI: Immune Checkpoint Inhibitor, ICIs: Immune Checkpoint Inhibitors, TMB-H: Tumor Mutational Burden-High, MSI-H: Microsatellite Instability-High, TOX: Thymocyte selection-associated high mobility group box protein Future directions and clinical implications Overcoming resistance to immunotherapy in CRC has emerged as a critical challenge in improving patient outcomes. The future directions in this field will harness advances in rational combination strategies, next-generation immunotherapeutic platforms, and biomarker-guided precision approaches aimed at broadening the efficacy of immunotherapy beyond currently responsive subgroups. Combination strategies to enhance immunotherapy response One promising avenue for this purpose lies in the integration of immunotherapy with traditional modalities such as chemotherapy, radiotherapy, and targeted therapies. Various studies have explored the association of chemotherapy and immunotherapy, showing enhanced immunotherapy efficacy [43]. However, side effects of chemotherapy can be intensified by immunotherapy, necessitating careful monitoring and dosage adjustments to optimize the advantages of combination therapy [214]. Moreover, the combination of ICIs and chemotherapy in MSS mCRC still has not produced clinically significant outcomes, limiting its current application in clinical practice [215]. Combining radiotherapy with immunotherapy is seen as a promising strategy to enhance responses to immunotherapy. It has also emerged as a complementary approach to ICI therapies in CRC, particularly in MSS tumors [216]. Various radiotherapy techniques have been explored, including stereotactic body radiotherapy (SBRT), low-dose radiotherapy (LDRT), spatially fractionated radiotherapy (SFRT), emerging ultra-high-dose-rate radiotherapy (FLASH-RT), and combination strategies such as LDRT plus SBRT or SFRT combined with FLASH-RT displaying synergistic efficacy, hinting at new mechanistic insights and therapeutic potential for radio-immunotherapy [217]. However, challenges like patient response variability and the need for predictive biomarkers persist [217]. The combination of immunotherapy with targeted therapies represents another promising avenue. Agents such as MEK inhibitors, tyrosine kinase inhibitors (TKIs), and aldehyde dehydrogenase (ALDH)-targeting compounds may enhance tumor immunogenicity and sensitize resistant tumors to immunotherapeutic interventions [218–220]. Integrative and multimodal immunotherapeutic approaches Integration of various immunotherapeutic modalities together presents a promising approach to overcoming resistance. For example, combining ICIs with antibody-based therapies, including EGFR-targeted antibodies can improve tumor antigen recognition and immune cell recruitment [70]. Dual ICIs combination therapy has shown better immunotherapy responses in clinical trials; however, some of them have demonstrated unacceptable toxicities or absence of response (e.g., durvalumab plus tremelimumab) [221]. Additionally, combinations of ICIs with CAR-T therapies may synergize to enhance anti-tumor immune responses [222]. Combining OVs therapy with other immunotherapies, including ICIs and CAR-T therapy, demonstrated significant potential [223]. Ongoing Clinical trial NCT05456165 is investigating the synergistic effects of combining neoantigen vaccines with ICIs (e.g., PD-1/PD-L1 inhibitors) to overcome resistance and improve response rate [224]. Moreover, combination therapy using Immunoadjuvants (e.g., Manganese Chloride) as a stimulator of humoral and cellular immune responses is under investigation [225]. From a clinical perspective, rational combination therapy should be selected according to the dominant resistance mechanism rather than applied empirically. Tumors with preserved antigen presentation but dominant checkpoint-mediated exhaustion may benefit most from ICIs or next-generation checkpoint combinations, whereas tumors with impaired MHC-I-restricted recognition may require MHC-independent strategies such as CAR-based therapies or BiTEs. In contrast, immune-desert or stromal-excluded MSS tumors may require priming or remodeling approaches, including radiotherapy, OVs, vaccines, STING/TLR agonists, cytokine-based agents, anti-angiogenic or anti-TGF-β combinations, and microbiome modulation. This resistance-matched approach may help avoid ineffective combinations, improve patient selection, and support biomarker-driven trial design in CRC [8, 47]. Next-generation cellular and immune-targeted therapeutics Other than combination therapy, efforts focus on emerging and improving novel immunotherapeutics. Emerging immune checkpoint targets such as LAG-3, TIM-3, TIGIT, and VISTA represent promising candidates for next-generation checkpoint blockade strategies [226]. Enhancing the immune response to CAR-T by improving targeting specificity, persistence, and engineering dual targeting CAR-T cells is another field of research that is gaining attention for further investigation [99, 100]. dual-targeted and armored CAR-T cells, such as those targeting both HER2 and HLA-A02, show promise in reducing off-tumor toxicity and increasing specificity in other solid tumors [227]. Additionally, Current studies are exploring Alternative Cell Types and Off-the-Shelf Therapies, such as NK cells and macrophage-based CAR therapies, as complementary approaches [228]. Moreover, minimizing TAM infiltration, altering their polarization, exploring neutralizing TAM methods, and combinatorial therapeutic strategies offer promising new avenues for therapy [229]. Biomarker-driven precision immuno-oncology In addition to using the methods mentioned above, researchers have paid attention to the development of robust biomarkers predicting response and resistance to immunotherapy as a critical future direction. Advanced multi-omics analyses, including genomics, transcriptomics, proteomics, and metabolomics, combined with artificial intelligence, can identify novel biomarkers linked to CRC molecular subtypes and the TME. These biomarkers can stratify patients for optimal and personalized therapies and enable real-time monitoring of treatment response and clonal evolution [230]. Liquid biopsies, including circulating tumor DNA (ctDNA), circulating miRNAs, and immune cell profiling, are emerging as minimally invasive tools for dynamic biomarkers [231]. Integrating biomarker-driven precision medicine into clinical trials and routine practice will improve patient selection, avoid unnecessary toxicities, and guide adaptive treatment strategies (Table 4). Table 4Clinical trials with immunotherapy in CRCTrial IDphasestatusEnrollmentinterventionArmsoutcomeRefNCT02380443Phase 2Completed13 patients with mCRCBiological: AlloStimProcedure: CryoablationDosing Schedule A (with cryoablation)OS = 97.0 days [232]Dosing Schedule B without cryoablationOS = 368.0 daysNCT03391232Phase 1/Phase 2Completed11 patients with CRC without MSI-H and BRAF mutationsBiological: PolyPEPI1018 CRC VaccineArm 1: Single Dose of PolyPEPI1018 CRC VaccineBoth CD4 + and CD8 + T cell responses in 4.0 ParticipantsCD4 + T cell responses in 4.0 ParticipantsCD8 + T cell responses 4.0 ParticipantsCD8 + T cell responses against three or more AG in 4.0 Participants [233]Arm 2: Multiple Doses of PolyPEPI1018 CRC VaccineBoth CD4 + and CD8 + T cell responses in5.0 ParticipantsCD4 + T cell responses in 6.0 ParticipantsCD8 + T cell responses 5.0 Participants CD8 + T cell responses against three or more AG 4.0 ParticipantsNCT03414983Phase 2/Phase 3Completed196 patients with mCRCBiological: NivolumabDrug: Oxaliplatin Drug: LeucovorinDrug: Fluorouracil Drug: BevacizumabArm A: NIV+mFOLFOX + BEVPFS per BICR = 11.86 MonthsORR per BICR = 60.6%DOR per BICR = 12.88 MonthsOS = 30.52 MonthsInvestigator-assessed ORR/PFS/DOR were similar. [234]Arm B: mFOLFOX + BEVPFS per BICR = 11.93 MonthsORR per BICR = 45.6%DOR BICR = 9.26 MonthsOS = 31.77MonthsInvestigator-assessed ORR/PFS/DOR were similar.NCT05061017Phase 2Completed13 patients with MSS mCRCDrug: PixatimodDrug: NivolumabDrug: Cyclophosphamide (low dose)Cohort 1: Pixatimod (PG545) + Nivolumab1-year OS = NRPD = 62.5%SD = 37.5% [235]Cohort 2 & Cohort 3: Pixatimod (PG545) + Nivolumab + Cyclophosphamide1-year OS = 19.0%PD = 60.0%SD = 40.0%NCT00103142Phase 2Completed74 patients with confirmed hepatic or pulmonary metastases secondary to adenocarcinoma of the colon and rectumBiological: FalimarevBiological: inalimarevBiological:SargramostimBiological:therapeutic autologous dendritic cellsArm 1: PANVAC-V + PANVAC-F + DCRecurrence-free Survival at 2 Years = 13.0participantsPositive ELISpot = 9.0participants [236]Arm 2: PANVAC-V + PANVAC-F + GM-CSFRecurrence-free Survival at 2 Years = 10.0ParticipantsPositive ELISpot = 4.0participantsNCT03642067Phase 2Completed59 patients with MSS Advanced CRCDrug: NivolumabDrug: RelatlimabCohort A: Composite PD-L1/Mucin (CPM) positive colorectal cancerORR = NR [237]Cohort B: Composite PD-L1/Mucin (CPM) negative colorectal cancerORR = 1.0ParticipantsCohort C: Colorectal cancer with no biomarker evaluation requiredORR = 2.0ParticipantsNCT03436563Phase 1/Phase 2Completed4 patients with MSI-H mCRCBiological: Anti-PD-L1/TGFbetaRII Fusion Protein M7824-OS = 4 PARTICIPANTSDFS = 3 PARTICIPANTSGrade 3 or Higher A/E Per CTCAE v4.03 = NR [238]NCT02834052Phase 1/Phase 2Completed42 patients with MRP mCRCDrug: pembrolizumabDrug: Poly-ICLCPhase 1: Poly-ICLC: 1 mgORR = NROS = NRDLT = 87.5%DOR = NR [239]Phase 1: Poly-ICLC: 2 mgORR = NROS of = NRDLT = 70.0%DOR = NRPhase 2: Poly-ICLC: 2 mgORR = 1.0 ParticipantsOS = NRDLT = 70.83%DOR = NRNCT03711058Phase 1/Phase 2Active, not recruiting48 patients with MSS CRCDrug: CopanlisibDrug: NivolumabPhase 1: Copanlisib and Nivolumab (De-Escalation)DLT = NRORR = NRDCR = NRDOR = NRPFS = NROS = NR [240]Phase 2/Arm A: P13K mutation/Copanlisib and NivolumabORR = 2.0 ParticipantsDCR = 5.0ParticipantsDOR = NRPFS = NROS = NRPhase 2/Arm B: P13K wild type/Copanlisib and NivolumabORR = NRDCR = 3.0ParticipantsDOR = NRPFS = NROS = NRNCT04491955Phase 2Completed32 patients with refractory/recurrent small bowel and CRCBiological: CV301Drug: MSB0011359CDrug: N-803Drug: NHS-IL12Arm 1: CEA/MUC1 Vaccines + M7824 + N-803 (Triple Therapy).(Quadruple Therapy)ORR = NRPFS = 1.8MonthsOS = 19.0MonthsDOR = NRPD = 1.0ParticipantsSerious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) = 12.0Participants [241]Arm 2 A: CEA/MUC1 Vaccines + M7824 + N-803 + NHSIL12ORR of DL 1 = 12.5ORR of DL 2 = NRPFS of DL 1 = 1.8 MonthsPFS of DL 2 = 1.8 MonthsOS of DL 1 = 19.0 MonthsOS of DL 2 = 4.5 MonthsDOR of DL 1 = NRDOR of DL 2 = 34.0 MonthsPD of DL 1 = 1.0 ParticipantsPD of DL 2 = NRSerious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) of DL 1 = 8.0ParticipantsSerious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) of DL 2 = 6.0ParticipantsArm 2B: CEA/MUC1 Vaccines + M7824 + N-803 + NHSIL12ORR = NRPFS = 0.92MonthsOS = 7.0 MonthsDOR = NRPD = NRSerious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) = 6.0ParticipantsNCT02981524Phase 2Completed17 patients with MMR-p Advanced CRCDrug: CYBiological: GVAXDrug: Pembrolizumab-OS = 213.0DaysORR = NRPFS = 82.0DaysNumber of Participants with Adverse Events = 6.0ParticipantsDOR = NR [242]NCT00773097Phase 2Completed46 patients with advanced colorectal adenomaBiological: MUC1 - Poly ICLC-Number of Participants with Anti Muc-1 Antibody = 39.0Number of Participants with Autoimmune Response to Muc-1 Vaccine = 39.0Number of Participants with Adverse Events Associated With the Study Agent = 39.0 [243]NCT03206073Phase 1/Phase 2Completed34 patients with refractory CRCDrug: DurvalumabDrug: TremelimumabBiological: Pexa-VecBiological: Pexa-VecArm A1: Pexa-Vec + DurvalumabOS = 6.2Months [244]Arm A2: Pexa-Vec +DurvalumabOS = 8.6MonthsArm B: Pexa-Vec + Durvalumab +TremelimumabOS B1 = 6.0MonthsOS B2 = 5.2MonthsNCT01394939Phase 1/Phase 2Completed52 patients with Metastatic, Refractory Colorectal CarcinomaBiological: JX-594Drug: IrinotecanSingle Agent_ Cohort 1: JX-594 administeredDetermine Radiographic Response Rate of Patients Enrolled in the Phase 2a Portion of the Study = NR [245]Single Agent_Cohort: 2 JX-594 administeredDetermine Radiographic Response Rate of Patients Enrolled in the Phase 2a Portion of the Study = NRCombination_Cohort 3: JX-594 administered in combination with IrinotecanDetermine Radiographic Response Rate of Patients Enrolled in the Phase 2a Portion of the Study = NRCombination_Cohort 4: JX-594 administered intravenously weekly in combination with IrinotecanDetermine Radiographic Response Rate of Patients Enrolled in the Phase 2a Portion of the Study = NRmCRC: metastatic Colorectal Cancer, MSI-H: Microsatellite Instability-High, OS: Overall Survival, PFS: Progression-Free Survival, ORR: Objective Response Rate, DOR: Duration of Response, BICR: Blinded Independent Central Review, BEV: Bevacizumab, MSS: Microsatellite Stable, PD: Progressive Disease, SD: Stable Disease, DFS: Disease-Free Survival, A/E: Adverse Events, CTCAE: Common Terminology Criteria for Adverse Events, ELISpot: Enzyme-Linked ImmunoSpot, MUC: Mucin, PD-L1: Programmed Death-Ligand 1, TGFbetaRII: Transforming Growth Factor beta Receptor II, DLT: Dose-Limiting Toxicity, GVAX: Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF) gene-transfected tumor vaccine, CY: Cyclophosphamide, GM-CSF: Granulocyte-Macrophage Colony-Stimulating Factor, Pexa-Vec: Pexastimogene Devacirepvec (oncolytic virus therapy), JX-594: A type of oncolytic poxvirus (same as Pexa-Vec), NR: Not Reached or Not Reported, CEA: Carcinoembryonic Antigen (tumor marker), NHS-IL12: necrosis-targeted IL-12 Clinical trials with immunotherapy in CRC Biological: AlloStim Procedure: Cryoablation Both CD4 + and CD8 + T cell responses in 4.0 Participants CD4 + T cell responses in 4.0 Participants CD8 + T cell responses 4.0 Participants CD8 + T cell responses against three or more AG in 4.0 Participants Both CD4 + and CD8 + T cell responses in 5.0 Participants CD4 + T cell responses in 6.0 Participants CD8 + T cell responses 5.0 Participants CD8 + T cell responses against three or more AG 4.0 Participants Biological: Nivolumab Drug: Oxaliplatin Drug: Leucovorin Drug: Fluorouracil Drug: Bevacizumab PFS per BICR = 11.86 Months ORR per BICR = 60.6% DOR per BICR = 12.88 Months OS = 30.52 Months Investigator-assessed ORR/PFS/DOR were similar. PFS per BICR = 11.93 Months ORR per BICR = 45.6% DOR BICR = 9.26 Months OS = 31.77Months Investigator-assessed ORR/PFS/DOR were similar. Drug: Pixatimod Drug: Nivolumab Drug: Cyclophosphamide (low dose) 1-year OS = NR PD = 62.5% SD = 37.5% 1-year OS = 19.0% PD = 60.0% SD = 40.0% Biological: Falimarev Biological: inalimarev Biological: Sargramostim Biological: therapeutic autologous dendritic cells Recurrence-free Survival at 2 Years = 13.0 participants Positive ELISpot = 9.0 participants Recurrence-free Survival at 2 Years = 10.0 Participants Positive ELISpot = 4.0 participants Drug: Nivolumab Drug: Relatlimab ORR = 1.0 Participants ORR = 2.0 Participants OS = 4 PARTICIPANTS DFS = 3 PARTICIPANTS Grade 3 or Higher A/E Per CTCAE v4.03 = NR Drug: pembrolizumab Drug: Poly-ICLC ORR = NR OS = NR DLT = 87.5% DOR = NR ORR = NR OS of = NR DLT = 70.0% DOR = NR ORR = 1.0 Participants OS = NR DLT = 70.83% DOR = NR Drug: Copanlisib Drug: Nivolumab DLT = NR ORR = NR DCR = NR DOR = NR PFS = NR OS = NR ORR = 2.0 Participants DCR = 5.0 Participants DOR = NR PFS = NR OS = NR ORR = NR DCR = 3.0 Participants DOR = NR PFS = NR OS = NR Biological: CV301 Drug: MSB0011359C Drug: N-803 Drug: NHS-IL12 Arm 1: CEA/MUC1 Vaccines + M7824 + N-803 (Triple Therapy). (Quadruple Therapy) ORR = NR PFS = 1.8 Months OS = 19.0 Months DOR = NR PD = 1.0 Participants Serious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) = 12.0 Participants ORR of DL 1 = 12.5 ORR of DL 2 = NR PFS of DL 1 = 1.8 Months PFS of DL 2 = 1.8 Months OS of DL 1 = 19.0 Months OS of DL 2 = 4.5 Months DOR of DL 1 = NR DOR of DL 2 = 34.0 Months PD of DL 1 = 1.0 Participants PD of DL 2 = NR Serious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) of DL 1 = 8.0 Participants Serious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) of DL 2 = 6.0 Participants ORR = NR PFS = 0.92 Months OS = 7.0 Months DOR = NR PD = NR Serious and/or Non-serious Adverse Events Assessed by (CTCAE v5.0) = 6.0 Participants Drug: CY Biological: GVAX Drug: Pembrolizumab OS = 213.0 Days ORR = NR PFS = 82.0 Days Number of Participants with Adverse Events = 6.0 Participants DOR = NR Number of Participants with Anti Muc-1 Antibody = 39.0 Number of Participants with Autoimmune Response to Muc-1 Vaccine = 39.0 Number of Participants with Adverse Events Associated With the Study Agent = 39.0 Phase 1 /Phase 2 Drug: Durvalumab Drug: Tremelimumab Biological: Pexa-Vec Biological: Pexa-Vec OS = 6.2 Months OS = 8.6 Months OS B1 = 6.0 Months OS B2 = 5.2 Months Phase 1 /Phase 2 Biological: JX-594 Drug: Irinotecan mCRC: metastatic Colorectal Cancer, MSI-H: Microsatellite Instability-High, OS: Overall Survival, PFS: Progression-Free Survival, ORR: Objective Response Rate, DOR: Duration of Response, BICR: Blinded Independent Central Review, BEV: Bevacizumab, MSS: Microsatellite Stable, PD: Progressive Disease, SD: Stable Disease, DFS: Disease-Free Survival, A/E: Adverse Events, CTCAE: Common Terminology Criteria for Adverse Events, ELISpot: Enzyme-Linked ImmunoSpot, MUC: Mucin, PD-L1: Programmed Death-Ligand 1, TGFbetaRII: Transforming Growth Factor beta Receptor II, DLT: Dose-Limiting Toxicity, GVAX: Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF) gene-transfected tumor vaccine, CY: Cyclophosphamide, GM-CSF: Granulocyte-Macrophage Colony-Stimulating Factor, Pexa-Vec: Pexastimogene Devacirepvec (oncolytic virus therapy), JX-594: A type of oncolytic poxvirus (same as Pexa-Vec), NR: Not Reached or Not Reported, CEA: Carcinoembryonic Antigen (tumor marker), NHS-IL12: necrosis-targeted IL-12 Taken together, these therapeutic approaches highlight both the remarkable progress and the persistent challenges in the field. Despite significant clinical advances, issues such as treatment resistance, safety concerns, and patient-specific variability remain unresolved. These considerations underscore the need for continued investigation, which is addressed in the concluding section. Conclusions CRC is a heterogeneous disease defined by distinct molecular subtypes that strongly influence prognosis and therapeutic outcomes. A critical distinction exists between MSI-H/dMMR and MSS/pMMR tumors, as it dictates the differential efficacy of immunotherapies. ICIs have revolutionized the management of MSI-H/dMMR CRC by producing durable clinical responses. However, the far larger group of MSS/pMMR patients continues to exhibit resistance, underscoring the urgent need for innovative and combinatorial strategies. Multiple immunotherapeutic modalities are currently under investigation. These include checkpoint blockade beyond PD-1 and CTLA-4 pathways, adoptive cell therapies such as CAR-T cells, bispecific antibody platforms, and cancer vaccines designed to augment antitumor immunity. OVs and microbiome-modulating therapies further expand the immunotherapy landscape by reshaping the TME and enhancing T-cell infiltration. In parallel, engineered cytokines, immunocytokines, and adjuvant approaches such as TLR and STING agonists are emerging as potent adjuncts capable of converting immune-resistant CRC into an immunologically active state. The integration of biomarkers is equally vital. Tumor mutational burden, MSI/MMR status, immune-related gene signatures, and gut microbiota composition represent some of the most promising predictors of response and resistance. Refining biomarker-driven patient stratification will be essential to optimizing therapeutic decisions and guide rational combinations. Future clinical trials are poised to define the next era of CRC immunotherapy, emphasizing combinatorial regimens, earlier application in neoadjuvant and adjuvant settings, and personalized approaches informed by molecular and immune profiling. In summary, CRC immunotherapy is progressing beyond single-agent checkpoint blockade toward multidimensional, biomarker-guided strategies. Long-term success will rely on the integration of novel therapeutic platforms, validation of predictive biomarkers, and translation of clinical trial findings into durable benefits across all CRC subtypes. Acknowledgements None. Author contributions Mahdi Jafarlou: Conceptualization, Supervision, Writing – original draft, Writing – review & editing. **Vahid Jafarlou: ** Investigation, Writing – review & editing. **Samaneh Nouri: ** Investigation, Writing – review & editing. **Nazila Alizadeh: ** Writing – original draft. **Zahra Azimzadeh Tabrizi: ** Visualization **. Mahsa Iravani: ** Investigation, Methodology, Validation. **Farid Ghorbaninezhad: ** Investigation, Methodology, Writing – review & editing. **Deniz Shakiba: ** Writing – review & editing. **Behzad Baradaran: ** Validation. Data availability The data that support the findings of this study are available on request from the corresponding author. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Samaneh Nouri and Nazila Alizadeh contributed equally. References Associated Data Data Availability Statement The data that support the findings of this study are available on request from the corresponding author.