Long non-coding RNAs in the tumor immune microenvironment of non-small cell lung cancer: mechanisms and clinical translational perspectives Authors: Weiwei Wang, Zhibiao Jiang, Kun Zhao, Yuanyin Teng, Lingling Wu, Ting Yu, Anquan Shang Published: 2025 ABSTRACT AbstractBackgroundImmune checkpoint inhibitors (ICIs) have improved outcomes for patients with non-small cell lung cancer (NSCLC), yet their efficacy is hindered by the immunosuppressive and therapy-resistant tumor immune microenvironment (TIME). Long non-coding RNAs (lncRNAs) have emerged as critical regulators of immune signaling, ferroptosis, and cellular interactions within the TIME, highlighting their potential roles in shaping ICI response.Main bodyLncRNAs modulate multiple layers of antitumor immunity in NSCLC. They regulate immune evasion through PD-1/PD-L1 signaling (SNHG12, SChLAP1, PCAT1, NEAT1), influence T-cell survival and differentiation (NKILA, MIR17HG), and affect macrophage phagocytic checkpoints (KCTD21–AS1–CD47). LncRNAs also integrate ferroptosis, stemness, and metabolic rewiring, as exemplified by RGMB-AS1 and AC026356.1. Several lncRNAs—including ADAMTS9–AS2, PCBP1–AS1, LY6K-AS, FAM207BP, and LINC01833—exhibit prognostic or predictive significance for immunotherapy. Exosomal lncRNAs further offer minimally invasive tools for liquid biopsy and longitudinal monitoring. Translational opportunities include nucleic acid therapeutics, innovative delivery systems, and rational combinations of ICIs with ferroptosis inducers or metabolic modulators. Multi-omics and spatial transcriptomics are essential to clarify mechanisms and identify high-confidence clinical candidates.ConclusionsLncRNAs are central regulators of NSCLC antitumor immunity and determinants of ICI sensitivity. Their dual potential as biomarkers and therapeutic targets underscores the need for mechanistic validation, multi-omics integration, and prospective clinical trials to accelerate their translation into NSCLC immunotherapy.Keywords: Long non-coding RNA, Non-small cell lung cancer, Tumor immune microenvironment, Immune checkpoint inhibitors, Ferroptosis Introduction Lung cancer remains the leading cause of cancer incidence and mortality worldwide. According to the latest IARC GLOBOCAN 2022 estimates, there were approximately 2.5 million new lung cancer cases in 2022 (accounting for 12.4% of all new cancers) and 1.8 million deaths (18.7% of all cancer deaths), underscoring both the immense public health burden and the regional disparities [1]. In particular, many Asian countries continue to face a disproportionately high burden due to the persistent prevalence of smoking and exposure to air pollution [1, 2]. Histologically, non-small cell lung cancer (NSCLC) represents~85% of all lung cancers, comprising subtypes such as adenocarcinoma, squamous cell carcinoma, and large-cell carcinoma, a distribution that has remained stable in recent epidemiological and clinical studies [3]. In recent years, ICIs have transformed the therapeutic landscape of NSCLC. In metastatic disease, anti- PD-1/PD-L1 monotherapy or combinations with chemotherapy have significantly improved both overall survival and progression-free survival. In resectable disease, perioperative immunotherapy-such as neoadjuvant nivolumab plus chemotherapy in the CheckMate-816 trial-has raised pathological complete response (pCR) and event-free survival (EFS), driving the treatment paradigm into earlier-stage NSCLC [4, 5]. For patients with unresectable stage III disease, immunotherapy-based combination strategies are also under active investigation, though results indicate that therapeutic “add-ons” are not universally beneficial. Despite these advances, durable responses and objective response rates remain limited, with substantial primary and acquired resistance. Low tumor immunogenicity, an immunosuppressive tumor microenvironment, and genomic alterations such as STK11/KEAP1 and EGFR, often in the context of low tumor mutational burden (TMB), are associated with poor benefit from ICIs. In real-world settings, after progression on first-line chemoimmunotherapy, only about one-third to fewer than half of patients are eligible for subsequent therapies. These challenges highlight the urgent need to deepen our understanding of, and to therapeutically remodel, the tumor immune microenvironment (TIME) [6, 7]. Within this framework, long non-coding RNAs (lncRNAs) have rapidly emerged as a new “hub” in TIME research [8]. Acting as competing endogenous RNAs (ceRNAs), scaffolds for chromatin or transcriptional complexes, and mediators of RNA-protein interactions, lncRNAs orchestrate regulatory networks across epigenetic, post-transcriptional, and protein interaction levels. Their functional spectrum is shaped by both subcellular localization (nuclear vs. cytoplasmic) and modification status (e.g., m6A) [9]. Importantly, lncRNAs can be stably secreted into body fluids via exosomes, offering a promising liquid biopsy strategy that bridges mechanistic insights and clinical translation. A major bottleneck of current immunotherapy lies in the dynamic immunosuppression and insufficient immunogenicity of the tumor immune microenvironment (TIME). Increasing evidence indicates that long non-coding RNAs (lncRNAs) occupy upstream or convergent positions in multiple “immune hub pathways,” thereby serving as key regulators shaping TIME [10]. For example, at the immune checkpoint level, various lncRNAs can upregulate PD-L1 through transcriptional or post-transcriptional mechanisms: SNHG12 stabilizes PD-L1 mRNA via HuR/USP8, while SChLAP1 inhibits AUF1-mediated mRNA destabilization; others, such as PCAT1→SOX2 and NEAT1↔DNMT1, indirectly suppress the cGAS/STING pathway, thereby weakening antitumor innate immunity and in situ T cell activation [11, 12]. Notably, the AhR-HNRNPH1–INCR1 axis reveals a novel regulatory layer by which IFN-γ induces PD-L1 and IDO expression through the JAK/STAT signaling pathway [13]. Beyond checkpoints, lncRNAs also directly affect T cell function. For instance, NKILA inhibits NF-κB, making tumor-specific CTLs/Th1 cells more prone to activation-induced cell death (AICD), thereby undermining effector immunity [14]. Meanwhile, some lncRNAs modulate macrophage phagocytic activity via the CD47-SIRPα “don’t eat me” signal, helping tumors evade immune clearance [15]. In the realm of metabolism and programmed cell death, ferroptosis has emerged as a critical entry point to overcome immunoresistance. RGMB-AS1 suppresses HMOX1 ubiquitination and activates NAA10, thereby enhancing ferroptosis and antitumor effects, highlighting a druggable connection among “lncRNA-ferroptosis-immunity” [16, 17]. Furthermore, the interplay between RNA epigenetic modifications-particularly m6A-and lncRNAs has attracted widespread attention. m6A writers/readers/erasers (such as the IGF2BP family) can stabilize immune-related lncRNAs or their target mRNAs, remodel immune phenotypes, and affect therapeutic sensitivity [18, 19]. Despite these rapidly accumulating studies, translating lncRNA mechanisms into real clinical benefit still faces multiple obstacles: (i) most current studies are limited to in vitro systems or small cohorts, lacking cross-cohort validation and prospective evaluation; (ii) delivery efficiency and specificity of nucleic acid therapeutics such as ASOs and siRNAs remain the “last mile” of clinical translation; (iii) companion diagnostics and predictive models based on lncRNAs (e.g., combined with PD-L1 expression, TMB, and STK11/KEAP1 status) still require standardization and prospective validation. Based on recent advances in the study of long non-coding RNAs (lncRNAs) within the tumor immune microenvironment (TIME) of non-small cell lung cancer (NSCLC), this review aims to systematically summarize the molecular mechanisms and network hierarchies by which lncRNAs contribute to immune evasion, ferroptosis, immune regulation, and their roles as prognostic or immune-infiltration biomarkers. Particular emphasis is placed on druggable nodes and cross-pathway coupling (e.g., the interactions among PD-L1, cGAS/STING, and ferroptosis), with representative studies (such as SNHG12, SChLAP1, PCAT1, NEAT1, NKILA, RGMB-AS1, and the AhR-INCR1 axis) highlighted to closely link molecular mechanisms with tumor phenotypes. Table 1 provides an integrated overview of representative lncRNAs that modulate immune regulation, ferroptosis, and therapeutic response in NSCLC. Table 1Representative lncRNAs regulating immune escape and immunotherapy response in NSCLClncRNAMajor Target/AxisMechanism/PathwayImmune/ClinicalRelevanceReference (PMID)SNHG12HuR–USP8–PD-L1Stabilizes PD-L1 mRNA and enhances immune evasionCorrelates with CD8+ T-cell exhaustion and poor ICI response35658874SChLAP1AUF1–PD-L1Inhibits AUF1-mediatePromotes tumor33904361d mRNA decay, upregulates PD-L1immune escape and PD-L1 stabilizationLINC01140miR-30d-5p/PDK1–m TORSponges miRNAs to enhance PD-L1 and c-MycSilencing restores antitumor immunity in vivo34446576PCAT1SOX2–cGAS/STINGSuppresses STING signaling via SOX2 activationReduces antigen presentation and T-cell chemotaxis35415876NKILANF-κB/IκBInhibits NF-κB survival signaling, promoting AICDDrives T-cell exhaustion; potential CAR-T target30224822MIR17HGmiR-17-5p/RUNX3SuppressesRUNX3-depen dent cytotoxic programsPromotes Treg dominance and immune tolerance—KCTD21-AS1miR-519d-5p/CD47– TIPRLm6 A-stabilized ceRNA enhancing CD47, inhibiting autophagyInhibits macrophage phagocytosis; promotes ‘cold’ TIME38383737NEAT1DNMT1/p53/cGAS/STINGEpigenetically suppresses innate immune pathwaysDecreases T-cell infiltration; poor survival32296457ZNRD1–AS1YTHDC2/m6 Am6 A reader–depend ent stabilization promoting angiogenesisLinks m6 A regulation to immune activation36581942AC026356.1Wnt/Stemnessm6 A-dependen t activation ofSustains CSC phenotype37142269Wnt pathwayand ferroptosis toleranceRGMB-AS1HMOX1/NAA10Inhibits HMOX1 ubiquitination and activates ferroptosisEnhances immunogeni city and ICI sensitivity38574881LINC01614Glutamine transporter networkCAF-specific lncRNA enhancing glutamine uptakeRemodels TMEmetabolism; immune evasion36209111Mir100hg(exosomal)H3K14 lactylationExosomal transfer induces epigenetic activationPromotes metastasis and immune suppression40022086LINC01356(exosomal)Blood–brain barrier remodelingAlters endothelial tight junctionsFacilitates brain metastasis in NSCLC35574344ADAMTS9-AS2Immune infiltration networkAssociated with CD8+ T-cell and DC infiltrationPredicts prognosis and ICI benefit34849000LY6K-ASCell-cycle/mitotic progressionCorrelates with proliferation and ‘cold’ TIMEIndependent prognostic biomarker33674747FAM207BPImmune scorecorrelationPromotes proliferation and migrationImmune-rela ted prognostic factor33434533PCBP1–AS1ITGAL (LFA-1)Downregulates lymphocyte adhesion moleculePredicts immune infiltration and survival37256932 Representative lncRNAs regulating immune escape and immunotherapy response in NSCLC 356588 74 339043 61 344465 76 354158 76 302248 22 Suppresses RUNX3-depen dent cytotoxic programs KCTD21-AS 1 383837 37 DNMT1/p53/cGAS/ST ING 322964 57 365819 42 371422 69 385748 81 362091 11 Mir100hg (exosomal) 400220 86 LINC01356 (exosomal) 355743 44 ADAMTS9- AS2 348490 00 336747 47 Immune score correlation 334345 33 372569 32 On this basis, the review further addresses the clinical significance of lncRNAs, evaluating their feasibility as diagnostic, prognostic, and predictive biomarkers, as well as potential therapeutic targets. Special attention is given to the prospects of exosomal lncRNAs in liquid biopsy and dynamic monitoring, as well as their value in combination strategies with ICIs, radiotherapy, and targeted therapies. Finally, we outline future research directions, including the integration of single-cell and multi-omics technologies, deeper investigation into RNA epigenetic modifications, the establishment of standardized paradigms for ferroptosis research, and overcoming technological bottlenecks in nucleic acid drug delivery systems. By addressing these aspects, this review seeks to provide a mechanistic framework for understanding the causal chain linking lncRNAs, TIME, and immunotherapy, and to propose actionable targets and strategies for designing novel combination therapies that can enhance immune responses and overcome therapeutic resistance. Overview of the roles of lncRNAs in the tumor immune microenvironment of lung cancer Basic characteristics and functional modes of lncRNAs Long non-coding RNAs (lncRNAs) generally refer to transcripts longer than 200 nucleotides that do not encode canonical proteins. Based on their genomic locations, they can be classified into intergenic (lincRNAs), antisense, intronic, and enhancer-associated (eRNA-like) subtypes [20]. Their sequences are often highly specific but show limited evolutionary conservation, with functions largely dependent on particular domains and secondary structures. The activity of lncRNAs is closely associated with their subcellular localization and modifications. In the nucleus, lncRNAs typically act as scaffolds or guides, recruiting epigenetic modification complexes (such as EZH2/PRC2, DNMT1) or transcription factors to specific loci, thereby regulating chromatin accessibility and transcription initiation. For example, NEAT1 interacts with DNMT1 to suppress the p53 and cGAS/STING axes, ultimately reshaping antigen presentation and innate immunity [21]. In the cytoplasm, lncRNAs predominantly regulate gene expression through competing endogenous RNA (ceRNA) mechanisms by sponging miRNAs, or by modulating mRNA stability and translation via RNA-binding proteins (RBPs) such as HuR, AUF1, and IGF2BP. They can also directly influence protein ubiquitination and phosphorylation homeostasis. For instance, SNHG12 stabilizes PD-L1 mRNA through HuR/USP8, whereas SChLAP1 upregulates PD-L1 by suppressing AUF1-mediated mRNA destabilization [11]. Meanwhile, RNA epigenetic modifications, particularly N6-methyladenosine (m6A), profoundly affect lncRNA processing, localization, and half-life. Conversely, lncRNAs themselves can modulate the activity of m6A writers, readers, and erasers. For example, ZNRD1–AS1 is positively regulated by the m6A reader YTHDC2, while the m6 A-dependent upregulation of AC026356.1 sustains cancer stemness and activates the Wnt pathway [22]. LINC00470 recruits the m6 A writer METTL3 to facilitate m6 A modification and degradation of PTEN mRNA [23], suggesting that lncRNAs may determine the substrate selectivity of methyltransferase complexes. Similarly, KIAA1429-mediated m6A deposition can be scaffolded by specific lncRNAs to enhance target specificity [24]. These interactions underscore the reciprocal regulatory loop between lncRNAs and m6A modification within the epitranscriptomic network. Overall, the molecular modes of action of lncRNAs can be summarized into four categories: (i) as decoys, they competitively bind transcription factors, miRNAs, or RBPs (e.g., XIST/miR-34a-5p/PD-L1; LINC01140 sponging multiple miRNAs to promote PD-L1 expression); (ii) as scaffolds, they assemble epigenetic complexes and signaling components (e.g., PCAT1 activating SOX2 to suppress cGAS/STING); (iii) as guides, they recruit modifying enzymes to specific loci (e.g., NEAT1 directing DNMT1); and (iv) as signals, they respond to cellular stress or inflammation and transmit cues to downstream effectors (e.g., the AhR-HNRNPH1–INCR1 axis converting IFN-γ signaling into PD-L1/IDO induction). In addition, lncRNAs perform two “logistical” functions: maintaining protein homeostasis and degradation (e.g., RGMB-AS1 inhibiting HMOX1 ubiquitination and cooperating with NAA10 to regulate ferroptosis) and mediating secretion and transport via exosomes, which establish long-range communication between tumors and immune cells. These features not only deepen the complexity of the TIME but also provide new avenues for liquid biopsy and clinical intervention. Figure 1 illustrates the fundamental features of lncRNAs and their diverse molecular modes of action, including Decoy, Scaffold, Guide, and Signal, as well as ancillary functions such as exosomal transport and protein homeostasis maintenance, highlighting how lncRNAs regulate target gene expression within the nucleus and cytoplasm. Basic features and functional modes of lncRNAs in tumor cells Major components of the tumor immune microenvironment (TIME) and lncRNA connection points Within the tumor immune microenvironment, lncRNAs exert multilayered regulatory effects across diverse immune cell populations, profoundly shaping the efficacy of antitumor immune responses [25, 26]. For effector (CD8+) and helper T cells (CD4+), lncRNAs act through three principal layers. At the checkpoint layer, lncRNAs such as SNHG12, SChLAP1, SOX2-OT, XIST, and LINC01140 directly or indirectly upregulate PD-L1, thereby impairing T-cell recognition and cytotoxicity [11, 27, 28]. At the life-death layer, NKILA suppresses NF-κB signaling, rendering tumor-specific CTLs and Th1 cells more prone to activation-induced cell death (AICD), which results in transient and easily exhausted immune responses [14]. At the in situ activation layer, PCAT1 and NEAT1 inhibit the cGAS/STING pathway, reducing type I interferon and chemokine (CXCL9/10) production, thereby weakening dendritic cell-mediated T-cell priming and local expansion [12]. Moreover, MIR17HG enhances the immunosuppressive function of regulatory T cells (Tregs) via the miR-17-5p/RUNX3 axis, altering the effector-to-suppressor T-cell ratio and creating an immunosuppressive milieu [29]. In the context of antigen presentation, cGAS-STING signaling is central to dendritic cell cross-presentation and T-cell activation. When this pathway is suppressed by nuclear lncRNAs such as PCAT1 and NEAT1, both antigen presentation and type I interferon cascades are compromised, reducing the plasticity of immunotherapy. For myeloid cells, m6A-modified KCTD21–AS1 enhances CD47 expression, strengthening the “don’t eat me” signal and impairing the phagocytic capacity of tumor-associated macrophages [15]. Similarly, the CDK1-related lncRNA/CXCL8 module recruits polymorphonuclear MDSCs (PMN-MDSCs) and neutrophils via IL-8, fostering a “cold tumor” phenotype characterized by high IL-8 and low T-cell infiltration. For natural killer (NK) cells, although direct evidence is limited, studies suggest that lncRNAs may indirectly dampen NK cell degranulation and cytotoxicity through PD-L1, TGF-β, and metabolic suppressive pathways involving adenosine and lactate. At the vascular and stromal level, ZNRD1–AS1 regulates the miR-942/TNS1 axis to promote migration and angiogenesis, indirectly altering immune cell trafficking and infiltration efficiency [22]. lncRNAs associated with cancer-associated fibroblasts (CAFs) may remodel the extracellular matrix into an immunosuppressive barrier through TGF-β/IL-6 networks. In terms of metabolism and programmed cell death, RGMB-AS1 suppresses HMOX1 ubiquitination and activates NAA10, thereby promoting ferroptosis through dual mechanisms. This not only directly restrains tumor growth but may also enhance immunogenicity by releasing DAMPs and altering iron-lipid metabolism [16]. In contrast, the m6 A-dependent upregulation of AC026356.1 sustains cancer stemness and activates Wnt signaling, a phenomenon often concomitant with immune tolerance, suggesting a coupling between stemness maintenance and immune “coldness” [30]. Finally, several lncRNAs-including ADAMTS9–AS2, PCBP1–AS1, LY6K-AS, FAM207BP, and LINC01833-are strongly associated with immune infiltration patterns and patient prognosis, highlighting their potential as biomarkers for risk stratification and predictors of immunotherapy response [31–34]. Figure 2 depicts the major cellular components of the tumor immune microenvironment (TIME) in lung cancer and the associated lncRNA regulatory networks, including T cells, Tregs, dendritic cells (DCs), tumor-associated macrophages/myeloid-derived suppressor cells (TAMs/MDSCs), NK cells, vascular endothelial cells, and cancer-associated fibroblasts (CAFs), illustrating their roles in immune suppression, immune activation, and metabolic/cell death processes. Tumor immune microenvironment (TIME) composition and lncRNA regulatory network LncRNAs and immune evasion mechanisms Regulation of the PD-1/PD-L1 Axis In NSCLC, PD-L1 expression is not only controlled by classical signaling pathways such as IFN-JAK/STAT, PI3K-AKT-mTOR, and HIF-1α, but is also finely tuned by multiple lncRNAs at post-transcriptional and translational levels [35–37]. Overall, lncRNAs enhance PD-L1-mediated immune evasion through four main mechanisms: (i) mRNA stability regulation: SNHG12 acts as a scaffold to facilitate HuR-USP8-mediated stabilization of PD-L1 mRNA, while SChLAP1 prolongs PD-L1 half-life by preventing AUF1 binding to its 3′UTR. Both mechanisms have been validated in vitro and in vivo, demonstrating significant suppression of CD8+ T-cell function [11]; (ii) ceRNA networks: XIST sponges miR-34a-5p to release its inhibitory effect on PD-L1, whereas LINC01140 simultaneously protects PD-L1 and c-Myc transcripts, thereby promoting immune evasion. In immune cell-based mouse models, silencing LINC01140 restores antitumor immunity [38, 39]; (iii) Amplification of upstream signaling: SOX2-OT upregulates PD-L1 via the miR-30d-5p/PDK1/mTOR axis, while PCAT1 and NEAT1 suppress the cGAS/STING-IFN-I pathway through SOX2 and DNMT1, respectively, thereby weakening antigen presentation and chemotactic signaling to sustain immune tolerance [28]; (iv) IFN-induced noncanonical regulation: The antisense transcript INCR1 interacts with HNRNPH1 to relieve translational repression of PD-L1, JAK2, and IDO1, forming the AhR-INCR1–HNRNPH1 axis that links environmental cues and IFN responses to immune checkpoint upregulation [13]. These mechanisms have been validated by qPCR, dual-luciferase assays, RIP/CLIP, mRNA half-life analysis, and immune cell co-culture or xenograft models. Some lncRNAs (e.g., LINC01140, PCAT1) also showed synergistic effects with immunotherapy or radiotherapy. From a translational perspective, targeting lncRNAs such as SNHG12, SChLAP1, SOX2-OT, XIST, and LINC01140 can directly reduce PD-L1 expression, whereas restoring cGAS/STING activity may “heat up” the tumor immune microenvironment by enhancing antigenicity and T-cell infiltration-even if accompanied by transient PD-L1 upregulation-which generally synergizes with ICIs. Potential therapeutic strategies include nucleic acid-based drugs targeting these lncRNAs, small molecules disrupting the INCR1–HNRNPH1 complex, and rationally timed combinations of STING agonists, radiotherapy, and ICIs. Clinically, the expression of these lncRNAs correlates with PD-L1 levels, immune infiltration, and treatment response, suggesting their potential as companion diagnostics and stratification markers [11, 27, 28, 38, 39]. However, integration with established biomarkers such as PD-L1 IHC, TMB, and STK11/KEAP1 status will be necessary. Importantly, some pathways exert cross-regulatory effects-for example, STING activation can transiently induce PD-L1, but still enhances overall antitumor immunity. Thus, optimal scheduling and dosing will be critical in clinical translation [21]. From a translational perspective, targeting lncRNAs such as SNHG12, SChLAP1, SOX2-OT, XIST, and LINC01140 may suppress PD-L1 expression, whereas restoring cGAS/STING signaling could “heat up” the tumor immune microenvironment by promoting antigenicity and T-cell infiltration, even if accompanied by temporary PD-L1 upregulation that synergizes with ICIs. Potential strategies include nucleic acid-based agents against oncogenic lncRNAs, small molecules disrupting the INCR1–HNRNPH1 complex, and rationally timed combinations of STING agonists, radiotherapy, and ICIs. Clinically, these lncRNAs correlate with PD-L1 levels, immune infiltration, and therapy response, suggesting their promise as companion diagnostics, although integration with established biomarkers such as PD-L1 IHC, TMB, and STK11/KEAP1 status will be essential. In summary, lncRNAs act as multilayered regulators and amplifiers of PD-L1 signaling while interacting with innate immune hubs like cGAS–STING; nevertheless, their mechanistic diversity, context dependence, and limited clinical validation underscore the need for multi-omics and prospective cohort studies to substantiate their translational potential in NSCLC immunotherapy. Figure 3 illustrates the multilayered regulatory mechanisms of lncRNAs on the PD-1/PD-L1 axis in NSCLC, including mRNA stability, ceRNA/miRNA axes, upstream signal amplification, and non-canonical IFN-mediated regulation, along with the associated combinatorial therapeutic strategies. Multilayered regulation of the PD-1/PD-L1 axis by lncRNAs in NSCLC T-cell apoptosis and functional suppression In the tumor immune microenvironment (TIME) of NSCLC, T-cell dysfunction arises not only from “external brakes” such as PD-1/PD-L1 checkpoints but also from “internal imbalance” involving metabolism, stress, and programmed cell death [40–42]. Two representative lncRNA circuits-NKILA and MIR17HG-shape immune hyporesponsiveness in NSCLC by acting on distinct yet complementary axes: the survival threshold of effector T cells (CTL/Th1) and the brake imposed by Treg dominance. Activation-induced cell death (AICD), triggered by repeated antigen stimulation through the FAS-FASL/TRAIL-DR5-Caspase-8/3 pathway, is physiologically designed to prevent excessive immune responses [43, 44]. Tumors exploit this brake to drive premature exhaustion of tumor-specific CTL/Th1 cells. NKILA binds the NF-κB/IκB complex, blocking IKK-mediated IκB phosphorylation and NF-κB nuclear translocation, thereby suppressing NF-κB-dependent survival genes (e.g., BCL2L1, CFLAR/c-FLIP, BIRC family) and antioxidant/stress-buffering programs. The net result is that effector T cells cross the AICD threshold more readily under equivalent TCR stimulation, marked by increased FASL, heightened caspase activation, and diminished persistence of cytokine secretion and cytotoxicity. Within the TIME, NKILA is induced by TGF-β, chronic antigen load, and oxidative stress, leading to enrichment of short-lived, NKILA^high TILs that fail to sustain responses. Functionally, silencing NKILA prolongs CTL survival, enhances tumor clearance, and promotes memory differentiation; however, systemic long-term inhibition could compromise the “physiological safeguard” of AICD [14], raising risks of immune-related toxicity. Thus, the most viable translational approaches are transient inhibition ex vivo before adoptive transfer (e.g., TIL/CAR-T) or localized/timed delivery strategies to minimize systemic exposure. MIR17HG, the host lncRNA of the miR-17~92 cluster, elevates miR-17-5p to repress RUNX3, a transcription factor central to CD8+ cytotoxic lineage commitment and Th1 differentiation that drives effector programs such as IFNG, GZMB, and PRF1. The MIR17HG↑-miR-175p↑-RUNX3↓ axis effectively downshifts CTL/Th1 cytotoxicity and IFN-γ output while fostering a Treg-favored environment. On one hand, weakened cytokine and chemokine profiles reduce effector-Treg antagonism; on the other, RUNX3 suppression stabilizes FOXP3 and reorganizes chemokine axes (e.g., CCL22), indirectly enhancing Treg recruitment and suppressive potency, ultimately lowering the effector/Treg ratio. Targeting this pathway by inhibiting MIR17HG or antagonizing miR-17-5p can “release the brake” on RUNX3, thereby restoring CTL function while attenuating Treg dominance-an “all-in-one” immune reprogramming strategy. Nonetheless, given the physiological roles of the miR-17~92 cluster in hematopoiesis and cardiovascular biology, precise dosing and targeted delivery are critical for safety [29]. Collectively, these two axes exemplify how tumors co-opt lncRNAs to lower the resilience and functional capacity of effector T cells while reinforcing Treg-mediated suppression. Yet, the strength of evidence is uneven: most findings derive from cell-based or xenograft studies, and robust patient-level validation of NKILA or MIR17HG expression and immune phenotype correlation remains limited. The translational feasibility of lncRNA-targeted immunomodulation also faces delivery and toxicity challenges, and conflicting reports exist regarding MIR17HG expression across NSCLC subtypes. Integrating multi-omics profiling with functional assays in patient-derived TILs and adoptive transfer models is thus essential to delineate causality and safety boundaries. From a therapeutic perspective, rational combination strategies—such as transient NKILA inhibition to extend TIL persistence, coupled with MIR17HG/miR-17-5p blockade to restore RUNX3-driven cytotoxicity—may jointly enhance T-cell quantity and quality, particularly when integrated with STING agonists or radiotherapy to boost antigenicity and chemotaxis [29]. Figure 4 summarizes these lncRNA-mediated pathways underlying T-cell apoptosis and functional suppression and outlines possible pharmacologic and delivery strategies to restore effector function and limit Treg activity in NSCLC. LncRNA-mediated T cell apoptosis and functional suppression in NSCLC Macrophage/Phagocytosis regulation In NSCLC, tumor cells frequently evade immune clearance by upregulating CD47, which engages the macrophage receptor SIRPα to deliver a “don’t eat me” signal to myeloid cells [45, 46]. Recent studies have identified the m6A-modified lncRNA KCTD21–AS1 as an upstream regulator of this pathway. Stabilized through METTL14-mediated m6 A modification, KCTD21–AS1 acts as a ceRNA to sponge miR-519d-5p, thereby releasing its repression of CD47 and TIPRL. This dual derepression exerts two key effects: (i) elevated CD47 expression strengthens the CD47–SIRPα inhibitory axis, raising the phagocytic threshold of macrophages and fostering resistance to clearance; and (ii) TIPRL upregulation suppresses autophagy markers such as LC3-II/I, decreasing autophagic flux and enhancing tumor survival and metastatic potential [15]. Second, TIPRL upregulation suppresses autophagy-related marker LC3-II/I, diminishing tumor cell autophagy and further promoting survival and metastatic potential. Both in vitro and xenograft models confirm that high KCTD21–AS1 expression inhibits macrophage phagocytosis, while either miR-519d-5p mimic supplementation or KCTD21–AS1 knockdown restores phagocytosis and autophagy and suppresses tumor growth [15]. Functional assays—including luciferase reporter, RIP/pull-down, mRNA half-life, autophagy flux, and macrophage co-culture—support this model. However, most of these data derive from cellular and xenograft systems, and clinical validation remains preliminary. Although strong mechanistic coherence exists, the sample sizes are limited, and inter-patient heterogeneity of KCTD21–AS1 and CD47 expression has not been thoroughly evaluated [15]. Moreover, the relative contribution of TIPRL-mediated autophagy inhibition versus CD47–SIRPα signaling to immune evasion remains unclear, as both pathways intersect with other “don’t eat me” cues such as MHC-I, PD-L1, and CD24. Further multi-omics profiling and single-cell analyses will be required to disentangle these overlapping regulatory networks and verify causality in patient-derived macrophage–tumor co-cultures. Translationally, targeting this axis can follow three strategies: (i) directly inhibiting KCTD21–AS1 or delivering miR-519d- 5p mimics to downregulate CD47 and TIPRL; (ii) combining with CD47/SIRPα blockade agents (e.g., magrolimab, evorpacept) to potentially enhance efficacy while lowering the required dose and improving the safety window; (iii) first relieving the KCTD21–AS1-mediated phagocytic brake, then combining with PD-1/PD-L1 inhibitors to augment T-cell responses. It is important to note that widespread CD47 expression on red blood cells can lead to anemia and hemolysis, prompting the development of strategies such as low-dose “priming” regimens and Fc-inactivated fusion proteins to mitigate toxicity [47, 48]. To overcome this, Fc-silent fusion constructs, local delivery systems, or low-dose “priming” regimens are being explored. Additionally, sequential regimens—first alleviating KCTD21–AS1-mediated phagocytic suppression, then combining with PD-1/PD-L1 blockade—may synergistically enhance both innate and adaptive immunity. In summary, KCTD21–AS1 integrates epitranscriptomic (m6 A) and ceRNA regulation to upregulate CD47 and TIPRL, suppress macrophage phagocytosis, and attenuate autophagy, collectively driving an immunologically “cold” tumor phenotype. While preclinical results are compelling, further clinical evidence, toxicity assessment, and delivery optimization are essential to confirm whether targeting this axis can safely “reheat” the tumor immune microenvironment and sensitize NSCLC to combination immunotherapies. Figure 5 schematically illustrates these mechanisms and their therapeutic implications. KCTD21–AS1-mediated regulation of macrophage phagocytosis in NSCLC LncRNAs in cell death and metabolic pathways Ferroptosis is a form of regulated cell death characterized by iron-dependent lipid peroxidation, orchestrated by a balance between “brake” and “accelerator” networks. On one hand, the System Xc-(SLC7A11/SLC3A2)-GSH-GPX4 axis constitutes a primary antioxidant defense, while FSP1-CoQ₁₀ and DHODH-CoQH₂ provide parallel rescue pathways. On the other hand, ACSL4–LPCAT3 promotes the incorporation of polyunsaturated fatty acids into membranes, LOX/ALOX catalyzes their peroxidation, and expansion of the labile iron pool via Fenton reactions further amplifies lipid peroxidation cascades [17, 49, 50]. Transcriptional stress axes such as NRF2–HMOX1, HIF-1α, and p53-SLC7A11 exert context-dependent, bidirectional effects on ferroptosis [51, 52]. In NSCLC, ferroptosis not only directly suppresses tumor growth but also enhances immunogenicity through the release of DAMPs and oxidized lipids, thereby increasing sensitivity to radiotherapy, targeted therapy, and immunotherapy. This makes ferroptosis a potential lever to convert “cold” tumors into “hot” tumors [53, 54]. Recent studies highlight RGMB-AS1 as a key driver in this process. It promotes ferroptosis via two complementary mechanisms: first, by inhibiting HMOX1 ubiquitination, RGMB-AS1 stabilizes HMOX1, expands the labile iron pool, and enhances lipid peroxidation; second, it activates NAA10, weakening GPX4-mediated protection and raising the lipid peroxidation threshold [16]. Functional assays demonstrate that RGMB-AS1 overexpression increases LIP, MDA, and 4-HNE levels while decreasing GPX4 activity, sensitizing cells to ferroptosis inducers such as erastin and RSL3, whereas knockdown reverses this phenotype. Importantly, HMOX1 may play dual roles depending on the oxidative or inflammatory context; RGMB -AS1-mediated stabilization tends to favor pro-ferroptotic conditions, highlighting its unique druggability. Ferroptosis and immunity also interact bidirectionally. IFN-γ from CD8+ T cells suppresses SLC7A11, reduces GSH/GPX4 defenses, and sensitizes tumor cells to ferroptosis. Conversely, ferroptotic tumor cells release HMGB1 and oxidized lipids, promoting dendritic cell maturation, enhancing cross-presentation, and upregulating CXCL9/10, thereby improving T-cell infiltration and ICI response. However, this effect is context-dependent: lipid peroxidation in T cells or DCs can impair immunity, emphasizing the need for tumor-selective and temporally optimized ferroptosis induction. In contrast, the lncRNA AC026356.1, upregulated in LUAD via m6 A-dependent stabilization, maintains stemness and activates the Wnt/β-catenin pathway, which is associated with immune suppression and ferroptosis resistance. High AC026356.1 expression is often accompanied by upregulation of SLC7A11/GPX4 and lipid metabolic remodeling, raising the ferroptosis threshold and weakening immune responses [30]. Therefore, targeting AC026356.1 or its m6 Amediated stabilization may disrupt Wnt-mediated immune suppression while lowering the ferroptosis barrier, creating a favorable metabolic environment for synergistic effects with RGMB-AS1, ICIs, and radiotherapy. From a translational perspective, lncRNAs in the ferroptosis-immunity network suggested several intervention strategies: in early therapy, short-term RGMB-AS1 induction can “borrow” ferroptosis to enhance antigenicity, while inhibition under high oxidative stress can prevent collateral damage. ASO, siRNA, or CRISPRi targeting AC026356.1 may lower the ferroptosis threshold and relieve immune suppression. Pharmacologically, combining System Xc−, GPX4, or FSP1 inhibitors with PD-1/PD-L1 blockade, fine-tuning iron pools with iron supplements or chelators, and modulating lipid metabolism all represent feasible combination approaches [55–57]. Delivery technologies such as nanocarriers, liposomes, or exosome mimetics enable tumor-specific targeting, and temporal optimization-priming (ferroptosis induction), amplification (STING agonist/radiotherapy/targeted therapy), and maintenance (ICIs)helps protect immune effector cells. Finally, integrating RGMB-AS1 and AC026356.1 expression with ferroptosis scores, iron metabolism readouts, and immune indicators (PD-L1, TMB, STK11/KEAP1 status) can guide patient stratification and efficacy prediction. In summary, RGMB-AS1 drives ferroptosis and enhances immunogenicity via dual “iron increase-antioxidant decrease” pathways, whereas AC026356.1 elevates the ferroptosis threshold and suppresses immune responses through a “stemnessWnt-immune cold” framework. Strategically timed and targeted modulation of the lncRNA-ferroptosis-immunity axis holds promise for converting “cold” TIME into “hot” TIME, representing a critical avenue to overcome immune resistance in NSCLC. Figure 6 illustrates how lncRNAs RGMB-AS1 and AC026356.1 bidirectionally regulate ferroptosis and tumor-immune interactions in NSCLC, including the promotion or inhibition of ferroptosis, modulation of TIME “hot” or “cold” states, and associated clinical translational strategies. lncRNA-mediated regulation of ferroptosis and immune crosstalk in NSCLC LncRNAs as immune-related biomarkers In NSCLC, multiple lncRNAs have emerged as potential immune-related biomarkers that mirror the cellular and molecular complexity of the TIME [58, 59]. At the single-gene level, ADAMTS9–AS2 is consistently associated with the infiltration of CD8+ T cells, macrophages, and dendritic cells across multiple LUAD cohorts. Higher ADAMTS9–AS2 expression correlates positively with overall survival, as supported by TCGA-based Kaplan–Meier analysis (HR = 0.68, 95% CI: 0.51–0.91, p = 0.01). Furthermore, multivariate Cox regression confirms that ADAMTS9–AS2 is an independent prognostic factor in LUAD (HR = 0.626, 95% CI: 0.397–0.986, p = 0.043), suggesting that it may characterize a more “hot” tumor immune microenvironment through regulation of stromal–immune interactions and inflammatory pathways [31]. PCBP1–AS1 regulates lymphocyte adhesion and infiltration by mediating downregulation of ITGAL (LFA-1); its expression in patient samples aligns with immune infiltration patterns and prognostic indicators, providing potential for risk stratification [32]. Y6K-AS is co-expressed with proliferation-related genes and is upregulated in highly proliferative, immunologically “cold” tumors, serving as an indicator of poor prognosis that reflects the coupling of cell cycle activity and immune suppression [33]. FAM207BP, a pseudogene-derived lncRNA, promotes tumor proliferation and migration and correlates with immune scores from ESTIMATE and TIMER, representing another immune-related prognostic factor [34]. LINC01833, in integrated TCGA/GEO analyses, correlates strongly with immune infiltration and OS and co-regulates chemokine, inflammatory, and antigenpresentation modules, making it a promising single-gene biomarker for preclinical evaluation [60]. Collectively, these five lncRNAs capture diverse facets of the TIME—ranging from immune infiltration density and cell adhesion to proliferation-driven immunosuppression—demonstrating translatable potential. Nevertheless, while bioinformatic correlations are robust, most evidence remains associative rather than causal. Many analyses rely on retrospective datasets, where tumor purity, coexisting inflammation, and batch effects can confound immune infiltration estimates. Cohort heterogeneity (e.g., histological subtypes, driver mutations, ethnic composition) and methodological inconsistencies across algorithms such as CIBERSORT, TIMER, xCell, and ssGSEA also limit reproducibility. Functional validations using CRISPRi, siRNA, or ASO models have begun to bridge this gap, but the scope and scale of in vivo or clinical confirmation remain limited, and many reported mechanisms have yet to be replicated across independent laboratories or multi-cohort studies [61, 62]. To improve reliability, recent studies have integrated multi-omics and immunotherapy cohorts, applying LASSO-Cox regression and machine-learning frameworks to construct composite immune-related lncRNA risk signatures that incorporate PD-L1-, cGAS-STING-, and ferroptosis-associated lncRNAs, thereby enhancing the predictive accuracy for immunotherapy response [63–65]. These models generate risk scores validated across multiple centers for OS and PFS and combined with PD-L1, TMB, STK11/KEAP1, and HLA LOH to construct multidimensional nomograms. However, potential overfitting, incomplete clinical annotation, and limited generalizability remain major barriers to clinical translation. Moreover, pre-analytical variability in FFPE, fresh tissue, and exosome-based assays—together with inconsistent normalization and quality control—affects detection reproducibility and complicates cross-platform validation. From a translational perspective, lncRNA-based diagnostics offer conceptual promise but require a standardized validation pipeline that integrates: (i) discovery in large, harmonized cohorts; (ii) independent validation in ICI-treated clinical datasets; (iii) mechanistic confirmation using immunocompetent and humanized models; and (iv) analytical standardization for assay development compliant with CLIA/LDT guidelines. Importantly, biomarker studies should be evaluated and reported in accordance with REMARK recommendations to ensure consistency in statistical reporting, cohort characterization, and analytical transparency across studies. Only through this closed loop—from computational discovery to mechanistic validation and clinical qualification—can these candidates evolve from correlational biomarkers to clinically actionable precision tools. In summary, while ADAMTS9–AS2, PCBP1–AS1, LY6K-AS, FAM207BP, and LINC01833 delineate key immune–tumor interactions and prognostic landscapes, their reproducibility, inter-cohort consistency, and translational feasibility remain to be rigorously established through multicenter prospective studies. Clinical translation and therapeutic implications From a translational viewpoint, lncRNAs offer dual opportunities as biomarkers and therapeutic targets in NSCLC [66–68]. Exosomal lncRNAs are emerging as highly promising biomarkers for the dynamic monitoring of tumor–immune interactions and treatment response, owing to their remarkable stability, tissue specificity, and detectability in circulation. Recent preclinical and clinical studies have provided compelling evidence linking exosomal lncRNAs to immune modulation, metabolic reprogramming, and therapeutic outcomes across multiple cancer types [69–71]. For instance, cancer-associated fibroblast (CAF)-specific LINC01614 enhances glutamine uptake and metabolic reprogramming in lung adenocarcinoma, thereby promoting tumor growth and immune evasion [72]. Moreover, exosomal Mir100hg, secreted by lung cancer stem cells, was recently shown to activate histone H3K14 lactylation in non-stem cancer cells, boosting their metastatic and invasive capacity [73]. This finding provides a striking example of how exosomal lncRNAs mediate epigenetic crosstalk between cancer stem and non-stem compartments. In addition, exosomal LINC01356, derived from brain-metastatic NSCLC cells, can remodel the blood–brain barrier (BBB) by downregulating tight junction proteins, facilitating brain metastasis and immune cell exclusion [74]. Together, these studies demonstrate that exosomal lncRNAs such as LINC01614, Mir100hg, and LINC01356 not only serve as key mediators of metabolic and epigenetic communication within the tumor microenvironment but also hold promise as prognostic and predictive biomarkers for metastasis and therapeutic response in NSCLC. Integrating such exosomal lncRNAs into multi-parametric liquid biopsy panels could enable early detection of metastasis, dynamic monitoring of immunotherapy efficacy, and stratified patient selection. Although most current evidence stems from preclinical or early clinical studies, extending similar analyses to NSCLC and integrating exosomal lncRNAs such as NEAT1, XIST, MALAT1, and HOTAIRM1 into longitudinal patient monitoring pipelines could significantly improve the precision and timeliness of immunotherapy evaluation. However, broad clinical application will require prospective multicenter validation to ensure analytical consistency and cost-effectiveness. Therapeutically, lncRNAs can be modulated through several translatable strategies. Inhibiting oncogenic or immunosuppressive lncRNAs (e.g., SNHG12, SChLAP1, SOX2-OT, XIST, NEAT1, PCAT1, KCTD21–AS1) restores antitumor immunity by reducing PD-L1, reactivating cGAS–STING signaling, and enhancing macrophage phagocytosis. Conversely, reactivating tumor-suppressive lncRNAs such as RGMB-AS1 promotes ferroptosis and boosts immunogenicity. Preclinical studies using nanoparticles, viral vectors, or exosome-mimetic carriers have demonstrated improved delivery and immune activation in lung cancer models [75, 76]. Combination strategies with ICIs represent the most promising translational direction. Suppressing PD-L1-stabilizing lncRNAs (SNHG12, XIST, LINC01140) or innate immunity suppressors (PCAT1, NEAT1) synergizes with PD-1/PD-L1 blockade, while targeting KCTD21–AS1 or restoring miR-519d-5p downregulates CD47 and enhances antigen presentation. Ferroptosis-promoting lncRNAs (e.g., RGMB-AS1) may further enhance antigenicity, forming a positive feedback loop with ICIs [16]. These findings support a rational, time-sequenced “ignite–amplify–maintain” strategy combining innate activation, checkpoint blockade, and lncRNA modulation. Despite rapid progress, several barriers remain: (i) limited in vivo stability and biodistribution of nucleic acid drugs; (ii) off-target effects from overlapping transcripts or isoforms; and (iii) immune activation or toxicity from oligonucleotide delivery [66, 67]. Ongoing work explores inhalable or local delivery, TME-responsive systems, and patient stratification based on lncRNA expression to improve safety and efficacy. Future clinical trials may stratify NSCLC patients by lncRNA expression or immune phenotype and test rational combinations (e.g., ICIs ± lncRNA inhibition, ICIs ± ferroptosis induction). Endpoints could include response rate and PFS, while translational readouts assess immune infiltration and PD-L1/ferroptosis/STING activity. In summary, lncRNAs bridge mechanistic discovery and precision immunotherapy, offering a foundation for clinically actionable biomarker development and combinatorial therapeutic design. Future perspectives Advancing lncRNA research in the tumor immune microenvironment (TIME) of lung cancer requires an integrated and clinically focused roadmap that bridges basic discovery with translational application [77–79]. Future progress will rely on coordinated innovation across technological, mechanistic, and clinical dimensions. Technological advances in single-cell and spatial multi-omics—such as scRNA-seq, scATAC-seq, CITE-seq, and spatial transcriptomics—are redefining our understanding of lncRNA regulation within immune and stromal compartments, enabling visualization of spatial “hot–cold” gradients marked by PD-L1, CD47, and CXCL9/10 expression [80, 81]. Complementary approaches including long-read sequencing, m6 A epitranscriptomic profiling, and integration with proteomics or metabolomics will further resolve isoform diversity and illuminate lncRNA–metabolism–immunity crosstalk, ultimately facilitating the construction of multi-scale atlases linking lncRNA expression to immune architecture, ferroptosis sensitivity, and resistance phenotypes. Mechanistically, key questions remain regarding the interplay between lncRNA, m6 A modification, metabolic stress, and immune escape. It is essential to elucidate how m6 A “writers/readers/erasers” control lncRNA stability and immune responsiveness, how ferroptosis-related lncRNAs such as RGMB-AS1 enhance immunogenicity, and how the AC026356.1–Wnt–stemness circuit confers ferroptosis tolerance and T-cell suppression. Clarifying the hierarchical relationships among STING–PD-L1, IDO–CD47, and GPX4–SLC7A11 pathways under oncogenic or immunotherapy-resistant contexts (EGFR/ALK, STK11/KEAP1, or HLA LOH) will be crucial for prioritizing therapeutic targets and designing rational combination regimens. On the translational front, a next-generation framework should integrate multi-omics-based patient stratification, window-of-opportunity neoadjuvant trials, and precision delivery systems. Early-phase trials combining pre- and post-treatment biopsies with exosomal monitoring can link molecular modulation to immune outcomes, while adaptive Bayesian designs (e.g., ICIs ± lncRNA inhibition, ICIs ± ferroptosis induction, or anti-CD47 ± KCTD21–AS1 blockade) may optimize treatment efficacy with endpoints such as ORR, PFS, and immune infiltration dynamics. In parallel, advances in targeted nanoparticle or exosome-mimetic delivery—particularly via inhalation or intratumoral administration—offer opportunities to enhance tissue specificity while minimizing systemic toxicity [82–84]. Additionally, bioinformatic frameworks including graph neural networks and multimodal transformers will play a pivotal role in integrating single-cell, spatial, and liquid biopsy data to construct interpretable models of lncRNA–immune–metabolic interactions. Over the next three to five years, research should focus on completing a “map–measure–modulate” cycle: mapping the lncRNA–TIME landscape using multi-omics and spatial tools, establishing reproducible detection systems for tissue and exosomal biomarkers, and implementing nucleic acid–based modulation of validated targets within stratified clinical trials. Through this integrated approach, lncRNAs can ultimately evolve from mechanistic insights into clinically actionable tools that enhance the precision and efficacy of immunotherapy in lung cancer. Conclusion Integrating current evidence, lncRNAs have evolved from passive bystanders to central regulators within the tumor immune microenvironment (TIME) of lung cancer, orchestrating immune evasion, ferroptosis, and cellular cross-talk through multilayered mechanisms. Representative lncRNAs regulate PD-1/PD-L1 expression (e.g., SNHG12, SChLAP1, XIST, LINC01140), suppress innate sensing pathways (PCAT1, NEAT1-cGAS/STING), modulate T-cell exhaustion and apoptosis thresholds (NKILA, MIR17HG), control macrophage phagocytic signaling (KCTD21–AS1–CD47), and bridge metabolic cell death with immunogenicity (RGMB-AS1-ferroptosis; AC026356.1-Wnt-stemness axis) [71, 85–87]. Collectively, these networks define the “hot–cold” TIME continuum and influence responsiveness to immune checkpoint blockade. Despite these advances, critical knowledge gaps remain. Most mechanistic studies rely on in vitro or murine models, with limited validation in patient cohorts or clinical settings. The causal hierarchy linking specific lncRNA alterations to TIME remodeling and immunotherapy response remains incompletely defined [26, 88]. Moreover, the heterogeneity of lncRNA isoforms, m6 A modifications, and spatial expression across immune and stromal compartments complicates biomarker translation. Addressing these limitations will require systematic multi-omics profiling, high-resolution spatial mapping, and integration of lncRNA expression into prospective immunotherapy trials. From a translational perspective, lncRNAs hold dual potential as biomarkers and therapeutic sensitizers. Immune-associated lncRNAs such as ADAMTS9–AS2, PCBP1–AS1, LY6K-AS, FAM207BP, and LINC01833 demonstrate strong correlations with immune infiltration and clinical outcomes, supporting their value in risk stratification and treatment prediction, especially when combined with PD-L1, TMB, STK11/KEAP1, and HLA status. On the therapeutic side, a rational “ignite–amplify–maintain” model offers a framework for intervention: RGMB-AS1–mediated ferroptosis to ignite antigenicity, PCAT1/NEAT1 inhibition to amplify innate signaling, and SNHG12/SChLAP1/XIST/LINC01140 suppression to maintain T-cell activity. Additionally, targeting the KCTD21–AS1–CD47 axis, optimized through dose-adjusted anti-CD47/anti-SIRPα therapy, may enhance phagocytosis and antigen presentation while mitigating hematologic toxicity 15]. Looking forward, the next translational milestones will depend on completing an integrated “map–measure–modulate” loop: (1) constructing high-resolution TIME maps using spatial multi-omics and single-cell tools; (2) establishing standardized detection pipelines for tissue and exosomal lncRNAs as companion diagnostics; and (3) implementing well-designed, biomarker-stratified clinical trials to validate safety, efficacy, and predictive accuracy. Achieving these goals will transform lncRNAs from mechanistic insights into clinically actionable nodes—bridging molecular discovery and therapeutic innovation to improve long-term outcomes for NSCLC patients. Acknowledgements Not applicable. Author contributions Weiwei Wang and Zhibiao Jiang contributed equally to this work. Kun Zhao, and Yuanyin Teng were responsible for study conception, literature review, and manuscript drafting. Yuanyin Teng and Lingling Wu performed data collection, analysis, and figure preparation. Ting Yu and Anquan Shang supervised the study, revised the manuscript critically for important intellectual content, and approved the final version for submission. All authors read and approved the final manuscript. Funding This work has been supported by the Jiangsu Provincial Health and Health Commission Research Project (K2024084, JSZJ20241204); the Lianyungang City. 521 High-level Talent Training Project (LYG065212024074); the Lianyungang City Health Science and Technology Project (202321); the Jiangsu University Medical Education Cooperation Project (JDYY2023085). Data availability The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare that there are no conflicts of interest. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Weiwei Wang and Zhibiao Jiang contributed equally and are shared as first authors. Contributor Information Ting Yu, Email: 13611711634@163.com. Anquan Shang, Email: shanganquan@tongji.edu.cn. References Associated Data Data Availability Statement The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.