Cellular immunotherapy for lung cancer: current clinical evidence, translational challenges, and future perspectives: a narrative review

Abstract

Background and Objective

Lung cancer remains a leading cause of cancer-related mortality. Cellular immunotherapy has achieved major advances in hematologic malignancies, but translation to lung cancer is constrained by antigen heterogeneity, limited trafficking and persistence, an immunosuppressive tumor microenvironment (TME), manufacturing complexity, and on-target off-tumor toxicity. This narrative review summarizes current clinical evidence, major translational barriers, and future development strategies for cellular immunotherapy in lung cancer.

Methods

PubMed, Web of Science, Embase, and ClinicalTrials.gov were searched from database inception to 1 June 2026 using Medical Subject Headings (MeSH) and free-text terms. Evidence on chimeric antigen receptor T-cell (CAR-T), T-cell receptor-engineered T-cell (TCR-T), tumor-infiltrating lymphocyte (TIL), natural killer (NK), chimeric antigen receptor natural killer (CAR-NK), and cancer vaccine-based strategies was reviewed, with priority given to non-small cell lung cancer (NSCLC) and small-cell lung cancer (SCLC).

Key Content and Findings

Most lung-cancer cellular therapy studies remain early phase. TIL therapy has demonstrated clinical activity in NSCLC, while EGFR, MSLN, CEA, MUC1, PD-L1, B7-H3, and DLL3 are representative targets under investigation for CAR-based approaches; NY-ESO-1 and KK-LC-1 are representative TCR-T targets. Major barriers include antigen escape, poor tumor infiltration, HLA restriction or loss, immunosuppressive signaling, cellular exhaustion, limited persistence, manufacturing burden, and treatment-related toxicity. Current optimization strategies include gene editing, chemokine-receptor engineering, armored or logic-gated receptors, safety switches, optimized lymphodepletion, and rational combinations with other anticancer therapies.

Conclusions

Cellular immunotherapy is a promising but still developing strategy for lung cancer. Future studies should prioritize biomarker-guided patient selection, safer and more persistent cell products, scalable manufacturing, and prospective clinical validation of rational combination strategies.

Introduction

Lung cancer remains a major cause of cancer-related morbidity and mortality worldwide (1,2). It is broadly classified into small-cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), with NSCLC representing the predominant subtype (2). Current treatment includes surgery, chemotherapy, radiotherapy, molecular targeted therapy, and immunotherapy (3). Although these approaches have improved outcomes in selected patients, prognosis remains unsatisfactory, with the five-year overall survival rate still below 20% in many reported populations (4,5). Many patients are also diagnosed at an advanced stage and lose the opportunity for radical surgery (6). Thus, novel strategies are needed for progressive, treatment-resistant, or recurrent disease.

Cellular immunotherapy enhances or redirects immune effector cells to recognize and eliminate malignant cells. Major platforms include chimeric antigen receptor T-cell (CAR-T) therapy, T-cell receptor-engineered T-cell (TCR-T) therapy, tumor-infiltrating lymphocyte (TIL) therapy, natural killer (NK) cell therapy, CAR-NK cell therapy, and cancer vaccine-based strategies. Their success in hematologic malignancies has stimulated increasing interest in solid tumors, including lung cancer.

Translation to lung cancer remains difficult because of marked tumor heterogeneity, variable antigen expression, abnormal vasculature, stromal barriers, insufficient immune-cell trafficking and infiltration, an immunosuppressive tumor microenvironment (TME), and limited cellular persistence. Potential on-target off-tumor toxicity is also important because several candidate antigens, including epidermal growth factor receptor (EGFR), mesothelin (MSLN), carcinoembryonic antigen (CEA), human epidermal growth factor receptor 2 (HER2), and programmed death-ligand 1 (PD-L1), may be expressed in normal tissues.

Strategies under investigation include armored cellular products, chemokine receptor modification, logic-gated CARs, inducible safety switches, gene editing, optimized lymphodepletion, and combinations with immune checkpoint inhibitors, radiotherapy, targeted therapy, antibody-drug conjugates, or cancer vaccines. Although early-phase studies in NSCLC and SCLC provide preliminary evidence, robust clinical validation remains limited.

This review integrates CAR-T, TCR-T, TIL, NK/CAR-NK, and cancer vaccine-based strategies in lung cancer, emphasizing representative targets, clinical evidence, safety, shared translational barriers, and future priorities. We present this article in accordance with the Narrative Review reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1627/rc).

Methods

This narrative review summarizes clinical evidence and translational progress in cellular immunotherapy for lung cancer. PubMed, Web of Science, Embase, and ClinicalTrials.gov were searched from database inception to 1 June 2026. The search strategy summary is presented in Table 1, with a detailed PubMed strategy in Table S1.

Table 1. The search strategy summary.

Items	Specification
Date of search	1 June 2026
Databases and other sources searched	PubMed, Web of Science, Embase, and ClinicalTrials.gov were searched. Reference lists of selected key articles and high-quality reviews were also reviewed when necessary
Search terms used	Search terms included combinations of MeSH terms and free-text terms related to lung cancer and cellular immunotherapy, including “lung cancer”, “Lung Neoplasms”, “non-small cell lung cancer”, “NSCLC”, “small-cell lung cancer”, “SCLC”, “cellular immunotherapy”, “adoptive cell therapy”, “CAR-T”, “chimeric antigen receptor T cell”, “TCR-T”, “T-cell receptor-engineered T cell”, “tumor-infiltrating lymphocyte”, “TIL”, “NK cell therapy”, “CAR-NK”, “cancer vaccine”, and “clinical trial”. A detailed PubMed search strategy is provided in Table S1
Timeframe	From database inception to 1 June 2026
Inclusion and exclusion criteria	Clinical trials, translational studies, mechanistically relevant preclinical studies, and high-quality reviews related to cellular immunotherapy for lung cancer were considered. Studies directly involving NSCLC or SCLC were prioritized. Evidence from other solid tumors was included only when it provided important platform-level insights into cellular therapy engineering, safety optimization, regulatory progress, or vaccine-cell therapy combinations. Articles not relevant to lung cancer, cellular immunotherapy, or the major therapeutic platforms discussed in this review were excluded
Selection process	Literature selection was primarily conducted by N.M. Key references and cited evidence were checked by both authors. Any uncertainty regarding article relevance was resolved through discussion and consensus
Additional considerations	This article is a narrative review rather than a systematic review or meta-analysis. Therefore, a comprehensive search of all literature in the field, formal risk-of-bias assessment, and quantitative synthesis were not performed. The included literature was narratively summarized according to therapeutic platform, target selection, clinical evidence, safety concerns, translational barriers, and future development strategies

NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer.

Search terms included MeSH and free-text terms for “lung cancer”, “Lung Neoplasms”, “non-small cell lung cancer”, “NSCLC”, “small-cell lung cancer”, “SCLC”, “cellular immunotherapy”, “adoptive cell therapy”, “CAR-T”, “chimeric antigen receptor T cell”, “TCR-T”, “T-cell receptor-engineered T cell”, “tumor-infiltrating lymphocyte”, “TIL”, “NK cell therapy”, “CAR-NK”, “cancer vaccine”, and “clinical trial”. Clinical trials, translational studies, mechanistically relevant preclinical studies, and high-quality reviews were considered. Priority was given to NSCLC- or SCLC-specific evidence; other solid-tumor studies were included only for relevant platform-level insights into engineering, safety, regulatory progress, or vaccine-cell combinations. Literature was screened by title/abstract and full text, with disagreements resolved by consensus. Because this was a narrative review, no formal quantitative synthesis was performed.

CAR-T cell therapy in lung cancer

CAR-T therapy genetically modifies T cells to express chimeric antigen receptors that recognize and eliminate cells bearing corresponding surface antigens (Figure 1). The approach has produced high objective and complete response rates in relapsed/refractory B-cell lymphoma, multiple myeloma, and follicular lymphoma (7,8). These successes have driven development in thoracic oncology, although solid-tumor barriers make lung-cancer translation more difficult.

Figure 1.

Schematic of cellular immunotherapy for lung cancer, including CAR-T, TCR-T, TILs, and NK cell therapies. This flowchart outlines four cellular immunotherapeutic strategies for lung cancer. CAR-T and TCR-T therapies begin with peripheral blood T cell collection, followed by genetic modification (via CAR or TCR) and ex vivo expansion. TIL therapy involves isolating and expanding lymphocytes from resected tumor tissue. NK/CAR-NK therapy utilizes NK cells from various sources, optionally engineered with CARs. Following ex vivo expansion, all cell products are infused to target tumors and overcome the immunosuppressive microenvironment. This figure was created using scientific elements from SciDraw and Bioicons, integrated and refined with Adobe Illustrator (version 29.4; Adobe Inc., San Jose, CA, USA). ADCC, antibody-dependent cellular cytotoxicity; CAR-NK, chimeric antigen receptor natural killer; CAR-T, chimeric antigen receptor T-cell; CEA, carcinoembryonic antigen; MDSC, myeloid-derived suppressor cell; EGFR, epidermal growth factor receptor; HLA, human leukocyte antigen; IL-2, interleukin-2; iPSCs, induced pluripotent stem cells; MSLN, mesothelin; MUC1, mucin 1; NK, natural killer; PB, peripheral blood; PD-L1, programmed cell death-ligand 1; TCR-T, T-cell receptor-engineered T-cell; TIL, tumor-infiltrating lymphocyte; TME, tumor microenvironment; Tregs, regulatory T cells.

Most lung-cancer CAR-T studies are Phase I or non-randomized Phase I/II trials, frequently embedded in heterogeneous solid-tumor cohorts. Only a limited number have reported lung-specific results, and clinical activity has generally been modest (9). Frequently investigated targets include EGFR, MSLN, CEA, mucin 1 (MUC1), PD-L1, B7-H3, and DLL3.

EGFR is a major NSCLC target, with approved inhibitors including gefitinib, afatinib, and osimertinib. In chemotherapy-refractory or relapsed EGFR-positive (>50%) NSCLC, EGFR-directed CAR-T cells incorporating 4-1BB and CD3ζ signaling showed EGFR-specific antitumor activity without unacceptable toxicity in NCT01869166 (10). A Phase I study (NCT03182816) further demonstrated the feasibility and safety of EGFR-specific CAR-T cells generated using the non-viral piggyBac system in advanced relapsed/refractory NSCLC (11). NCT05060796 and NCT04153799 are evaluating CXCR5-modified EGFR CAR-T cells to improve trafficking in advanced NSCLC.

MSLN is a glycosylphosphatidylinositol-anchored protein with limited normal-tissue expression and overexpression in mesothelioma, pancreatic cancer, and lung cancer (12,13). Lung-cancer studies have evaluated MSLN CAR-T cells and products incorporating programmed cell death protein 1 (PD-1)/cytotoxic T-lymphocyte-associated protein 4 (CTLA-4)-targeting components (NCT04489862, NCT06248697). In NCT03054298, low-dose MSLN CAR-T cells, with or without cyclophosphamide preconditioning, showed favorable safety and feasibility, whereas severe pulmonary toxicity occurred in the high-dose group. Safety strategies include an inducible caspase-9 system (NCT02414269) (14) and human leukocyte antigen (HLA)-gated designs (15); preclinical CCR2b/CCR4 co-expression has also improved migration and infiltration (16).

CEA is highly expressed in several epithelial malignancies, including lung cancer, and serum CEA has been associated with brain metastasis in advanced disease (17). CEA-directed CAR-T therapy has entered lung-cancer testing, including Phase I trials NCT02349724 and NCT04348643. Preclinically, CRISPR-mediated deletion of the E3 ubiquitin ligase Cbl-b delayed CAR-T exhaustion and increased IFN-γ and TNF-α secretion against CEA-positive tumors (18).

MUC1 is a transmembrane mucin that is frequently overexpressed and aberrantly glycosylated in lung cancer (19). MUC1-directed CAR-T cells have shown activity in MUC1-positive xenograft models (20), and anti-PD-1 blockade enhanced MUC1/PSCA CAR-T activity against NSCLC (21). NCT02587689, NCT03356808, and NCT04025216 are evaluating MUC1 CAR-T therapy in lung cancer, while NCT03525782 is testing MUC1 CAR-T cells combined with PD-1 knockout in advanced NSCLC.

PD-L1-directed CAR-T cells, using a single-chain variable fragment derived from atezolizumab, can eliminate PD-L1-positive NSCLC cells preclinically (22); dual-targeting designs may further improve activity and reduce escape (23). Clinically, NCT03330834 enrolled only one patient after severe cytokine release syndrome (CRS) following the first infusion, while NCT03060343 is evaluating PD-L1-positive relapsed/refractory NSCLC. CTLA-4 expression is also observed in NSCLC, although its prognostic value remains uncertain (24), and several Phase I/II programs are evaluating EGFR- or MSLN-directed CAR-T cells co-expressing anti-CTLA-4 or anti-PD-1 components. HER2 exon 20 alterations represent another potential NSCLC target (25), and glypican-3 (GPC3) overexpression has been associated with poor differentiation, metastasis, and prognosis (26). CD276/B7-H3 is an additional emerging CAR-T target (27); NCT03198052 and NCT04864821 are ongoing. In 2026, CRISPR knock-in B7-H3 CAR-T cells co-engineered with c-JUN enhanced killing of SCLC models, including tumors with low B7-H3 density, and supported Good Manufacturing Practice (GMP)-scale manufacturing (28).

For SCLC, delta-like ligand 3 (DLL3) is a particularly relevant lineage-associated target. DLL3-directed CAR-T strategies incorporating Tmod logic gating illustrate how disease-specific antigens can be paired with control circuits to improve malignant-versus-normal discrimination (29). Other engineering approaches include off-the-shelf products and armored cells, while inducible caspase-9 safety switches allow therapeutic cells to be eliminated after severe toxicity (30-32).

A central challenge is balancing antitumor potency with antigen specificity and systemic safety. Single-target therapy can promote antigen escape, while high-dose MSLN-directed products illustrate the risk of pulmonary toxicity. Current approaches include chemokine-receptor engineering such as CXCR5 or CCR2b to improve tumor homing and logic-gating systems such as Tmod to exploit tumor-specific loss of heterozygosity. Recent NSCLC preclinical work further showed that fifth-generation folate receptor alpha (FRα)-directed CAR-T cells engineered to secrete an anti-PD-L1 single-chain variable fragment enhanced expansion and cytotoxicity against FRα/PD-L1-positive lung-cancer cells, although further validation is required (33).

Target specificity and normal-tissue safety considerations

Target specificity is a central determinant of the safety of CAR-T therapy in solid tumors because many candidate surface antigens are tumor-associated rather than strictly tumor-specific. Even low-level expression in normal tissues may become clinically relevant when highly active or persistent CAR-T cells are administered. This issue is particularly important in lung cancer, where recognition of antigens expressed by normal bronchial, pulmonary epithelial, or mesothelial cells may result in potentially serious pulmonary toxicity.

c-MET illustrates the challenge associated with targeting a tumor-associated antigen that is not completely restricted to malignant cells. MET expression has been detected in normal human lung tissue and bronchial epithelium, although its expression may be increased in subsets of NSCLC (34). Therefore, c-MET-directed CAR-T cells carry a theoretical risk of on-target off-tumor recognition of normal epithelial tissues. Direct clinical safety data for c-MET CAR-T therapy in lung cancer remain limited. However, a phase 0 study using intratumorally administered, transient mRNA-engineered c-MET CAR-T cells in patients with metastatic breast cancer reported no treatment-related adverse events above grade 1 (35). A subsequent phase I study of intravenously administered RNA-electroporated c-MET-directed CAR-T cells in patients with metastatic melanoma or triple-negative breast cancer similarly reported no grade 3–5 treatment-related adverse events (36). These findings provide preliminary evidence that transient CAR expression may reduce normal-tissue toxicity, but they should not be directly extrapolated to persistently expressed c-MET CAR constructs in lung cancer. Careful assessment of antigen density in both tumor and normal pulmonary tissues therefore remains essential.

EGFRvIII may offer a more favorable target-specificity profile because it represents a tumor-associated mutant form of EGFR containing a novel junctional epitope that is not detectable in normal tissues, including normal lung. This distinguishes EGFRvIII from wild-type EGFR, which is expressed in multiple normal epithelial tissues and therefore carries a greater risk of on-target off-tumor toxicity. In a first-in-human study of intravenously administered EGFRvIII-directed CAR-T cells in patients with recurrent glioblastoma, no evidence of off-tumor toxicity or CRS was observed (37). Nevertheless, these safety data were obtained outside the lung-cancer setting, and the applicability of EGFRvIII-directed CAR-T therapy to lung cancer remains insufficiently established. Furthermore, the potential safety advantage of EGFRvIII does not eliminate efficacy-related challenges such as heterogeneous target expression and antigen loss (37).

MSLN represents a particularly relevant example of the trade-off between therapeutic activity and normal-tissue safety in thoracic malignancies. MSLN is highly expressed in several malignancies but is also physiologically expressed by mesothelial tissues, and low-level expression can occur in benign pulmonary epithelial cells. Importantly, a phase I study of highly active MSLN-directed CAR-T cells reported two cases of severe pulmonary toxicity after intravenous administration in a high-dose cohort (1–3 ×10⁸ cells/m²). Both patients developed progressive hypoxemia within 48 hours, and one patient ultimately experienced grade 5 respiratory failure. In the fatal case, postmortem examination demonstrated acute lung injury, extensive T-cell infiltration, and CAR-T-cell accumulation in the lungs, while molecular analyses detected low-level MSLN expression in benign pulmonary epithelial cells. These observations provide direct clinical evidence that low-level antigen expression in nonmalignant lung tissue can become clinically relevant when potent CAR-T products are administered at sufficiently high doses (38).

Collectively, these findings emphasize that target selection for lung cancer CAR-T therapy should consider not only the frequency and intensity of antigen expression in tumor cells but also its distribution and inducibility in normal pulmonary tissues. Potential approaches to reduce normal-tissue injury include transient CAR expression, regional delivery, affinity tuning, logic-gated receptor designs, inducible suicide switches, and cautious dose escalation. In addition, patients with pre-existing inflammatory or fibrotic lung disease may warrant particular attention when targeting antigens such as MSLN whose expression in benign pulmonary tissue may increase under pathological conditions.

TCR-T cell therapy in lung cancer

Similar to CAR-T cell therapy, TCR-T cell therapy involves the genetic modification of T cells to express T-cell receptors (TCRs) that specifically recognize tumor antigens, thereby directing T cells to target and eliminate tumors. In contrast to the HLA-independent recognition of CAR-T cells, TCR-T therapy necessitates that TCR-T cell recognition depends on intracellular antigens presented by antigen-presenting cells (APCs) via HLA molecules. This mechanism confers TCR-T cells the potential to target a broader repertoire of tumor antigens, demonstrating superior applicability in the treatment of solid tumors such as lung cancer (39) (Figure 1). Currently, numerous studies and clinical trials are underway to evaluate the anti-tumor efficacy and safety of TCR-T cell therapy in lung cancer.

In recent years, TCR-T therapy has demonstrated considerable potential in solid tumors. In a Phase I pilot open-label clinical trial (NCT02588612), autologous NY-ESO-1-specific TCR-engineered T cells (letetresgene autoleucel; GSK3377794) were evaluated in patients with advanced metastatic NSCLC whose tumors expressed NY-ESO-1/LAGE-1a and who had compatible HLA-A*02 alleles. The study enrolled 10 patients and evaluated the safety and efficacy of this approach following lymphodepleting chemotherapy. NY-ESO-1-specific TCR-T strategies had previously demonstrated clinical activity in synovial sarcoma, supporting further investigation of this therapeutic platform in lung cancer (40). Furthermore, NCT03029273 and NCT03709706 are evaluating NY-ESO-1-specific TCR-T cells in patients with progressive or recurrent NSCLC.

Beyond NY-ESO-1, KK-LC-1 has also been identified as a promising target antigen for lung cancer treatment (41). Several clinical studies are evaluating KK-LC-1-directed or related TCR-T strategies in advanced lung cancer and other solid tumors. NCT05035407 is a Phase I study evaluating KK-LC-1 TCR-T cells following non-myeloablative lymphodepletion with cyclophosphamide and fludarabine and supported by high-dose aldesleukin. NCT05483491 is another Phase I study evaluating KK-LC-1 TCR-T cells with aldesleukin in patients with metastatic KK-LC-1-positive cancers, including lung cancer. NCT03778814 is also evaluating TCR-T-cell immunotherapy in advanced lung cancer and other solid tumors, such as pancreatic cancer, synovial sarcoma, and advanced soft tissue sarcoma (42-44). These successes further bolster the developmental prospects of TCR-T cell therapy in the field of thoracic oncology.

The preliminary clinical successes of NY-ESO-1- and KK-LC-1-targeted therapies in advanced NSCLC have validated the efficacy of the TCR-T therapeutic pathway. However, the downregulation or loss of HLA expression, frequently observed in lung cancer, remains a primary bottleneck leading to therapeutic resistance. In recent years, breakthrough advancements in this field have focused on the development of personalized TCR-T therapies targeting driver gene mutations (such as KRAS G12D/C). Clinical data have confirmed that genetically engineered high-affinity TCRs can precisely recognize mutation-derived peptides presented by HLA, inducing durable objective responses in NSCLC patients who have previously failed targeted therapies or immune checkpoint inhibitors. Current optimization strategies include integrating single-cell sequencing technology to screen for dominant clones, and combining TCR-T with small-molecule targeted inhibitors or RNA vaccines to further enhance anti-tumor activity and promote the formation of immunological memory in vivo (45).

To facilitate the widespread clinical application of TCR-T therapy in lung cancer, future research must focus on several critical avenues. First, continued preclinical investigations are required to screen and identify TCR sequences characterized by high affinity and high specificity. Second, emphasis should be placed on developing personalized TCR-T products targeting driver gene mutations. Third, it is essential to explore combined strategies—such as the integration of epigenetic modulators—to reverse the downregulation or loss of HLA expression in tumor cells. These approaches aim to overcome current therapeutic bottlenecks, ultimately providing survival benefits to a broader population of lung cancer patients who have failed standard-of-care treatments.

TIL cell therapy in lung cancer

TIL therapy was clinically pioneered by Rosenberg and colleagues in the late 1980s. In a 1988 preliminary clinical study, 20 patients with metastatic melanoma received adoptive transfer of autologous TILs together with interleukin-2 (IL-2) after cyclophosphamide pretreatment. Objective tumor regression was observed in 11 of 20 patients (55%), providing early clinical evidence for the feasibility and antitumor activity of TIL-based adoptive cell therapy (46) (Figure 1).

To date, over 70 clinical trials have been initiated to investigate TIL therapy in lung cancer, demonstrating promising application prospects. In a single-arm, open-label Phase I study (NCT03215810), 20 patients with advanced NSCLC who progressed after nivolumab treatment received autologous TIL infusion; the results indicated that the therapy possessed a favorable safety profile and clinical activity (47). Furthermore, a recent clinical trial (NCT04614103) reported that TIL therapy achieved an objective response rate (ORR) of 26.1% and a disease control rate (DCR) of 82.6% in patients with NSCLC.

The primary advantage of TIL therapy lies in its inherent multi-target recognition capability, which enables it to induce durable remissions in NSCLC patients who have developed resistance to nivolumab. Nevertheless, its clinical translation is hindered by significant challenges, including the complexity of isolation and expansion techniques, rapid T-cell exhaustion post-infusion, metabolic restrictions within the TME, and insufficient in vivo persistence (48). To overcome these bottlenecks, various optimization strategies are being explored. At the cellular engineering level, CRISPR-mediated gene editing (e.g., knocking out CISH or Cbl-b genes) aims to bolster TIL resistance to functional exhaustion. A recently initiated clinical trial (NCT05566223) is currently evaluating the safety and efficacy of CISH-knockout engineered TILs in patients with metastatic NSCLC. At the therapeutic strategy level, administering lymphodepletion preconditioning prior to TIL infusion helps alleviate immunosuppression, thereby enhancing the proliferation and activity of TILs in vivo (49). The development of milder yet more potent lymphodepletion regimens (such as 60Cy/125Flu) represents a key research direction. Concurrently, pharmacological inhibition of the AKT signaling pathway during ex vivo expansion facilitates the enrichment of TILs with a memory phenotype, extending their survival post-infusion and augmenting anti-tumor immune responses (50). Culturing memory-phenotype TILs via AKT inhibition is a core strategy for prolonging their persistence in vivo.

Overall, TIL cell therapy holds promising prospects for the treatment of lung cancer. By integrating strategies such as gene editing, optimized preconditioning regimens, and refined cell culture processes, the therapeutic efficacy and clinical scope of TIL therapy are expected to expand further.

NK cell therapy in lung cancer

NK cells play a pivotal role in the innate immunosurveillance against viral, bacterial, and malignant threats. They are capable of rapidly eradicating aberrant cells through direct cytotoxicity and orchestrating the activation of other immune cells, thereby initiating adaptive immune responses (Figure 1). NK cells can be derived and expanded from a diverse range of sources, including peripheral blood mononuclear cells (PBMCs), umbilical cord blood (UCB), induced pluripotent stem cells (iPSCs), hematopoietic stem and progenitor cells (HSPCs), and immortalized cell lines such as NK-92. These diverse origins provide a sustainable and stable reservoir for NK-cell-based cancer immunotherapy (51,52).

NK cell therapy has been extensively explored in the clinical treatment of various malignancies. Results from a Phase I trial (NCT02030561) demonstrated that 22 patients with refractory HER2-positive solid tumors exhibited favorable tolerance and preliminary anti-tumor activity following treatment with trastuzumab combined with autologous peripheral blood-derived expanded NK cells (53). In the field of lung cancer, recent clinical studies have highlighted the potential of combining NK cell therapy with immune checkpoint inhibitors. For instance, the combination of anti-PD-1 antibodies and autologous NK cells for Stage IIIB/IIIC or IV NSCLC showed substantial anti-tumor activity and an acceptable safety profile (NCT03958097) (54). Additionally, the combination of allogeneic NK cells with cryoablation for advanced NSCLC has validated preliminary safety and efficacy (NCT02843815) (55). These attributes confer a superior safety profile in combination therapies. However, while CAR-NK cells have demonstrated anti-tumor effects in both preclinical and clinical studies for hematologic and solid tumors, clinical research in lung cancer remains in its infancy (e.g., NCT06454890, NCT05507593, NCT02839954). Similar to CAR-T therapy, CAR-NK therapy faces challenges such as the immunosuppressive TME, insufficient in vivo persistence, and antigen escape (56). A core obstacle currently is the short lifespan of NK cells post-infusion and their susceptibility to functional dysfunction (anergy) induced by the suppressive TME.

To address the aforementioned challenges, future optimization efforts are primarily focused on advanced engineering strategies. Utilizing iPSC technology to develop “off-the-shelf“ CAR-NK products holds significant potential for standardized production and cost control. Furthermore, “armored” modifications represent a critical pathway for enhancing therapeutic efficacy. For instance, genetically engineering CAR-NK cells to co-express cytokines such as IL-15 or IL-21 enables autocrine signaling, thereby reducing reliance on exogenous cytokines and effectively bolstering their survival and cytolytic functions within the complex lung cancer microenvironment. Recent preclinical evidence in NSCLC also showed that targeted integration of a humanized ROBO1 CAR into the PD-1 locus of NK-92 cells enhanced long-term cytotoxicity and inhibited tumor growth in a lung cancer xenograft model, providing another strategy to combine target recognition with resistance to inhibitory signaling (57).

Cancer vaccines and lung cancer

Cancer vaccines activate tumor-antigen-specific immunity and may be used alone or with other immunotherapies. In metastatic NSCLC, the Phase II trial NCT05077709 evaluated the indoleamine 2,3-dioxygenase (IDO)/PD-L1-targeting IO102-IO103 vaccine with pembrolizumab and reported a DCR exceeding 80% (58). In CLDN6-positive relapsed/refractory solid tumors, CLDN6-targeted CAR-T cells plus a CLDN6-encoding RNA vaccine achieved a DCR above 60% in NCT04503278 (59), supporting the concept that vaccination can amplify adoptively transferred cells.

Lung-cancer vaccine platforms include peptide/protein, DNA, vector-based, dendritic-cell (DC), RNA, and personalized neoantigen approaches.

Among peptide/protein vaccines, CIMAvax-EGF, MAGE-A3, NY-ESO-1, and the MUC1-targeting liposomal vaccine BLP25 have entered NSCLC investigation. CIMAvax-EGF is an epidermal growth factor (EGF)-P64 conjugate that induces anti-EGF antibodies and limits EGF-EGFR signaling (60). A Phase II randomized trial in stage IIIB/IV NSCLC after first-line chemotherapy demonstrated favorable safety and specific immune responses, supporting its potential in advanced disease.

DNA vaccines encode tumor antigens, sometimes together with immune-modulatory molecules. The pVAX1-MAGE-A3-sPD1 construct induced MAGE-A3-specific responses, inhibited tumor growth, and prolonged survival preclinically. In NCT00199849, a NY-ESO-1 DNA vaccine was evaluated in solid tumors including NSCLC and induced antigen-specific T-cell responses (61).

Vector-based vaccines use attenuated or replication-deficient vectors to deliver recombinant antigens (62). TG4010 uses modified vaccinia Ankara to express MUC1 and IL-2; Phase II studies in stage IIIB/IV NSCLC showed improved outcomes when combined with first-line chemotherapy (63,64). TG4050 represents a personalized neoantigen strategy delivering up to 30 patient-specific neoantigens. In the Phase I NCT04183166 study presented at AACR 2024, TG4050 elicited tumor-specific immune responses and showed potential to reduce recurrence in HPV-negative head and neck cancer. Although not lung-cancer-specific, this study provides platform-level support for personalized neoantigen vaccine development.

DC vaccines exploit professional APCs and can be loaded with tumor-derived or personalized antigens to induce tumor-specific T-cell responses (65). In a pilot study of personalized neoantigen-pulsed autologous DC vaccination in 12 heavily pretreated patients with metastatic lung cancer, all treatment-related adverse events were grade 1–2; the ORR was 25%, the DCR was 75%, median progression-free survival was 5.5 months, and median overall survival was 7.9 months (66). These findings support the feasibility and potential clinical activity of personalized DC vaccine strategies in advanced lung cancer.

Vaccine efficacy remains limited by antigen downregulation, loss of major histocompatibility complex expression, and an immunosuppressive TME (67). Current strategies therefore emphasize novel adjuvants and combinations with immune checkpoint inhibitors or cellular therapies; NCT05950139 is evaluating such a combination in advanced ALK-positive NSCLC (68). Lung-cancer vaccines are consequently moving from fixed-antigen platforms toward RNA, viral-vector, and personalized approaches. Because vaccine monotherapy may be insufficient to overcome tumor immune suppression, combinations that prime or amplify cellular responses are increasingly important. Non-lung vaccine studies are discussed here only as platform-level evidence for future lung-cancer development.

Table 2 summarizes the recognition mechanisms, representative targets or approaches, major advantages and limitations, potentially applicable populations, clinical development stages, and future translational priorities of the major cellular immunotherapy strategies.

Table 2. Comparative characteristics and clinical development of major cellular immunotherapy strategies in lung cancer.

Strategy	Recognition/therapeutic mechanism	Representative targets/approaches	Major advantages	Major limitations	Potentially applicable populations	Current clinical development in lung cancer
CAR-T	HLA-independent recognition of cell-surface antigens through engineered CARs	EGFR, MSLN, CEA, MUC1, PD-L1, B7-H3; DLL3 in SCLC	Direct antigen-specific cytotoxicity; flexible receptor engineering; logic-gated and armored designs feasible	Antigen heterogeneity; antigen escape; poor trafficking/infiltration; immunosuppressive TME; on-target off-tumor toxicity	Mainly patients with advanced or relapsed/refractory antigen-positive NSCLC; DLL3-directed approaches may be relevant to SCLC	Predominantly Phase I or early Phase I/II studies; preliminary feasibility and activity reported
TCR-T	HLA-dependent recognition of intracellular antigen-derived peptides presented by HLA molecules	NY-ESO-1, KK-LC-1; mutation-derived neoantigens such as KRAS-associated epitopes	Access to intracellular antigen repertoire; potential for highly specific targeting of tumor-associated or mutation-derived antigens	HLA restriction; HLA downregulation/loss; individualized target selection; potential T-cell exhaustion	Patients with advanced/recurrent NSCLC expressing defined antigens or neoantigens and compatible HLA alleles	Early clinical development, including Phase I/II and Phase II studies
TIL	Polyclonal recognition of naturally occurring tumor-associated and neoantigen-derived epitopes	Autologous tumor-reactive TILs	Multi-antigen recognition; lower dependence on a single predefined target; potential activity after checkpoint inhibitor resistance	Complex isolation and expansion; individualized manufacturing; lymphodepletion requirement; exhaustion and limited persistence	Selected patients with advanced NSCLC, particularly after progression on immune checkpoint inhibitors, when sufficient tumor tissue is available	Early-phase clinical studies with preliminary antitumor activity reported
NK cell/CAR-NK	Innate tumor recognition or CAR-mediated antigen-specific recognition	Autologous/allogeneic NK cells; iPSC-derived CAR-NK; engineered IL-15/IL-21-expressing CAR-NK	Favorable safety potential; lower risk of severe alloreactivity; potential for allogeneic/off-the-shelf manufacturing	Short in vivo persistence; TME-induced dysfunction; antigen escape for CAR-NK; limited lung-cancer-specific clinical evidence	Mainly patients with advanced NSCLC; potentially suitable for combination approaches and standardized allogeneic treatment	NK-cell combinations have entered clinical studies; CAR-NK development in lung cancer remains at an early investigational stage
Cancer vaccines/vaccine-cell combinations	Induction or amplification of tumor-antigen-specific immune responses	CIMAvax-EGF, MAGE-A3, NY-ESO-1, MUC1-based vaccines, DC vaccines; RNA vaccine-CAR-T combinations	Broad immune priming; potential induction of immune memory; may enhance cellular therapy persistence and activity	Antigen downregulation; impaired MHC expression; immunosuppressive TME; limited efficacy of vaccine monotherapy	Patients with antigen-defined or molecularly selected advanced/metastatic NSCLC; potentially useful as combination partners with cellular therapies or ICIs	Multiple vaccine platforms have reached clinical trials; vaccine-cell combinations remain emerging and investigational

B7-H3, B7 homolog 3; CAR-NK, chimeric antigen receptor natural killer; CAR-T, chimeric antigen receptor T-cell; CEA, carcinoembryonic antigen; DC, dendritic cell; DLL3, delta-like ligand 3; EGFR, epidermal growth factor receptor; HLA, human leukocyte antigen; ICI, immune checkpoint inhibitor; iPSC, induced pluripotent stem cell; MSLN, mesothelin; MUC1, mucin 1; NK, natural killer; NSCLC, non-small cell lung cancer; PD-L1, programmed death-ligand 1; SCLC, small-cell lung cancer; TCR-T, T-cell receptor-engineered T-cell; TIL, tumor-infiltrating lymphocyte; TME, tumor microenvironment.

Vaccine-cell therapy combination strategies

Beyond their role as standalone immunotherapies, cancer vaccines may also serve as immunological amplifiers for adoptively transferred immune cells. By providing sustained antigen presentation and stimulating antigen-specific immune responses in vivo, vaccine platforms may promote the expansion, persistence, and functional activity of transferred lymphocytes. Among current approaches, the combination of CAR-T cells with antigen-encoding RNA vaccines provides the most direct clinical proof of concept. In the phase 1 BNT211-01 trial, CLDN6-specific CAR-T cells were evaluated with or without a CLDN6-encoding CAR-T-cell-amplifying RNA vaccine (CARVac) in patients with relapsed or refractory CLDN6-positive solid tumors. Although this study was not specific to lung cancer, it demonstrated the feasibility of using vaccination to support CAR-T-cell expansion and persistence in vivo and therefore provides an important translational framework for future lung cancer studies (59).

Similar principles may also be relevant to TCR-T therapy. Because TCR-T cells recognize peptide-HLA complexes, vaccines encoding tumor-associated antigens or patient-specific neoantigens could theoretically provide repeated antigenic stimulation after cell infusion and thereby support the expansion and maintenance of transferred TCR-T cells. This concept is consistent with emerging strategies discussed in the TCR-T field, including combinations with RNA vaccines. However, direct clinical evidence for TCR-T-vaccine combinations in lung cancer remains limited, and their clinical value should therefore be considered investigational rather than established.

The relationship between vaccination and other cellular platforms is less well defined. Personalized neoantigen or dendritic-cell vaccines could potentially broaden endogenous antitumor T-cell responses and complement TIL-based treatment, whereas direct evidence supporting vaccine-NK or vaccine-CAR-NK combinations in lung cancer remains insufficient. Therefore, at present, vaccine-cell therapy combinations should be regarded primarily as an emerging strategy for improving immune-cell expansion, persistence, and antigenic breadth rather than as an established treatment paradigm in lung cancer. Further prospective studies are required to determine which cellular platforms, vaccine modalities, and treatment sequences provide clinically meaningful synergy.

Comparative overview of cellular immunotherapy strategies in lung cancer

The major cellular immunotherapy platforms for lung cancer differ substantially in their antigen-recognition mechanisms, target requirements, therapeutic advantages, limitations, candidate patient populations, and stages of clinical development. CAR-T therapy enables HLA-independent recognition of cell-surface antigens and allows extensive receptor engineering, but its application in lung cancer remains limited by antigen heterogeneity, insufficient tumor trafficking, and the risk of on-target off-tumor toxicity. TCR-T therapy expands the targetable antigen repertoire to intracellular tumor-associated or mutation-derived peptides, although its efficacy depends on appropriate HLA expression and may be compromised by HLA downregulation or loss. TIL therapy provides naturally occurring polyclonal tumor recognition and may therefore partially overcome single-antigen escape, but individualized manufacturing, ex vivo expansion, and limited cellular persistence remain important practical constraints. NK and CAR-NK therapies offer potential safety and manufacturing advantages, particularly for the development of allogeneic or off-the-shelf products, although their clinical evaluation in lung cancer remains at an early stage and insufficient in vivo persistence remains a major limitation. Cancer vaccines, although mechanistically distinct from adoptive cell therapies, can induce or amplify tumor-specific immune responses and may serve as complementary partners for cellular therapies, particularly by enhancing antigen-specific T-cell activation and persistence.

From a clinical perspective, these platforms are also being developed for different patient populations. Current CAR-T studies have predominantly enrolled patients with advanced or relapsed/refractory antigen-positive NSCLC, whereas DLL3-directed strategies may provide a potential avenue for SCLC. TCR-T therapy is particularly relevant to patients whose tumors express defined tumor antigens or mutation-derived epitopes in the context of compatible HLA alleles. TIL therapy may be particularly attractive for patients with advanced NSCLC who have progressed after immune checkpoint inhibitor therapy and from whom adequate tumor tissue can be obtained for TIL isolation and expansion. NK-cell-based approaches have mainly been investigated in advanced NSCLC, including combination strategies with immune checkpoint inhibitors or local therapies, while CAR-NK approaches remain largely investigational. Vaccine-based approaches have reached clinical evaluation in advanced or metastatic NSCLC and may be particularly valuable as combination partners rather than as standalone cellular immunotherapies. Overall, most cellular immunotherapy strategies for lung cancer remain in early-phase clinical development, and no single platform has yet established a definitive therapeutic advantage across unselected patients. Their major comparative characteristics are summarized in Table 2.

Shared translational challenges of cellular immunotherapy in lung cancer

Despite platform-specific differences, several barriers are shared across cellular immunotherapies: antigen heterogeneity and immune escape, an immunosuppressive TME, inadequate trafficking, limited persistence, and manufacturing and patient-selection constraints.

Inter- and intratumoral antigen heterogeneity can leave antigen-low or antigen-negative clones after CAR-T or CAR-NK therapy. TCR-T and vaccine activity can also be impaired by HLA downregulation or defective antigen presentation. Multi-antigen targeting, personalized target selection, and logic-gated recognition may reduce escape, although they add design and safety complexity.

The TME creates a second barrier through abnormal vasculature, stromal restriction, metabolic stress, and inhibitory signaling that reduce migration, infiltration, and effector function. Chemokine-receptor engineering, armored products, and combinations with immune checkpoint inhibitors are intended to improve tumor access and local activity, but their clinical benefit in lung cancer remains to be established.

Limited persistence and progressive exhaustion further restrict durability. Prolonged expansion can reduce TIL fitness, NK/CAR-NK cells often have relatively short in vivo persistence, and sustained antigen stimulation can impair CAR-T and TCR-T function. Strategies that preserve memory-like states, optimize lymphodepletion, modulate exhaustion pathways, or provide controlled cytokine support are therefore important.

Manufacturing and patient selection also constrain scalability. Autologous CAR-T, TCR-T, and TIL therapies require individualized collection, expansion, and quality control; TCR-T additionally requires HLA compatibility, and TIL therapy depends on adequate tumor tissue. Standardized allogeneic or off-the-shelf products may reduce logistical barriers, although durability and efficacy remain uncertain. Biomarker-guided selection and standardized manufacturing will be essential for reproducible clinical benefit.

These interconnected challenges require coordinated target selection, cellular engineering, biomarker-based stratification, manufacturing standards, and rational combinations.

Toxicity profiles and management of cellular immunotherapy

The safety profiles of cellular immunotherapies differ according to the immune-cell platform, target antigen, cellular engineering strategy, conditioning regimen, and route of administration. Major treatment-related toxicities may arise from excessive immune activation, recognition of antigens expressed in normal tissues, cross-reactivity of engineered receptors, or supportive treatments such as lymphodepletion and cytokine administration. Therefore, toxicity prevention, early recognition, and platform-specific management are essential components of the clinical development of cellular immunotherapy for lung cancer.

CAR-T-cell therapy is most prominently associated with CRS, immune effector cell-associated neurotoxicity syndrome (ICANS), and target-dependent on-target off-tumor toxicity. CRS and ICANS can be graded using the consensus criteria developed by the American Society for Transplantation and Cellular Therapy (ASTCT) (69). Management depends on severity and includes close clinical monitoring and supportive care; tocilizumab, an IL-6 receptor antagonist, is widely used for clinically significant CAR-T-cell-associated CRS, while corticosteroids play an important role in the management of severe or refractory CRS and remain a major therapeutic approach for ICANS (70). In lung cancer, target-dependent pulmonary injury requires particular attention because several candidate antigens may also be expressed at low levels in normal pulmonary tissues. In addition to careful antigen selection and dose escalation, safety-oriented engineering approaches such as transient CAR expression, logic-gated receptors, affinity optimization, and inducible suicide switches may help reduce severe toxicity. The target-specific pulmonary safety considerations for MSLN, c-MET, and EGFR/EGFRvIII are discussed above.

TCR-T-cell therapy presents a different spectrum of target-related risks. Because engineered TCRs recognize peptide-HLA complexes, toxicity may result not only from recognition of the intended antigen in normal tissues but also from unintended cross-reactivity with structurally similar peptides. Historical experience with affinity-enhanced MAGE-A3-specific TCRs demonstrated that cross-recognition of a titin-derived peptide could result in severe and fatal cardiac toxicity, illustrating the importance of extensive preclinical specificity screening before clinical translation (71). Strategies to reduce these risks include comprehensive assessment of antigen expression in normal tissues, evaluation of peptide cross-reactivity, conservative receptor-affinity optimization, cautious dose escalation, and incorporation of controllable safety mechanisms.

For TIL therapy, a substantial proportion of treatment toxicity may arise from the treatment regimen rather than from the infused TIL product alone. TIL-based protocols typically involve lymphodepleting chemotherapy and may include cytokine support after cell infusion. Clinical experience with lifileucel shows that its safety profile is strongly influenced by the expected toxicities of nonmyeloablative lymphodepletion and high-dose IL-2 support, with hematologic and infectious complications requiring careful monitoring and supportive care (72). This consideration is especially relevant to patients with lung cancer who may have limited pulmonary reserve following previous systemic therapy or thoracic radiotherapy.

NK- and CAR-NK-cell therapies may have a comparatively favorable acute toxicity profile. In an early clinical study of cord-blood-derived CD19-directed CAR-NK cells for lymphoid malignancies, CRS, neurotoxicity, and graft-versus-host disease were not observed, providing proof-of-concept for a potentially favorable safety profile (73). However, these data were generated in hematologic malignancies rather than lung cancer, and comparable safety cannot be assumed for all CAR-NK constructs or solid-tumor settings. Patients should therefore still undergo appropriate monitoring, particularly when CAR-NK products incorporate cytokine-supporting modifications or are administered after lymphodepletion.

Overall, toxicity management in lung cancer cellular immunotherapy should be considered a component of treatment design rather than merely a response to adverse events. Optimized target selection, pre-infusion risk assessment, careful dose escalation, standardized toxicity grading, early intervention for immune-mediated toxicities, and incorporation of controllable safety mechanisms will be essential for improving the therapeutic index of these emerging treatments.

Strengths and limitations

This review integrates major platforms and distinguishes lung-cancer-specific from platform-level evidence. Limitations include heterogeneous early-phase studies and non-quantitative narrative synthesis.

Conclusions

Cellular immunotherapy is developing unevenly across lung-cancer histologies. NSCLC has the broader evidence base, with CAR-T, TCR-T, TIL, NK-cell approaches, and cancer vaccines all entering clinical evaluation. CAR-T studies targeting EGFR, MSLN, PD-L1, B7-H3, and other antigens show feasibility but remain predominantly early-phase, with efficacy balanced against antigen escape and target-related toxicity. TCR-T expands the target repertoire to intracellular tumor-associated and mutation-derived antigens but remains constrained by HLA restriction and loss of antigen presentation. Among the approaches reviewed, TIL therapy currently provides one of the more direct lung-cancer-specific clinical signals, including activity after checkpoint-inhibitor failure and an ORR of 26.1% with a DCR of 82.6% in NCT04614103. NK-cell therapies and vaccines remain promising, but their optimal roles and combinations are not yet defined.

SCLC cellular immunotherapy is less mature and concentrated on a narrower range of disease-associated targets. DLL3-directed engineered-cell therapy is particularly relevant because DLL3 is linked to the neuroendocrine lineage and may support disease-specific targeting. Nevertheless, efficacy, durability, and safety require prospective validation. The development strategy for lung cancer should therefore be histology- and biomarker-guided rather than based on a single cellular platform for all patients.

For NSCLC, immediate priorities include further validation of TIL-based therapy in selected patients and optimization of engineered T-cell specificity, persistence, and trafficking. For SCLC, DLL3-directed cellular therapy warrants focused development. Across both histologies, personalized or multi-antigen targeting, improved cellular durability, and rational combinations are likely to be more informative than single-target cellular monotherapy.

Priority research questions for clinical translation

Several issues should receive priority in the next stage of research. First, target selection must balance tumor coverage with normal-tissue safety. Antigen-density assessment, multi-antigen recognition, logic-gated receptors, affinity tuning, and personalized targeting may reduce both on-target off-tumor injury and antigen escape, but each requires prospective validation.

Second, biomarker-defined patient selection should become central to clinical development. Histology, antigen density and heterogeneity, HLA background for TCR-based approaches, previous treatment exposure, tumor accessibility for TIL generation, and relevant tumor-microenvironment features may all influence feasibility and response. Prospective trials should incorporate biomarker-defined enrollment and correlative analyses.

Third, durable benefit requires improved tumor homing, persistence, and resistance to immunosuppression without simply escalating infused-cell dose. Engineering should preserve memory-like states, reduce exhaustion, and enhance selective tumor trafficking and local function.

Fourth, the field must progress from proof-of-concept studies to reproducible trials with clearly defined populations, standardized manufacturing and toxicity-management procedures, clinically meaningful endpoints, and adequate follow-up. Comparative or rational combination studies will be needed to determine incremental benefit over current standards.

The central goal is a coherent translational framework linking safe target selection, informed patient selection, durable cellular function, scalable manufacturing, and rigorous clinical validation.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.