Tertiary lymphoid structures predict the neoadjuvant chemoimmunotherapy response of stage III non-small cell lung cancer

Introduction

Lung cancer remains the leading cause of cancer-related mortality all over the world, with non-small cell lung cancer (NSCLC) accounting for 85% of cases. Approximately 20–25% of NSCLC patients present with stage III disease at diagnosis, a heterogeneous group with variable prognosis depending on resectability and nodal involvement (1). In the era of traditional therapy, the 5-year survival rate for stage III NSCLC was historically reported to be approximately 13% to 36% (1,2), and thus optimized therapeutic strategies are urgently needed. The management of stage III NSCLC has evolved from conventional current chemoradiotherapy (CRT) to incorporating immune checkpoint inhibitors (ICIs). For unresectable disease, the PACIFIC regimen (durvalumab consolidation post-CRT) remains the standard, demonstrating a 5-year overall survival (OS) of 42.9% (3). However, for resectable stage III NSCLC, neoadjuvant chemoimmunotherapy has emerged as a transformative strategy. The NADIM trial (4) reported a 5-year OS of 69.3% with nivolumab plus chemotherapy, while CheckMate 816 (5) confirmed improved pathological complete response (pCR) rates (24% vs. 2.2%) and event-free survival (EFS) (31.6 vs. 20.8 months) with neoadjuvant immunotherapy/chemotherapy. Despite promising prospects, several challenges persist in clinical application, with variable efficacy observed in neoadjuvant chemoimmunotherapy. A critical contributing factor is the lack of robust biomarkers for patient selection and response prediction.

In recent years, tertiary lymphoid structures (TLS) have emerged as a focal point in cancer research, with a growing number of studies demonstrating their close association with tumor development, therapy, and prognosis (6,7). TLS are ectopic lymphoid aggregates that develop in non-lymphoid tissues, including tumors, in response to chronic inflammation or persistent antigen exposure (8). Structurally resembling secondary lymphoid organs (SLO), TLS are organized into distinct T-cell and B-cell zones, often containing germinal centers (GC), dendritic cells, and high endothelial venules (HEV) (9,10). Unlike SLO, TLS arise de novo in pathological settings, such as cancer, autoimmune diseases, and chronic infections, reflecting sustained immune activation (10,11). In cancer, TLS are increasingly recognized as critical regulators of antitumor immunity. Over the past decades, advances in multiplex immunohistochemistry (mIHC) and single-cell RNA sequencing have elucidated their cellular composition and functional significance in cancer (12-14). The presence of TLS correlates with improved patient survival in multiple malignancies, including melanoma, lung cancer, and breast cancer, suggesting a protective role (13,15,16). TLS support intratumoral lymphocyte infiltration, enhance cytotoxic T-cell activity, and foster antibody production against tumor-associated antigens (7,17).

The immunological functions of TLS—particularly their role in sustaining T-cell memory and tertiary lymphoid neogenesis post-treatment—may critically influence long-term survival outcomes (18,19). While studies have established TLS as prognostic markers in early-stage NSCLC (6,20,21), their dynamic interaction with neoadjuvant chemoimmunotherapy in locally advanced disease remains poorly characterized. Addressing this research gap could refine patient stratification and identify mechanisms of treatment resistance.

This study aims to evaluate TLS as predictive biomarkers for therapeutic efficacy and survival in stage (IIIA/IIIB/IIIC) NSCLC patients undergoing neoadjuvant chemoimmunotherapy, with particular focus on their spatial organization, immune cell composition, and association with therapeutic response. We further attempt to develop an integrative TLS scoring model to quantify the predictive utility in this clinical context. We present this article in accordance with the REMARK and TRIPOD (22) reporting checklists (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0253/rc).

Methods

Study design and patients

A total of 102 patients with stage III (IIIA–IIIC) primary lung cancer who received neoadjuvant chemoimmunotherapy and subsequent surgery at Tianjin Medical University Cancer Institute and Hospital between January 2021 and December 2021 were recruited, tumor tissues were collected, and the presence of TLS was assessed by hematoxylin and eosin (H&E) staining. Patients who were TLS-negative were excluded. The eligibility criteria were as follows: (I) complete clinical data; (II) age 18–75 years; (III) pathologically confirmed NSCLC; (IV) TLS positive; (V) clinical stage III; (VI) surgical resection of R0; and (VII) received standardized neoadjuvant chemo-immunotherapy treatment. Finally, a total of 44 patients are recruited, the clinical characteristics of these patients are provided in Table 1. For the neoadjuvant chemoimmunotherapy regime, patients were administered two courses of neoadjuvant paclitaxel 175 mg/m² plus carboplatin [area under the curve (AUC) 5; 5 mg/mL per minute (min)] for squamous cell carcinoma and pemetrexed 500 mg/m² plus carboplatin for adenocarcinoma, with anti-programmed cell death protein 1 (PD-1) antibody administered by intravenous infusion at a recommended dose of 200 mg once every 3 weeks. Surgery was performed 4–8 weeks following the completion of two cycles of neoadjuvant therapy. A full dose of chemotherapy drugs was administered if the leukocyte count was within the normal range and a reduced dose if myelosuppression was present. All patients received four cycles of consolidation chemotherapy postoperatively, using the regimen described above. No other treatment was administered. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Tianjin Medical University Cancer Institute and Hospital. As this was a retrospective analysis of medical records obtained during previous clinical treatment, the requirement for individual informed consent was waived.

H&E staining

Tumor tissues were collected from the 102 patients. We cut the formalin-fixed paraffin-embedded (FFPE) NSCLC tissues into serial 4 µm sections and used four serial sections for H&E staining. The H&E staining was performed with the assistance of the department of Pathology in our hospital. The whole slides were scanned at ×10 magnification using an Olympus BX63 microscope (Tokyo, Japan). TLS is defined as aggregates of T cells, B cells and dendritic cells. In H&E staining they are identified as aggregates at least 50 lymphocytes (23). Subsequent analysis revealed the presence of TLS in 44 out of 102 analyzed tumor samples.

mIHC staining and multispectral analysis

mIHC was performed on tumor tissues from the 44-patient cohort to evaluate TLS using the PerkinElmer Opal 7-color kit, according to the manufacturer’s instructions. After dewaxing, rehydration, and antigen retrieval in Tris-ethylenediaminetetraacetic acid (EDTA) buffer (pH 9.0), sections were blocked and sequentially incubated overnight at 4 ℃ with the following primary antibodies: CD20 (1:500, #ab9475, Abcam, Abcam, Cambridge, UK), CD3 (1:150, #ab11089, Abcam), CD4 (1:700, #ab133616, Abcam), CD8 (1:1,000, #ab237709, Abcam), CD21 (1:200, #ab75985, Abcam), Bcl-6 (1:300, #89369S, CST, Danvers, MA, USA), PD-1 (1:300, #PA583301, Invitrogen, Carlsbad, CA, USA), programmed death-ligand 1 (PD-L1) (1:500, # MA527896, Invitrogen). For each antibody, incubation was followed by horseradish peroxidase (HRP)-conjugated secondary antibody application, tyramide signal amplification (TSA), and microwave-mediated antibody stripping. After sequential labeling of all targets, nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI). Two multiplex immunofluorescence staining panels were employed. Slides were imaged using a TissueFAXS Spectra system and quantified with StrataQuest software (TissueGnostics) according to previously described methods (24,25).

Under the supervision of two experienced pathologists, tumor regions and the tumor-stroma interface (extending 500 µm from the tumor nest) were delineated to localize TLS. Molecular expression and spatial relationships within a total of 406 analyzed fields (containing 174 TLS) were assessed using contextual tissue cytometry. A standardized algorithm with uniform thresholds was applied across all fluorescence channels. Spatial proximity was quantified by calculating the mean number of cells within a 30 µm radius from each reference cell’s centroid. TLS maturity was graded from 1 (naïve) to 3 (most mature) as reported in our previous work (14,24), TLS density (number per tissue area) and area were measured. Patients were stratified into high/low groups based on median values for area and density. PD-L1 proportion score (PS) was calculated as (PD-L1⁺ cells/total DAPI⁺ cells) ×100 and similarly stratified.

Pathologic assessment of response to neoadjuvant chemoimmunotherapy

Pathological response was assessed by two blinded pathologists in accordance with established methodologies (26,27). All H&E-stained sections were reviewed. Viable tumor, necrosis, and stroma were quantified on each slide (1–5 slides per patient; mean 2.7). The mean percentage of viable tumor relative to total tissue area across all slides determined the response. Major pathological response (MPR) was defined as ≤10% viable tumor.

Statistical analysis

All statistical analyses were performed with SPSS version 22.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism version 8.0.0 (GraphPad Software, San Diego, USA). Measurement data were expressed as mean ± standard deviation (SD), the t-test or Mann-Whitney U test was used to compare differences in the counting data. Survival distributions were estimated by the Kaplan-Meier method, and differences between groups were evaluated with the log-rank test. Associations between clinicopathological variables were assessed using χ² test. Progression-free survival (PFS) was calculated from the date of surgery to the date of documented tumor recurrence or distant metastasis. Univariable and multivariable Cox proportional hazard regression analysis were performed by the R version 4.3.3 (The R Foundation, Vienna, Austria) to identify predictors of prognosis, and variables with P<0.05 in multivariable Cox proportional hazard regression analyses were finally incorporated into the prognostic model. Furthermore, variables that were significant in the univariable analysis (P<0.05) were entered into the multivariable Cox regression model. In addition, PD-L1 expression was included a priori due to its established clinical relevance as a predictive biomarker for immunotherapy response in NSCLC. All statistical tests were two-sided, and P<0.05 was considered statistically significant.

Results

TLS are associated with early clinical stages and favorable prognosis

We first evaluated the features, maturity and spatial distribution of TLS in the NSCLC tissues, the representative images were showed in Figure 1. H&E staining was first performed on archival tumors to screen for TLS, which were recognized as circumscribed, densely packed lymphocytic aggregates (Figure 1A) (9,11). Only patients whose tumors contained TLS were retained for subsequent analyses; TLS-negative cases were excluded. mIHC staining was then applied to characterize both the maturation stage and the spatial distribution of TLS within the tumor microenvironment (TME). According to the mIHC results, TLS were stratified into three maturation stages. Stage I, designated early TLS (E-TLS), comprised densely packed lymphocytic infiltrates that lacked both follicular dendritic cell (FDC) networks (CD21 negative) and active germinal-center activity (Bcl-6 negative) (Figure 1B). Stage II, termed primary-follicle-like TLS (PFL-TLS), exhibited a central meshwork of CD21-positive FDC yet remained devoid of germinal-center reactions (Bcl-6 negative) (Figure 1C). Stage III, referred to as secondary-follicle-like TLS (SFL-TLS), displayed fully developed germinal-center reactions positive for both CD21 and Bcl-6 (Figure 1D). On the basis of this classification, patients were assigned to three ordinal grades: grade 1 tumors harbored exclusively E-TLS without any PFL- or SFL-TLS; grade 2 tumors contained E-TLS together with at least one PFL-TLS but no evidence of SFL-TLS; and grade 3 tumors encompassed at least one SFL-TLS within the neoplastic tissue (Table 2).

Figure 1 Identification of TLS features and maturity in stage III NSCLC. (A) Representative images of TLS in H&E staining (magnification, ×10 and ×20). (B-D) Representative images of TLS in mIHC staining (magnification, ×200). The slides were stained with CD3 (green), CD20 (cyan), CD21 (magenta), Bcl-6 (red), and DAPI (blue). E-TLS was considered as both FDC and Bcl-6 were negative (B). PFL-TLS were considered as FDC positive and Bcl-6 negative (C). SFL-TLS were considered as both FDC and Bcl-6 were positive (D). E-TLS, early TLS; FDC, follicular dendritic cell; H&E, hematoxylin and eosin; mIHC, multiplex immunohistochemistry; NSCLC, non-small cell lung cancer; PFL-TLS, primary-follicle-like TLS; SFL-TLS, secondary-follicle-like TLS; TLS, tertiary lymphoid structures.

We observed a significant difference in the maturity of TLS in tumor tissues among patients with different stages (Figure 2, Table 3). Patients with earlier T-stage exhibited higher TLS maturity, whereas those with more advanced T-stage showed less mature TLS (P=0.02). The relationship between clinical stage and TLS maturity followed a pattern similar to that of T stage (Figure 2C, Table 3), with patients at earlier clinical stages exhibiting higher TLS maturity in tumor tissues. However, possibly due to the limited sample size, this difference did not reach statistical significance (P=0.08). Interestingly, TLS maturity was not associated with N-stage (Figure 2B, Table 3), this may suggest that TLS have limited influence on lymphatic metastasis. To further elucidate the prognostic relevance of TLS, we evaluated the association between TLS characteristics and patient survival outcomes, results were showed in Figure 3. Kaplan-Meier survival analysis revealed that patients exhibiting higher TLS maturity, greater TLS density, larger TLS area, and a stromal TLS location had significantly prolonged PFS. Specifically, patients in the G3 TLS maturity group had a median PFS (mPFS) of 33.1 months, compared to 20.6 and 12.3 months for the G2 and G1 groups, respectively (P=0.004; Figure 3A). Similarly, the mPFS was 30.0 months for patients with higher TLS density versus 16.5 months for those with lower density (P=0.004; Figure 3B). Regarding TLS size, patients with a larger TLS area experienced a significantly longer mPFS of 33.1 months, compared to 23.5 months for patients with a smaller TLS area (P=0.006; Figure 3C). Furthermore, the mPFS was 27.9 months for patients with TLS located in the tumor stroma, in contrast to 20.6 months for patients with TLS in the tumor parenchyma (P=0.02; Figure 3D). Collectively, these results strongly indicate that TLS characteristics exert a considerable influence on clinical outcomes and may serve as valuable prognostic indicators for patient survival.

Figure 2 Distribution of patients according to TLS maturity across T stage (A), N stage (B), and clinical stage (C) of NSCLC. Patient numbers are annotated on each part. N, lymph node; NSCLC, non-small cell lung cancer; T, tumor; TLS, tertiary lymphoid structures.

Figure 3 Association of TLS status with stage III NSCLC patient survival. (A) Survival analysis based on TLS maturity. (B) Survival analysis based on TLS density. (C) Survival analysis based on TLS area. (D) Survival analysis based on TLS location. CI, confidence interval; G, grade; mo, months; mPFS, median PFS; NSCLC, non-small cell lung cancer; PFS, progression-free survival; TLS, tertiary lymphoid structures.

The TLS status significantly influence the response of neoadjuvant chemo-immunotherapy

We focus on the prognostic significance on neoadjuvant chemoimmunotherapy response of TLS. Based on the assessment of therapeutic response, patients with NSCLC who received neoadjuvant chemoimmunotherapy were stratified into the MPR and non-MPR groups. A significant association was observed between TLS characteristics and treatment response. Patients who achieved a MPR were characterized by a combined high TLS density and advanced maturity, suggesting that both metrics are salient features of an immune-supportive TME. The maturity and density, but not the area and the location, of TLS were significantly associated with neoadjuvant chemoimmunotherapy response (Figure 4A, Table 4). A mature, grade 3 TLS phenotype was predominant in the MPR group (61%), whereas the non-MPR group was characterized by less mature, grade 1 TLS (44%; grade 1 frequency: 8% vs. 44%, respectively, P=0.01). A parallel disparity was observed in TLS density: high TLS density was prevalent in the MPR group (64% vs. 36% with low density), a pattern that was inverted in the non-MPR group (33% high vs. 67% low density), P=0.03. As for the TLS area and location, observed were not statistically significant. We next characterized the immune cell composition within TLS across the different response groups. Comparative analysis revealed that TLS in the MPR group harbored significantly higher proportions of CD20⁺ B cells, CD3⁺ T cells, and CD8⁺ T cells compared to those in the non-MPR group. Notably, the proportion of CD4⁺ T cells was comparable between the two groups (Figure 4B). Given the established role of PD-1 as a key marker of T cell exhaustion, we next quantified the prevalence of PD-1-expressing T lymphocytes in the TME. However, our analysis revealed no significant differences in the density of either PD-1⁺ CD8⁺ or PD-1⁺ CD4⁺ T cells between the response groups (Figure 4B). These results suggest that TLS in patients with good response are characterized by both increased maturity, density and a greater abundance of potent cytotoxic T lymphocytes, pointing to critical structural and compositional differences underpinning the effective anti-tumor response. We next sought to determine the spatial distribution between T cells and CD20⁺ B cells inside TLS. The bivariate K(r) function (28) was used to characterize the spatial distributions and a search radius of 30 µm was applied for this calculation, as this distance is widely recognized for evaluating the spatial associations between two different cell populations (Figure 4C, Figure S1). The TLS microenvironment in MPR patients exhibited a specific enrichment of CD8⁺ T cells, but not CD4⁺ T cells in the immediate vicinity of CD20⁺ B cell clusters (Figure 4D). This distinct spatial proximity, which was significantly more pronounced than in non-MPR patients (P<0.001), implies a greater frequency of direct cell-cell interactions between these two immune populations. Given the central role of CD8⁺ T cells in tumor cell killing and the antigen-presenting capacity of B cells, their close spatial association likely facilitates the activation of CD8⁺ T cells. Consequently, enhanced CD8⁺ T cell-B cell interactions may promote a more robust anti-tumor immune response, representing a key mechanism underlying the improved clinical response to neoadjuvant therapy. In contrast, the spatial relationships between PD-1 positive CD8⁺ T or PD-1 positive CD4⁺ T cells and CD20⁺ B cells did not differ significantly between MPR and non-MPR patients (Figure 4D).

Figure 4 Association of TLS and TIL with response to neoadjuvant chemoimmunotherapy in stage III NSCLC. (A) Comparison of different TLS maturity, density, area and location percentage for stage III NSCLC patients who displayed MPR or non-MPR treated with neoadjuvant chemoimmunotherapy. (B) Comparison of TIL subsets for stage III NSCLC patients who displayed MPR or non-MPR treated with neoadjuvant chemoimmunotherapy. (C) Illustration of methodology for spatial analyses performed. Densities of cells of interest within a certain radius (30 μm) of a reference cell were calculated. (D) Density of TIL subsets in stage III NSCLC patients treated with neoadjuvant chemoimmunotherapy. Data are presented as means ± standard deviation. *, P<0.05; ***, P<0.01. MPR, major pathological response; ns, not significant according to an unpaired two-tailed Student’s t-test; NSCLC, non-small cell lung cancer; TIL, tumor-infiltrating lymphocyte; TLS, tertiary lymphoid structures.

To comprehensively assess the impact of TLS status on the efficacy and survival of NSCLC patients receiving neoadjuvant chemoimmunotherapy, we constructed a TLS status scoring system. Four key TLS metrics (maturity, density, area and location) were integrated by assigning weighted scores based on their individual states, with the total sum representing the final TLS status score (Figure 5A). A higher score indicates a more favorable TLS profile. Patients were stratified into TLS_poor, TLS_moderate, and TLS_favorable groups according to this score. Our analysis revealed a significant correlation between the TLS status score and treatment response. Patients with higher TLS status scores exhibited significantly better efficacy following neoadjuvant chemoimmunotherapy (P=0.01, Figure 5B). Furthermore, a higher TLS status score was associated with a significant increase in intratumoral infiltration of CD8⁺ T cells (P=0.02) and a concurrent decrease in PD1⁺CD8⁺ T cells (P=0.051, Figure 5C). This observation suggests that a favorable TLS status promotes the infiltration and potentially enhances the functionality of T cells within the tumor. Survival analysis revealed markedly divergent PFS outcomes among patients with different TLS statuses. Patients with a favorable TLS status had a mPFS of 33.1 months, compared to 20.6 and 12.3 months for those with moderate and poor status, respectively (P<0.001, Figure 5D).

Figure 5 Development and prognostic value of the TLS status score. (A) Schematic of the TLS status scoring system integrating four histological parameters: maturity, density, area, and location. Patients were stratified into three categories based on cumulative scores: poor (score 0–1), moderate (score 2–3), and favorable (score 4–5). For each group, the minimum and maximum total scores are shown; ellipses (…) represent intermediate possible scores within that group. This diagram is intended to convey the scoring logic and does not depict individual patients or all possible score combination. (B) Comparison of patients with different TLS status score percentage for NSCLC patients who displayed MPR or non-MPR treated with neoadjuvant chemoimmunotherapy. (C) Quantification of CD8⁺ and PD-1⁺CD8⁺ T cells infiltration across different TLS status groups. (D) Survival analysis based on TLS status score. (E) Curve analysis evaluating the diagnostic performance of the TLS status score for predicting patient prognosis. Data are presented as means ± standard deviation. *, P<0.05. AUC, area under the curve; CI, confidence interval; G, grade; MPR, major pathological response; ns, not significant according to an unpaired two-tailed Student’s t-test; NSCLC, non-small cell lung cancer; TLS, tertiary lymphoid structures.

To further quantify the predictive value of the TLS status score for patient prognosis, we performed receiver operating characteristic (ROC) curve analysis (Figure 5E). Patients with PFS greater than the mPFS were defined as having survival benefit, while others were defined as survival loss. The analysis demonstrated that the TLS status score had a substantial AUC value of 0.82 (P<0.001), indicating its accuracy in predicting post-treatment survival outcomes. An optimal cutoff value of 4 for the TLS status score was determined, yielding a sensitivity of 72.7% and a specificity of 81.8%. Given a survival benefit prevalence of 50% (22/44) in our cohort, the positive predictive value (PPV) was 80.0% and the negative predictive value (NPV) was 75.0%, indicating that the scoring system has favorable utility for both identifying patients likely to benefit from neoadjuvant chemoimmunotherapy and excluding those who may not. In conclusion, these results suggests that a TLS status score greater than 4 can effectively predict survival benefit in NSCLC patients undergoing neoadjuvant chemoimmunotherapy (P<0.001).

To further exclude the potential confounding effect of clinical stage on survival loss and TLS status in our results, we conducted a chi-square test to examine the association of survival loss with clinical stage and TLS status (Table S1). Results showed that survival loss was significantly associated with TLS status (P=0.001) but not with clinical stage (P=0.38). A subgroup survival analysis restricted to patients with favorable TLS status showed that no significant difference was observed between stage IIIA and stage IIIB patients (P=0.16, no stage IIIC patients were included in this cohort, Figure S2). These findings confirmed that the predictive value of TLS status is independent of clinical stage.

The PD-1 and PD-L1 expression after neoadjuvant chemoimmunotherapy related to TLS

To better elucidate the TME after neoadjuvant therapies, we employed mIHC to investigate the expression of PD-1 and PD-L1 in NSCLC. Data indicated that both PD-1 and PD-L1 was expressed in TME and TLS, the representative images were provided in Figure 6. We assessed the PD-L1 expression in the TME using PS system, based on the PS of PD-L1, patients were stratified into high- and low-PS groups. Our research revealed a statistically significant association between PD-L1 PS and clinical stage in NSCLC, where earlier-stage disease was inversely associated with PD-L1 expression levels (Figure 7A). Furthermore, PD-L1 PS held prognostic value for therapeutic response; patients who obtained a MPR following neoadjuvant chemoimmunotherapy exhibited significantly lower PD-L1 PS compared to non-MPR (Figure 7B). Conversely, assessment of survival outcomes indicated that PD-L1 PS was not a predictor of PFS, the mPFS was 27.9 months in the PD-L1 PS^high group versus 25.2 months in the PD-L1 PS^low group (P=0.74, Figure 7C), no significant difference was observed. We next investigated the potential association between PD-L1 expression and TLS status. Comprehensive analysis revealed that TLS maturity exhibited a significant inverse correlation with PD-L1 PS, indicating that patients with more mature TLS structures generally demonstrated lower PD-L1 expression levels. In contrast, PD-L1 expression showed only non-significant correlations with both TLS density, area and location within the TME (Figure 7D). To determine whether the inverse correlation between TLS maturity and PD-L1 expression is independent of clinical stage, we performed stage-stratified analyses (Figure S3). In stage IIIA patients (n=30), an inverse trend was observed between TLS maturity and PD-L1 expression (P=0.07). In stage IIIB (n=10) and stage IIIC (n=4) patients, sample sizes were too small to draw meaningful conclusions. These stratified results suggest that the overall inverse correlation is not simply driven by stage imbalance, although larger cohorts are needed for definitive multivariable analysis. These suggest that TLS maturity, rather than their density or area, may possess the capacity to modulate the immune checkpoint expression in the TME. To further investigate the relationship between TLS status and the immune landscape of the TME, we analyzed the prevalence of PD-1 positive T cells. We found that the frequency of PD-1⁺CD8⁺ T cells was inversely correlated with TLS maturity; patients with more mature TLS had a lower proportion of PD-1⁺CD8⁺ T cells (Figure 7E). Notably, the area or density of TLS did not influence PD-1⁺CD8⁺ T cell levels. In contrast, the frequency of PD-1⁺CD4⁺ T cells remained relatively stable and showed no significant association with TLS features (Figure 7F). Collectively, these results indicate that TLS maturation, rather than its mere presence or size, is linked to a more favorable immune contexture within the TME, suggesting a state of reduced T cell dysfunction and more positive anti-tumor immunity. To provide a more comprehensive evaluation of TLS-related factors, Cox proportional hazards regression analyses were performed (Table 5). In the univariable model, more mature TLS (G3; HR =0.19, P=0.002) and a more favorable TLS status score (HR =0.12, P<0.001) were identified as significant protective factors for survival. Conversely, lower TLS density (HR =7.58, P<0.001) and smaller TLS area (HR =2.95, P=0.01) were significant risk factors. In the multivariable model, which adjusted for potential confounders including PD-L1 expression, lower TLS density (HR =15.8, P=0.01) and smaller TLS area (HR =12.8, P=0.01) remained independent risk factors for poorer survival. Notably, while TLS maturity and TLS status score showed a trend toward protective effects (HR =0.30 and 0.25, respectively), they did not retain statistical significance after multivariable adjustment, suggesting that their prognostic value may be partially confounded by other covariates—particularly TLS density and area, which were closely correlated with maturity. These findings indicate that among all TLS characteristics, density and area are the most robust independent predictors of survival in this patient cohort.

Figure 6 Characterization of the peri-TLS immune microenvironment and PD-1/PD-L1 expression in NSCLC in mIHC staining. Magnification, ×200. mIHC, multiplex immunohistochemistry; NSCLC, non-small cell lung cancer; PD-1, programmed cell death protein 1; PD-L1, programmed death-ligand 1; TLS, tertiary lymphoid structures.

Figure 7 Association of PD-L1 expression with clinical outcomes and TLS characteristics in stage III NSCLC patients receiving neoadjuvant chemoimmunotherapy. Correlation of PD-L1 PS with clinical stage (A), therapeutic response (B), survival (C) and TLS status (D) in stage III NSCLC treated with neoadjuvant chemoimmunotherapy. Association of PD-1-expressing CD8⁺ T cells (E) and PD-1-expressing CD4⁺ T cells (F) with TLS status in stage III NSCLC treated with neoadjuvant chemoimmunotherapy. *, P<0.5; ***, P<0.01. CI, confidence interval; G, grade; mo, months; mPFS, median PFS; ns, not significant according to an unpaired two-tailed Student’s t-test.; NSCLC, non-small cell lung cancer; PD-1, programmed cell death protein 1; PD-L1, programmed death ligand 1; PFS, progression-free survival; PS, proportion score; TLS, tertiary lymphoid structures.

Discussion

Stage III NSCLC represents a highly heterogeneous and clinically challenging disease entity, the prognosis remains suboptimal. A pivotal shift in the therapeutic landscape has been the integration of ICI in both resectable and unresectable NSCLC (29-31). While consolidation immunotherapy with durvalumab following CRT has become the standard of care for unresectable disease, there is growing emphasis on the potential of neoadjuvant chemoimmunotherapy—administering ICI with platinum-based chemotherapy prior to surgery (30). This approach aims to enhance pathological responses and improve long-term survival by targeting micrometastatic disease early and eliciting a robust, systemic anti-tumor immune response. Major clinical trials have demonstrated that neoadjuvant chemoimmunotherapy significantly increases rates of MPR and pCR, establishing its importance in the current treatment armamentarium for resectable stage III NSCLC (32). However, a significant clinical challenge persists: the efficacy of neoadjuvant chemoimmunotherapy exhibits considerable inter-patient heterogeneity. A substantial proportion of patients derive minimal benefit, experiencing considerable treatment-related toxicity without a corresponding pathological response, thus underscoring the limitations of current patient selection criteria (33,34). Consequently, there is an urgent, unmet need to identify and validate novel biomarkers to accurately assess and predict therapeutic efficacy. The discovery of such biomarkers, potentially derived from the tumor immune microenvironment, peripheral blood, or radiomic features, is critical for refining patient stratification, maximizing therapeutic benefit, and ultimately improving survival outcomes in this complex patient population.

TLS have emerged as a biomarker for predicting treatment response and prognosis in NSCLC due to their crucial role within the TME. Functionally, TLS serve as local hubs for anti-tumor immunity by supporting the in-situ priming, expansion, and differentiation of tumor-specific T and B cells (35). This local activity is crucial for generating a robust, targeted immune response directly within the TME, potentially bypassing the need for systemic lymphocyte activation (12). Recently, TLS have emerged as a specific positive biomarker for response to immunotherapy, particularly ICI. Their density and maturity often predict better treatment outcomes (7). In renal cell carcinoma, TLS-positive tumors treated with ICI exhibited IgG-producing plasma cells and apoptotic malignant cells, correlating with better survival (36). Similarly, TLS promote apoptosis of malignant cells and enhance immunotherapy response in nasopharyngeal carcinoma (37). In high-grade serous ovarian cancer, developed TLS associate with increased immune cell activity and prognostic benefit (38). Therapeutic strategies to induce or mature TLS are being explored to convert “cold” tumors into “hot” ones. In our study of resectable stage III NSCLC, the presence of mature TLS was significantly associated with superior response to neoadjuvant chemoimmunotherapy. Our proposed TLS status score demonstrated robust predictive performance for PFS, with an AUC of 0.82 and a well-balanced sensitivity (72.7%) and specificity (81.8%). The calculated PPV of 80.0% and NPV of 75.0% further support the clinical utility of this scoring system for identifying patients likely to benefit from neoadjuvant chemoimmunotherapy in stage III NSCLC. Nevertheless, given the relatively modest sample size of this retrospective study, these performance estimates should be interpreted with caution. External validation in larger, independent prospective cohorts is warranted to confirm the clinical utility of this TLS scoring system. In conclusion, patients with mature TLS exhibited significantly improved survival outcomes, identifying TLS maturity as a promising biomarker for both treatment efficacy and prognosis. Our findings are broadly consistent with two recent studies investigating TLS in neoadjuvant chemoimmunotherapy for NSCLC. The NADIM trial demonstrated that neoadjuvant chemoimmunotherapy induces TLS formation, with complete responders exhibiting more functional mature TLS and enhanced antigen presentation and humoral responses (39). Similarly, another study reported that neoadjuvant chemoimmunotherapy-treated NSCLCs showed higher TLS maturation and abundance, both significantly correlating with MPR and DFS (40). Our study complements and extends these observations in stage III NSCLC by providing detailed histopathological characterization of TLS (maturity, density, area, location, and spatial distribution) integrated into a composite scoring system, and by demonstrating that superior TLS features correlate with improved PFS while TLS maturity inversely correlates with PD-L1 expression. Collectively, these converging lines of evidence support TLS as a robust predictive biomarker for neoadjuvant chemoimmunotherapy response in resectable NSCLC.

Additionally, we observed a significant inverse correlation between mature TLS and PD-L1 expression in the TME following neoadjuvant chemoimmunotherapy, suggesting that well-developed TLS may modulate adaptive immune resistance. PD-L1 is a key immune checkpoint protein that binds to PD-1 on activated T cells, transmitting an inhibitory signal that leads to T cell exhaustion (41,42). The significance of PD-L1 expression is twofold. Firstly, it serves as a key biomarker of an adaptive immune resistance mechanism. Tumors upregulate PD-L1 specifically in response to inflammatory signals, particularly interferon-gamma (IFN-γ) released by tumor-infiltrating lymphocytes, effectively creating a shield against the ongoing immune attack. Secondly, the level of PD-L1 expression has profound implications for cancer immunotherapy (42,43). The development of ICI, particularly monoclonal antibodies that block the PD-1/PD-L1 axis, represents a landmark advancement in oncology. These therapies reinvigorate the exhausted T cells, restoring their capacity to recognize and eliminate tumor cells. Consequently, PD-L1 expression levels are often used as a predictive biomarker to help identify patients who are more likely to benefit from the immunotherapy (44).

However, the influence of TLS on PD-L1 expression within the TME remains largely unexplored, it was suggested that TLS and PD-L1 expression may be independent predictive factors in research in 2021 (45). Interestingly, our results in stage III NSCLC reveal a potential link between TLS and PD-L1 expression following neoadjuvant chemoimmunotherapy. Specifically, we observed an inverse correlation, wherein the presence of mature TLS was associated with relatively lower levels of PD-L1 in the TME. We hypothesize that this phenomenon may be attributed to the role of mature TLS as organized anti-tumor “field headquarters” (46). These structures are enriched with a multitude of T cells and other effector immune cells. We postulate that these activated lymphocytes can directly recognize and eliminate PD-L1-positive tumor cells through potent cytotoxic effects, effectively clearing these immunosuppressive variants before they have the opportunity to become widely established. This direct cytolytic activity could thereby reduce the overall detectable PD-L1 burden in the TME. This mechanism provides a plausible explanation for our previously reported finding that NSCLC patients with mature TLS experience superior clinical outcomes and prolonged survival after neoadjuvant chemoimmunotherapy, as the very structure that fosters a robust immune response may also preemptively counteract a key resistance pathway. Moreover, the inverse correlation between TLS maturity and PD-L1 expression, along with their association with tumor stage, suggests that early-stage tumors harbor a less immunosuppressive microenvironment that supports TLS maturation, whereas advanced tumors exhibit stress-driven PD-L1 upregulation and impaired TLS formation. These findings imply that TLS maturity and PD-L1 expression may be jointly shaped by the evolving TME rather than directly linked causally. Further mechanistic studies are warranted to delineate this relationship.

A key limitation of this study is the lack of insight into the molecular mechanisms governing the interplay between TLS and PD-L1 expression. Elucidating these potential pathways, including cytokine dynamics, specific TLS immune cell subsets, and other co-inhibitory molecules, is a crucial goal for future research. A deeper understanding of this crosstalk is essential for developing novel strategies to potentiate anti-tumor immunity. Furthermore, all TLS analyses were performed on post-treatment surgical specimens, and whether pre-treatment TLS status can similarly predict response remains unknown. To translate these observations into a clinically actionable tool for treatment selection, future studies should focus on developing non-invasive approaches—such as radiomics or liquid biopsies—or prospectively validate TLS assessment using pre-treatment biopsy samples. For the TLS scoring system, our findings are hypothesis-generating and still require independent external validation.

Collectively, our study elucidates the prognostic and predictive roles of TLS in patients with stage III (IIIA–IIIC) NSCLC undergoing neoadjuvant chemoimmunotherapy. A TLS status scoring model was developed, wherein a score of 4 emerged as a strong independent predictor of better survival. Further investigation is warranted to elucidate the role of TLS in predicting therapeutic response and its potential mechanistic influence on PD-L1 expression within the TME.

Conclusions

TLS status, encompassing maturity, density, area, and stromal location, serves as a predictive biomarker for neoadjuvant chemoimmunotherapy response in resectable stage III NSCLC. The composite TLS score robustly identifies patients likely to achieve superior PFS and MPR. The inverse correlation between TLS maturity and PD-L1 expression suggests TLS modulate the immune TME through mechanisms partially independent of the PD-1/PD-L1 axis. Routine TLS assessment may optimize patient selection and guide personalized therapeutic strategies in this setting.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the REMARK and TRIPOD reporting checklists. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0253/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0253/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0253/prf

Funding: This study was supported by Tianjin Municipal Education Commission (No. 2024ZXZD008), National Natural Science Foundation of China (Nos. 82303196 and 82460580), and Tianjin Key Medical Discipline (Specialty) Construction Project (No. TJYXZDXK-009A).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0253/coif). The authors have no conflicts of interest to declare.

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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Tianjin Medical University Cancer Institute and Hospital. As this was a retrospective analysis of medical records obtained during previous clinical treatment, the requirement for individual informed consent was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Goldstraw P, Chansky K, Crowley J, et al. The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2016;11:39-51. [Crossref] [PubMed]
2. Liu SY, Wu YL. Unmet Clinical Demand for Patients With Unresectable Stage III NSCLC Having Actionable Genetic Alterations. J Thorac Oncol 2021;16:712-4. [Crossref] [PubMed]
3. Spigel DR, Faivre-Finn C, Gray JE, et al. Five-Year Survival Outcomes From the PACIFIC Trial: Durvalumab After Chemoradiotherapy in Stage III Non-Small-Cell Lung Cancer. J Clin Oncol 2022;40:1301-11. [Crossref] [PubMed]
4. Provencio M, Nadal E, Insa A, et al. Perioperative chemotherapy and nivolumab in non-small-cell lung cancer (NADIM): 5-year clinical outcomes from a multicentre, single-arm, phase 2 trial. Lancet Oncol 2024;25:1453-64. [Crossref] [PubMed]
5. Forde PM, Spicer J, Lu S, et al. Neoadjuvant Nivolumab plus Chemotherapy in Resectable Lung Cancer. N Engl J Med 2022;386:1973-85. [Crossref] [PubMed]
6. Giatromanolaki A, Chatzipantelis P, Contrafouris CA, et al. Tertiary Lymphoid Structures, Immune Response, and Prognostic Relevance in Non-Small Cell Lung Cancer. Cancer Invest 2023;41:48-57. [Crossref] [PubMed]
7. Peyraud F, Guegan JP, Vanhersecke L, et al. Tertiary lymphoid structures and cancer immunotherapy: From bench to bedside. Med 2025;6:100546. [Crossref] [PubMed]
8. Sautès-Fridman C, Petitprez F, Calderaro J, et al. Tertiary lymphoid structures in the era of cancer immunotherapy. Nat Rev Cancer 2019;19:307-25. [Crossref] [PubMed]
9. Fridman WH, Meylan M, Pupier G, et al. Tertiary lymphoid structures and B cells: An intratumoral immunity cycle. Immunity 2023;56:2254-69. [Crossref] [PubMed]
10. Dong Y, Wang T, Wu H. Tertiary lymphoid structures in autoimmune diseases. Front Immunol 2024;14:1322035.
11. Wang S, Wang H, Li C, et al. Tertiary lymphoid structures in cancer: immune mechanisms and clinical implications. MedComm (2020) 2024;5:e489.
12. Schumacher TN, Thommen DS. Tertiary lymphoid structures in cancer. Science 2022;375:eabf9419. [Crossref] [PubMed]
13. He M, He Q, Cai X, et al. Intratumoral tertiary lymphoid structure (TLS) maturation is influenced by draining lymph nodes of lung cancer. J Immunother Cancer 2023;11:e005539. [Crossref] [PubMed]
14. Luo J, Shi X, Liu Y, et al. Immune checkpoint ligands expressed on mature high endothelial venules predict poor prognosis of NSCLC: have a relationship with CD8(+) T lymphocytes infiltration. Front Immunol 2024;15:1302761. [Crossref] [PubMed]
15. Cabrita R, Lauss M, Sanna A, et al. Tertiary lymphoid structures improve immunotherapy and survival in melanoma. Nature 2020;577:561-5. [Crossref] [PubMed]
16. Wang Q, Sun K, Liu R, et al. Single-cell transcriptome sequencing of B-cell heterogeneity and tertiary lymphoid structure predicts breast cancer prognosis and neoadjuvant therapy efficacy. Clin Transl Med 2023;13:e1346. [Crossref] [PubMed]
17. Schumacher TN, Schreiber RD. Neoantigens in cancer immunotherapy. Science 2015;348:69-74. [Crossref] [PubMed]
18. Bindea G, Mlecnik B, Tosolini M, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 2013;39:782-95. [Crossref] [PubMed]
19. Bruni D, Angell HK, Galon J. The immune contexture and Immunoscore in cancer prognosis and therapeutic efficacy. Nat Rev Cancer 2020;20:662-80. [Crossref] [PubMed]
20. Xin S, Wen S, He P, et al. Density of tertiary lymphoid structures and their correlation with prognosis in non-small cell lung cancer. Front Immunol 2024;15:1423775. [Crossref] [PubMed]
21. Liu YM, Zhao K, Xie M, et al. Density and maturity ratio of tertiary lymphoid structures in stage II-III non-small cell lung cancer predict postoperative recurrence risk. Transl Lung Cancer Res 2025;14:3349-62. [Crossref] [PubMed]
22. Collins GS, Reitsma JB, Altman DG, et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ 2015;350:g7594. [Crossref] [PubMed]
23. KoksoyAASalvatierraME; MOSAIC Team, Solis Soto LM, Haymaker C. TLS_Finder: An algorithm for Identifying Tertiary Lymphoid Structures Using Immune Cell Spatial Coordinates. Preprint. bioRxiv 2024;2024.12.26.630405.
24. Luo J, Pang S, Hui Z, et al. Blocking Tim-3 enhances the anti-tumor immunity of STING agonist ADU-S100 by unleashing CD4(+) T cells through regulating type 2 conventional dendritic cells. Theranostics 2023;13:4836-57. [Crossref] [PubMed]
25. Wang T, Zhang J, Li N, et al. Spatial distribution and functional analysis define the action pathway of Tim-3/Tim-3 ligands in tumor development. Mol Ther 2022;30:1135-48. [Crossref] [PubMed]
26. Hellmann MD, Chaft JE, William WN Jr, et al. Pathological response after neoadjuvant chemotherapy in resectable non-small-cell lung cancers: proposal for the use of major pathological response as a surrogate endpoint. Lancet Oncol 2014;15:e42-50. [Crossref] [PubMed]
27. Travis WD, Dacic S, Wistuba I, et al. IASLC Multidisciplinary Recommendations for Pathologic Assessment of Lung Cancer Resection Specimens After Neoadjuvant Therapy. J Thorac Oncol 2020;15:709-40. [Crossref] [PubMed]
28. Carstens JL, Correa de Sampaio P, Yang D, et al. Spatial computation of intratumoral T cells correlates with survival of patients with pancreatic cancer. Nat Commun 2017;8:15095. [Crossref] [PubMed]
29. Garassino MC, Mazieres J, Reck M, et al. Durvalumab After Sequential Chemoradiotherapy in Stage III, Unresectable NSCLC: The Phase 2 PACIFIC-6 Trial. J Thorac Oncol 2022;17:1415-27. [Crossref] [PubMed]
30. Hamada A, Soh J, Hata A, et al. Neoadjuvant Concurrent Chemo-Immuno-Radiation Therapy Followed by Surgery and Adjuvant Immunotherapy for Resectable Stage III N2 NSCLC: Primary Results From the SQUAT Trial (WJOG 12119L). J Thorac Oncol 2025;20:1098-107. [Crossref] [PubMed]
31. Orosz Z, Kovács Á. The role of chemoradiotherapy and immunotherapy in stage III NSCLC. Pathol Oncol Res 2024;30:1611716. [Crossref] [PubMed]
32. Ricciuti B, Fusco F, Cooper A, et al. Neoadjuvant PD-1 and PD-L1 Blockade With Chemotherapy for Borderline Resectable and Unresectable Stage III Non-Small Cell Lung Cancer. JAMA Oncol 2025;11:735-41. [Crossref] [PubMed]
33. Uprety D, Mandrekar SJ, Wigle D, et al. Neoadjuvant Immunotherapy for NSCLC: Current Concepts and Future Approaches. J Thorac Oncol 2020;15:1281-97. [Crossref] [PubMed]
34. West HJ, Kim JY. Rapid Advances in Resectable Non-Small Cell Lung Cancer: A Narrative Review. JAMA Oncol 2024;10:249-55. [Crossref] [PubMed]
35. Chen Y, Wu Y, Yan G, et al. Tertiary lymphoid structures in cancer: maturation and induction. Front Immunol 2024;15:1369626. [Crossref] [PubMed]
36. Meylan M, Petitprez F, Becht E, et al. Tertiary lymphoid structures generate and propagate anti-tumor antibody-producing plasma cells in renal cell cancer. Immunity 2022;55:527-541.e5. [Crossref] [PubMed]
37. Liu Y, Ye SY, He S, et al. Single-cell and spatial transcriptome analyses reveal tertiary lymphoid structures linked to tumour progression and immunotherapy response in nasopharyngeal carcinoma. Nat Commun 2024;15:7713. [Crossref] [PubMed]
38. MacFawn IP, Magnon G, Gorecki G, et al. The activity of tertiary lymphoid structures in high grade serous ovarian cancer is governed by site, stroma, and cellular interactions. Cancer Cell 2024;42:1864-1881.e5. [Crossref] [PubMed]
39. Chen S, Crabill GA, Pritchard TS, et al. Mechanisms regulating PD-L1 expression on tumor and immune cells. J Immunother Cancer 2019;7:305. [Crossref] [PubMed]
40. Gil González MLÁ, Molina-Alejandre M, Sierra-Rodero B, et al. 734 Tertiary lymphoid structures as hallmarks of response to neoadjuvant chemoimmunotherapy in locally advanced non-small cell lung cancer NADIM trial. J Immunother Cancer 2025;13:A1-A1534.
41. Sun X, Liu W, Sun L, et al. Maturation and abundance of tertiary lymphoid structures are associated with the efficacy of neoadjuvant chemoimmunotherapy in resectable non-small cell lung cancer. J Immunother Cancer 2022;10:e005531. [Crossref] [PubMed]
42. Ai L, Xu A, Xu J. Roles of PD-1/PD-L1 Pathway: Signaling, Cancer, and Beyond. Adv Exp Med Biol 2020;1248:33-59. [Crossref] [PubMed]
43. Khosravi GR, Mostafavi S, Bastan S, et al. Immunologic tumor microenvironment modulators for turning cold tumors hot. Cancer Commun (Lond) 2024;44:521-53. [Crossref] [PubMed]
44. Lee D, Cho M, Kim E, et al. PD-L1: From cancer immunotherapy to therapeutic implications in multiple disorders. Mol Ther 2024;32:4235-55. [Crossref] [PubMed]
45. Vanhersecke L, Brunet M, Guégan JP, et al. Mature tertiary lymphoid structures predict immune checkpoint inhibitor efficacy in solid tumors independently of PD-L1 expression. Nat Cancer 2021;2:794-802. [Crossref] [PubMed]
46. Xie Y, Peng H, Hu Y, et al. Immune microenvironment spatial landscapes of tertiary lymphoid structures in gastric cancer. BMC Med 2025;23:59. [Crossref] [PubMed]