Sequential Immune Activation of Effector T Cells as Biomarkers of Response to Durvalumab in Patients with Locally Advanced NSCLC

Abstract

Purpose:

Durvalumab therapy following concurrent chemoradiotherapy (cCRT) improves progression-free survival (PFS) in patients with unresectable locally advanced non–small cell lung cancer. In this prospective observational study, we evaluated the changes in peripheral blood immune cell counts to elucidate the immunologic mechanisms underlying cCRT and durvalumab therapy.

Experimental Design:

Peripheral blood mononuclear cell (PBMC) samples were collected at four time points: before cCRT, after cCRT, at the start of durvalumab, and 8 weeks after the start of durvalumab, and analyzed by multicolor flow cytometry.

Results:

Of the 149 enrolled patients, 115 received durvalumab consolidation therapy after cCRT. The median PFS in the overall population was 24.2 months, and the 3-year PFS rate was 38.9%. PBMC analysis showed an increased effector fraction of CD4⁺ T cells before and after cCRT but no change in CD8⁺ T cells. Following durvalumab therapy, the effector fraction ratio of CD8⁺ T cells (CD62L^low CD8⁺ T cells) increased and positively correlated with increased CD62L^low CD4⁺ T cells during cCRT. Patients whose proportion of CD62L^low CD4⁺ T cells exceeded the threshold for cCRT had better PFS than those below the threshold. Patients whose CD62L^low CD8⁺ T-cell proportion exceeded the threshold after durvalumab therapy showed prolonged PFS compared with those below the threshold.

Conclusions:

cCRT promotes an increase in effector CD4⁺ T cells, and the subsequent increase in CD8⁺ T cells following durvalumab therapy prolongs PFS. Peripheral blood effector–type CD4⁺ and CD8⁺ T cells are potential biomarkers for evaluating the immune status of patients and predicting treatment efficacy.

Translational Relevance.

This study elucidates the temporal dynamics of peripheral immune cell populations during chemoradiotherapy (cCRT) and subsequent durvalumab consolidation therapy in patients with locally advanced non–small cell lung cancer. A sequential expansion of CD62L^low CD4⁺ and CD8⁺ T cells was observed, with Th7R CD4⁺ T cells emerging as key modulators of CD8⁺ T-cell proliferation. These findings suggest that CD4⁺ T-cell priming during cCRT may facilitate the expansion of stem-like precursor exhausted CD8⁺ T cells upon programmed cell death ligand 1 blockade. The correlation between Th7R frequency and clinical outcomes highlights its potential as a predictive biomarker. Furthermore, dynamic changes in programmed cell death protein 1⁺ dendritic cells may influence T-cell activation and treatment efficacy. This work provides mechanistic insight into the cancer immunity cycle and identifies peripheral effector–type T cells as promising biomarkers for stratifying patients and optimizing immunotherapy strategies.

Introduction

The treatment of locally advanced non–small cell lung cancer (LA-NSCLC) requires a multimodal approach. Following concurrent chemoradiotherapy (cCRT), the administration of durvalumab, an anti–programmed cell death ligand 1 (PD-L1) antibody, as consolidation therapy significantly improves progression-free survival (PFS) and overall survival (OS; refs. 1, 2). Supported by robust long-term outcomes and accumulating real-world evidence, it has become an established standard of care (3–6). In a subgroup analysis of the PACIFIC study, patients randomized within 2 weeks of completing radiotherapy and those for whom durvalumab treatment was initiated early demonstrated favorable PFS and OS outcomes (1–3), suggesting that immunologic activation induced by radiotherapy enhances the antitumor efficacy of durvalumab. This immunologic activation is lost relatively quickly after the completion of radiotherapy. However, the mechanisms underlying such immunologic synergistic effects and the predictive biomarkers of durvalumab treatment efficacy remain unclear.

Radiotherapy is crucial for the treatment of solid tumors and exerts cytotoxic effects through two primary mechanisms: direct DNA damage caused by high-energy photons and the generation of reactive oxygen species (7, 8). Beyond its direct effects on cancer cells, radiation-induced DNA damage triggers a cascade of biological responses that contribute to immune regulation via the Stimulator of Interferon Genes /cyclic GMP–AMP synthase (STING/cGAS) pathway, potentially enhancing antitumor T-cell immunity (9, 10). One of the most well-known effects is the upregulation of danger signals such as HMG-1, which promotes the activation of antigen-presenting cells via Toll-like receptor 4 signaling (11, 12). Antigen-presenting cells in the tumor microenvironment (TME), such as dendritic cells (DC), take up cancer antigens derived from radiation-induced apoptotic cancer cells while receiving activation stimuli from danger signals, migrate to the draining lymph nodes, and promote the priming and clonal proliferation of cancer antigen-specific T cells. However, the “abscopal effect,” in which proliferated T cells target lesions outside the irradiated area, has been widely reported (13–15). As the frequency of the abscopal effect increases with the use of immune checkpoint inhibitors (ICI), such as programmed cell death protein 1 (PD-1) blockade therapy, T cells proliferated by radiotherapy are suggested to be under the control of PD-1 (12, 16, 17). We aimed to identify the specific T-cell subpopulations involved in each mechanism by observing the T-cell subsets that increase with cCRT and anti–PD-L1 therapy. Our findings clarify the mechanisms linking the efficacy of radiation and durvalumab consolidation therapy and identify biomarkers predictive of the efficacy of durvalumab consolidation therapy. Although studies have been conducted on immune responses and predictive biomarkers in radiotherapy alone, few have investigated biomarker dynamics when durvalumab consolidation therapy is added after cCRT (18–21). To address this gap, a prospective study involving patients with LA-NSCLC scheduled to receive definitive cCRT was conducted. Peripheral blood samples were used to evaluate immune responses and to identify potential predictive biomarkers that could aid in optimizing immunotherapy strategies.

Materials and Methods

Study design and patient enrollment

This multicenter, prospective, observational study was conducted across 15 institutions, targeting a cohort of 150 patients with stage III NSCLC. This study aimed to identify biomarkers predictive of the response to consolidation durvalumab following cCRT. Patient enrollment was initiated in January 2020. Eligibility criteria included a histologically confirmed diagnosis of NSCLC and a clinical stage III classification based on the eighth edition of the tumor–node–metastasis (TNM) system, including cases of postsurgical recurrence (22). Participants were required to be scheduled for cCRT integrating platinum-based chemotherapy and definitive thoracic radiation, followed by planned consolidation therapy with durvalumab with a dosing regimen of 10 mg/kg every 2 weeks. Additionally, the patients had to be at least 20 years of age and were required to provide written informed consent. Patients with a history of treatment with PD-1/PD-L1 inhibitors or anti–cytotoxic T-lymphocyte associated antigen-4 (CTLA-4) antibodies were excluded. Radiologic assessment of the tumors was conducted according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 (23). This study was registered with the UMIN Clinical Trials Registry (UMIN000039167).

Ethical compliance

The study was approved by the Regional Ethics Review Committee of the SMUIMC. All patients adhered to the ethical principles of the Declaration of Helsinki about medical research involving human subjects. Informed written consent was obtained from all participants prior to enrollment.

Sample collection and peripheral blood mononuclear cell analysis

Case report forms were electronically collected as digital input, incorporating comprehensive information on patient background, including age, histology, performance status, and TNM classification. Data on the chemotherapy regimen, radiation dose and frequency, durvalumab administration timing, treatment efficacy, and laboratory parameters, such as white blood cell count, lactate dehydrogenase, and Krebs von den Lungen-6, were recorded. The records included instances of pneumonitis, immune-related adverse events, and the use of corticosteroids or immunosuppressive agents.

Blood collection was scheduled at four specific time points: before the initiation of cCRT, upon completion of cCRT, within 7 days before the start of durvalumab administration, and at week 8 (±7 days) following the initiation of durvalumab treatment. Samples were collected in BD Vacutainer CPT tubes (heparin, 8 mL), and peripheral blood mononuclear cells (PBMC) were isolated for biomarker analysis. These samples were collected in heparinized CPT Vacutainer tubes (Becton Dickinson Vacutainer Systems) and centrifuged at 1,500 × g for 20 minutes at room temperature to isolate PBMCs from granulocytes. Erythrocytes were separated within 2 hours of collection, washed twice with phosphate buffer, and subjected to flow cytometry. Isolated PBMCs were cryopreserved at −80°C using CELLBANKER 2 (Nippon Zenyaku Kogyo Co. Ltd.) and promptly transferred to a liquid nitrogen storage system. PBMC samples obtained from patients were analyzed using Fortessa (FACS Diva Software LSR II, BD Biosciences) and CyTOF (LSR Fortessa flow microfluorometer, Becton Dickinson) to simultaneously quantify up to 50 different proteins per cell in suspension. Lymphocytes were gated based on forward and side scatter parameters, and four-color flow cytometry was used to evaluate PBMC phenotypes and antigen expression. Key differentiation markers, CCR7, CD45RA, CD45RO, and CD62L, were assessed to determine T-cell memory status, homing capacity, and migratory potential. The differentiation of naïve T cells (CCR7⁺CD45RA⁺) into effector T cells, characterized by high-affinity T-cell receptors and cytokine responsiveness, was examined. Cellular fractions were stratified using surface marker expression, and relative proportions were quantified. Circulating DCs were analyzed, and CD123⁻CD11c⁺ myeloid DC (mDC) and CD123⁺CD11c⁻ plasmacytoid DC (pDC) subsets were identified. Within CD4⁺ T cells, regulatory T cells expressing CD25^high and the transcription factor forkhead box P3 were quantified. The expression of CXCR3, CCR4, and CCR6 was used to determine T helper (Th) subset polarization, which is crucial for the regulation of immune responses. CD4⁺ T-cell subsets were classified into Th1 (CXCR3⁺CCR4⁻CCR6⁻), Th2 (CXCR3⁻CCR4⁺CCR6⁻), and Th17 (CXCR3⁻CCR4⁺CCR6⁺) populations (24, 25). Beyond these traditional classifications, a subset of CXCR3⁺CCR4⁻CCR6⁺ cells, as well as CXCR3⁻CCR4⁻CCR6⁺ single-positive T cells, was identified (26). Moreover, the expression of inducible T-cell costimulator, CTLA-4, lymphocyte activation gene-3, CD80 (B7-1⁺), CD86 (B7-2⁺), CD26, B7-1, CD127, CD40L, T-cell immunoglobulin, CXCR4, PD-L1, and PD-1 was evaluated within DC, CD4⁺, and CD8⁺ T-cell clusters. Median cell proportion values were used to classify the populations into high- and low-expression groups. The detailed gating workflows for both the CyTOF and Fortessa instruments used in PBMC profiling are shown in Supplementary Fig. S1. All antibodies utilized in the panel design, along with the corresponding fluorochromes or metal tags, are listed in Supplementary Tables 1A and 1B to facilitate methodologic reproducibility.

Statistical analysis

PFS was defined as the time from the start of cCRT and durvalumab administration until disease progression or death. OS was defined as the time from cCRT initiation to durvalumab administration until death. Patients who were alive without documented disease progression or death were censored at the last follow-up date, and those without documented death were censored at the last follow-up. The predictive performance of CD62L^low CD4⁺ T cells, including differentiated CD4⁺ T-cell subsets and CD8⁺ T cells, was assessed using this cutoff value. Survival curves were estimated using the Kaplan–Meier method. Tests for differences and hazard ratios (HR) were performed using log-rank tests and the Gehan–Breslow–Wilcoxon test, respectively. The nonparametric Mann–Whitney U test was used for age. Differences before and after chemoradiotherapy and durvalumab administration were assessed using the Student t test. Spearman rank correlation and linear regression analyses were used to determine correlations between each cell cluster and PFS and between cell populations, respectively. All P values were two-sided, and values less than 0.05 were considered statistically significant. Statistical analyses were performed using Prism 9 (RRID: SCR_002798; GraphPad Software) and SAS, version 9.4 (SAS Institute Inc.).

Results

Enrollment and baseline characteristics

Between January 2020 and May 2021, among the patients registered via the electronic data capture system, 149 individuals were considered fully eligible for analysis after excluding three patients who withdrew their informed consent or met the exclusion criteria. Of these, three patients who did not receive cCRT and an additional three whose samples exhibited a viable cell fraction below 30% were excluded from the biomarker evaluation. Finally, 143 patients were included in the biomarker analysis. Paired blood samples collected before and after chemoradiotherapy were available for 135 (94.4%) patients. Among those included in the biomarker analysis cohort, 115 patients (80.4%) received consolidation therapy with durvalumab; of these, paired specimens obtained before and after durvalumab administration were adequately collected in 101 cases (70.6%). Figure 1 illustrates the flowchart of patient inclusion, including enrollment and sample availability. The baseline clinical characteristics are presented in Table 1. Among the patients who received durvalumab consolidation therapy, the median age was 71.0 years (range: 36–86). The majority were males (n = 95, 82.6%), and a substantial proportion had a history of smoking (n = 103, 89.6%). The histologic subtypes included adenocarcinoma (n = 57, 49.6%) and squamous cell carcinoma (n = 48, 41.7%). The disease stage distribution at diagnosis was stage IIIA (n = 50, 43.5%), stage IIIB (n = 47, 40.9%), and stage IIIC (n = 14, 12.2%). Among the patients with available PD-L1 tumor proportion score (TPS) data, 24 (20.9%) had TPS <1%, 21 (18.3%) had TPS between 1% and 49%, and 26 (22.6%) had TPS ≥50%. The most commonly administered cCRT regimen was carboplatin plus paclitaxel (47.0%).

Figure 1.

CONSORT diagram of patient inclusion and immune cell analysis. Among the 149 initially enrolled patients, 3 who did not receive cCRT were excluded, resulting in 146 treated cases. After excluding three patients whose PBMC samples contained <30% viable cells, 143 patients were included in the biomarker cohort. Of these, 115 patients received durvalumab consolidation therapy, whereas 28 did not. Immune cell dynamics before and after cCRT were evaluable in 135 patients and paired analyses before and after durvalumab administration were feasible in 101 patients.

Table 1.

Baseline characteristics.

Different variables	Patients included in the analysis	Patients receiving durvalumab
	143, n (%)	115, n (%)
Age	​	​
Median (range), years	71 (36–88)	71 (36–86)
Sex	​	​
Male	117 (81.8)	95 (82.6)
Female	26 (18.2)	20 (17.4)
Smoking history	​	​
Current or former	127 (88.8)	103 (89.6)
Never	16 (11.2)	12 (10.4)
Histology	​	​
Adeno	67 (46.9)	57 (49.6)
Squamous	62 (43.4)	48 (41.7)
Others	14 (9.8)	10 (8.7)
ECOG PS	​	​
0	69 (48.3)	58 (50.4)
1	68 (47.6)	53 (46.1)
2	6 (4.2)	4 (3.5)
Staginga	​	​
IIIA	60 (42)	50 (43.5)
IIIB	60 (42)	47 (40.9)
IIIC	19 (13.3)	14 (12.2)
Others	4 (2.8)	4 (3.5)
Tb	​	​
T1	23 (16.1)	20 (17.4)
T2	30 (21)	24 (20.9)
T3	41 (28.7)	33 (28.7)
T4	47 (32.9)	36 (31.3)
N	​	​
N0	10 (7)	9 (7.8)
N1	19 (13.3)	17 (14.8)
N2	76 (53.1)	58 (50.4)
N3	38 (26.6)	31 (27)
Driver oncogenec	​	​
EGFR	8 (5.6)	7 (6.1)
Wild-type	74 (51.7)	58 (50.4)
Unknown	61 (42.7)	50 (43.5)
ALK	5 (3.5)	4 (3.5)
Wild-type	70 (49)	55 (47.8)
Unknown	68 (47.6)	56 (48.7)
PD-L1 status	​	​
<1	26 (18.2)	24 (20.9)
1–49	27 (18.9)	21 (18.3)
≥50	38 (26.6)	26 (22.6)
Unknown	52 (36.4)	44 (38.3)
Chemotherapy	​	​
CBDCA + PTX	66 (45.5)	54 (47)
Low-dose CBDCA	36 (25.2)	28 (24.3)
CDDP/CBDCA + VNR	21 (14.7)	16 (13.9)
CDDP/CBDCA + S-1	14 (9.8)	12 (10.4)
CDDP + PEM	4 (2.8)	2 (1.7)
CDDP + DTX	2 (1.4)	1 (0.9)
Others	2 (1.4)	2 (1.7)

Abbreviations: ALK, anaplastic lymphoma kinase; ECOG PS, Eastern Cooperative Oncology Group Performance Status.

^aClinical stages were entered into the electronic case report form (eCRF) based on the eighth edition.

^bTwo patients were excluded from each treatment group.

^cIn addition, one HER2 exon 20 insertion, one BRAF V600E mutation, and one KRAS G12C were included.

Dynamic changes in immune cell populations before and after cCRT and durvalumab consolidation

First, we examined dynamic changes in cellular populations during cCRT and subsequent durvalumab consolidation therapy. No significant overall shifts in the total populations of mDCs or pDCs were observed. However, the proportion of PD-L1–expressing mDCs increased significantly following cCRT (mean difference: 4.7 ± 17.3, t = 3.2, df = 135, P = 0.002) and subsequently decreased significantly after durvalumab administration (mean difference: −8.5 ± 11.1, P < 0.0001; Fig. 2A and B). Similarly, the proportion of PD-L1–expressing pDCs increased significantly after cCRT (mean difference: 11.3 ± 22.9, t = 5.8, df = 135, P < 0.0001) and decreased significantly following durvalumab treatment (mean difference: −19.2 ± 18.7, P < 0.0001; Fig. 2C and D). Among T lymphocytes, CD62L^low CD4⁺ T cells increased following cCRT and remained elevated after durvalumab therapy (mean difference: cCRT, 8.8 ± 10.2, t = 10.1, df = 135, P < 0.0001; durvalumab, 1.6 ± 7.1, t = 2.2, df = 104, P = 0.028; Fig. 2E and F). Conversely, CD62L^low CD8⁺ T cells showed no significant change after cCRT but demonstrated a notable increase following durvalumab treatment (mean difference: cCRT, -2.0 ± 12.6, t = 1.8, df = 135, P = 0.069; durvalumab, 6.8 ± 9.3, t = 7.5, df = 104, P < 0.0001; Fig. 2G and H). Analysis of the correlation between cCRT-induced increases in CD62L^low CD4⁺ T cells and CD62L^low CD8⁺ T cell proportions before and after durvalumab administration revealed a statistically significant moderate positive association (r = 0.33, 95% confidence interval (CI), 0.14–0.50, P = 0.0008; Y = 0.5436 * X + 0.000; Fig. 2I).

Figure 2.

Longitudinal changes in immune cell subset proportions across treatment phases. Immune cell dynamics were assessed at four time points: before and after cCRT administration and before and after durvalumab administration. Blue dots and lines represent measurements prior to each intervention, and red dots and lines indicate posttreatment values. A and B, PD-L1 expression dynamics in CD123^-CD11c⁺ mDCs. C and D, PD-L1 expression dynamics in CD123⁺ CD11c− pDCs. E and F, Changes in CD62L^low CD4⁺ T-cell subsets. G and H, Changes in CD62L^low CD8⁺ T-cell subsets. I, Correlation between CD62L^low CD4⁺ T-cell changes after cCRT and CD62L^low CD8⁺ T-cell changes after durvalumab. The left y-axis indicates the absolute proportion of each immune cell subset, whereas the right y-axis represents the magnitude of change, defined as the difference between the mean postintervention proportion and the mean preintervention proportion. CRT, chemoradiotherapy; Dur, durvalumab.

Detailed analyses revealed dynamic changes in CD4⁺ T-cell subsets before and after cCRT (Supplementary Fig. S2A–S2D). Among CD8⁺ T cells, durvalumab administration resulted in a significant reduction in CCR7⁺CD45RA⁺ (naïve) and CCR7⁺CD45RA⁻ (central memory) CD8⁺ T cells, whereas CCR7⁻CD45RA⁺ (effector) CD8⁺ T cells increased significantly (Supplementary Fig. S3A–S3D).

Association between T-cell subset proportions after cCRT and CD8⁺ T-cell proportions following durvalumab treatment

We performed correlation analyses between post-cCRT CD4⁺ T-cell subsets and CD8⁺ T-cell frequencies after durvalumab administration to identify the immune cell populations that influence the expansion of CD8⁺ T cells from cCRT through durvalumab treatment. Th7R cell frequency after cCRT and CD8⁺ T-cell proportions following durvalumab therapy were correlated (P = 0.006, r = 0.27, 95% CI, 0.08–0.44; Fig. 3A–D). Post-cCRT CD62L^low Th7R cell proportions correlated with post-durvalumab CD62L^low CD8⁺ T cells (P = 0.01, r = −0.24, 95% CI, 0.05–0.42), PD-1⁺CD62L^high CD8⁺ T cells (P = 0.007, r = 0.26, 95% CI, 0.07–0.43), and CD8⁺ effector memory T cells (P = 0.003, r = 0.29, 95% CI, 0.11–0.46; Fig. 4A–C).

Figure 3.

Correlation between CD4⁺ T-cell subsets after chemoradiotherapy and CD8⁺ T-cell effector populations following durvalumab therapy. This figure illustrates the relationship between CD4⁺ T-cell subsets measured immediately after cCRT and CD62L^low CD8⁺ T cells after durvalumab therapy. Each dot represents an individual patient, and correlation coefficients and P values are shown. A, Th1 CD4⁺ T cells vs. CD8⁺ T cells. B, Th2 CD4⁺ T cells vs. CD8⁺ T cells. C, Th17 CD4⁺ T cells vs. CD8⁺ T cells. D, Th7R CD4⁺ T cells vs. CD8⁺ T cells.

Figure 4.

Correlation between post-cCRT Th7R CD4⁺ T cells and effector CD8⁺ T-cell subsets following durvalumab therapy. Th7R CD4⁺ T cells (CXCR3⁺CCR4⁻CCR6⁺ and CXCR3⁻CCR4⁻CCR6⁺) measured after cCRT were correlated with effector CD8⁺ T-cell subsets after durvalumab. Each dot represents one patient, and the correlation coefficients and statistical significance were annotated. A, Th7R CD4⁺ T cells vs. CD62L^high CD8⁺ T cells. B, Th7R CD4⁺ T cells vs. PD-1⁺CD62L^high CD8⁺ T cells. C, Th7R CD4⁺ T cells vs. CCR7^-CD45RA⁺ CD8⁺ T cells. EM, effector memory  CD8⁺ T cells (CCR7^-CD45RA^-).

PFS and OS in the total cohort and biomarker analysis subgroup

The median observation period was 42.5 months [95% CI, 35.1–not reached (NR)] in the overall cohort and 44.1 months (95% CI, 39.1–NR) in patients receiving durvalumab. The median PFS in the total analysis cohort and the durvalumab-treated subgroup was 20.2 months (95% CI, 13.4–27.9) and 24.2 months (95% CI, 15.7–32.6), respectively (Fig. 5A and B). The median OS was 42.4 months (95% CI, 35.0–NR) in the entire cohort and 43.1 months (95% CI, 37.3–NR) in the durvalumab-treated group (Supplementary Fig. S4A and S4B).

Figure 5.

PFS analysis and biomarker stratification. Kaplan–Meier curves depict PFS in the overall and stratified subgroups. Biomarker-defined subsets were based on treatment-induced changes in effector T-cell proportions and baseline cellular frequencies prior to durvalumab treatment. Patients were dichotomized using predefined cutoffs, and statistical significance was assessed using the log-rank and Wilcoxon tests. A, PFS of the entire cohort. B, PFS of patients receiving durvalumab. C, PFS stratified by CD62L^low CD4⁺ T-cell changes after cCRT. D, PFS stratified by CD62L^low CD8⁺ T-cell changes after durvalumab treatment. E, PFS stratified by Th7R CD4⁺ T-cell frequency before durvalumab.

In CD62L^low CD4⁺ T cells, which showed a significant increase following cCRT, a cutoff value of 3.9 was used to stratify patients into high and low groups. Kaplan–Meier analysis revealed a median PFS of 28.5 months (95% CI, 20.4–38) in the high group and 11.0 months (95% CI, 5.3–21.2) in the low group (log-rank P = 0.057, Wilcoxon P = 0.038; Fig. 5C). Similarly, for CD62L^low CD8⁺ T cells, which increased significantly after durvalumab treatment, a cutoff value of 6.3 was used. Patients with a higher increase exhibited a median PFS of 32.6 months (95% CI, 24.5–NR), whereas those with a lower increase had a median PFS of 15.7 months (95% CI, 8.3–25.6) (log-rank P = 0.053, Wilcoxon P = 0.023; Fig. 5D). The corresponding OS curves for these stratifications are presented in Supplementary Fig. S4C and S4D. Within the CD62L^low CD4⁺ T-cell compartment, the Th7R subset, previously implicated in CD62L^low CD8⁺ T-cell modulation, was evaluated. Patients were divided based on the post-chemoradiotherapy Th7R cell frequency using a cutoff value of 10.1%. Kaplan–Meier analysis revealed that patients with a value above the cutoff exhibited a median PFS of 43.6 months, whereas those below the cutoff had a median PFS of 22.8 months (HR, 0.58; 95% CI, 0.33–0.99; log-rank P = 0.044; Wilcoxon P = 0.095; Fig. 5E). The OS analysis for this Th7R stratification is shown in Supplementary Fig. S4E. This difference was more pronounced in adenocarcinoma cases (Supplementary Fig. S5A–S5D). To reduce potential confounding, we excluded 14 patients with stage IIIC disease—whose survival curves were distinct from the rest—and repeated the analyses in patients with stage IIIA and IIIB disease, whose survival outcomes were comparable. In this restricted cohort, differences in survival between high and low groups for (i) changes in CD62L^low CD4⁺ T cells before and after chemoradiotherapy, (ii) changes in CD62L^low CD8⁺ T cells before and after durvalumab, and (iii) baseline Th7R cell proportions prior to durvalumab were further accentuated (Supplementary Fig. S6A–S6H).

Discussion

In this study, a potential correlation was identified between an increase in CD62L^low CD4⁺ T cells following cCRT and an increase in CD62L^low CD8⁺ T cells following durvalumab administration. Patients exhibiting these immunologic changes may derive greater benefits from durvalumab-enhanced therapy. To our knowledge, this is the first report to clarify the temporal dynamics of antitumor immune cell populations during cCRT and subsequent durvalumab therapy. Although the evaluation of immune cell dynamics at disease progression and during long-term disease control would have been informative, feasibility and funding constraints limited sampling. We therefore focused on PBMC dynamics at four time points before and after chemoradiotherapy and durvalumab.

Radiotherapy can induce danger signals, such as HMGB1 and ATP, in the TME, leading to tumor cell apoptosis, which in turn triggers the supply of tumor antigens, activation of DCs, and promotion of T-cell priming and clonal proliferation in the draining lymph nodes (27–31). The increase in cancer antigen–specific T cells can induce the abscopal effect, which is an antitumor effect at nonradiation treatment sites (13–15). The increase was primarily attributed to CD8⁺ T cells. However, in this study, the increase was limited to CD4⁺ T cells, and no increase in CD8⁺ T cells was observed during cCRT. A likely explanation for this is that CD8⁺ T cells proliferate rapidly. Platinum-based agents and taxane-based anticancer drugs used in cCRT target the M-phase of the cell cycle, rendering cells that have entered the cell cycle more susceptible to the effects of anticancer drugs (32, 33). CD8⁺ T cells that exhibit vigorous proliferation in response to antigen stimulation may be the most affected. On the other hand, CD4⁺ T cells, which primarily perform helper functions and lack direct cytotoxic activity, do not require significant clonal proliferation and thus may be less affected (24, 34).

A proliferation burst of CD8⁺ T cells occurs in the peripheral blood of patients with lung cancer treated with PD-1 inhibitors, which is correlated with antitumor effects (35–38). This finding is consistent with the results of this study, which demonstrated an increase in CD8⁺ T cells following treatment with durvalumab, an anti–PD-L1 antibody drug. The significant expansion of CCR7⁻CD45RA⁺ effector CD8⁺ T cells highlights the activation of antitumor immunity in response to treatment. Normal effector CD8⁺ T cells expressing PD-1 lose their self-renewal capacity, and the origin of cells with proliferation bursts has long remained unclear. However, single-cell RNA sequencing analysis of the TME revealed the existence of stem-like memory cells called precursor exhausted CD8⁺ T cells (Tpex), expressing PD-1 along with the IL7 receptor and TCF7, and retaining self-renewal capacity (39–42). These cells are now considered to be the origin of the proliferation burst induced by PD-1 inhibitors. The reprogramming of Tpex requires the formation of a “Triad” state, in which DCs, CD4⁺ T cells, and CD8⁺ T cells directly contact each other. In this study, CD4⁺ T-cell proliferation induced by prior cCRT correlated with CD8⁺ T-cell proliferation following durvalumab treatment. This finding suggests that an increase in CD4⁺ T cells induced by cCRT may promote Tpex expansion and facilitate the proliferation burst induced by the subsequent durvalumab treatment.

In the context of the cancer immunity cycle, naïve T cells undergo clonal expansion and activation upon antigen presentation through priming by antigen-presenting cells (43). CD4⁺ lymphocytes are pivotal as commanders, orchestrating immune responses, whereas CD8⁺ T cells directly attack cancer cells by recognizing tumor antigens restricted by MHC class I molecules. CD4⁺ T cells functionally differentiate to fulfill multiple roles, enabling efficient elimination of target non–self-antigens. CD4⁺ T cells undergo functional differentiation to efficiently eliminate non–self-antigens (34).

Among the CD4⁺ T-cell subsets, only post-cCRT Th7R T cells showed a significant association with the proportion of CD8⁺ T cells that expanded after durvalumab treatment. Th7R cells are IL7R^highCCR6⁺ Th1-like CD4⁺ T cells that coexist with Tpex within the TME and play a crucial role in sustaining antitumor immunity (25). These cells express lymphotoxin-β and CXCL13, contributing to high endothelial venule and tertiary lymphoid structure formation, correlating with Tpex abundance, and potentially serving as predictors of response to ICI therapy (44, 45). If this correlation holds, Th7R T cells may regulate CD8⁺ T-cell expansion following durvalumab treatment. Further in-depth correlation analysis revealed a significant association between post-cCRT Th7R T cells and CD62L^high CD8⁺ T cells, a subset indicative of Tpex status, suggesting that Th7R may be pivotal in maintaining antitumor immunity by orchestrating CD8⁺ T-cell dynamics. Furthermore, in this study, patients with post-cCRT Th7R T-cell proportions exceeding a predefined cutoff exhibited significantly prolonged PFS and OS compared with those below the threshold, underscoring the prognostic relevance of Th7R cells in the context of anti–PD-L1 therapy. Although we did not analyze the TME directly, prior studies indicate that circulating Tpex-like CD8⁺ T cells mirror intratumoral states and track with responses to PD-(L)1 blockade. We therefore regard post-cCRT Th7R expansion and subsequent CD62L^high CD8⁺ T-cell increases as plausible peripheral surrogates of a Tpex-supportive niche rather than definitive evidence of intratumoral dynamics.

The proportions of mDCs and pDCs remained stable throughout cCRT and durvalumab treatment; however, PD-L1⁺ DCs significantly increased after radiation and decreased following durvalumab treatment. This may be driven by the release of Damage-Associated Molecular Patterns (DAMPs) and HSPs, which is a novel finding. Oh and colleagues (46) reported that PD-L1+ expression on DCs negatively affects T-cell activation, fostering an immunosuppressive TME. The increase in PD-L1⁺ DCs after radiation may impair CD4⁺ T-cell activation, thereby reducing CD8⁺ T-cell priming. Our findings suggest a sequential immune response in which cCRT induces PD-L1⁺ DC expansion and CD62L^low CD4⁺ T-cell increase, followed by durvalumab-driven CD62L^low CD8⁺ T-cell expansion. Patients with higher levels of activated CD4⁺ and CD8⁺ T cells exhibit better prognostic outcomes. This study systematically delineates the effects of cCRT and durvalumab within the cancer immunity cycle and identifies potential biomarkers for therapeutic response.

A limitation of this study is the lack of direct analysis of the TME, including genomic alterations, spatial clustering of immune effector populations within the tissue, and T-cell receptor repertoire profiling. The correlation and homology between T cells influencing antitumor effects in the circulating blood and immune effector cells within the microenvironment merit further evaluation. This study aimed to identify biomarkers for ICI therapy, but a subset of patients, estimated at 20%, achieved curative outcomes with cCRT alone, potentially limiting the statistical significance in log-rank analyses. The observed changes in CD62L^low CD4⁺ and CD62L^low CD8⁺ T cells following treatment remain inconclusive, specifically whether they drive therapeutic effects or simply result from them. To clarify this aspect, future studies should examine whether introducing antigen-specific expanded cells can directly enhance treatment efficacy. In addition, the role of homeostatic proliferation in lymphocyte fluctuations remains unclear (47). CRT reduces absolute lymphocyte counts, but whether the observed expansion represents self-antigen–driven compensatory proliferation is unclear. Notably, such expansions reportedly influence the efficacy of anti–PD-1 therapies (48). Furthermore, the influence of systemic corticosteroids cannot be ignored. Among the 115 patients treated with durvalumab, there was no significant difference in PFS [21.2 vs. 26.2 months; HR 1.15 (95% CI, 0.67–1.97), P = 0.60] or OS between patients who did and did not receive corticosteroids. However, at 8 weeks after durvalumab initiation, patients who had received corticosteroids showed significantly lower proportions of mDCs, pDCs, effector CD8⁺ T cells, and CD27^-CD62L^low Th7R cells. Given the limited sample size, further investigation is required, and a separate study is currently underway to evaluate the immunologic impact of corticosteroids. Further investigation should establish their relevance to immunotherapy outcomes.

In summary, cCRT promoted an increase in effector CD4⁺ T cells, and the subsequent increase in CD8⁺ T cells following durvalumab therapy led to prolonged PFS. Peripheral blood effector-type CD4⁺ and CD8⁺ T cells may be potential biomarkers for evaluating patients’ immune status and predicting treatment efficacy. This finding is considered an important insight, enabling treatment interventions tailored to the immune status.

Supplementary Material

Supplementary Figure 1

Supplementary Figure 1.

Supplementary Figure 2

Supplementary Figure 2.

Supplementary Figure 3

Supplementary Figure 3.

Supplementary Figure 4

Supplementary Figure 4.

Supplementary Figure 5

Supplementary Figure 5.

Supplementary Figure 6

Supplementary Figure 6.

Supplementary Table 1

Supplementary Table 1.

Data Availability

The datasets generated and analyzed during the current study are not publicly available due to patient privacy and institutional ethical restrictions, but deidentified data may be made available from the corresponding author upon reasonable request.

Authors’ Disclosures

A. Mouri reports grants and personal fees from AstraZeneca during the conduct of the study, as well as personal fees from AstraZeneca, Bristol Myers Squibb, Chugai Pharmaceutical, Eli Lilly, and Ono Pharmaceutical outside the submitted work. H. Kenmotsu reports grants and personal fees from AstraZeneca K.K. during the conduct of the study, as well as grants and personal fees from Ono Pharmaceutical Co. Ltd., Novartis Pharma K.K., Eli Lilly K.K., and Bristol Myers Squibb; grants from Loxo Oncology and the Japan Agency for Medical Research and Development; and personal fees from Amgen Inc., Bayer, Boehringer Ingelheim, Chugai Pharmaceutical Co. Ltd., Daiichi Sankyo Co. Ltd., Eli Lilly K.K./Kyowa Hakko Kirin Co. Ltd., Merck, MSD, Pfizer, Taiho Pharmaceutical, Takeda Pharmaceutical Co. Ltd., Janssen Pharmaceutical K.K., Thermo Fisher Scientific Inc., and Guardant Health outside the submitted work. H. Kagamu reports personal fees from AstraZeneca, MSD, Bristol Myers Squibb, Ono Pharmaceutical, and Chugai Pharmaceutical and grants from Boehringer Ingelheim outside the submitted work. K. Azuma reports personal fees from AstraZeneca, Bristol Myers Squibb, Chugai Pharmaceutical, MSD, Ono Pharmaceutical, Takeda Pharmaceutical, Taiho Pharmaceutical, Amgen, and Janssen Pharmaceutical outside the submitted work. R. Saito reports personal fees from AstraZeneca, Chugai Pharmaceutical, Daiichi Sankyo, Merck, MSD, Ono Pharmaceutical, and Taiho Pharmaceutical outside the submitted work. H. Akamatsu reports personal fees from AstraZeneca during the conduct of the study, as well as personal fees from Amgen Inc., Boehringer Ingelheim Inc., Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly Japan K.K., MSD K.K., Nippon Kayaku Co. Ltd., Novartis Pharma K.K., Ono Pharmaceutical Co. Ltd., Pfizer Inc., Takeda Pharmaceutical Co. Ltd., Taiho Pharmaceutical Co. Ltd., Janssen Pharmaceutical K.K., and Sandoz and grants and personal fees from Chugai Pharmaceutical Co. Ltd. outside the submitted work. K. Yonesaka reports grants, personal fees, and other support from Daiichi Sankyo Co. Ltd. and personal fees from Amgen, AstraZeneca, Chugai Pharmaceutical, and Pfizer outside the submitted work. H. Nagashima reports personal fees from Asahi Kasei Pharma, AstraZeneca, Boehringer Ingelheim, Daiichi Sankyo, KYORIN Pharmaceutical, MIYARISAN Pharmaceutical, Shionogi, Pfizer, GSK, Chugai Pharmaceutical, MSD, Ono Pharmaceutical, Eli Lilly, Sanofi, and Insmed outside the submitted work. S. Takahashi reports personal fees from AstraZeneca, Bristol Myers Squibb, Chugai Pharmaceutical, MSD, Nippon Kayaku, Ono Pharmaceutical, Daiichi Sankyo, Takeda Pharmaceutical, Taiho Pharmaceutical, Kyowa Kirin, Eli Lilly Japan, Pfizer, Janssen Pharmaceutical, and Amgen outside the submitted work. N. Yanagitani reports personal fees from AstraZeneca, Chugai Pharmaceutical Co. Ltd., Ono Pharmaceutical Co. Ltd., Bristol Myers Squibb, Eli Lilly and Company, Pfizer Inc., and Takeda Pharmaceutical Co. Ltd. outside the submitted work. K. Ninomiya reports personal fees from AstraZeneca, Boehringer Ingelheim, Kyowa Kirin, Eli Lilly Japan, Chugai Pharmaceutical, Nippon Kayaku, Taiho Pharmaceutical, MSD K.K., Ono Pharmaceutical, Takeda Pharmaceutical, Pfizer Inc., Bristol Myers Squibb, Elekta K.K., Janssen Pharmaceuticals, Daiichi Sankyo, Amgen K.K., Novartis, and Guardant Health outside the submitted work. Y. Nishioka reports personal fees from AstraZeneca K.K., MSD K.K., and Ono Pharmaceutical Co. Ltd. and grants and personal fees from Chugai Pharmaceutical Co. Ltd. during the conduct of the study, as well as grants and personal fees from Asahi Kasei Pharma Corporation, Otsuka Pharmaceutical Co., Ltd., KYORIN Pharmaceutical Co. Ltd., and Nippon Boehringer Ingelheim Co. Ltd.; personal fees from Amgen Inc., Eli Lilly Japan K.K., Insmed GK, Eisai Co. Ltd., GSK K.K., Sanofi K.K., Takeda Pharmaceutical Co. Ltd., Pfizer Inc., Bristol Myers Squibb Company, Regeneron Pharmaceuticals, Nippon Kayaku Co. Ltd., and Taiho Pharmaceutical Co. Ltd. outside the submitted work. K. Mori reports personal fees from Daiichi Sankyo, Chugai Pharmaceutical, Eli Lilly, and Ono Pharmaceutical outside the submitted work. S. Kitano reports grants from AstraZeneca during the conduct of the study, as well as grants and personal fees from AstraZeneca, Pfizer, Nippon Boehringer Ingelheim, Taiho Pharmaceutical, MSD, Eisai, Astellas Pharma, Ono Pharmaceutical Co. Ltd., Bristol Myers Squibb, GSK, Daiichi Sankyo, Chugai Pharmaceutical, Merck KGaA, Takeda Pharmaceuticals, and Kyowa Kirin; personal fees from Novartis, Rakuten Medical, and United Immunity; and grants from Takara Bio Inc., Incyte, Eli Lilly/Loxo Oncology, AbbVie, Syneos Health Clinical, Moderna, and Medpace outside the submitted work. K. Tamada reports grants, personal fees, and other support from Noile-Immune Biotech Inc. and grants from Hitachi Ltd., Chugai Pharmaceutical Co. Ltd., and Ajinomoto Co. Inc. outside the submitted work. N. Yamamoto reports grants and personal fees from AstraZeneca, Chugai Pharmaceutical, and Merck Sharp & Dohme and personal fees from Daiichi Sankyo, Eli Lilly, Ono Pharmaceutical, Taiho Pharmaceutical, and Takeda Pharmaceutical during the conduct of the study. T. Mitsudomi reports grants and personal fees from AstraZeneca during the conduct of the study, as well as personal fees from MSD, Ono Pharmaceutical, Bristol Myers Squibb, Novocure, Daiichi Sankyo, Taiho Pharmaceutical, Eli Lilly, and Johnson & Johnson outside the submitted work. No disclosures were reported by the other authors.

Authors’ Contributions

A. Mouri: Data curation, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. H. Kenmotsu: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. H. Kagamu: Conceptualization, data curation, formal analysis, supervision, funding acquisition, methodology, writing–original draft, project administration, writing–review and editing. K. Azuma: Resources, investigation, writing–review and editing. R. Saito: Resources, investigation, writing–review and editing. H. Akamatsu: Resources, investigation, writing–review and editing. K. Yonesaka: Resources, investigation, writing–review and editing. M. Kakegawa: Resources, investigation, writing–review and editing. H. Nagashima: Resources, investigation, writing–review and editing. S. Takahashi: Resources, investigation, writing–review and editing. M. Fujita: Resources, investigation, writing–review and editing. N. Yanagitani: Resources, investigation, writing–review and editing. K. Ninomiya: Resources, investigation, writing–review and editing. Y. Nishioka: Resources, investigation, writing–review and editing. K. Mori: Data curation, formal analysis, investigation, writing–review and editing. S. Kitano: Conceptualization, data curation, formal analysis, supervision, funding acquisition, methodology, project administration, writing–review and editing. K. Tamada: Conceptualization, data curation, formal analysis, supervision, funding acquisition, methodology, project administration, writing–review and editing. N. Yamamoto: Conceptualization, supervision, funding acquisition, methodology, project administration, writing–review and editing. A. Gemma: Conceptualization, supervision, funding acquisition, methodology, project administration, writing–review and editing. T. Mitsudomi: Conceptualization, supervision, funding acquisition, methodology, project administration, writing–review and editing.