Association of FDG PET/CT and immune profiling with pathologic complete response after neoadjuvant immunotherapy in stage II–III non-small cell lung cancer Authors: Eun Seong Lee, Saem Mul Park, Seunghun Lee, Sehui Kim, Juwhan Choi, Jun Hee Lee, Jae Seon Eo, Hyun Koo Kim, Hwan Seok Yong, Chun-Jen J Chen, Heidi Robinson, P Rod Dunbar, Sung Yong Lee Published: 2026 ABSTRACT AbstractBackgroundNeoadjuvant immunotherapy has emerged as a transformative approach for resectable non-small cell lung cancer (NSCLC). Pathologic complete response (pCR) after neoadjuvant therapy serves as a robust surrogate marker for early treatment efficacy and long-term outcomes. However, reliable noninvasive predictors of pCR before surgery remain an unmet clinical need. This study investigated the association between fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) parameters and tumor-infiltrating immune cell subsets in pre-treatment biopsies with pCR in patients with stage II–III NSCLC treated with immune checkpoint inhibitor (ICI)-based neoadjuvant therapy.MethodsWe retrospectively analyzed 13 patients with stage II–III NSCLC who underwent ICI-based neoadjuvant therapy and baseline FDG PET/CT. pCR was assessed in resected surgical specimens. PET/CT parameters, including tumor-to-liver ratio (TLR) and maximum standardized uptake value (SUVmax), were evaluated before and after neoadjuvant treatment. Multiplex immunofluorescence (mIF) was performed on pre-treatment tumor biopsies to quantify tumor-infiltrating immune cell subsets and functional immune markers.ResultsEight patients (62%) achieved pCR. Compared with the non-pCR group, patients with pCR showed significantly lower post-treatment TLR (median 1.51 vs. 2.14, P=0.004) and a greater relative reduction in TLR (median 0.82 vs. 0.58, P=0.004). Pre-treatment biopsies from patients with pCR demonstrated higher infiltration of CD4+ T cells (P=0.04) and increased TIGIT expression on both CD8+ and CD4+ T cells. Pre-treatment TLR positively correlated with CD68+ macrophages and Ki67+ CD8+ T cells, while changes in TLR were significantly associated with TIGIT+ tumor-infiltrating T cells. During follow-up, recurrence occurred in only one patient in the non-pCR group.ConclusionsDynamic changes in PET/CT-derived TLR and immune biomarkers from pre-treatment biopsies were associated with pCR following neoadjuvant immunotherapy in stage II–III NSCLC. These findings support the complementary role of metabolic imaging and immune profiling in improving patient selection and perioperative decision-making.Keywords: Fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT), neoadjuvant immunotherapy, non-small cell lung cancer (NSCLC), pathologic complete response (pCR), immune biomarkers Highlight box Key findings • In this retrospective cohort of patients with stage II–III non-small cell lung cancer (NSCLC) treated with immune checkpoint inhibitor-based neoadjuvant therapy, pathologic complete response (pCR) was associated with lower post-treatment tumor-to-liver ratio (TLR) and greater relative reduction in TLR on fluorodeoxyglucose positron emission tomography (PET)/computed tomography (CT). • Pre-treatment tumor biopsies from patients who achieved pCR showed higher CD4+ T cell infiltration and increased TIGIT expression on tumor-infiltrating T cell subsets. • Pre-treatment TLR positively correlated with CD68+ macrophages and Ki67+ CD8+ T cells, while changes in TLR were significantly associated with TIGIT+ tumor-infiltrating T cells. What is known and what is new? • Neoadjuvant immunotherapy has improved pathologic response and event-free survival in resectable NSCLC, and pCR is increasingly recognized as an early surrogate marker of favorable long-term outcomes. • Dynamic changes in PET/CT-derived TLR and immune biomarkers from pre-treatment biopsies were associated with pCR following neoadjuvant immunotherapy in stage II–III NSCLC. What is the implication, and what should change now? • Integrated metabolic imaging and immune profiling may improve preoperative response assessment and patient stratification in resectable NSCLC. • These findings support further validation of PET/CT-derived TLR and tumor-infiltrating immune markers as complementary biomarkers in larger prospective cohorts. Introduction Resectable non-small cell lung cancer (NSCLC) remains a major therapeutic challenge, with 5-year recurrence-free survival rates falling to 50% in stage II and 34% in stage III disease despite curative-intent surgery and perioperative therapy (1). Recent landmark trials, including CheckMate 816 and KEYNOTE-671, have demonstrated that adding immune checkpoint inhibitors (ICIs) to neoadjuvant chemotherapy improves pathologic complete response (pCR) rates and prolongs event-free survival (EFS) (2,3). Currently, pCR assessed in resected specimens after neoadjuvant ICI therapy is recognized as a strong surrogate endpoint for long-term survival in resectable NSCLC. However, accurately predicting pCR before surgery remains an unmet clinical need, as clinicians still lack robust, noninvasive biomarkers to identify patients most likely to benefit from neoadjuvant ICIs. Traditional clinicopathologic factors and programmed death-ligand 1 (PD-L1) expression alone are insufficient to guide treatment decisions (4). Functional imaging with 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) offers whole-tumor assessment beyond morphology and may reflect both tumor metabolism and immune activity in the tumor microenvironment (5). At the same time, tumor-infiltrating lymphocytes and the immune checkpoint markers they express provide complementary insight into T cell-mediated antitumor immunity (6). Integrating metabolic and immunologic biomarkers has the potential to refine patient stratification, optimize surgical timing, and improve perioperative outcomes. Therefore, in this study we retrospectively evaluated patients with stage II–III NSCLC treated with ICI-based neoadjuvant therapy. By analyzing FDG PET/CT parameters, particularly the tumor-to-liver ratio (TLR), alongside multiplex immunofluorescence (mIF) profiling of T cell subsets, we aimed to identify potential indicators of pCR. Our findings highlight a novel metabolic-immunologic signature that may help address the pressing unmet need for preoperative predictors of neoadjuvant ICI treatment response in resectable NSCLC. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-0628/rc). Methods Study design and patient selection We retrospectively reviewed patients with histologically confirmed stage II–III NSCLC who received neoadjuvant therapy containing an ICI at Korea University Guro Hospital, a tertiary academic hospital in Korea, between December 2017 and November 2023. Eligibility criteria included availability of baseline 18F-FDG PET/CT performed prior to treatment initiation, as well as pathologic confirmation of residual disease or pCR after neoadjuvant treatment completion. A pCR was confirmed by the pathologist when no viable residual tumor cells were present in both the resected primary lung tumor and all sampled regional lymph nodes (ypT0N0) after completion of neoadjuvant therapy. A patient without pathological assessment after neoadjuvant therapy was excluded (Figure 1). This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of Korea University Guro Hospital (IRB No. 2023GR0562, approval date: 12.14, 2023). The requirement for informed consent was waived by the Institutional Review Board due to the retrospective nature of the study and the use of de-identified data. Patient selection. Flowchart of screening and enrollment of 16 patients treated with neoadjuvant ICI, as well as the reasons for exclusion and the numbers of patients included in the post-treatment PET and immune marker analyses. ICI, immune checkpoint inhibitor; PET, positron emission tomography. Treatment and clinical data collection All patients received a programmed cell death protein 1 (PD-1) targeting monoclonal antibody (PD-1 inhibitor) combined with platinum-based chemotherapy, except for one patient who was treated with an ipilimumab-containing regimen (Table 1). Clinical information, including age, sex, smoking history, histologic subtype, PD-L1 tumor proportion score, clinical nodal status, complete blood counts, and survival outcomes, was extracted from medical records. Table 1Baseline characteristics in patients with or without pCR (n=13)CharacteristicspCR group (n=8)Non-pCR group (n=5)P valueAge at diagnosis (years)68 [54–73]62 [57–70]0.22Sex0.38 Male84 Female01Smokers0.38 Yes84 No01Smokers (pack-year)32.5 [10–100]40 [0–60]0.94Histology>0.99 Squamous cell carcinoma43 Adenocarcinoma42PD-L1 SP26365.00 [15–100]5.50 [0–20]†0.008*cN2 stage ≥2>0.99 Yes43 No42Serum inflammatory marker NLR2.87 [1.41–7.36]1.60 [1.36–9.55]0.35 dNLR2.03 [1.06–5.49]1.08 [1.04–5.25]0.44 SII1043 [507–2,200]435 [193–7,445]0.22PET inflammatory parameters SUVmean of bone marrow (spine)1.42 [1.05–2.39]1.44 [0.94–2.38]>0.99 SUVmean of spleen1.62 [1.35–2.33]1.33 [1.16–2.45]0.62 SUVmean of liver2.03 [1.62–2.46]1.37 [1.28–3.20]0.62Regimen0.03* Nivolumab + gemcitabine + platinum03 Nivolumab/pembrolizumab + taxol + platinum40 Nivolumab + ipilimumab10 Nivolumab/pembrolizumab + pemetrexed + platinum32Data are presented as median [range] or number. Variables include age, sex, smoking history, histology, PD-L1 status, clinical stage, serum inflammatory markers, PET inflammatory parameters and treatment regimen. †, available for 4 out of 5 patients. *, P<0.05. cN, clinical node; dNLR, derived neutrophil-to-lymphocyte ratio; NLR, neutrophil-to-lymphocyte ratio; pCR, pathologic complete response; PD-L1, programmed death-ligand 1; PET, positron emission tomography; SII, systemic immune-inflammation index; SUVmean, mean standardized uptake value. Baseline characteristics in patients with or without pCR (n=13) Data are presented as median [range] or number. Variables include age, sex, smoking history, histology, PD-L1 status, clinical stage, serum inflammatory markers, PET inflammatory parameters and treatment regimen. †, available for 4 out of 5 patients. *, P<0.05. cN, clinical node; dNLR, derived neutrophil-to-lymphocyte ratio; NLR, neutrophil-to-lymphocyte ratio; pCR, pathologic complete response; PD-L1, programmed death-ligand 1; PET, positron emission tomography; SII, systemic immune-inflammation index; SUVmean, mean standardized uptake value. PET/CT acquisition and analysis All PET/CT scans were performed under a routine clinical protocol. Imaging was conducted on either a Gemini TF 16-channel PET/CT scanner (Philips Medical Systems, Cleveland, OH, USA) or Discovery MI PET/CT (GE Healthcare, Waukesha, WI, USA). A nuclear medicine physician delineated volumes of interest (VOIs) for the primary tumor and suspected lymph nodes using MIM software (version 7.3.8) and maximum standardized uptake value (SUVmax) was measured. Mean SUVs were also obtained from the liver (30 mm VOI), spleen (20 mm VOI), and bone marrow (10 mm VOI on the L3, L4, and L5 spines). TLR was calculated as the ratio of tumor SUVmax to liver SUVmean. For patients with post-treatment PET/CT, changes in PET parameters were calculated as: (pre-PET parameter − post-PET parameter)/pre-PET parameter. Tissue staining with mIF Formalin-fixed paraffin-embedded (FFPE) pre-treatment tumor biopsy specimens were subjected to mIF to assess immune cell subsets using multiple antibody panels designed to detect the expression of CD3, CD4, CD8, CD20, CD68, CD45, FOXP3, Ki67, granzyme B (GZB), TIA1, and TIGIT. FFPE tissue sections were dewaxed in xylene, rehydrated through graded ethanol concentrations, and subjected to heat-mediated antigen retrieval using Tris-ethylenediaminetetraacetic acid (EDTA) buffer (Abcam, Cambridge, UK). Sections were then blocked with 0.25% casein and 10% human serum, followed by incubation with primary antibodies and corresponding secondary antibodies (Table S1). All slides were counterstained with 4',6-diamidino-2-phenylindole (DAPI), mounted using ProLong Gold (Invitrogen, Waltham, MA, USA), and scanned on the PhenoImager HT instrument (Quanterix/Akoya Biosciences, Billerica, MA, USA). After whole-slide scanning, multispectral images of tumor regions were acquired at 20× magnification (0.5 µm/pixel) using multispectral filters. Acquired multispectral images were unmixed, and cell segmentation and phenotyping were performed using inForm software (Akoya Biosciences) following the workflow we previously described (7). Phenotype summaries were then analyzed using the phenoptr and phenoptrReports R packages (version 0.3.1). Immune cell densities were indicated as percentages of total nucleated cells or of specific parental cell subsets. CD4 T cells were defined as CD3+ CD8− T cells. Statistical analysis Categorical variables were compared between the pCR and non-pCR groups using Fisher’s exact test, and continuous variables were compared using the Mann-Whitney U test. Correlations between PET parameters and immune cell subsets were evaluated using Spearman’s rank correlation coefficient. Survival analysis was performed using the log-rank test. A two-sided P value <0.05 was considered statistically significant. Statistical analyses were performed using IBM SPSS Statistics (version 20). Results Patient characteristics Thirteen patients with stage II–III NSCLC who underwent ICI-based neoadjuvant therapy were included. The median age was 66 years (range, 54–73 years), and most were male (92%) and current or former smokers (92%). Adenocarcinoma and squamous cell carcinoma were equally represented. Twelve patients proceeded to surgery, while one underwent chemoradiotherapy after residual disease was confirmed on biopsy. pCR was achieved in 8 patients (62%), whereas 5 had residual viable tumor (Table 1). Expression of PD-L1 was significantly higher in the pCR group than in the non-PCR group. PET/CT parameters and pathologic response Pre-treatment TLR did not differ significantly between the pCR and non-pCR groups (median 8.04 vs. 5.77, P=0.065). Post-treatment PET/CT (n=11) demonstrated significantly lower TLR in patients with pCR (median 1.51 vs. 2.14, P=0.004) and greater relative reduction in TLR (median 0.82 vs. 0.58, P=0.004). SUVmax reduction was also more pronounced in the pCR group (median 0.83 vs. 0.49, P=0.009) (Table 2, Figure 2). Table 2Pre- and post-treatment PET/CT parameters and pathologic responsePET parameterspCR group (n=8)Non-pCR group (n=5)P valueTLR Pre-treatment (n=13)8.04 (6.28 to 11.22)5.77 (5.08 to 7.96)0.07 Post-treatment (n=11)†1.51 (0.93 to 1.99)2.14 (2.11 to 9.70)0.004* Changes (n=11)†0.82 (0.72 to 0.86)0.58 (−0.22 to 0.70)0.004*SUVmax of total tumor Pre-treatment (n=13)16.36 (13.43 to 19.04)7.90 (6.61 to 23.23)0.62 Post-treatment (n=11)†2.57 (2.24 to 5.38)4.72 (3.36 to 17.94)0.052 Changes (n=11)†0.83 (0.72 to 0.86)0.49 (0.23 to 0.77)0.009*Data are presented as median (range). Comparison of TLR and SUVmax between patients achieving pCR and non-pCR. †, available for 6 patient in pCR group. *, P<0.05. CT, computed tomography; pCR, pathologic complete response; PET, positron emission tomography; SUVmax, maximum standardized uptake value; TLR, tumor-to-liver ratio. Pre- and post-treatment PET/CT parameters and pathologic response Data are presented as median (range). Comparison of TLR and SUVmax between patients achieving pCR and non-pCR. †, available for 6 patient in pCR group. *, P<0.05. CT, computed tomography; pCR, pathologic complete response; PET, positron emission tomography; SUVmax, maximum standardized uptake value; TLR, tumor-to-liver ratio. Comparison of PET/CT parameters between pCR and non-pCR groups. (A) Pre-treatment TLR. (B) Post-treatment TLR. (C) Relative reduction in TLR. pCR, pathologic complete response; PET/CT, positron emission tomography/computed tomography; TLR, tumor-to-liver ratio. Immune biomarker analysis mIF was performed in 10 patients with available pre-treatment biopsy specimens (Figure 3A-3F). The pCR group exhibited significantly higher infiltration of CD4+ T cells (median 9.36% vs. 4.77%, P=0.04; Figure 3G), although total CD3+ T cell density was not significantly different (P=0.07; Table 3). GZB+ CD8 T cells and TIA1+ CD8 T cells were also not significantly different in the pCR group (Figure 3H, Table 3). TIGIT expression was observed on both CD8 and CD4 T cells (Figure 3E,3F), and pCR cases had significantly higher proportions of TIGIT+ CD8 T cells among CD8 T cells (Figure 3I). Furthermore, the proportions of TIGIT+ CD8 T cells, TIGIT+ CD4 T cells, or TIGIT+ CD3 T cells among total cells detected were also significantly higher in the pCR group (P=0.04 in each case) (Table 3, Figure 3J-3L). Tumor-infiltrating immune cells in pre-treatment biopsies were visualized using mIF. Biopsy samples were stained with antibody panels to assess the presence of tumor-infiltrating immune cells including T cells, B cells, and macrophages (A,B) and functional markers expressed by T cells (C-F). Pan-CK was used as a tumor marker. Biopsies from the pCR group contained greater infiltration of CD4 T cells (A) compared with those from the non-pCR group (B) as shown in Table 3. GZB-expressing T cell subpopulations were present in both the pCR (C) and non-pCR (D) groups. The pCR group (E) exhibited higher infiltration of TIGIT+ T cell subpopulations compared with the non-pCR group (F). Scale bars represent 50 µm. Quantitative analyses demonstrated a trend toward higher proportions of GZB+ CD8 T cells in the pCR group compared with the non-pCR group (H), whereas CD4 T cells (G) and TIGIT+ T cell subpopulations were significantly enriched in the pCR group (I-L). CK, cytokeratin; GZB, granzyme B; mIF, multiplex immunofluorescence; pCR, pathologic complete response. Table 3Immune biomarker analysis in pre-treatment biopsiesImmune markerspCR group (n=6)Non-pCR group (n=4)P valueCD20/total cells (%)0.75 (0.25–3.71)0.47 (0.05–8.68)0.76CD4/total cells (%)9.36 (5.59–11.41)4.77 (3.03–6.85)0.04*CD8/total cells (%)4.62 (1.97–22.25)2.04 (1.07–5.14)0.17CD3/total cells (%)13.53 (7.97–34.82)6.68 (4.18–12.00)0.07CD68/total cells (%)5.25 (0.80–11.67)4.54 (2.04–8.11)0.76CD45/total cells (%)13.69 (6.55–56.28)12.52 (6.01–24.88)0.61FOXP3/total cells (%)0.44 (0.24–2.85)0.42 (0.10–0.97)0.76Ki67+ CD8 cells/total cells (%)0.31 (0.01–2.27)0.09 (0.00–0.32)0.11Ki67+ CD8 cells/CD8 cells (%)6.30 (0.82–16.18)1.89 (0.23–23.68)0.35GZB+ CD8 cells/total cells (%)3.12 (0.18–8.03)0.25 (0.14–1.90)0.07GZB+ CD8 cells/CD8 cells (%)49.58 (10.61–67.28)21.11 (4.38–37.21)0.07TIA1+ CD8 cells/total cells (%)2.04 (0.32–15.71)0.53 (0.08–1.98)0.17TIA1+ CD8 cells/CD8 cells (%)36.86 (18.37–78.72)18.99 (7.89–38.70)0.17TIGIT+ CD8 cells/total cells (%)0.84 (0.23–6.59)0.12 (0.00–0.43)0.04*TIGIT+ CD8 cells/CD8 cells (%)25.26 (10.68–47.49)5.81 (0.00–11.78)0.02*TIGIT+ CD3 cells/total cells (%)3.16 (1.12–12.58)0.63 (0.10–1.85)0.04*TIGIT+ CD3 cells/CD3 cells (%)30.85 (16.89–54.06)15.97 (6.38–20.16)0.07TIGIT+ CD4 cells/total cells (%)2.32 (0.9–5.99)0.51 (0.10–1.42)0.04*TIGIT+ CD4 cells/CD4 cells (%)40.65 (19.63–60.45)21.25 (10.71–26.28)0.11Data are presented as median (range). Expression levels of CD3+, CD4+, CD8+, TIGIT+ subsets, and cytotoxic markers (GZB, TIA1) stratified by pCR vs. non-pCR. *, P<0.05. GZB, granzyme B; pCR, pathologic complete response; TLR, tumor-to-liver ratio. Immune biomarker analysis in pre-treatment biopsies Data are presented as median (range). Expression levels of CD3+, CD4+, CD8+, TIGIT+ subsets, and cytotoxic markers (GZB, TIA1) stratified by pCR vs. non-pCR. *, P<0.05. GZB, granzyme B; pCR, pathologic complete response; TLR, tumor-to-liver ratio. Correlation between PET parameters and immune markers Pre-treatment TLR positively correlated with CD68+ macrophages (ρ=0.88, P=0.004) and Ki67+ CD8+ T cells (ρ=0.71, P=0.047) (Table 4). Changes in TLR were strongly associated with TIGIT-expressing T cells among total cells detected in pre-treatment biopsies (Table 4), including TIGIT+ CD3, CD4, and CD8 T cells (ρ=0.71, P=0.047 each). Post-treatment TLR inversely correlated with the proportion of TIA1+ CD8 cells among total CD8 T cells in pre-treatment biopsies (ρ=−0.74, P=0.037) (Table 4). Table 4Correlation between PET/CT parameters and immune biomarkers (n=8)Immune markersPre-treatment TLRPost-treatment TLRChange in TLRRhoPRhoPRhoPCD20/total cells (%)0.5000.21−0.0710.870.0480.91CD4/total cells (%)0.5030.20−0.6110.110.6950.056CD8/total cells (%)0.4760.23−0.5480.160.6430.09CD3/total cells (%)0.4760.23−0.5240.180.5950.12CD68/total cells (%)0.8810.004*−0.4050.320.4520.26CD45/total cells (%)0.2140.61−0.3810.350.2860.49FOXP3/total cells (%)−0.4290.29−0.0480.910.1190.78Ki67+ CD8 cells/total cells (%)0.7140.047*−0.5480.160.5710.14Ki67+ CD8 cells/CD8 cells (%)0.6190.10−0.1430.740.0950.82GZB+ CD8 cells/total cells (%)0.4790.23−0.6350.090.7070.050*GZB+ CD8 cells/CD8 cells (%)0.0950.82−0.6900.0580.6190.10TIA1+ CD8 cells/total cells (%)0.3810.35−0.6430.090.6900.058TIA1+ CD8 cells/CD8 cells (%)0.0950.82−0.7380.04*0.6900.058TIGIT+ CD8 cells/total cells (%)0.2860.49−0.6430.090.7140.047*TIGIT+ CD8 cells/CD8 cells (%)0.2860.49−0.6900.0580.7860.02*TIGIT+ CD3 cells/total cells (%)0.2860.49−0.6430.090.7140.047*TIGIT+ CD3 cells/CD3 cells (%)0.0950.82−0.4760.230.5950.12TIGIT+ CD4 cells/total cells (%)0.2860.49−0.6430.090.7140.047*TIGIT+ CD4 cells/CD4 cells (%)−0.0240.96−0.3810.350.4760.23*, P<0.05. Spearman correlation coefficients between TLR dynamics (pretreatment, post-treatment, reduction) and immune cell subsets. PET/CT, positron emission tomography/computed tomography; TLR, tumor-to-liver ratio. Correlation between PET/CT parameters and immune biomarkers (n=8) *, P<0.05. Spearman correlation coefficients between TLR dynamics (pretreatment, post-treatment, reduction) and immune cell subsets. PET/CT, positron emission tomography/computed tomography; TLR, tumor-to-liver ratio. Recurrence outcomes During a median follow-up of 26.4 months, 1 patient developed tumor recurrence. While the log-rank test did not reach statistical significance (P=0.11), the recurrence occurred in the non-pCR group within 1 year after therapy completion (Figure 4). Recurrence-free survival stratified by pathologic response. Early recurrence event is observed in the non-pCR group. Case 2 was an inoperable case after neoadjuvant therapy with pathologic confirmation of residual tumor on biopsy. cTLR, changes in tumor-to-liver ratio; GZB, granzyme B; NA, not available; pCR, pathologic complete response; PD-L1, programmed death-ligand 1. Discussion In this study, we demonstrated that changes in FDG PET/CT parameters, particularly TLR, together with the presence of several immune biomarkers in pre-treatment biopsies, are associated with pCR in patients with stage II–III NSCLC receiving neoadjuvant immunotherapy. To our knowledge, this is among the first studies to integrate metabolic imaging and immune profiling in this setting, providing proof-of-concept evidence for a complementary biomarker strategy. Our results are clinically meaningful because reliable preoperative predictors of pCR remain a critical unmet need in perioperative management. Comparison with existing clinical evidence The prognostic value of pCR has been firmly established in recent landmark trials. In CheckMate 816, neoadjuvant nivolumab plus chemotherapy improved both pCR and EFS compared with chemotherapy alone, validating pCR as a surrogate for long-term benefit (2). The KEYNOTE-671 trial confirmed these findings with perioperative pembrolizumab (3). Additional trials, such as NADIM I and II (8) and SAKK 16/14 (9), consistently demonstrated that patients achieving pCR or major pathologic response (MPR) have superior survival outcomes. Yet, none of these pivotal studies identified robust, noninvasive predictors of pCR before surgery beyond PD-L1 expression, which is known to be heterogeneous and sometimes unreliable (10). Our study addresses this gap by showing that PET-derived TLR dynamics, in combination with T cell biomarkers, are associated with pCR. Role of FDG PET/CT as a predictive tool Several studies have evaluated FDG PET/CT for predicting immunotherapy response. Tao et al. reported that higher baseline SUVmax predicted MPR to PD-1 blockade in resectable NSCLC (5). Chen and colleagues demonstrated that dynamic FDG PET/CT could identify major pathologic responders (11), while Zhuang et al. found that uptake values correlated with outcomes of chemo-immunotherapy (12). Other groups have suggested that volumetric measures such as metabolic tumor volume (MTV) or total lesion glycolysis (TLG) provide additional prognostic value (13,14). Recent study demonstrated that post-therapy PET/CT parameters and their changes are robust predictors of pCR (15). Our study expands this literature by focusing on TLR, which normalizes tumor uptake against liver activity and thus further reduces variability. While SUVmax is a simple and widely accepted index, it is susceptible to multiple confounding factors, and previous study have demonstrated that further normalizing SUVmax against liver tissue may provide a superior prognostic indicator (16). As noted in the “Methods” section, paired longitudinal PET/CT images were utilized in this study; furthermore, due to a replacement of the PET scanner during the study period, two different scanners were used. Given our relatively small patient cohort, minimizing both inter-scanner variability and intra-patient variability may improve the result’s reliability. Importantly, both post-treatment TLR and the degree of TLR reduction were significantly associated with pCR and yielded a better P value compared to SUVmax in our cohort, highlighting the potential of TLR as a simple yet robust metric. In this small study, the largest reductions in TLR over the course of neoadjuvant treatment were observed in patients whose excised tumors were subsequently confirmed to have achieved a pCR. Integration with immune biomarkers A key novelty of our study is the integration of PET parameters with immune profiling. CD4 T cells, and TIGIT+ T cells (including both CD8 and CD4 subsets of TIGIT+ T cells) were significantly enriched in pre-treatment biopsies from pCR cases. These results align with prior evidence that T cell activation is essential for effective ICI response (17). Expression of TIGIT, often considered an “exhaustion” marker with negative prognostic impact, may paradoxically reflect ongoing immune activation (18,19). Ku et al. previously showed that distinct immune infiltrates correlate with outcomes to ICIs in advanced NSCLC (17), supporting our findings in the perioperative context. Correlation analyses demonstrated that higher pre-treatment TLR was positively associated with CD68+ macrophage infiltration and Ki67+ CD8+ T cell abundance in baseline tumor biopsies. Our observation that pre-treatment TLR correlates positively with both macrophage density and proliferating cytotoxic T cells suggests that increased metabolic activity on FDG PET-CT may reflect not only intrinsic tumor glycolysis but also an immune-engaged tumor microenvironment. Elevated glucose uptake in tumors has been reported to associate with higher levels of immune-cell infiltration, including macrophage populations and T-lymphocytes, indicating that PET-derived metabolic parameters may capture the composite metabolic demands of both cancer cells and infiltrating immune cells rather than tumor burden alone (20). Furthermore, tumors exhibiting heightened metabolic signatures on PET often harbor enriched inflammatory signatures and activated immune subsets, supporting the concept that metabolically “hot” tumors may simultaneously display increased glucose consumption by immune cells such as proliferating CD8+ T cells. By demonstrating strong correlations between TLR dynamics and T cell functional exhaustion markers, our study further bridges metabolic imaging with immune biology. In particular, the associations between pre-treatment PET uptake and markers of T cell proliferation (Ki67) and activation (TIGIT) suggest that FDG uptake may, in part, reflect ongoing immune activation in addition to tumor metabolic activity, providing a coherent mechanistic explanation for the immunometabolic phenotype observed in our analysis. Clinical implications These findings have several potential applications in clinical practice. First, they may help identify patients before surgery who are most likely to achieve pCR, informing surgical decision-making and perioperative planning. For instance, patients with profound TLR reduction and favorable immune signatures may proceed directly to surgery with high confidence in therapeutic benefit. Conversely, patients lacking these features may be candidates for intensified regimens, closer monitoring, or enrollment in clinical trials testing novel perioperative strategies. Second, our integrated approach may reduce dependence on PD-L1 as the sole biomarker, addressing its limitations related to spatial and temporal heterogeneity. Third, by combining imaging and immune profiling, clinicians can gain complementary insights: PET/CT offers a whole-tumor perspective, while immune assays provide mechanistic understanding at the cellular level. Strengths and novelty of this study Several elements enhance the novelty of our study. In addition to prior PET studies that primarily analyzed SUVmax, we demonstrated that TLR—a normalized parameter—may be a more robust parameter in predicting pCR. Moreover, we incorporated immune markers into the analysis, showing that imaging and tissue biomarkers can complement and reinforce each other. The triangulation of PET/CT, immunofluorescence, and pathologic outcomes, even in a small cohort, provides a multidimensional perspective that strengthens the validity of our findings. Limitations and strategies to address them While our findings reveal intriguing correlations between pre-treatment PET/CT metrics, immune cell subsets, and pCR, several limitations must be acknowledged. Most notably, our sample size is small (n=13), which inherently limits statistical power and increases the risk of spurious correlations. Consequently, these associations should be interpreted with caution and are strictly hypothesis-generating. Future studies with larger validation cohorts are required to confirm these relationships. To mitigate this, we performed paired analyses (11 patients with pre- and post-treatment PET/CT) and included multiplex immune profiling (10 patients), with 8 patients contributing both datasets for integrated analysis. Even with these measures, the findings should be interpreted with caution. Second, this was a retrospective, single-institution study, potentially introducing selection bias. We minimized this by including all consecutive eligible patients and applying standardized imaging protocols. Third, immune profiling relied on pre-treatment biopsies, which may not fully capture intratumoral heterogeneity. Additionally, our analysis did not differentiate between the tumor core and stroma because the available tissue area was extremely limited with highly variable proportions of stroma across different samples. We addressed this by pairing tissue data with PET/CT, which reflects global tumor characteristics. Fourth, the follow-up duration was short (median 26.4 months), precluding definitive survival analyses. Nevertheless, recurrence data were systematically captured, and the recurrence occurred within 1 year, consistent with known relapse dynamics. Fifth, external validation was lacking, but rigorous statistical testing and cross-validation across imaging and immunologic data provide internal consistency. Lastly, heterogeneity in neoadjuvant treatment regimens represents an important limitation of this study. Although patients received different therapeutic combinations, all regimens were based on PD-1 inhibitor-centered neoadjuvant treatment with or without platinum-based chemotherapy, reflecting contemporary clinical practice rather than fundamentally different therapeutic classes. Accordingly, our findings should be interpreted as demonstrating associations between PET/CT-derived metabolic biomarkers, pre-treatment immune microenvironment characteristics, and pCR within the context of PD-1 inhibitor-based neoadjuvant therapy as a whole, rather than the independent effects of individual agents. Because chemotherapy itself may influence both tumor metabolism and immune remodeling, the relative contribution of each treatment component cannot be distinguished in this small exploratory cohort. Future prospective studies with larger sample sizes and more homogeneous treatment regimens are warranted to validate these findings and determine regimen-specific biomarker performance. Future directions Future prospective multicenter trials are required to validate these findings. Embedding PET/CT metrics such as TLR into ongoing perioperative immunotherapy studies could test their predictive utility in larger cohorts. Moreover, serial imaging and immune profiling during therapy could provide dynamic insights into treatment response, enabling adaptive treatment strategies. Development of integrated predictive models that combine metabolic and immunologic signatures may ultimately facilitate personalized perioperative therapy in NSCLC. Conclusions In conclusion, our study demonstrates that FDG PET/CT-derived TLR and immune biomarkers are strong candidates for complementary predictors of pCR in stage II–III NSCLC treated with neoadjuvant immunotherapy. By addressing limitations of current biomarkers and integrating imaging with immunology, this approach provides a novel pathway toward improving patient selection and optimizing perioperative outcomes in thoracic oncology. At present, our results offer promising potential but remain hypothesis-generating, highlighting the clear necessity for future large-scale, multi-center studies. Supplementary The article’s supplementary files as