Impact of spatial distribution of M2 macrophages on prognosis and neoadjuvant chemotherapy resistance in gastric cancer

Introduction

Gastric cancer (GC) is one of the most common malignancies worldwide, characterized by high incidence and mortality rates. In 2022, there were approximately one million new cases of GC globally, and its mortality rate ranked fifth among all malignant tumors (1). This is primarily because most GC cases are diagnosed at an advanced stage, leading to a poor prognosis. Neoadjuvant chemotherapy (NAC) is currently a critical therapeutic approach for locally advanced GC. It aims to reduce tumor volume, eradicate micrometastases, and improve the R0 resection rate through preoperative chemotherapy. However, clinical observations reveal that approximately 30–40% of GC patients exhibit primary resistance to NAC, resulting in suboptimal pathological complete response (pCR) rates (2,3). These patients not only fail to achieve preoperative downstaging but may also experience delayed surgical timing and endure unnecessary toxic side effects. Ultimately, they face an exceptionally high risk of postoperative recurrence and a dismal long-term prognosis. Furthermore, developing precise preoperative evaluation models, such as those assessing lymph node predictive value, has been shown to be crucial for stratifying patient risks and guiding therapeutic decisions (4). Therefore, there is an urgent clinical need to identify biomarkers capable of predicting chemosensitivity and long-term patient prognosis, as well as elucidating the microenvironmental mechanisms underlying chemoresistance.

Tumor-associated macrophages (TAMs) constitute the most abundantly infiltrated and functionally plastic immune cell population within the tumor microenvironment (TME). They are primarily polarized into M1 (anti-tumor) and M2 (pro-tumor) phenotypes; in the TME, the vast majority are induced to differentiate into the M2 phenotype. Consequently, M2 macrophages act as key stromal cells driving malignant progression, immune evasion, and angiogenesis in GC. Furthermore, numerous bulk RNA-sequencing (Bulk RNA-seq) studies have demonstrated that high infiltration of M2 macrophages is significantly correlated with shortened survival and reduced chemosensitivity in GC patients (5-7). They are also widely recognized in the academic community as an important independent risk factor for poor prognosis in GC. Although the pro-tumor role of M2 macrophages is fundamentally established, traditional infiltration assessments often overlook the spatial structural heterogeneity of the TME. From anatomical and pathological perspectives, the TME exhibits profound spatial heterogeneity. The intratumoral core is the central region of dense tumor cell growth, whereas the peritumoral stroma is primarily composed of fibroblasts, blood vessels, and immune cells. These two compartments differ fundamentally in physical barriers, nutrient supply, and immune infiltration. Therefore, whether M2 macrophages of similar abundance infiltrate the tumor core or are sequestered in the peritumoral stroma may exert distinctly different impacts on their functional roles (8-10).

Based on this, we hypothesize that the spatial distribution of M2 macrophages within the GC microenvironment is a pivotal factor determining the efficacy of NAC and patient prognosis. In this study, we established a “pure baseline” cohort using public databases such as The Cancer Genome Atlas (TCGA). Through transcriptomic analysis, we confirmed the correlation between M2 macrophages and GC prognosis and revealed a highly significant positive correlation between the M2 marker MRC1 and the core pro-fibrotic factor TGFB1. This elucidates, at the whole-genome level, the central role of M2 macrophages in driving stromal remodeling and constructing resistance barriers. Subsequently, we enrolled 10 GC patients and employed multiplex immunofluorescence (mIF) to spatially quantify M2 macrophages in both the intratumoral core and the peritumoral stroma. By calculating the tumor-to-peritumoral ratio (TPR) and integrating the internationally recognized Mandard tumor regression grade (TRG) (11), we investigated the association between the spatial distribution of M2 macrophages, chemotherapy response, and prognosis. This study aims to provide a novel spatial perspective on the mechanisms of NAC resistance in GC and offer a theoretical basis for personalized chemotherapy decision-making and clinical prognostic evaluation. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0398/rc).

Methods

Bioinformatics data sources and analysis

Data acquisition

Data were downloaded from the TCGA database, comprising transcriptomic, miRNA, lncRNA, and survival data, along with corresponding clinical characteristics. The dataset included 408 gastric adenocarcinoma tissues and 36 adjacent normal gastric mucosa tissues.

The inclusion criteria were: (I) tissue samples with a complete transcriptomic expression matrix; (II) primary lesion samples pathologically confirmed as gastric adenocarcinoma (treatment-naïve); and (III) availability of core clinical characteristics (e.g., age, sex, stage) as well as definitive overall survival (OS) and survival status records.

The exclusion criteria were: (I) patients with a follow-up time of less than 30 days; (II) samples with severe deficiencies in clinicopathological information; and (III) patients who had received any form of anti-tumor therapy, such as NAC, radiotherapy, or targeted/immunotherapy, prior to tissue sampling.

Consequently, 408 treatment-naïve samples were filtered to establish a “pure baseline” cohort, devoid of pharmacological interventions. This cohort was utilized to validate the natural multi-omic association between M2 macrophages and the stromal barrier, providing a reference for subsequent pathological validation in the clinical cohort.

Selection of M2 macrophage markers

Based on literature reviews and research experience, CD163 and CD206 were selected as signature markers for M2 macrophages. High expression levels of CD163 and CD206 effectively evaluate the M2 polarization status of TAMs.

Association analysis of clinicopathological characteristics

Stratified analysis of GC patients with complete clinicopathological data in The Cancer Genome Atlas Stomach Adenocarcinoma (TCGA-STAD) cohort was conducted using the UALCAN database. The relationship between the expression levels of M2 markers and patients’ clinical stage and tumor grade was investigated to observe whether M2 markers exhibited an upregulation trend concomitant with increased tumor malignancy, thereby revealing their potential role in GC progression.

Survival and prognostic evaluation

After excluding patients with missing overall survival data, a total of 371 GC patients with complete OS data were screened from the TCGA database utilizing the Kaplan-Meier Plotter online tool. Based on the expression levels of CD163 and MRC1, and to avoid statistical bias introduced by artificially set thresholds (such as the median), the system’s built-in “Auto-select best cutoff” algorithm was employed. By continuously traversing the expression values of all patients, the expression level that maximized the significance of the survival difference (log-rank test) between the two groups was selected as the optimal cutoff. Patients were accordingly dichotomized into high- and low-expression groups, and Kaplan-Meier survival curves were generated to evaluate the predictive value of M2 TAM infiltration abundance on the OS of GC patients.

Correlation validation between M2 macrophage markers and pro-fibrotic factors

Utilizing the Gene_Corr module of TIMER2.0, the expression correlation between M2 markers and core pro-fibrotic factors was deeply mined, after excluding cases without complete gene expression matrices, across 375 treatment-naïve gastric adenocarcinoma samples from the TCGA-STAD cohort. This analysis aimed to clarify the potential driving role of M2 macrophages in the formation of the GC stromal barrier. To establish the synergistic relationship between M2 macrophage infiltration and GC stromal remodeling from an omics perspective, MRC1 (encoding CD206) and TGFB1 were selected as representatives of the M2 marker and the pro-fibrotic effector molecule, respectively. The Spearman’s rank correlation coefficient was used to evaluate their co-expression levels in GC tissues.

Functional enrichment and hub network analysis of the core gene set

Using Spearman correlation analysis with MRC1 as the anchor, a top 100 core gene set highly synergistic with M2 macrophages was identified genome-wide. Subsequently, utilizing the Weishengxin platform (https://bioinformatics.com.cn/) and the clusterProfiler algorithm package (with Homo sapiens as the reference), Gene Ontology (GO) functional enrichment analysis encompassing biological process (BP), cellular component (CC), and molecular function (MF) was performed. A multiple-panel dot plot was generated to highlight significantly enriched pathways (P<0.001), thereby revealing the potential molecular mechanisms by which M2 macrophages participate in microenvironmental remodeling.

Clinical sample collection and patient data

Study design and inclusion criteria

Formalin-fixed, paraffin-embedded (FFPE) tissue specimens were collected from 10 patients who underwent surgical resection for GC at Hunan Cancer Hospital between January 2024 and June 2025. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Informed consent was obtained from all enrolled patients, and the study protocol was approved by the Institutional Review Board of Hunan Cancer Hospital (No. SBQLL-2024-176).

The inclusion criteria were: (I) patients with newly diagnosed primary gastric adenocarcinoma confirmed by histopathology, referencing the Chinese Society of Clinical Oncology (CSCO) Guidelines for the Diagnosis and Treatment of Gastric Cancer (2025) (12); (II) patients who underwent radical gastrectomy (R0 resection) with complete postoperative pathological data; (III) aged 18–75 years, regardless of sex.

The exclusion criteria were: (I) severe cardiac, pulmonary, hepatic, or renal dysfunction; (II) history of other malignancies or synchronous multiple primary cancers; (III) incomplete clinicopathological data.

Sample grouping and tissue collection

The 10 patients were divided into a neoadjuvant therapy group and an upfront surgery group based on preoperative intervention status. Within the neoadjuvant therapy group, patients were further sub-stratified into a chemosensitive group and a chemoresistant group according to the postoperative Mandard TRG. Baseline clinicopathological data [e.g., age, sex, tumor location, histological differentiation, clinical tumor-node-metastasis (TNM) stage] are detailed in Table 1. Statistical analysis revealed no significant differences in these baseline characteristics between the chemosensitive and chemoresistant groups (P>0.05), indicating good clinical comparability. The characteristics of the upfront surgery group are presented in the table as a control for the natural progression state. Morphological validation was performed on all enrolled tissue specimens to ensure the presence of intact tumor parenchyma and peritumoral stroma, satisfying the requirements for subsequent spatial analysis. All clinical data involved in the study were strictly anonymized.

Surgically resected specimens were routinely dehydrated and paraffin-embedded to produce 4-µm-thick continuous sections. Tumor and adjacent tissues for each patient were collected to prepare independent sections. For spatial quantitative analysis, these were delineated into the “intratumoral core” and “peritumoral stroma” for comparative assessment, exploring the spatial heterogeneity of the TME.

Core marker selection and mIF staining

To comprehensively evaluate the microenvironmental status, two core markers, CD206 (encoded by MRC1) and CD163, were selected. mIF was employed for in situ co-staining of the tissue sections, enabling the observation of M2 macrophage spatial distribution on the same slide.

Specific primary antibodies were utilized: mouse anti-human CD163 monoclonal antibody (Clone: UMB154, dilution 1:500, Abcam, Cambridge, UK) and rabbit anti-human CD206 monoclonal antibody (Clone: EPR23418-26, dilution 1:1,000, Abcam).

Staining procedure

• Hematoxylin and eosin (H&E) staining and region identification: the first section was stained with H&E. Under light microscopy, the specimen was confirmed as tumor or adjacent tissue, structural integrity was verified, and valid tissue regions for subsequent fluorescence analysis were identified (excluding necrosis, hemorrhage, and fragmented edges).
• Immunofluorescence staining: adjacent sections were deparaffinized, hydrated, and subjected to antigen retrieval.
• Primary antibody incubation: a mixture of mouse anti-human CD163 and rabbit anti-human CD206 monoclonal antibodies was applied and incubated overnight at 4 ℃ in a wet chamber.
• Secondary antibody incubation: corresponding species-specific fluorescent secondary antibodies (e.g., Cy3-conjugated for CD163, FITC-conjugated for CD206) were applied and incubated in the dark.
• Counterstaining: nuclei were counterstained with DAPI.
• Controls: to exclude non-specific staining and fluorescence channel cross-talk, a negative control was established by replacing the primary antibody with phosphate-buffered saline (PBS).

Image acquisition and panoramic objective quantitative assessment

Sections subjected to mIF staining were scanned using a PANNORAMIC whole-slide scanner (3DHISTECH, Budapest, Hungary) to generate high-resolution digital slides. Digital slides were browsed using SlideViewer 2.5 software (3DHISTECH). At 200× magnification, avoiding necrotic and hemorrhagic areas, three representative fields of view (FOVs) with dense positivity or typical tissue architecture were captured per slide. Three FOVs were quantitatively analyzed using QuPath 0.5.1 and Image-Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA) software. The arithmetic mean of the three FOVs served as the final measurement for the sample. The specific metrics are defined as follows:

• Positive rate (PR) reflects the relative abundance of M2 TAMs at the cellular level (analyzed by QuPath 0.5.1).
• Mean optical density (MOD) reflects the average protein expression intensity of M2 markers within the positive areas (analyzed by Image-Pro Plus 6.0).
• Areal density (AD) comprehensively reflects the “total expression” of M2 markers across the entire FOV, integrating both expression intensity and infiltration extent (analyzed by Image-Pro Plus 6.0).
• Positive area percentage intuitively reflects the extent of M2 TAM infiltration (analyzed by Image-Pro Plus 6.0).

TRG and spatial infiltration ratio model

Pathological response evaluation for patients receiving NAC was based on the internationally recognized Mandard TRG system (11). TRG ratings were obtained from postoperative routine pathological diagnostic reports. TRG 1–2 was defined as an effective chemotherapeutic response (sensitive group), while TRG 3–5 was defined as a poor response (resistant group).

Simultaneously, leveraging the quantitative data acquired via QuPath, the TPR was constructed:

TPR=ADinintratumoralcoreADinperitumoralstroma[1]

Statistical analysis

Data collation and statistical analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism 9.0 (GraphPad Software, San Diego, CA, USA). Continuous variables lacking normal distribution were described as median (interquartile range, IQR). Comparisons between two groups were conducted using the non-parametric Mann-Whitney U test, while multi-group comparisons employed the Kruskal-Wallis test. Categorical variables were expressed as frequencies. Given the small sample size of the clinical validation cohort, between-group differences were compared using Fisher’s exact test. For survival analysis, Kaplan-Meier curves were plotted, and the log-rank test evaluated prognostic differences. Bivariate correlation was assessed using Spearman’s rank correlation to comprehensively evaluate the robustness of the TPR metric across different clinical backgrounds. All statistical tests were two-sided, and P<0.05 was considered statistically significant.

Results

High expression of M2 macrophages is significantly associated with malignant progression and poor prognosis in GC

Based on the TCGA-STAD large-sample public cohort, differential expression assessment and survival analysis were conducted for the M2 core marker CD163. The differential analysis revealed that the expression level of CD163 exhibited a continuously increasing trend with escalating GC malignancy (Figure 1A). Notably, in poorly differentiated (grade 3) GC tissues, the expression of CD163 was significantly higher than in well/moderately differentiated tissues. Furthermore, Kaplan-Meier survival analysis demonstrated that the OS of patients with high CD163 expression was significantly shorter than that of patients with low expression (hazard ratio =1.59, 95% confidence interval: 1.05–2.41, log-rank P=0.03) (Figure 1B). These data indicate that M2 macrophages are a critical factor driving malignant progression and poor prognosis in GC.

Figure 1 Clinical significance of M2 macrophage markers in the TCGA-STAD cohort. (A) Differential distribution of M2 marker expression levels across varying histological grades (grades 1–3); CD163 expression increased with escalating grade. (B) Kaplan-Meier OS curves for groups with high versus low expression of the M2 macrophage marker, indicating a significantly poorer prognosis in the high-expression group (P=0.03). CI, confidence interval; HR, hazard ratio; OS, overall survival; STAD, stomach adenocarcinoma; TCGA, The Cancer Genome Atlas.

Co-expression association between the M2 macrophage marker and core pro-fibrotic factors

To validate the driving role of M2 macrophages in constructing the physical stromal barrier of tumors, the expression correlation between M2 markers and pro-fibrotic factors was analyzed in the treatment-naïve GC cohort. As shown in Figure 2, a highly significant positive correlation exists between the M2 macrophage signature gene MRC1 (encoding CD206) and TGFB1, the core factor driving stromal remodeling. This result directly confirms at the transcriptomic level that the infiltration abundance of M2 macrophages is closely accompanied by the trend of tumor stromalization, providing solid omics evidence for their status as core regulators of the microenvironmental barrier.

Figure 2 Correlation analysis between the M2 macrophage marker MRC1 and the pro-fibrotic factor TGFB1 (P<0.001). TPM, transcripts per million.

Functional enrichment analysis of M2 core synergistic genes

Based on the co-expression relationship between M2 macrophages and pro-fibrotic factors, multi-dimensional GO enrichment analysis was further performed on the top 100 synergistic gene set to explore the molecular pathways through which they remodel the microenvironment. As illustrated in Figure 3, at the BP level, “positive regulation of cytokine production” and “myeloid leukocyte activation” were significantly enriched; CCs were primarily localized to the “secretory granule membrane” and the “external side of plasma membrane”; and the MF dimension exhibited significant features of “immune receptor activity” and “carbohydrate binding” (all P<0.001). These multi-dimensional results confirm that M2 macrophages are not in a resting state but actively participate in the stromal remodeling of the local microenvironment through vigorous membrane receptor-mediated processes and cytokine secretion functions.

Figure 3 Gene Ontology functional enrichment analysis of M2 macrophage core synergistic genes. Dot plots of enrichment for (A) biological process; (B) cellular component; and (C) molecular function.

Baseline characteristics of the clinical cohort and precise delineation of mIF spatial compartments

During the clinical validation phase, specimens from 10 patients who underwent surgical resection for GC were enrolled. Seven patients received preoperative NAC, while the remaining 3 patients (P07, P08, P09) underwent upfront radical surgery. The cohort consisted of 7 males and 3 females, with a median age of 67 years. Based on routine postoperative pathological evaluation, 4 patients in the NAC group who exhibited favorable histopathological responses—P01 (TRG 1) along with P04, P05, and P06 (all TRG 2)—were classified into the chemosensitive group. Conversely, P02 and P03 (both TRG 3), who showed obvious primary chemoresistance, along with P10—who lacked a specific TRG rating but demonstrated definitive non-response characteristics based on postoperative residuum and clinical evaluation—were collectively categorized into the chemoresistant group. There were no statistically significant differences between the two groups regarding baseline clinicopathological characteristics such as age, sex, tumor location, and TNM stage (all P>0.05) (Table 1), effectively excluding interference from baseline confounding factors.

Multidimensional measurement and spatially heterogeneous distribution of M2 macrophages

Whole-slide scanning of the sections was performed via mIF. The tissue sections were segmented into the intratumoral core and the peritumoral stroma to quantify key features within the TME and further clarify the spatial distribution relationship of tumor-infiltrating immune cells. Multi-parameter quantitative extraction of CD206 and CD163 dual-positive macrophages within both compartments was conducted. The PR, positive area percentage, MOD, and AD were utilized as core evaluation metrics for comprehensive assessment.

The measurement results indicated that patients with different chemotherapeutic responses possessed distinct local infiltration characteristics. In chemoresistant specimens (TRG 3), M2 macrophages were severely depleted within the tumor core, exhibiting a median AD of 0.219, an MOD of 70.11, a PR as low as 3.07%, and a positive area of 0.303. Conversely, in the chemosensitive group (TRG 1–2), the median AD of M2 macrophages in the core surged to 0.329, the MOD significantly enhanced to 86.40, the PR reached 8.11%, and the positive area increased to 0.412 (Figure 4). This demonstrates that the chemosensitive tumor core recruited more M2 macrophages, with a more active protein expression abundance per single cell. Calculations based solely on the overall total infiltration across the entire slide fail to accurately reflect such profound microenvironmental disparities.

Figure 4 Quantitative comparison of M2 macrophage infiltration in the tumor core between the neoadjuvant chemotherapy sensitive and resistant groups. Four objective metrics were acquired via digital image analysis to evaluate the absolute infiltration level of M2 macrophages (CD206⁺). (A) Positive rate; (B) positive area; (C) MOD; (D) AD. Data are presented as median and interquartile range. The central horizontal lines represent the median, and the box edges represent the 25th and 75th percentiles. AD, areal density; MOD, mean optical density; TRG, tumor regression grade.

Construction of the TPR and comparison of M2 spatial distribution with chemotherapeutic efficacy

By integrating the average values from the tumor core and the peritumoral stroma, we constructed a TPR model for each patient. Based on the TPR, patients were divided into two spatial distribution patterns: the peritumoral-enriched pattern (TPR <1) and the intratumoral infiltration pattern (TPR >1) (Table 2). In the peritumoral-enriched pattern, the M2 density in the stroma was significantly higher than in the core, with TPR values ranging from 0.49 to 0.87. In the intratumoral infiltration pattern, the M2 density in the core surpassed that in the stroma, with TPR values ranging from 1.08 to 7.18.

Comparing the spatial distribution characteristics of M2 with the postoperative TRG revealed that in the non-responding resistant group (TRG 3), excluding individual heterogeneous cases (e.g., Patient 10), the vast majority of patients exhibited massive accumulation of M2 macrophages in the peritumoral stroma, forming a typical peritumoral-enriched pattern. The overall TPR in this group remained low, with a median of 0.50. Conversely, in the sensitive group achieving favorable pathological regression (TRG 1–2), M2 macrophages demonstrated centripetal remodeling characteristics, with significantly elevated density in the tumor core for most patients, yielding a median TPR of 6.67 (Figure 5A).

Figure 5 Correlation analysis between spatial distribution characteristics of M2 macrophages and neoadjuvant chemotherapy efficacy. (A) Paired plot of spatial distribution trends: displays the spatial displacement of M2 macrophage AD between the peritumoral stroma and the tumor core across 7 patients receiving NAC. Blue lines (TRG 1–2) predominantly show an ascending centripetal trend; red lines (TRG 3) exhibit peripheral retention or a descending trend. (B) Correlation plot between TPR and clinical efficacy. The distribution relationship between the TPR and the TRG in the full cohort (n=10). AD, areal density; TPR, tumor-to-peritumoral ratio; TRG, tumor regression grade.

To further validate the quantitative relationship between this spatial distribution pattern and clinical efficacy, rank correlation analysis was introduced. The results revealed a significant negative correlation between the TPR values of the full cohort (n=10) and the clinical TRG rating (r_s=−0.65, P=0.043). These objective data robustly prove that the spatial retention of M2 macrophages in the peritumoral stroma (low TPR) possesses a highly robust clinical correspondence with primary resistance to NAC (Figure 5B).

M2 spatial distribution patterns in the natural baseline state of GC

Further analysis revealed that the peritumoral-enriched pattern (TPR <1) appeared not only in chemoresistant patients but also manifested a similar spatial pattern in patients from the upfront surgery group devoid of chemotherapy intervention.

For Patient 07, the stromal M2 AD was 0.242, the core AD was 0.170, and the TPR value was 0.70. For Patient 08, the stromal M2 AD was 0.272, the core AD was 0.135, and the TPR value was 0.49.

The spatial distribution of M2 macrophages in the natural baseline state predominantly exhibited a peritumoral-enriched pattern. This indicates that during the natural, untreated growth process of GC, the TME of certain patients spontaneously forms dense stromal components, thereby restricting the infiltration of immune cells into the tumor core. Meanwhile, Patient 09 in the same group retained relatively robust core infiltration characteristics (TPR =1.08).

This spatial distribution heterogeneity further elucidates the role of NAC: in patients belonging to the sensitive group who received NAC, effective chemotherapy likely breached this physical stromal defense barrier, triggering intense microenvironmental remodeling and a massive centripetal influx of M2 macrophages.

Discussion

Establishment of the M2-driven stromal barrier mechanism based on public omics data

In exploring the mechanisms underlying primary resistance to NAC, rather than directly utilizing post-chemotherapy public sequencing cohorts, we selected the treatment-naïve TCGA cohort to establish a “pure baseline.” This approach aimed to eliminate the confounding interference of secondary tissue damage induced by chemotherapeutic agents. Cohort analysis confirmed a highly significant synergistic co-expression between the M2 marker MRC1 and the core fibroblast activation gene TGFB1. Subsequent functional enrichment analysis further demonstrated that M2-related gene clusters were primarily enriched in pathways associated with stromal remodeling and connective tissue development. These lines of evidence indicate that M2 macrophages, acting as core regulators, are intrinsically involved in constructing the physical barrier. This transcriptomic foundation provides a robust basis for our subsequent evaluation of the correlation between the TPR and chemoresistance using clinical samples.

Peritumoral M2 enrichment and the “immune-exclusive” resistance barrier

Quantitative analysis of clinical samples revealed a phenomenon of “hindered centripetal infiltration” of M2 macrophages in patients with primary resistance to NAC, specifically manifesting as an inverted distribution characterized by “high peritumoral and low intratumoral” density. This spatial distribution aligns with the immune-exclusive microenvironment model proposed in recent tumor immunology, wherein a specific barrier sequesters recruited immune cells at the invasive margin of the tumor. Similar characteristics of “stromal retention and parenchymal depletion” have been observed in studies of gastrointestinal Krukenberg tumors and colorectal cancer liver metastases; such spatial segregation can concurrently drive malignant progression and treatment failure, leading to a significant reduction in OS (13,14). Compared to absolute cell quantification across the whole tissue, spatial profiling effectively elucidates the distribution of immune cells within the TME, allowing for a more accurate assessment of immune evasion status and refined risk stratification for long-term prognosis, thereby conferring higher clinical predictive value (15,16).

We hypothesize a functional interplay between M2 macrophages in the peritumoral stroma and cancer-associated fibroblasts (CAFs). Xiao et al. demonstrated that the dense extracellular matrix (ECM) network constructed by FAP⁺ CAFs physically restricts effector T cells from entering the tumor nest, corroborating this hypothesis (17). The extremely low TPR values (median 0.50) in the resistant group also suggest that M2 macrophages may secrete pro-fibrotic factors, such as TGF-β, to induce stromal remodeling and stiffening, thereby constructing a physical barrier (17-19). Under this obstruction, chemotherapeutic agents struggle to penetrate the tumor core, sparing deep-seated tumor cells from complete eradication. Protected by the stromal barrier, these residual tumor cells often cause early postoperative recurrence and ultimately compromise the prognosis. This spatial stromal barrier was also corroborated in our upfront surgery patients (Patients 07/08), indicating that a low TPR is not entirely induced by chemotherapy. Rather, during the natural progression of certain GCs, the TME spontaneously forms a barrier, dictating these patients’ innate resistance to therapy and poor prognosis (20).

Chemotherapy-induced secondary tissue remodeling

In the intratumoral core of chemosensitive patients (TRG 1–2), a high density of M2 macrophages was observed. Integrating this with the pathological regression evaluation system established by Mandard et al., we offer a more rational explanation from a dynamic developmental perspective (11,21). The high TPR in the core region is not the trigger for chemosensitivity; rather, it is the pathological footprint left by massive tumor regression following effective chemotherapy. Upon successfully penetrating the peripheral barrier, chemotherapeutic agents induce tumor cell pyroptosis or necrosis, releasing abundant damage-associated molecular patterns (DAMPs), such as HMGB1 and ATP, into the microenvironment (22,23). Peripheral monocytes are recruited by these signals and migrate into the tumor core. Among them, macrophages primarily execute efferocytosis: on the one hand, they phagocytize and clear necrotic debris; on the other hand, they secrete anti-inflammatory cytokines and induce fibrotic scar formation, thereby completing the repair and remodeling of damaged tissues (24-27). This tissue-repairing capacity is also conceptually consistent with the broader physiological roles of specific pathways in gastroprotection and the maintenance of gastric mucosal homeostasis (28). Therefore, M2 macrophages in the sensitive group act more as “scavengers” clearing the site. This finding highlights the necessity of avoiding the cognitive bias associated with static histological sections when evaluating the post-chemotherapy microenvironment, emphasizing the need to fully consider spatial cellular migration and functional remodeling triggered by therapeutic intervention.

Limitations and future perspectives

By introducing the dimension of spatial distribution, we provide a novel perspective for research on chemoresistance in GC; however, certain limitations remain. First, due to stringent inclusion criteria and the difficulty of specimen acquisition, the sample size of the NAC group was relatively limited, which may constrain the statistical power and generalizability of the conclusions.

Therefore, these results should be interpreted with caution until reproduced on a wider scale. Additionally, owing to the retrospective nature of the study, comprehensive clinical data regarding Helicobacter pylori or Epstein-Barr virus (EBV) infections, as well as specific TME features such as microsatellite instability (MSI) status and tumor-infiltrating lymphocyte (TIL) density, were not fully available for all patients. Future studies incorporating large-scale cohorts with detailed histological and comprehensive profiling are required to systematically evaluate the spatial interplay between M2 macrophages, TILs, and specific pathogen infections within the TME. Future validation of the clinical reliability and universal applicability of the TPR metric in large-scale prospective cohorts is warranted. Furthermore, this study primarily focused on M2 macrophages and has yet to systematically elucidate the spatial interplay between CD8⁺ T cells and M2 cells within the TME. The TME is a multi-component synergistic entity; future studies should leverage higher-plex mIF to systematically evaluate the spatial interaction patterns of diverse immune cells. Additionally, a few heterogeneous cases were identified during spatial pattern analysis: Patient 06 in the sensitive group exhibited a peritumoral-enriched pattern (TPR =0.87), while Patient 10 in the resistant group, despite having an extremely high clinical malignant burden (T4N2Mx), displayed partial core infiltration spatially (TPR =1.86). This suggests significant individual heterogeneity within the GC microenvironment and implies that the spatial distribution of M2 macrophages may be regulated by other unquantified stromal cells, necessitating further clarification through multi-marker panoramic landscape profiling in the future.

Conclusions

In conclusion, the clinical role of M2 macrophages in the gastric cancer microenvironment is heavily dictated by their spatial distribution. A peritumoral-enriched pattern (TPR <1) constructs a physical stromal barrier that mediates primary resistance to neoadjuvant chemotherapy, whereas high intratumoral infiltration (TPR >1) signifies tissue remodeling secondary to effective chemotherapy. The proposed TPR offers a novel and precise spatial metric for assessing chemosensitivity, highlighting the clinical necessity of shifting from absolute macrophage counting to spatial architecture profiling to guide personalized neoadjuvant strategies.

Acknowledgments

We would like to thank all the participants and staff who contributed to this study for their expert technical assistance and valuable suggestions. During the preparation of this work, the authors utilized large language model (LLM) tools to assist in translating the original Chinese draft into English and polishing the academic language to improve readability. After using these tools, the authors comprehensively reviewed, edited, and validated all contents, taking full responsibility for the final publication.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0398/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0398/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0398/prf

Funding: This work was supported by the Science and Technology Innovation 2030 - Major Project of the National Health Commission of China (Grant No. 2024ZD0521301); the Natural Science Foundation of Hunan Province (Grant No. 2025JJ50568); the Clinical Application Discipline Leader Program of the Hunan Provincial Administration of Traditional Chinese Medicine (Grant No. Xiangzhongyiyao [2022] 4); and the Key Laboratory of Integrated Traditional Chinese and Western Medicine for Tumor Prevention and Treatment of the Hunan Provincial Administration of Traditional Chinese Medicine (Grant No. 2023 [Xiangcaiyu] 0001).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0398/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 Institutional Review Board of Hunan Cancer Hospital (No. SBQLL-2024-176). Informed consent was obtained from all the patients.

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. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Calice G, Calabrese C, Notarangelo T. Innovative CAR-T approaches targeting Claudin 18.2 to counteract drug resistance in gastric cancer. Biomed Pharmacother 2025;193:118863. [Crossref] [PubMed]
3. Yufeng Z, Le X, Gong C, et al. Prediction of neoadjuvant therapy efficacy in gastric cancer: the interplay between biomarkers and radiomics and its potential for clinical translation. Front Oncol 2025;15:1642398. [Crossref] [PubMed]
4. Lamprecht CB, Kashuv T, Lucke-Wold B. Understanding metastatic patterns in gastric cancer: Insights from lymph node distribution and pathology. World J Gastrointest Oncol 2025;17:103709. [Crossref] [PubMed]
5. Hu J, Ma Y, Ma J, et al. M2 Macrophage-Based Prognostic Nomogram for Gastric Cancer After Surgical Resection. Front Oncol 2021;11:690037. [Crossref] [PubMed]
6. Hu X, Ma Z, Xu B, et al. Glutamine metabolic microenvironment drives M2 macrophage polarization to mediate trastuzumab resistance in HER2-positive gastric cancer. Cancer Commun (Lond) 2023;43:909-37. [Crossref] [PubMed]
7. Su P, Jiang L, Zhang Y, et al. Crosstalk between tumor-associated macrophages and tumor cells promotes chemoresistance via CXCL5/PI3K/AKT/mTOR pathway in gastric cancer. Cancer Cell Int 2022;22:290. [Crossref] [PubMed]
8. Wan L, Li B, Ma T, et al. An Immune-privileged niche mediates immunotherapy resistance in esophageal carcinoma. J Immunother Cancer 2026;14:e013959. [Crossref] [PubMed]
9. Karjula T, Elomaa H, Väyrynen SA, et al. Multiplexed analysis of macrophage polarisation in pulmonary metastases of microsatellite stable colorectal cancer. Cancer Immunol Immunother 2024;73:59. [Crossref] [PubMed]
10. Pavlov S, Ali E, Ambrozkiewicz F, et al. Macrophages in Colorectal Cancer: from Normal Mucosa to Distant Metastasis: Beyond the M1/M2 Paradigm. J Cancer 2026;17:157-76. [Crossref] [PubMed]
11. Mandard AM, Dalibard F, Mandard JC, et al. Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma. Clinicopathologic correlations. Cancer 1994;73:2680-6.
12. Guidelines Working Committee of CSCO. Chinese Society of Clinical Oncology (CSCO) Diagnosis and Treatment Guidelines for Gastric Cancer 2025. Beijing: People's Medical Publishing House; 2025.
13. Huang Z, Li B, Qin H, et al. Invasion characteristics and clinical significance of tumor-associated macrophages in gastrointestinal Krukenberg tumors. Front Oncol 2023;13:1006183. [Crossref] [PubMed]
14. Khanduri I, Maru DM, Parra ER. Exploratory study of macrophage polarization and spatial distribution in colorectal cancer liver metastasis: a pilot study. Front Immunol 2023;14:1223864. [Crossref] [PubMed]
15. Shen R, Huang Y, Kong D, et al. Spatial distribution pattern of immune cells is associated with patient prognosis in colorectal cancer. J Transl Med 2024;22:606. [Crossref] [PubMed]
16. Cancer Genome Atlas Research Network. Comprehensive molecular characterization of gastric adenocarcinoma. Nature 2014;513:202-9. [Crossref] [PubMed]
17. Xiao Z, Todd L, Huang L, et al. Desmoplastic stroma restricts T cell extravasation and mediates immune exclusion and immunosuppression in solid tumors. Nat Commun 2023;14:5110. [Crossref] [PubMed]
18. Wu Q, Huang Q, Jiang Y, et al. Remodeling Chondroitin-6-Sulfate-Mediated Immune Exclusion Enhances Anti-PD-1 Response in Colorectal Cancer with Microsatellite Stability. Cancer Immunol Res 2022;10:182-99. [Crossref] [PubMed]
19. Ong CJ, Zhao JJ, Liu Y, et al. Spatial Heterogeneity, Stromal Phenotypes, and Therapeutic Vulnerabilities in Colorectal Cancer Peritoneal Metastasis. Clin Cancer Res 2025;31:2515-29. [Crossref] [PubMed]
20. Long F, Zhong W, Zhao F, et al. DAB2 (+) macrophages support FAP (+) fibroblasts in shaping tumor barrier and inducing poor clinical outcomes in liver cancer. Theranostics 2024;14:4822-43. [Crossref] [PubMed]
21. Abboretti F, Mantziari S, Didisheim L, et al. Prognostic value of tumor regression grade (TRG) after oncological gastrectomy for gastric cancer. Langenbecks Arch Surg 2024;409:199. [Crossref] [PubMed]
22. Wang JB, Gao YX, Ye YH, et al. Comprehensive multi-omics analysis of pyroptosis for optimizing neoadjuvant immunotherapy in patients with gastric cancer. Theranostics 2024;14:2915-33. [Crossref] [PubMed]
23. Hänggi K, Li J, Gangadharan A, et al. Interleukin-1α release during necrotic-like cell death generates myeloid-driven immunosuppression that restricts anti-tumor immunity. Cancer Cell 2024;42:2015-2031.e11. [Crossref] [PubMed]
24. An M, Mehta A, Min BH, et al. Early Immune Remodeling Steers Clinical Response to First-Line Chemoimmunotherapy in Advanced Gastric Cancer. Cancer Discov 2024;14:766-85. [Crossref] [PubMed]
25. Mantovani A, Allavena P, Marchesi F, et al. Macrophages as tools and targets in cancer therapy. Nat Rev Drug Discov 2022;21:799-820. [Crossref] [PubMed]
26. Srivastava S, Furlan SN, Jaeger-Ruckstuhl CA, et al. Immunogenic Chemotherapy Enhances Recruitment of CAR-T Cells to Lung Tumors and Improves Antitumor Efficacy when Combined with Checkpoint Blockade. Cancer Cell 2021;39:193-208.e10. [Crossref] [PubMed]
27. Myers KV, Pienta KJ, Amend SR. Cancer Cells and M2 Macrophages: Cooperative Invasive Ecosystem Engineers. Cancer Control 2020;27:1073274820911058. [Crossref] [PubMed]
28. Venkatesan S, Lucke-Wold B. Mind the gut: Navigating the complex landscape of gastroprotection in neurosurgical patients. World J Gastroenterol 2025;31:102959. [Crossref] [PubMed]