INTRODUCTION
Pancreatic ductal adenocarcinoma (PDAC) is regarded as one of the most formidable tumors, with a poor prognosis despite aggressive multimodal therapy[1,2]. Its relentless aggressiveness and heterogeneous clinical behavior underscore the urgent need for novel biomarkers to refine prognostic stratification and guide personalized treatment. While traditional tumor-node-metastasis staging provides a framework[3,4], it fails to capture the biological complexity of PDAC; particularly the dynamic crosstalk between cancer cells and tumor microenvironment (TME). Emerging research highlights the TME as a critical orchestrator of PDAC progression, with tumor budding (TB) implicated as a histomorphological definition of epithelial-mesenchymal transition (EMT), desmoplastic reaction (DR) mediating chemoresistance and immune evasion, and tumor-infiltrating lymphocytes (TILs) potentially modulating antitumor immunity[5,6]. Nevertheless, the prognostic significance of these features, individually or synergistically, remains incompletely defined, hindering their integration into routine clinical practice.
In the World Journal of Gastrointestinal Oncology, Alpsoy et al[7] addressed this gap with a meticulously designed retrospective cohort study. They rigorously evaluated the triad of TB, DR, and TILs in 100 surgically resected PDAC specimens, using standardized international consensus criteria [International Tumor Budding Consensus Conference (ITBCC) for TB, and Ueno classification for DR] to ensure reproducibility. The ITBCC standard is currently the mainstream method for evaluating TB[8]. It has the advantages of standardization and repeatability, as well as clear prognostic value. However, it also has limitations such as the risk of overcorrection and insufficient applicability of biopsy specimens. The Ueno classification serves as the basis for ITBCC grading, simplifying the grading process to facilitate its rapid clinical application. However, it has the drawbacks of an unstandardized field of view area and reliance on the experience of pathologists. In the future, improvements may be made through digital pathology and artificial intelligence (AI)-assisted analysis, as well as immunohistochemical staining assistance. Their findings are significant. TB [both peritumoral (PTB) and intratumoral budding (ITB)] emerges as a powerful independent predictor of survival, strongly correlating with lymph node metastasis and aggressive histopathological features. TIL density inversely associates with high-grade TB and serves as an independent favorable prognostic factor. Immature DR (DR3) is linked to high-grade TB, suggesting a stromal landscape permissive to EMT-driven dissemination. Crucially, the authors have demonstrated that ITB assessment in biopsies, a clinically feasible tool, may mirror the prognostic value of PTB in resected specimens, offering a window into tumor biology for unresectable tumors[7]. This bridges a vital translational gap, proposing actionable histopathological markers even in nonoperative settings. This study validated TB and TILs as indispensable prognosticators and challenged the field to standardize their evaluation in PDAC pathology protocols. As stroma-targeted and immunomodulatory therapies advance, Alpsoy et al[7] have provided a compelling rationale for incorporating TME-based biomarkers into the prognostic arsenal for PDAC, a crucial step toward precision oncology. Although TB, DR, and TILs are included in pathological reporting standards for solid tumors such as colorectal cancer, their application in PDAC remains contentious. Based on recent evidence, this editorial analyzes the predictive value and clinical translation pathways of these three factors for PDAC prognosis.
CORE DISCOVERIES AND CLINICAL IMPLICATIONS
TB: An independent prognostic factor requiring standardization
TB, associated with EMT and diminished survival in various cancers including PDAC[6], was confirmed by Alpsoy et al[7] as a robust adverse prognostic factor. PTB and ITB were negative predictors for poor survival (significant hazard ratios, P < 0.001), significantly associated with lymph node metastasis and lymphovascular invasion. Patients with high-grade budding (TB3) exhibited a median survival 62% shorter than those with low-grade budding (TB1) (15.2 months vs 39.8 months, P < 0.001). The ITBCC scoring system proved applicable to PDAC. Evaluation of TB, representing EMT, may serve as a histological feature for selecting tumors for advanced molecular analyses to identify poor-prognosis and drug-resistant subtypes[8]. While multicenter studies indicate variable reproducibility in TB assessment[9], suggesting potential for machine learning applications to improve interobserver agreement, TB has also been validated to be closely associated with prognosis in PDAC patients receiving neoadjuvant therapy[10,11], consistent with findings in resectable patients although requiring further elucidation in advanced disease.
PTB showed a high correlation with ITB (r = 0.89), suggesting that ITB assessment in biopsy specimens could potentially substitute for PTB evaluation in surgical specimens. This offers a novel approach for risk stratification in patients ineligible for surgery and challenges the traditional view equating TB solely with distant metastasis. The authors invoke the partial EMT theory - in which cells coexpress both epithelial and mesenchymal characteristics, prevalent in cancer and contributing to malignant progression[12] - to explain this phenomenon, emphasizing the need for immunohistochemical verification of the EMT phenotype. Partial EMTs are intermediate phenotypes in the EMT continuum (epithelial state to partial EMT to mesenchymal state), possessing both epithelial and mesenchymal characteristics. Their core features include: Phenotypic plasticity, dynamic reversibility, and close correlation with the invasiveness and drug resistance of PDAC. Partial EMT is the core hub of PDAC invasion and metastasis, and its interaction network with the TME provides new opportunities for precise treatment. By clarifying the theoretical framework of partial EMT and its functions in PDAC-TME, the clinical translational value of the research can be significantly enhanced, and a new perspective for targeting tumor plasticity can be provided. Current assessment relies on hematoxylin and eosin staining, while detailed immunohistochemical assays have confirmed a relationship between EMT and TB in the tumor-host interface. Combined assessment of lymph node ratio and EMT within tumor buds may identify a subset of patients with a more favorable survival[13]. TB is an established independent negative prognostic factor in PDAC; however, standardized assessment protocols are essential. Incorporating TB assessment, particularly ITB, into routine PDAC pathological reports could aid decisions regarding postoperative adjuvant therapy. The strong correlation between ITB and PTB supports evaluating ITB in biopsy specimens to guide systemic treatment for unresectable patients.
DR: Stromal complexity and therapeutic implications
The DR, a hallmark of PDAC progression, involves significant alterations in stromal-derived extracellular matrix that critically drive cancer progression[14,15]. Alpsoy et al[7] revealed an association between TB and DR: High PTB/ITB significantly correlated with immature stroma (DR3) (P < 0.05), reinforcing the role of TB as an EMT morphological marker and deepening understanding of the invasive biology of PDAC. Immature DR creates a physical barrier restricting drug penetration. Targeting pancreatic stellate cell activation or collagen crosslinking represents a potential strategy to improve chemotherapy response[16-18]. Although the DR classification itself did not show independent prognostic significance in this study, optimizing stratification may require combining DR assessment with molecular markers (e.g., transforming growth factor-β and lysyl oxidase) or integrating it into a combined TME biomarker model. As TME in PDAC is full of extracellular matrix, which is mainly derived from cancer-associated fibroblasts, Zhang et al[19] have reported that single-cell RNA sequencing (scRNA-seq) facilitates understanding of the diverse cancer-associated fibroblast subtypes. In the future, the comprehensive analysis of CA heterogeneity form different visions will generate innovative solutions for cancer diagnosis and treatment.
TILs: Prognostic potential and assessment challenges
Accumulating evidence supports the predictive significance of TILs in multiple tumors[20]. In this study by Alpsoy et al[7], a semiquantitative stromal TIL score (sTILs, medium/high density) emerged as an independent protective factor (hazard ratio = 0.42) and was inversely correlated with low TB grade (P < 0.001). The study innovatively compared TIL scoring methods: Semiquantitative scoring (sTILs) was an independent prognostic factor, while percentage quantification lacked significance, highlighting the critical need for method standardization. TILs exhibit high heterogeneity, encompassing diverse functional subsets. Reports suggest prognostic implications for CD3+, CD4+, and CD8+ TILs[21], but not necessarily for FOXP3+ cells or composite immune cell scores. Intratumoral infiltration by OX40+ lymphocytes was regarded as a novel favorable prognostic biomarker for resected PDAC[22]. Tumor-associated tertiary lymphoid structures, as functional immune niches, whose role in PDAC is not fully understood[23], are associated with survival benefits. However, DR may impede T-cell infiltration, necessitating spatial transcriptomics to dissect immune-stromal interactions[24]. Referencing guidelines like those from the International Immuno-Oncology Biomarkers Working Group for TILs assessment in breast cancer[25] is advisable, but PDAC-specific standards must be developed. Multicolor immunohistochemistry is essential to characterize functional TIL subpopulations (e.g., CD8+ T cells and regulatory T cells) and capture the complexity of the immune microenvironment[21,25].
TME component interplay: Synergy and therapeutic opportunities
The interaction between TME components plays a crucial role in PDAC[14]. Beyond the TB-DR association, Alpsoy et al[7] observed a negative correlation between TB and TILs: TIL density was significantly reduced in high-TB groups (P < 0.001). This suggests that an immunosuppressive microenvironment may synergistically promote tumor progression[18], providing a rationale for combined therapeutic strategies targeting the TME. Future research should explore multiparameter models combining TB, TILs, and DR with molecular profiling (e.g., CDKN2A mutation status, TLS presence) to achieve more precise prognostic prediction. It is necessary to expand the diverse methodologies for TME assessment: Computational deconvolution techniques (CIBERSORTx, using DNA methylation data through analysis of Infinium arrays); single-cell and spatial omics integration (scRNA-seq combined with T cell receptor-sequencing, spatial transcriptomics); and multiomics prognostic models (immune-CSC-TB spectrum, PRECOG database). Further development of biopsy-applicable strategies can be pursued: Such as the standardization of ITB based on biopsy samples combined with liquid biopsy. Construct dynamic monitoring models to track the new antigen-T cell receptor interactions through longitudinal samples and analyze the impact of immune editing on budding. In summary, the ME assessment needs to integrate the three dimensions of standardized morphology (ITBCC/Ueno), spatial multi-omics (CODEX/scRNA-seq), and computational biology (deconvolution/AI).
LIMITATIONS AND FUTURE DIRECTIONS
While impactful, this study had limitations inherent to its design. As a single-center retrospective cohort analysis of only 100 surgical patients (excluding 176 receiving neoadjuvant therapy), it risked selection bias, potentially limiting generalization to the broader PDAC population, particularly advanced or unresectable cases. Retrospective analysis cannot fully control confounding variables (e.g., variations in postoperative treatments), which may affect survival analysis accuracy. Although meeting basic statistical requirements, the sample size may be insufficient to capture the full spectrum of biological heterogeneity in highly variable PDAC (e.g., molecular subtype effects), potentially limiting statistical power in multivariate analyses. Dichotomizing continuous variables (e.g., age and tumor size) using mean/median cutoffs may result in information loss. Furthermore, univariate analysis of multiple variables without multiplicity correction (e.g., Bonferroni) increases the risk of false positives.
Assessment of TB, DR, and TILs relies on subjective pathological interpretation. Despite adherence to international consensus (ITBCC), the lack of PDAC-specific standards may compromise result reproducibility. The absence of immunohistochemical validation (e.g., E-cadherin for EMT) weakens the link between TB and EMT. The study assessed overall lymphocyte density using sTILs and percentage quantification) methods but did not distinguish functional TIL subpopulations (e.g., CD8+ T cells, and regulatory T cells). This oversight, requiring multicolor immunohistochemistry[22,26], failed to capture immune microenvironment complexity, potentially explaining why TIL-DR interactions lacked significance (possibly related to immunosuppressive stroma).
The focus on morphological indicators (TB, DR, and TILs) without integrating key PDAC molecular characteristics (e.g., KRAS mutations and SMAD4 deletion)[2,3] hinders elucidation of TME-tumor intrinsic driver interactions. DR classification was limited to hematoxylin and eosin-based maturity grading (mature/intermediate/immature). Collagen density (e.g., via Masson staining) or fibroblast activity (e.g., α-smooth muscle actin staining) were not evaluated, potentially missing critical information on stromal heterogeneity.
Clinical translation faces hurdles. While proposing ITB assessment in biopsies as a surrogate for PTB in resection specimens, the study did not empirically validate consistency between biopsy and resection samples. Tumor heterogeneity and sampling limitations in small biopsies may affect ITB assessment accuracy. Although TB and DR are associated with chemotherapy resistance (e.g., gemcitabine), the study did not analyze links between these markers and adjuvant therapy response, limiting their utility for guiding individualized treatment. The association between TB and DR3 suggests EMT-stroma crosstalk, but lacks mechanistic data (e.g., transforming growth factor-β pathway activity). The negative TB-TILs correlation warrants exploration of causality: Does TB drive immune exclusion, or does immune deficiency promote TB formation?
CONCLUSION
Alpsoy et al’s study represents a significant advancement as the first to integrate the three major TME markers - TB, DR, and TILs - in PDAC[7]. It confirms TB as an independent prognostic factor and reveals interactions among TME components, providing a practical framework for pathological assessment. However, the retrospective design, limitations in pathological assessment methods, and absence of molecular mechanistic insights may affect the generalization of the conclusions. Future research must validate these findings through prospective multicenter studies, implement standardized pathological procedures (leveraging digital pathology and AI assistance), and integrate multiomics techniques (spatial transcriptomics and single-cell sequencing) to unravel the TME network. This integrated approach, spanning morphology to molecular mechanisms, is essential for achieving precise PDAC stratification and ultimately improving patient survival.
Provenance and peer review: Invited article; Externally peer reviewed.
Peer-review model: Single blind
Specialty type: Oncology
Country of origin: China
Peer-review report’s classification
Scientific Quality: Grade A
Novelty: Grade A
Creativity or Innovation: Grade A
Scientific Significance: Grade A
P-Reviewer: Wei R, PhD, Assistant Professor, Postdoctoral Fellow, China S-Editor: Bai SR L-Editor: A P-Editor: Zheng XM
