Introduction
Pancreatic ductal adenocarcinoma (PDAC) is a dismal disease with an average five-year survival rate of approximately 12%1. The tumor microenvironment (TME) accounts for approximately 40% to 80% of the total tumor mass in PDAC and is composed of cancer-associated fibroblasts and immune cells. PDAC is typically considered a non-immunogenic tumor, showing limited responsiveness to immunotherapy2. Nevertheless, the TME plays a crucial role in shaping disease progression and determining patient outcomes.
Aims & Methods
The aim of this study is to characterize the immunoprofiles of PDAC and analyze their association with tumor stage and patient prognosis.
The study cohort included 443 transcriptomes spanning multiple disease stages (Stage I=35, Stage II/III=244, and Stage IV=164). The contribution of the TME to overall survival (OS) prediction was assessed using four PDAC classifiers: two mono-compartment classifiers based solely on tumoral cell (Bailey3 and Chan-Seng-Yue4), and two multi-compartment classifiers (Puleo5 and MR-Gradient6) that incorporate TME components. The performance of each classifier's Cox regression model was evaluated using the Akaike Information Criterion (AIC). Hierarchical Clustering on Principal Components (HCPC) was applied to inferred immune cell populations (xCell and CIBERSORT) to identify immune clusters (IC). The gene profile of each IC was determined through differential expression analysis. IC assignment in the validation cohorts (TCGA-PAAD and Puleo) was performed using Gene Set Variation Analysis (GSVA) scores.
Results
The multi-compartment classifiers demonstrated superior performance, with the MR-Gradient showing the lowest AIC with hazard ratio (HR) of 0.68 (95% CI, 0.61–0.76; P<0.001), indicating the best model fit. Unsupervised clustering of HCPC revealed four distinct IC. Approximately 18.3% of patients were classified as IC1, characterized by enrichment of B cell and activated dendritic cells populations was associated with the better outcome (HR=0.57; 95% CI, 0.41–0.78; P<0.001), followed by IC2, predominated by CD8⁺ T cells (HR=0.70; 95% CI, 0.53–0.93; P=0.014). In contrast, IC3 and IC4 were defined by an immunosuppressive tumor microenvironment. The frequency of IC1 was highest in Stage I (45.7%) tumors and progressively decreased through Stage IV (17.1%), suggesting a stage-dependent shift in immune composition. These findings were validated in the TCGA-PAAD and Puleo cohorts, where IC1 consistently showed superior outcomes, with HR of 0.37 (95% CI, 0.17–0.78; P=0.009) and 0.43 (95% CI, 0.28–0.66; P<0.001), respectively.
Conclusion
These results validate that incorporating the immunological compartment alongside tumoral cell enhances the prediction of patient outcomes. Moreover, we identified four immunological clusters that span all tumor stages, each defined by distinct inferred immune cell compositions. Notably, patients whose TME was enriched in B cells and activated dendritic cells exhibited the most favorable survival outcomes. Overall, this study provides a framework for stratifying PDAC patients based on the immunological composition of their TME, potentially guiding more personalized therapeutic approaches.
References
1. Siegel, R. L., Giaquinto, A. N. & Jemal, A. Cancer statistics, 2024. CA Cancer J Clin 74, 12–49 (2024).
2. Brouwer, T. P. et al. Local and systemic immune profiles of human pancreatic ductal adenocarcinoma revealed by single-cell mass cytometry. J Immunother Cancer 10, e004638 (2022).
3. Bailey, P. et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature 531, 47–52 (2016).
4. Chan-Seng-Yue, M. et al. Transcription phenotypes of pancreatic cancer are driven by genomic events during tumor evolution. Nat Genet 52, 231–240 (2020).
5. Puleo, F. et al. Stratification of Pancreatic Ductal Adenocarcinomas Based on Tumor and Microenvironment Features. Gastroenterology 155, 1999–2013 (2018).
6. Fraunhoffer, N. A. et al. Multi-omics data integration and modeling unravels new mechanisms for pancreatic cancer and improves prognostic prediction. NPJ Precis Oncol 6, (2022).
