Discovery of a prognostic two-lncRNA panel correlated with immune-associated transcripts for forecasting overall survival in pancreatic cancer.
TL;DR
The presented findings offer novel insights into the correlation of immune-related lncRNAs with pancreatic cancer progression and provide a foundation for future risk stratification modeling.
Abstract
Background
Pancreatic cancer is among the most lethal cancers because of late diagnosis and inadequate therapeutic preferences. The tumor microenvironment and immune response have critical roles in disease progression and patient outcomes. LncRNAs are regulators of immune function and tumorigenesis, yet their prognostic value in pancreatic cancer remains partly clarified.
Methods
RNA-sequencing data and corresponding clinical features for 178 pancreatic cancer patients were obtained from TCGA. Immune-related mRNAs were retrieved from the MSigDB, and immune-related lncRNAs were identified based on a significant Pearson correlation with at least two immune-related mRNAs. Regression analyses were sequentially applied to construct a prognostic immune-related lncRNA set. The predictive performance of the set was assessed using time-dependent ROC curves and Kaplan-Meier survival analysis in both the TCGA training cohort and an independent external validation cohort (GSE224564).
Results
From 4059 lncRNAs, 1172 immune-related lncRNAs demonstrated strong associations with immune-related mRNAs. Sequential screening identified 34 survival-associated lncRNAs by univariate Cox regression (P < 0.01), 12 lncRNAs following LASSO regression, and ultimately two lncRNAs-CASC8 (HR = 1.7, P < 0.05) and LINC01004 (HR = 0.54, P < 0.05)-by multivariate Cox regression. A risk score signature was developed using these two lncRNAs. Time-dependent ROC analysis demonstrated moderate-to-good predictive precision in the training cohort, with area under the curve (AUC) values of 0.702, 0.774, and 0.84 for 1-, 3-, and 5-year overall survival, respectively. Patients in the high-risk group exhibited significantly worse survival compared to the low-risk group (P < 0.05). External validation in the GSE224564 cohort confirmed consistent risk stratification (P = 0.019), though with lower discriminative accuracy (AUC ∼ 0.60). Multivariate analysis demonstrated that the risk score acted as an independent prognostic factor in both cohorts.
Conclusion
The CASC8 and LINC01004 immune-related lncRNA profile serves as a potential prognostic marker for pancreatic cancer. This set demonstrates increasing predictive trends over time in the training dataset. The presented findings offer novel insights into the correlation of immune-related lncRNAs with pancreatic cancer progression and provide a foundation for future risk stratification modeling.