Pancreatic adenocarcinoma (PAAD) is among the most common malignant tumors of the gastrointestinal tract and has a poor prognosis. Research on its pathogenesis and treatment remains insufficient; therefore, the present study aimed to develop a novel model for predicting PAAD prognosis. Transcriptome data for pancreatic cancer were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus databases (GSE28735), with TCGA data serving as the training set and GSE28735 data as the validation set. Differential expression analysis of GSE28735 data was performed using the limma package, followed by the identification of potential prognostic genes through univariate Cox regression and least absolute shrinkage and selection operator regression. Patients were stratified into high- and low-risk groups based on risk scores, followed by survival analysis, nomogram construction, immune infiltration analysis and mutation analysis. Reverse transcription-quantitative PCR (RT-qPCR) experiments were performed for validation. A prognostic prediction model for PAAD was constructed using TCGA patient samples and comprised 12 genes. Kaplan-Meier analysis indicated that the model effectively distinguished between high- and low-risk groups, which corresponded with poor and favorable prognoses, respectively. Receiver operating characteristic curve analysis demonstrated high predictive accuracy. Additionally, significant differences in immune infiltration and mutation levels were observed between the two risk groups. RT-qPCR results further demonstrated that the expression of prognostic genes differed significantly between pancreatic cancer (PANC-1) and normal (HPDE-6) cell lines. In conclusion, the prognostic model developed in the present study may contribute to an improved understanding of PAAD prognosis and provide new tools for predicting prognosis and immune response in patients with PAAD.
Zhao et al. (Tue,) studied this question.