Pancreatic cancer, specifically pancreatic ductal adenocarcinoma (PDAC), represents one of the most aggressive malignancies. Lymph node (LN) status is considered as one of the most significant risk factors for survival of PDAC patients, and it is of great importance for making reasonable therapeutic strategies to individual patients. Imaging techniques are widely used in the evaluation of LN status in PDAC patients, however, their application are limited because of the inconsistent sensitivities and specificities findings.
Gene Set Enrichment Analysis (GSEA) and leading edge analysis were used to analyze the data from The Cancer Genome Atlas (TCGA) on 177 PDAC patients to identify genes associated with LN metastasis. The identified genes with LN metastasis were indexed by Spearman’s rank-correlation test to construct the risk score model. Risk scores were used to predict LN metastasis and overall survival (OS). For validation, we used 80 specimens from patients with PDAC diagnosed at Fudan University Shanghai Cancer Center.
A risk model consisting of three genes (MAPK9, ITGA5, AKT2) was developed. This model could correctly predict the LN metastasis evaluated by receiver operating characteristic (ROC) curves [area under curve (AUC) = 0.668, P = 0.001], and risk score positively associated with the number of metastatic lymph node (NLN; Spearman r = 0.3309, P < 0.0001), especially in the PDAC with greatest dimension ≤ 4 cm and total lymph nodes dissected (TLN) > 12 [AUC = 0.80, P = 0.003; Spearman r for MLN = 0.4237, P = 0.0004; Spearman r for lymph node rate (LNR) = 0.3171, P = 0.0089]. In the set of PDAC patients with TLN > 12, patients with high risk had a worse OS than that with low risk with hazard ratio (HR) of 2.657 (P = 0.0044) for all stage and HR of 2.548 (P = 0.045) for stage I and II. Patients from stage I and stage IIA with high risk scores had a similar OS with stage IIB (median survival = 21.13 months vs. 20.23 months, P = 0.8227). In the validation set, high risk scores could also effectively predict the LN metastasis and poor prognosis in resectable PDAC.
Our findings highlight three-gene signature with effective capacity for identification of PDAC patients with poor prognosis that are likely to suffer from LN metastasis.
Clinical trial identification
Legal entity responsible for the study
Fudan University Shanghai Cancer Center.
National Science Foundation for Distinguished Young Scholars of China.
All authors have declared no conflicts of interest.