Nomograms combined with SERPINE1-related module genes predict overall and recurrence-free survival after curative resection of gastric cancer: a study based on TCGA and GEO data
Xing-Chuan Li,Song Wang,Jia-Rui Zhu,Yu-Ping Wang,Yong-Ning Zhou
DOI: https://doi.org/10.21037/tcr-20-818
2020-01-01
Translational Cancer Research
Abstract:Background: Serpin peptidase inhibitor, clade E, member 1 (SERPINE1) has been investigated as an oncogene and potential biomarker in several cancers, including gastric cancer (GC). This study aimed to investigate SERPINE1 expression and its diagnostic and prognostic value by analyzing data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. Methods: A meta-analysis was performed to investigate SERPINE1 expression levels in GC tissues and adjacent normal tissues. Gene set enrichment, multi experiment matrix (MEM), and protein-protein interaction (PPI) network analyses were performed to identify the most enriched signaling pathways and SERPINE1-related module genes. A Cox regression model was used to develop a nomogram that was able to predict the overall survival (OS) and recurrence-free survival (RFS) of individual patients. Results: Meta-analyses revealed an elevated trend in SERPINE1 expression levels in TCGA [standard mean difference (SMD)=0.95; 95% confidence interval (CI), 0.53-1.36; P<0.001]. The diagnostic meta-analysis results indicated that the area under the curve (AUC) of the summary receiver operating characteristic (SROC) was 0.80 (95% CI, 0.77-0.84). The factors identified to predict OS were age >= 60 years [hazard ratio (HR), 2.14; 95% CI, 1.45-3.16; P<0.01], R2 margins (HR, 2.70; 95% CI, 1.41-5.14; P<0.05), lymph node-positive proportion (HR, 3.38; 95% CI, 2.03-5.63; P<0.001), patient tumor status (HR, 3.33; 95% CI, 2.28-4.87; P<0.001), and OS risk score (HR, 2.72; 95% CI, 1.82-4.05; P<0.05). The following variables were associated with RFS: male sex (HR, 2.55; 95% CI, 1.46-4.45; P<0.01), R2 margins (HR, 13.08; 95% CI, 4.26-40.15; P<0.001), lymph node-positive proportion (HR, 2.55; 95% CI, 1.20-5.45; P<0.05), and RFS risk score (HR, 2.70; 95% CI, 1.82-4.06; P<0.001). The discriminative ability of the final model for OS and RFS was assessed using C statistics (0.755 for OS and 0.745 for RFS). Conclusions: SERPINE1 was upregulated in GC, showed a high diagnostic value, and was associated with poorer OS and RFS. The OS and RFS risk for an individual patient could be estimated using these nomograms, which could lead to individualized therapeutic choices.