Proteomic and Metabolic Signaling Pathways in Different Prognosis of Gastric Cancer.
Guodong Lian,Daguang Wang,Chang Shu,Chengguo Wei,Fei Ye,David Y. Zhang,Jian Suo
DOI: https://doi.org/10.1200/jco.2014.32.15_suppl.e15087
IF: 45.3
2014-01-01
Journal of Clinical Oncology
Abstract:e15087 Background: Gastric cancer (GC) remains the fourth most common cancer and the second most common cause of cancer-related death worldwide. To tailor treatment regimens individually on the basis of patients’ chemosensitivity is a more effective strategy. Methods: After assessing the chemosensitivity status through survival time after surgery and adjuvant chemotherapy, 140 stage III (AJC) GC patients were split into two groups: chemosensitive (67 cases) and chemoresistant (73 cases). The differentially expressed proteins between those two groups were assessed using Protein Pathway Array (PPA). Data were analyzed by Significance Analysis of Microarrays (SAM) and BRB Array Tools software v.3.3.0. Ingenuity Pathway Analysis (IPA) was explored to further analysis the relevant signaling pathways. Results: Among 292 protein-specific or phosphorylation site-specific proteins tested, 147 were detected in cancer tissues. SAM analysis showed that 23 of these proteins/phosphorproteins had significant difference (q<5%) between chemosensitive and chemoresistant groups. We further carried out supervised K-fold cross validation (K=10) analysis using support vector machine (SVM), and 18 proteins (PLk1, FKHR, HDAC1, calpain 2, WT1, cdk4, β3 Tubulin, HMG-1, NMT1, Bcl-xL, V-ATPase H, tsg 101, calpastatin, P-cadherin, ADH, p-JNK, DACH1 and E-cadherin) were selected to separate gastric cancer of chemosensitive group from chemoresistant group with 94% correct rate, 94.8% sensitivity and 93% specificity. Six proteins (ERCC1, FKHR, PLK1, DACH1, E-Cadherin and ADH) were identified as independent predictors of chemoresistance. Using these 6 proteins, a risk score model was established to determine the risk of chemoresistance for each individual gastric cancer patient. Conclusions: Our current study indicated that some proteins differentially expressed in gastric cancer can be selected as clinically useful biomarkers. The risk score model is useful for determining patients’ risk of chemoresistance after radical surgery.