Prognostic Model Based on Immune Checkpoint Proteins Expression and Clinicopathological Factors to Predict Outcome of Patients with Gastric Cancer.
Ti Wen,Xiujuan Qu,Yi Li,Yuee Teng,Jingdong Zhang,Bo Jin,Mingfang Zhao,Xiaofang Che,Zhi Li,Yunpeng Liu
DOI: https://doi.org/10.1200/jco.2016.34.15_suppl.e15570
IF: 45.3
2016-01-01
Journal of Clinical Oncology
Abstract:e15570 Background: Gastric cancer is a leading cause of cancer death worldwide, and its prognostic factors remain unclear. TNM staging was used to predict the prognosis of GC, however, it cannot meet current demands. To improve prediction efficiency, a model that reflects both tumor-related factor and host's local immune response in tumor microenvironment (TME) is needed. Immune checkpoint proteins regulate T-cell function to promote cancer immune escape. PDL1/PD-1 pathway, one of the immune checkpoint pathways, has become a new target in cancer treatment. Here, we aimed to illuminate the prognostic roles of immune checkpoint protein in GC. Methods: Immune checkpoint proteins expressed in tumor microenvironment (TME) were detected by IHC staining. Survival analysis was performed by the Kaplan-Meier method. A prognostic model was constructed from 174 gastric cancer patients using a scoring system, which was based on hazard ratios in a Cox proportional hazard model. Results: We examined PD-L1, PD-1, CD8 protein expression in TME from 174 GC patients (pts), and explored their effects on pt’s overall survival (OS). Our results showed that TC PD-L1+ was associated with better clinical outcomes, including longer OS et al (P < 0.05). However, PD-L1+ or PD-1+ IC was correlated with poorer OS (P = 0.0244, P = 0.024). Furthermore, multivariate analysis identified PD-L1+ TC, PD-L1+ IC, PD-1+ IC, CD8high IC, lauren classify, lymphatic invasion and AJCC stage were independent prognostic factors. A prognostic model was established by combination of these factors. Patients were categorized into low-/median-/high-risk groups according to their scores, and the 5-year survival rates were 96%, 56.23% and 18.36%, respectively (P < 0.0001). Our seven-factors prognostic model yielded an area under the ROC curve (AUC) of 0.855 for prediction of mortality at 5 years, which was superior to TNM stage with an AUC of 0.685. Conclusions: This study showed a prognostic model based on different prognostic roles of immune checkpoint proteins, which was more accuracy than TNM staging and might provide guidance for immunotherapy options in GC.