Background: The diverse anticancer measures display varied efficacy in different
patients. Thus, appropriate therapy should be chosen for individual patients, and prognostic
prediction, based on biomarkers, is a prerequisite for personalized therapy.
Objective: In this study, the prognostic model was established based on the genes that were
significantly correlated with the survival time for patient death risk evaluation.
Method: Univariate Cox proportional hazards regression analysis was utilized for screening the
genes significantly correlated with the patients’ survival time. Multivariate Cox proportional hazards
regression analysis was utilized for establishing the model. Kaplan-Meier and ROC analyses were
used for the validation of the prognostic prediction potential of the constructed model.
Results: ROC analysis was conducted in the training and validation datasets, and their AUROC
values were 0.774 and 0.723, respectively. In comparison to the known prognostic biomarkers, our
prognostic biomarker model constituted by the combination of 6 genes displayed superiority in
Conclusions: These results indicated that our biomarker model could effectively stratify the risks in
gastric adenocarcinoma patients with high prognostic prediction accuracy and sensitivity.