Background: Colorectal cancer (CRC) is a kind of tumor with high incidence and its
treatment situation is still very difficult despite the constant renewal and development of treatment
Objective: To assist the prognosis, monitoring and survival of CRC patients with a model.
Methods: In this study, we established a new prognostic model for CRC. Four groups of CRC data
were accessed from the GEO database, and then differential analysis (logFoldChange>1, adjust-
P<0.05) was carried out by using the limma package along with the RobustRankAggreg package
used to identify the overlapping differentially expressed genes (DEGs). Univariate and multivariate
Cox regression analyses were performed on the DEGs to screen the genes related to the patient’s
prognosis, and a five-gene prognostic prediction model (including RPX, CXCL13, MMP10,
FABP4 and CLDN23) was constructed. Then, we further plotted ROC curves to evaluate the predictive
performance of the five-gene prognostic signature in the TCGA data sets (the AUC values
of 1, 3, 5-year survival were 0.68, 0.632, 0.675, respectively) and an external independent data set
GSE2962 (the AUC values of 1, 3, 5-year survival were 0.689, 0.702, 0.631, respectively).
Results: The results showed that the model could effectively predict the prognosis of CRC patients,
which provides a robust predictive model for the prognosis of CRC patients.
Conclusion: The model could effectively predict the prognosis of CRC patients, which provides a
robust predictive model for the prognosis of CRC patients.