Fig. 5

Construction and validation of a MRGs-based nomogram model. A-B. The selection of prognostic SCMRDEGs based on the optimal parameter 𝜆 that was obtained in LASSO analysis. C-D. K-M survival curves of high-MRGs and low-MRGs groups in training and validation cohorts. E-F. ROC curves of MRGs in training and validation cohorts. G. Multivariate cox regression analysis in the training cohort. H. A nomogram model was constructed with MRGs, age, and AJCC staging. I-K. Calibration curves were used to validate the predictive accuracy of our nomogram for predicting 1-, 3-, and 5-year OS. L. IHC assessed the protein expression level of GHRL, COL4A1, CAV1, GJA1, TIMP1, and IGFBP3.