Fig. 6: Predictive power of APC and CEG signatures for biochemical recurrence and metastasis of PCa. | Nature Communications

Fig. 6: Predictive power of APC and CEG signatures for biochemical recurrence and metastasis of PCa.

From: Spatial multi-omics identifies aggressive prostate cancer signatures highlighting pro-inflammatory chemokine activity in the tumor microenvironment

Fig. 6: Predictive power of APC and CEG signatures for biochemical recurrence and metastasis of PCa.

Shown are inverse Kaplan–Meier curves using APC (a, c, e, g) and CEG (b, d, f, h) signature activity for tertile grouping patients. Three external datasets were used TCGA (a, b, n = 485 cases), GSE116918 (c, d, n = 248 cases) and META855 (e–h, combined from the cohorts GSE72291, GSE79957, GSE62116 and GSE79915, n = 855 cases combined). Clinical endpoints are biochemical recurrence (a–f) and metastasis (g, h). Number at risk tables are given below each plot. Hazard ratios (HR) with their 95% confidence interval in parentheses are obtained from either a univariate Cox proportional hazards regression (Cox HR, a–d) or univariate (UV) and multivariate (MV) Fine-Gray subdistribution hazard model (sHR, e, f) and corresponding p values obtained from the hazard models (univariate Cox: Wald test, two-sided, sHR: Gray’s test, two-side, no multiple testing adjustment) are shown for the high vs. low tertile patients with the low tertile as reference. Significant HRs (p ≤ 0.05) are indicated by an asterisk.

Back to article page