Skip to main content
Figure 2 | Genome Medicine

Figure 2

From: A multi-gene signature predicts outcome in patients with pancreatic ductal adenocarcinoma

Figure 2

Performance assessment of PDAC classifier. (A) Kaplan–Meier survival analysis of patient risk groups identified with the training cohort using the 36-gene signature. (B) Kaplan–Meier survival analysis of the predicted risk group of patients in the merged validation cohort using the 36-gene signature. The hazard ratio (HR) was estimated using a Cox proportional hazards model, and curves were compared using a log-rank test. (C-F) Independent validation across all the individual datasets that make up the validation cohort. (G) Assessment of classification accuracy using sensitivity and specificity estimates. Patients in the validation cohort were dichotomised by a median survival of 20 months (grey line), and classed into low- and high-risk groups, dashed black and red curves respectively. The solid curves represent a patient’s predicted risk group. Comparison between the two sets of groups indicates an overall 64.68% classification accuracy. (H) Forest plots of the genes in the prognostic signature. A Cox proportional hazards model was fitted to the signature genes in a univariate context. The horizontal axis represents HR (black squares) and 95% CIs (solid line). The asterisks represent the significance of the difference in patient outcome between the low- and high-expression groups (***P < 0.001, **P < 0.01 and *P < 0.05; Wald test). CI, confidence interval; HR, hazard ratio; PDAC, pancreatic ductal adenocarcinoma.

Back to article page