Figure 13
From: Tissue Phenomics for prognostic biomarker discovery in low- and intermediate-risk prostate cancer
Performance evaluation of multi-variate stratification using different types of classifiers with the subset of 8 significant phenes selected using the FWER method of Holm-Bonferroni followed by CART-feature selection (Table 2). The Kaplan-Meier plots show groups “non-progression” (green) versus “progression” (blue) as predicted, with the events corresponding to true tumor progression. X-axis: disease-free survival time (DFS) in month, y-axis: disease-free survival probability. In the confusion matrices rows provide the true class and columns the predicted class (0: non-progression, 1: progression). (Note that for some patients with DFS times of up to 36 month the exact disease recurrence dates were not available and only the approximated DFS time of 36 month could be used, leading to the event aggregation at 36 month in the plot.
