Skip to main content
. 2012 Feb 8;7(2):e31129. doi: 10.1371/journal.pone.0031129

Figure 2. Predictive power of the prognostic score based on ERI in two dataset not used to derive the signature.

Figure 2

(A) the Miller breast cancer dataset [20] and (B) the Shedden lung cancer dataset [32]. Patients are divided into groups using the median score as a cutoff.