Skip to main content
. 2017 Dec 5;8(68):112992–113001. doi: 10.18632/oncotarget.22947

Figure 2. The process of finding important and significant (designated “Relevant”) habitats.

Figure 2

After adjusting for clinical covariates (age, Karnofsky performance score, tumor volume, and IDH1 mutation status), we identified important habitats (positive variable importance) via Random Forest survival analysis. Those habitats were then assessed for significance via Cox Proportional Hazards Regression to determine overall survival (OS). Only habitat 2,7, and 10 are both important and significant (i.e “relevant”) in determining OS.