Skip to main content
. 2021 Nov 17;11:769272. doi: 10.3389/fonc.2021.769272

Figure 1.

Figure 1

(A) The performance of radiomic features for the evaluation of tumor microenvironment immune type (TMIT)-I tumors. All the features showed moderate power for predicting the TMIT-I tumors, and PET features have better ability than CT features with higher area under the receiver operating characteristic (ROC) curve (AUC). Panel (B) was the ROC curve of the optimal PET (GLZLM_SZE) and CT (Conventional_HUmax) features to distinguish TMIT-I from other groups.