Skip to main content
. 2020 Mar 19;18:668–675. doi: 10.1016/j.csbj.2020.03.007

Table 3.

The classification performance of PreMSIm (%).

Cancer type g Accuracy Sensitivity Specificity AUC
Esophageal cancer 96 100 95 99
Rectum cancer 91 67 92 79
Uterine cancer 95 100 94 99
Gastric cancer (GSE13911) 90 89 90 89
Gastric cancer (GSE62254) 88 78 91 87
Colorectal cancer (GSE13067) 98 100 95 99
Colorectal cancer (GSE13294) 92 86 99 96
Colorectal cancer (GSE18088) 96 95 97 97
Colorectal cancer (GSE26682-GPL570) 98 83 99 93
Colorectal cancer (GSE26682-GPL96) 70 82 68 81
Colorectal cancer (GSE35896) 92 100 91 98
Colorectal cancer (GSE39084) 93 100 91 99
Colorectal cancer (GSE39582) 90 90 90 94
Colorectal cancer (GSE41258) 77 60 82 82
Colorectal cancer (GSE75316) 95 100 94 99
Colorectal cancer (GSE92921) 95 80 96 89
Colorectal cancer (GSE27544) 91 75 100 94
Colorectal cancer (GSE24550) 88 100 84 94
Colorectal cancer (GSE25071) 87 100 85 98

Note:

* These prediction results were obtained by the PreMSIm R package.

g

The Esophageal, Rectum, and Uterine cancer datasets were from TCGA and the others were from GEO.