Table 4 The performance of the 11 gene pairs’ signature in independent datasets.

From: Development of machine learning-based predictors for early diagnosis of hepatocellular carcinoma

Dataset

NO.HCC

NO.CwoHCC

mRMR + KNN

mRMR + SVM

MRMD + SVM

Sn

Sp

Sn

Sp

Sn

Sp

Testing set (biopsy)

29

48

1

1

1

1

1

0.8542

Testing set (surgery)

220

36

1

1

1

1

1

0.8889

GEO (biopsy)

 GSE121248

70

–

0.9286

–

0.9429

–

1

–

 GSE47197

61

–

0.3607

–

0.9836

–

1

–

GEO (surgery)

 GSE109211

140

–

0.7214

–

0.7786

–

0.9929

–

 GSE62743

132

–

0.6288

–

0.8636

–

1

–

 GSE46444

88

–

0.3409

–

0.5227

–

1

–

 GSE10141

80

–

0

–

0.9875

–

1

–

 GSE164760

53

–

0.0755

–

0.2453

–

1

–

 GSE19977

164

–

1

–

0.9939

–

1

–

 GSE112790

183

–

0.9836

–

0.9945

–

1

–

 GSE102079

152

–

0.9737

–

0.9934

–

1

–

 GSE76427

115

–

0.7826

–

0.9478

–

1

–

 GSE78737

103

–

0.2427

–

0.3301

–

0.8544

–

 GSE9843

91

–

0.9231

–

0.9231

–

1

–

 GSE43619

88

–

0.7273

–

0.8523

–

1

–

 GSE62232

81

–

0.9506

–

0.9753

–

1

–

 GSE39791

72

–

1

–

1

–

1

–

 GSE15765

70

–

0.9571

–

1

–

1

–

 GSE87630

64

–

0

–

1

–

1

–

 GSE36411

42

–

0.8095

–

0.881

–

1

–

 GSE89377

40

–

0

–

0.2

–

0.825

–

 GSE57957

39

–

1

–

0.9744

–

1

–

 GSE14323

38

–

0.0789

–

0.1579

–

0.5263

–

 GSE6764

35

–

0.8

–

0.8286

–

1

–

 GSE101685

24

–

1

–

1

–

1

–

 GSE84598

22

–

1

–

1

–

1

–

 GSE41804

20

–

0.9

–

1

–

1

–

 GSE17548

17

–

0.7059

–

0.7059

–

1

–

 GSE84402

13

–

0

–

0.0769

–

1

–

 GSE115018

12

–

0

–

0.9167

–

1

–

 GSE98383

11

–

0.9091

–

1

–

1

–

 GSE29721

10

–

0.8

–

0.8

–

1

–

 GSE116174

64

–

0.9063

–

1

–

1

–

ICGC (surgery)

243

–

–

–

–

–

–

–

TCGA (surgery)

371

–

–

–

–

–

–

–

  1. NO.HCC, number of HCC samples, NO.CwoHCC, number of CwoHCC samples, Sn sensitivity, Sp specificity.