Skip to main content

Table 4 Summary of thymoma subtype prediction by a LOOCV method

From: Using proteomic profiling to characterize protein signatures of different thymoma subtypes

 

Thymoma Type

AB

B2

B3

No. of Samples

8

8

5

No. of Unique DEPa

43

42

37

LOOCV mean accuracy rateb

83%

83%

96%

Sensitivity

0.89

0.73

0.50

Specificity

1.00

1.00

0.86

PPVc

100%

100%

38%

NPVd

97%

91%

91%

AUCe

0.99

0.93

0.94

  1. a DEP: differentially expressed protein
  2. bThe average classification accuracy rate was calculated in leave-one-out cross validation using differentially expressed proteins
  3. c PPV: positive predictive value, the proportion of positive results in a classifier that is truly positive in the experimental results
  4. d NPV: negative predictive value, the proportion of negative results in a classifier that is truly negative in the experimental results
  5. e AUC: Area under the curve, represents how well a classifier can distinguish one type from the others
Back to article page

BMC Cancer

ISSN: 1471-2407

Contact us