Table 4 Classification, methods and AUC of previous relevant studies and this study.

From: Differences between suspected keratoconus and subclinical keratoconus via multiparameter analysis in Chinese populations

Sequence

Study

Classification

criteria and sample size

Methods

AUC

SEN

SPE

ACC

1

Zhang, X. et al.[14]

VAE-NT eyes(59) vs. Normal eyes(142)

Combined tomographic and biomechanical parameters

0.849

0.754

0.871

/

2

Song, P. et al.[15]

SKC eyes(31) vs. Normal eyes(35)

Classifcation And Regression Tree (CART)

0.95

0.903

0.943

0.924

3

Smadja, D.et al[17]

FFKC eyes(47) vs. Normal eye(95)

Automated Decision Tree

/

0.936

0.972

/

4

Ambrósio, R. et al.[41]

VAE-NT eyes(551) vs. normal eyes(1680)

TBI_v.2

0.945

0.844

0.901

/

5

Lopes, B. T. et al.[42]

VAE-NT eyes(188) vs. stable(298)

Pentacam Random Forest Index(PRFI)

0.968

0.966

0.852

/

6

Our study

SuspectedKC(116) vs. Subclincia KC(28)

Combined tomographic and biomechanical parameters

0.964

0.963

0.888

0.902

  1. SEN: sensitivity; SPE: specificity; ACC: accuracy.