Skip to main content

Table 3 Comparison of demographics between training and validation dataset

From: Development of a classification system based on corneal biomechanical properties using artificial intelligence predicting keratoconus severity

  Training dataset Validation dataset P value
N 308 126 N.A.
Healthy 83 33
Mild KC 71 35
Moderate KC 73 35
Advanced KC 81 23
Age (years) 32.0 ± 8.0 32.3 ± 8.1 0.721
Km (D) 45.9 ± 3.6 45.9 ± 3.8 0.898
Kmax (D) 51.5 ± 6.7 51.2 ± 6.5 0.617
I-S value (D) 4.3 ± 4.0 3.8 ± 3.5 0.250
ARC (mm) 7.1 ± 0.7 7.1 ± 0.7 0.668
PRC (mm) 5.5 ± 0.7 5.6 ± 0.7 0.535
TCT (μm) 499 ± 54 497 ± 56 0.810
BAD-D 5.7 ± 4.4 5.4 ± 4.4 0.531
bIOP (mmHg) 14.5 ± 2.1 14.6 ± 2.1 0.659
  1. ARC anterior radius of curvature; BAD-D Belin/Ambrósio total deviation value; bIOP biomechanical intraocular pressure by Corvis ST; I-S inferior-superior; Km mean keratometry values; Kmax maximal keratometry value; N number of subjects; N.A. not applicable; PRC posterior radius of curvature; TCT thinnest corneal thickness. Bold type signifies P < 0.05