Table 7 of Versura, Mol Vis 2013; 19:1247-1257.

Table 7. Diagnostic performance of clinical tests and proteins.

Test variable Cut-off points Specificity Sensitivity LR+ PPV ROC curve analysis parameters
Area under the curve Standard error 95% confidence interval
Schirmer test I ≤10 mm / 5′ 84.4 37.7 2.43 89.6 0.594 0.049 0.523 – 0.662
TFBUT ≤10 s 76.6 75.5 3.13 91.5 0.87 0.035 0.816 – 0.914
Cornea NEI score >0 93.3 30.6 5,04 21,0 0,664 0,054 0,566 – 0,754
Conjunctiva NEI Score >0 66,7 68.7 10.3 97.3 0.824 0.03 0.765 – 0.874
ZAG-2 ≤0.4 mg / ml 87.5 56.9 2.03 81.8 0.759 0.041 0.691 – 0.819
ALB >10%versus total TP 90.9 42.5 4.7 97.1 0.729 0.05 0.657 – 0.793
LACTO ≤1.2 mg / ml 93.2 44.2 6.4 94.5 0.812 0.036 0.753 – 0.861
LIPOC-1 ≤1.1 mg / ml 94.9 65.5 12.3 97.1 0.843 0.034 0.787 – 0.888
TP ≤6.5 mg / ml 96.6 55 16.7 97.8 0.861 0.032 0.808 – 0.904
LYS-C ≤2.0 mg / ml 96.6 57.8 17.7 97.9 0.826 0.032 0.813 – 0.908
TRANSF ≤0.3 mg / ml 96.7 70.8 18 98.1 0.885 0,030 0.798– 0.903

Specificity, sensitivity, positive likelihood ratio (LR+) Positive Predictive Value (PPV+) and parameters of ROC curve analysis for each test or protein were calculated comparing normal control subjects versus DE patients. Data are listed on increasing value of LR+. Explanation in the text, sections Methods and Discussion

Versura, Mol Vis 2013; 19:1247-1257. http://www.molvis.org/molvis/v19/1247

©2013 Molecular Vision http://www.molvis.org/molvis/

ISSN 1090-0535