Skip to main content
. 2018 Jun 15;8:9172. doi: 10.1038/s41598-018-27538-3

Table 2.

Indices of test validity estimated for the evaluation of the predictive performance of PLS-DA models between BC vs CTRL (LVs: 3), BC vs MONITOR (LVs: 3) and MONITOR vs CTRL (LVs: 3) samples in the validation set.

LVs PLS-DA model
BC vs CTRL BC vs MONITOR CTRL vs MONITOR
3 3 3
AUROC 0.94 0.75 0.53
Sensitivity 81.8 (64.5–93.0)% 69.7 (51.3–84.4)% 45.4 (16.7–76.6)%
Specificity 90.9 (58.7–99.8)% 75.0 (64.4–83.8)% 76.0 (66.6–83.8)%
PLR 9.0 (1–4–58.7) 2.8 (1.8–4.3) 1.9 (0.9–3.9)
NLR 0.2 (0.00–0.42) 0.4 (0.2–0.7) 0.7 (0.4–1.2)

PLR: positive likelihood ratio; NLR: negative likelihood ratio.