Skip to main content
. 2023 Mar 31;9(4):e15123. doi: 10.1016/j.heliyon.2023.e15123

Fig. 4.

Fig. 4

Multi-ROC analysis of the combinations of 20 lipids in ME1 to discriminate patients with HEs. (A) ROC of the discovery group. The combined model of 20 lipids was the best to discriminate HEs. AUC = 0.804, 95% CI = 0.674–0.916. (B) ROC of the validation group. In the validation group, AUC for this combined model of 20 lipids was 0.737, 95% CI = 0.474–0.959. Abbreviations: AUC, area under curve; CI, confidence interval; HEs, hard exudates; ROC, receiver operating characteristic.