Using new indices to predict metabolism dysfunction-associated fatty liver disease (MAFLD): analysis of the national health and nutrition examination survey database

Table 6 The AUC data from train set, internal validation set and second validation set

	AUC	Sensitivity	Specificity	Cutoff	p-value¹
Train Set
AAU	0.844	0.814	0.727	0.308	ref
AGTR	0.782	0.767	0.648	0.312	< 0.001
AIP	0.759	0.702	0.686	0.305	< 0.001
UHR	0.750	0.716	0.660	0.306	< 0.001
AWB	0.902	0.837	0.803	0.286	< 0.001
BAAU	0.915	0.882	0.786	0.271	< 0.001
WAAU	0.930	0.878	0.824	0.288	< 0.001
BWAAU	0.930	0.875	0.826	0.293	< 0.001
Internal validation
AAU	0.824	0.761	0.739	0.334	ref
AGTR	0.773	0.680	0.721	0.368	< 0.001
AIP	0.749	0.687	0.707	0.317	< 0.001
UHR	0.728	0.684	0.651	0.309	< 0.001
AWB	0.894	0.841	0.772	0.268	< 0.001
BAAU	0.903	0.862	0.777	0.296	< 0.001
WAAU	0.920	0.843	0.840	0.359	< 0.001
BWAAU	0.920	0.852	0.831	0.343	< 0.001
Second validation
AAU	0.775	0.761	0.656	0.301	ref
AGTR	0.697	0.792	0.507	0.263	< 0.001
AIP	0.743	0.733	0.643	0.301	< 0.001
UHR	0.722	0.704	0.638	0.287	< 0.001
AWB	0.831	0.792	0.726	0.247	< 0.001
BAAU	0.830	0.778	0.737	0.303	< 0.001
WAAU	0.837	0.812	0.710	0.222	< 0.001
BWAAU	0.837	0.804	0.717	0.236	< 0.001

AWB AIP + Waist + BMI, AAU AGTR + AIP + UHR, BAAU BMI + AGTR + AIP + log(UHR), WAAU Waist + AGTR + AIP + log(UHR), BWAAU BMI + Waist + AGTR + AIP + log(UHR), AGTR albumin to glutamyl transpeptidase ratio, AIP plasma atherogenicity index, UHR serum uric acid to high-density lipoprotein cholesterol ratio
^*The logarithm is based on “e”
¹Delong test

ISSN: 1471-230X