Table 3. Evaluation metrics for ’Core’ dataset performance upon predicting all response using XGBoost with MICE and SMOTE.

Response	AUC	Accuracy	Sensitivity	Specificity	F1
MASL vs. MASH	0.719	0.663	0.658	0.667	0.661
At-Risk MASH	0.899	0.820	0.827	0.812	0.821
High Activity	0.801	0.723	0.720	0.734	0.724
Clinically Significant Fibrosis	0.852	0.778	0.767	0.789	0.775
Advanced Fibrosis (Histology confirmed)	0.960	0.895	0.909	0.880	0.896
Cirrhosis (Histology confirmed)	0.994	0.964	0.980	0.949	0.965
Advanced Fibrosis (Histology & Clinically confirmed)	0.961	0.901	0.915	0.888	0.903
Cirrhosis (Histology & Clinically confirmed)	0.993	0.960	0.973	0.947	0.960
At-Risk MASLD	0.921	0.846	0.856	0.835	0.847