Fig. 4: ML models for discriminating liver fibrosis regressors following short-term AVT.

a AUC values (95% CI) of ML models developed using five feature selection methods and four ML algorithms. These models were constructed from the training set and validated in the testing set derived from short-term patients in the 4D-DIA-MS discovery cohort. The AUC value (95% CI) of the optimal short-term panel is highlighted in dark red. Training set: n = 34 for R; n = 23 for NR. Testing set: n = 22 for R; n = 15 for NR. b Receiver operating characteristic (ROC) curve analysis of the selected short-term panel and ML model for monitoring fibrosis regression status in short-term patients from the PRM-MS validation cohort (IDH1 was not detected, which was imputed using the mean abundance of IDH1 in the training set). Data were presented as AUC values with 95% CI. PRM-MS validation cohort: n = 32 for R; n = 22 for NR. c AUC values (95% CI) of the selected optimal short-term panel and ML model across clinically relevant populations, including patients with baseline Ishak scores ≥ 3, non-steatosis patients at baseline, and those with ΔIshak ≥ 1 post-short-term AVT. These values are presented for both the 4D-DIA-MS discovery cohort (including the training set, testing set, and the total of both sets) and the PRM-MS validation cohort (IDH1 was not detected, which was imputed using the mean abundance of IDH1 in the training set). Detailed subgroup sample sizes are provided in Supplementary Table 1. Abbreviations: 4D-DIA-MS four-dimensional data-independent acquisition mass spectrometry, AUC area under the curve, PLS partial least squares, PRM-MS parallel reaction monitoring mass spectrometry, SVM support vector machine.