did not reach statistical significancep = 0.5232
In pre-specified pairwise comparisons versus the model group (Mann-Whitney U test with Holm adjustment; m = 5), AI scores were significantly lower in the Pred, TWG-H, and TWG-M groups (adjusted p = 0.0426 for each), whereas the TWG-L group did not reach statistical significance (adjusted p = 0.5232; Figure 1B ).