Additional file 6 of Machine learning-based prediction of short-term outcomes in aneurysmal subarachnoid hemorrhage: a multicenter study integrating clinical and inflammatory indicators

Additional file 6. Fig S5: Comparison of model performance and DeLong tests. ROC curves of three models (ModA, ModB, ModC) are shown in panel A, and decision curve analysis in panel B compares net clinical benefit across probability thresholds against treat-all and treat-none strategies. ModA incorporated all nine predictors (GCS score, modified Fisher grade, WFNS grade, PNI, NAR, CLR, SII, procalcitonin, and PLR). ModB included only traditional clinical scores (GCS score, modified Fisher grade, and WFNS grade). ModC was restricted to inflammation- and nutrition-related markers (PNI, NAR, CLR, SII, procalcitonin, and PLR).