A multimodal uncertainty-aware AI system optimizes ovarian cancer risk assessment workflow.

After correcting for multiple comparisons (Holm-Bonferroni), stepwise improvements in balanced metrics (F1 score and MCC) remained statistically robust across both internal and external cohorts, whereas the observed increases in AUC and sensitivity-focused F2 score—although nominally significant in certain cases (e.g., external cohort AUC nominal P = 0.016)—did not consistently survive multiple testing correction.

also in 7,732 other papers

Although increases in AUC (to 0.955, P = 0.606) and F2 score (to 0.866, P = 0.083) at this stage were observed, these did not reach statistical significance.

also in 111,027 other papers

Although positive trends were also noted for AUC (to 0.926, P = 0.940) and F2 score (to 0.757, P = 0.146), these changes did not achieve significance after adjusting for multiple comparisons.

also in 1,397 other papers