Milea Timbergen

137 Table 5. Performance of the random forest multilabel radiomics models for the DTF CTNNB1 mutation stratification based on; model 4: T1w imaging features, age and sex; model 5: T1w + T2w imaging features; and model 6: T1w + T1w post-contrast imaging features. Model 4 was evaluated for a single class (S45F, T41A, and WT) or the overall performance (All). Outcomes are presented with the 95% confidence interval. Model 4 - S45F T1w+ age +sex Model 4 - T41A T1w + age + sex Model 4 - WT T1w + age + sex Model 4 - All T1w + age + sex Model 5 - All T1w + T2w Model 5 - All T1w + T1w post-contrast AUC 0.61 [0.44, 0.77] 0.56 [0.43, 0.68] 0.74 [0.60, 0.87] 0.63 [0.54, 0.72] 0.63 [0.53, 0.72] 0.60 [0.50, 0.69] BCA 0.48 [0.35, 0.61] 0.53 [0.42, 0.64] 0.65 [0.54, 0.75] 0.56 [0.47, 0.64] 0.57 [0.48, 0.66] 0.53 [0.44, 0.61] Sensitivity 0.15 [<0, 0.37] 0.49 [0.27, 0.71] 0.56 [0.35, 0.77] NA NA NA Specificity 0.83 [0.67, 0.98] 0.59 [0.41, 0.76] 0.72 [0.55, 0.89] NA NA NA NPV 0.76 [0.70, 0.82] 0.65 [0.53, 0.77] 0.73 [0.64, 0.82] NA NA NA PPV 0.17 [<0, 0.45] 0.42 [0.28, 0.56] 0.59 [0.40, 0.77] NA NA NA T1w, T1-weighted MRI; T2w,: T2-weighted MRI; AUC, area under the receiver operator characteristic curve; BCA, balanced classification accuracy; PPV, positive predictive value; NPV, negative predictive value; WT, wild-type, NA, not applicable 5