Milea Timbergen

151 Supplemental Table 1. (continued) LoG (12*3=36 features) min max mean median std skewness kurtosis peak range energy quartile entropy Vessel (12*3=36 features) min max mean median std skewness kurtosis peak range energy quartile entropy LBP (6*3=18 features): mean std median kurtosis skewness peak Local phase (12*3=36 features) min max mean median std skewness kurtosis peak range energy quartile entropy Orientation (3 features): theta_x theta_y theta_z GLCM, gray level co-occurrence matrix; GLCMMS, GLCM multislice; NGTDM, neighbourhood gray tone difference matrix; GLSZM, gray level size zone matrix; GLRLM, gray level run length matrix; LBP, local binary patterns; LoG, Laplacian of Gaussian; std, standard deviation. Supplemental Table 2. Performance of the radiomics models for the DTF differential diagnosis based on T1w and T2w non-FatSat imaging features; T1w and T2w FatSat imaging features; T1w and T1w post- contrast non-FatSat imaging features; and T1w and T1w post-contrast FatSat imaging features. Outcomes are presented with the 95% confidence interval. T1w + T2 non-FatSat T1w + T2w FatSat T1w + T1w post- contrast non-FatSat T1w + T1w post- contrast FatSat AUC 0.83 [0.76, 0.89] 0.83 [0.77, 0.89] 0.80 [0.74, 0.85] 0.82 [0.75, 0.88] BCA 0.64 [0.58, 0.71] 0.66 [0.59, 0.72] 0.73 [0.67, 0.79] 0.72 [0.66, 0.79] Sensitivity 0.32 [0.19, 0.44] 0.34 [0.20, 0.47] 0.60 [0.49, 0.72] 0.59 [0.48, 0.70] Specificity 0.97 [0.92, >1] 0.97 [0.94, 1.00] 0.85 [0.79, 0.92] 0.86 [0.79, 0.94] NPV 0.74 [0.70, 0.77] 0.74 [0.70, 0.78] 0.79 [0.74, 0.84] 0.79 [0.74, 0.83] PPV 0.87 [0.68, >1] 0.88 [0.71, >1] 0.71 [0.60, 0.82] 0.72 [0.61, 0.84] T1w, T1-weighted images, T2w, T2-weighted images; AUC, area under the receiver operator characteristic curve; BCA, balanced classification accuracy; PPV, positive predictive value; NPV, negative predictive value 5