Table 3 The features selection methods and their corresponding parameters in the models
Models | Features | SelectKBest P value | SelectKBest Score | Select Percentile P value | Select Percentile Score | Lasso coefficient |
|---|---|---|---|---|---|---|
Clinicopathologic | p63 | 0.000216545 | 15.65154265 | NA | NA | NA |
nuclear grade | 0.023386159 | 5.432974831 | NA | NA | NA | |
Conventional MRI | heterogeneous enhancement pattern | 0.0000164 | 22.2414581 | NA | NA | NA |
peritumoral edema on T2WI | 0.015883232 | 6.185856224 | NA | NA | NA | |
DCE-MRI radiomics | a_wavelet-LHH_glszm_SmallAreaEmphasis | NA | NA | 0.010987487 | 6.920084579 | 0.001458018 |
a_wavelet-LHL_firstorder_Skewness | NA | NA | 0.017251576 | 6.023456627 | -0.522407251 | |
b_wavelet-HLH_glszm_SizeZoneNonUniformityNormalized | NA | NA | 0.001811074 | 10.73097267 | 0.441698208 | |
b_wavelet-HLH_glszm_SmallAreaEmphasis | NA | NA | 0.000770171 | 12.65733984 | 0.350393762 | |
c_wavelet-HHH_glcm_MaximumProbability | NA | NA | 0.005073203 | 8.510581486 | -0.006136234 | |
c_wavelet-HHH_glrlm_RunLengthNonUniformityNormalized | NA | NA | 0.038189923 | 4.507011169 | 0.275763724 | |
c_wavelet-HHL_glszm_SmallAreaEmphasis | NA | NA | 0.002185117 | 10.31798937 | 0.346955315 | |
c_wavelet-LHH_glcm_Imc1 | NA | NA | 0.033946605 | 4.726340662 | 0.02556862 | |
d_wavelet-HHL_glszm_SmallAreaEmphasis | NA | NA | 0.006421524 | 8.018453923 | 0.04114306 | |
e_gradient_glcm_Imc1 | NA | NA | 0.028745645 | 5.039358324 | 0.379574092 | |
e_wavelet-LHH_glcm_Imc1 | NA | NA | 0.031984038 | 4.837987471 | 0.089713532 |