Table 3 The final selected features as respective inputs into the five models
From: Predicting response to CCRT for esophageal squamous carcinoma by a radiomics-clinical SHAP model
Model | Feature (Filter _Feature Class_Feature) | Abbreviation |
|---|---|---|
Clinical | Maximum Diameter of the Lesion | MDL |
CTV | wavelet-HLH_firstorder_Mean | Wav-HLH_FMean |
wavelet-LLL_glcm_Imc1 | Wav-LLL_GI | |
square_glcm_InverseVariance | Squ_GIV | |
gradient_firstorder_10Percentile | Gra_F10P | |
CTV-Clinical | wavelet-LLL_glcm_Imc1 | Wav-LLL_GI |
wavelet-HLH_firstorder_Mean | Wav-HLH_FMean | |
square_glcm_InverseVariance | Squ_GIV | |
Maximum diameter of the lesions | MDL | |
GTV | original_shape_Elongation | Ori_SE |
original _shape_Flatness | Ori_SF | |
wavelet-HHH_glcm_Imc1 | Wav-HHH_GI | |
square_gldm_LargeDependenceHighGrayLevelEmphasis | Squ_GLDHGLE | |
GTV-Clinical | wavelet-HHH_glcm_Imc1 | Wav-HHH_GI |
original_shape_Elongation | Ori_SE | |
square_gldm_LargeDependenceHighGrayLevelEmphasis | Squ_GLDHGLE | |
Maximum diameter of the lesions | MDL |