Predicting the response to neoadjuvant chemotherapy for breast cancer: wavelet transforming radiomics in MRI

BMC Cancer

Table 2 Textures and performance (AUC, Accuracy, Sensitivity and Specificity) of six RF models

	RF Models
	I	II	III	IV	V	VI
Features	Volumetric	Volumetric + Peripheral	Wavelet	Volumetric + Wavelet	Peripheral + Wavelet	Volumetric + Peripheral + Wavelet
Selected Features	GLZSM_salgle GLCM_homo1 GLCM_diffEntro GLCM_dissimilar SHAPE_surfaceArea	GLZSM_salgle GLCM_homo1 GLCM_diffEntro GLCM_dissimilar Bndry_RL_rln Bndry_GLZSM_salgle Bndry_GLCM_contrast SHAPE_surfaceArea	LHH_GLZSM_zp LLH_GLCM_infoCorr1 HHH_GLCM_correlation	GLZSM_salgle GLCM_homo1 GLCM_diffEntro GLCM_dissimilar HHH_GLCM_correlation LHH_GLZSM_zp LLH_GLCM_infoCorr1 SHAPE_surfaceArea	Bndry_RL_rln Bndry_GLZSM_salgle Bndry_GLCM_contrast HHH_GLCM_correlation LHH_GLZSM_zp LLH_GLCM_infoCorr1	SHAPE_surfaceArea GLZSM_salgle GLCM_homo1 GLCM_diffEntro GLCM_dissimilar Bndry_RL_rln Bndry_GLZSM_salgle Bndry_GLCM_contrast HHH_GLCM_correlation LHH_GLZSM_zp LLH_GLCM_infoCorr1
AUC (mean ± SD)	0.816 ± 0.033	0.823 ± 0.020	0.888 ± 0.025	0.876 ± 0.015	0.885 ± 0.030	0.874 ± 0.019
Accuracy (mean ± SD)	0.747 ± 0.022	0.751 ± 0.0150	0.810 ± 0.030	0.781 ± 0.028	0.797 ± 0.032	0.787 ± 0.024
Sensitivity (mean ± SD)	0.676 ± 0.043	0.684 ± 0.043	0.762 ± 0.035	0.730 ± 0.049	0.770 ± 0.034	0.727 ± 0.045
Specificity (mean ± SD)	0.812 ± 0.023	0.798 ± 0.021	0.845 ± 0.031	0.818 ± 0.037	0.818 ± 0.047	0.830 ± 0.035

NOTE: Bndry_GLZSM_salgle, Bndry_GLCM_contrast and Bndry_RL_rln were peripheral texture features

ISSN: 1471-2407