Machine Learning Based on a Multiparametric and Multiregional Radiomics Signature Predicts Radiotherapeutic Response in Patients with Glioblastoma

Table 2

A summary of the high-throughput radiomics features extracted.

MRI sequences	Region	Group	Feature name	Type
T1WI	Whole tumour	Shape	MinorAxisLength	Origin
CE-T1WI	Edema	Texture	MeanAbsoluteDeviation	Wavelet-HHL
CE-T1WI	Enhancement	Texture	GLCM_JointEnergy	Wavelet-LLH
CE-T1WI	Enhancement	Texture	GLDM_DependenceNonUniformityNormalized	Wavelet-LLH
CE-T1WI	Enhancement	Intensity	90Percentile	Wavelet-LHH
CE-T1WI	Enhancement	Intensity	90Percentile	Wavelet-HLH
T2WI	Whole tumour	Intensity	90Percentile	Log-sigma-1-mm
T2WI	Nonenhancement	Texture	GLSZM-SizeZoneNonUniformity	Wavelet-LHH