Prognostic value of CT-based radiomics in grade 1–2 pancreatic neuroendocrine tumors

Cancer Imaging

Table 2 Radiomics score (R-score)

Feature group	Radiomics feature*	Coefficient
Morphologic feature	Volume	–0.047
	Flatness	–0.075
First-order, intensity-based statistical feature	Maximum (AP, N)	–0.027
	25th percentile (AP, N)	–0.196
	Median (AP, LOG)	0.049
	Robust mean absolute deviation (AP, LOG)	–0.069
Texture, GLCM	Cluster shade (AP, N)	0.085
	Difference entropy (AP, N)	–0.106
	Joint entropy (AP, N)	0.656
Texture, GLSZM	Large zone high gray level emphasis (AP, N)	0.104
Higher order, GLCM	Cluster shade (AP, LOG)	0.251
	Contrast (AP, LOG)	0.028
Higher order, GLSZM	Zone size variance (AP, LOG)	–0.013
Higher order, GLDZM	Large distance high gray level emphasis (AP, G)	0.116
Higher order, NGLDM	High dependence high gray level emphasis (AP, LOG)	–0.112

Note—radiomics score is expressed as the sum of each radiomics feature multiplied by its coefficient
*The information within parentheses indicates the scan phases and types of images used for radiomics feature extraction
AP, arterial phase; G, Gaussian; GLCM, gray-level co-occurrence matrix; GLDZM, gray-level distance-zone matrix; GLSZM, gray-level size-zone matrix; LOG, Laplacian-of-Gaussian; N, non-transformed raw image; NGLDM, neighboring gray-level dependence matrix

ISSN: 1470-7330