Table 2 The image features selected by LASSO Cox model for construction of multiparametric radiomics signature. The features were ranked by their contribution to the signature. The feature was named as modality_region_matrix_title, where title can be found in Supplementary Table 1. Note that there were two different calculations for GLCM_IMC, which can be found in4. Here GLCM_IMC2 indicated the second calculation listed in4.

From: A Fully-Automatic Multiparametric Radiomics Model: Towards Reproducible and Prognostic Imaging Signature for Prediction of Overall Survival in Glioblastoma Multiforme

No.

Feature

Weight

Extraction Parameters

C-Index on Validation Data

\({f}_{1}\)

T1_wholetumor_GLCM_IDMN

−0.67085113

VS = 1, QM = Ld, GL = 128

0.397

\({f}_{2}\)

T1C_solidecore_GLRLM_HGRE

0.66718691

VS = 1, QM = Ld, GL = 32

0.582

\({f}_{3}\)

T1_wholetumor_GLSZM_GLN

0.52140412

VS = 1, QM = Eq, GL = 32

0.605

\({f}_{4}\)

T1C_nonenhancing_GLCM_IMC2

-0.15004393

VS = 1, QM = Eq, GL = 64

0.401