Repeatability and reproducibility study of radiomic features on a phantom and human cohort

Abstract

The repeatability and reproducibility of radiomic features extracted from CT scans need to be investigated to evaluate the temporal stability of imaging features with respect to a controlled scenario (test–retest), as well as their dependence on acquisition parameters such as slice thickness, or tube current. Only robust and stable features should be used in prognostication/prediction models to improve generalizability across multiple institutions. In this study, we investigated the repeatability and reproducibility of radiomic features with respect to three different scanners, variable slice thickness, tube current, and use of intravenous (IV) contrast medium, combining phantom studies and human subjects with non-small cell lung cancer. In all, half of the radiomic features showed good repeatability (ICC > 0.9) independent of scanner model. Within acquisition protocols, changes in slice thickness was associated with poorer reproducibility compared to the use of IV contrast. Broad feature classes exhibit different behaviors, with only few features appearing to be the most stable. 108 features presented both good repeatability and reproducibility in all the experiments, most of them being wavelet and Laplacian of Gaussian features.

Prognostic Value of Radiomic Features of 18F-FDG PET/CT in Patients With B-Cell Lymphoma Treated With CD19/CD22 Dual-Targeted Chimeric Antigen Receptor T Cells

“…The radiomic features of PET and CT images were selected by the following procedure. Intraclass correlation coefficient (ICC > 0.75) was first performed to remove the redundant features (25,26). Subsequently, the retained features were further selected by the least absolute shrinkage and selection operator (LASSO) regression algorithm (27).…”

Radiomic Features of 18F-FDG PET in Hodgkin Lymphoma Are Predictive of Outcomes

“…A total of 15 patients were randomly selected to calculate the interobserver agreement of the feature extraction. e intraclass correlation coefficient (ICC) was used to determine the repeatability/reproducibility of features in our research, and ICC >0.75 was selected [24][25][26]. Subsequently, the least absolute shrinkage and selection operator (LASSO) COX regression model was used to select the most useful prognostic features with 10-fold cross validation for selecting the parameter Lambda in the training cohort [27,28] (adriamycin, bleomycin, vinblastine, and dacarbazine).…”

Radiomics of Patients with Locally Advanced Rectal Cancer: Effect of Preprocessing on Features Estimation from Computed Tomography Imaging

“…Robustness of radiomic features relies on reproducibility and repeatability of their estimates considering different aspects of the radiomic workflow [ 19 , 20 ]. Previous studies of CT imaging [ 21 – 24 ], as well as of MR [ 25 – 30 ] and NM [ 31 – 34 ] imaging, have assessed the reproducibility and repeatability of radiomic features estimation for various applications. Phantom and in vivo CT studies have reported dependence of radiomic feature estimates on various factors such as scanner type [ 35 , 36 ], tube current [ 37 – 39 ], acquisition voxel size [ 21 , 35 ], reconstruction kernel [ 40 – 43 ], and number of gray levels [ 21 ] or gray level discretization [ 35 ].…”