Skip to main content
Insights into Imaging
Fig. 1 | Insights into Imaging

Fig. 1

From: Robustness of radiomics features of virtual unenhanced and virtual monoenergetic images in dual-energy CT among different imaging platforms and potential role of CT number variability

Fig. 1

Study workflow. This study was composed of three steps, namely image acquisition, radiomics analysis and robustness analysis. A standardized phantom was scanned on ten platforms on seven DECT-capable scanners of three types with the same voxel and typical abdomen-pelvic examination parameters. Corresponding VUE images and VMI70keV were generated. Pyradiomics was employed to extract 19 first-order and 75 texture radiomics features from ROIs segmented with a rigid registration. The test–retest repeatability was evaluated by Bland–Altman analysis for repeated scans, and the hypothesis that the obtained biases of the radiomics feature values between the scan and rescan was equal to zero was tested by one-sample t test. The inter-platform reproducibility among VUE images, and that among VMI70keV images, were assessed by CV and QCD. Inter-platform reproducibility between two particular platform pairs were estimated by ICC and CCC to characterize inter-platform difference across DECT platforms. Since there were ten platforms, forty-five comparisons were performed within the VUE images and within the VMI70keV, respectively. CT number and their inter-platform reproducibility were calculated. The correlation between CT number variability of and percentage of robust radiomics features was investigated. dsDECT = dual-source dual-energy CT, rsDECT = rapid kV-switching dual-energy CT, dlDECT = dual-layer dual-energy CT

Back to article page