Robustness of Photon-Counting CT and Dect Radiomics Feature In Chest Tumor Phantom: A Two Center
Abstract
Purpose
This study systematically evaluated the robustness of radiomics features derived from photon-counting detector CT (PCD-CT) and three dual-energy CT (DECT) platforms under different reconstruction strategies and scanning conditions using chest tumor phantoms.
Methods
Twelve chest tumor phantoms with varying sizes and texture characteristics were constructed using a CRIS chest phantom. All phantoms were scanned on four energy CT systems: PCD-CT, dual-layer DECT (DL-DECT), dual-source DECT (DS-DECT), and rapid kV-switching DECT (RS-DECT). Images were reconstructed as conventional images (CI), virtual monoenergetic images (VMI; 40, 70, and 100 keV), and virtual non-contrast (VNC) images. Tumor-specific regions of interest were delineated, and radiomic features were extracted following standardized preprocessing. Inter-scanner robustness was assessed using the concordance correlation coefficient (CCC), coefficient of variation (CV), and quartile coefficient of dispersion (QCD).
Results
Radiomics robustness varied markedly across reconstruction types. In CI, the highest robustness for small spherical phantoms was observed between DS-DECT and RS-DECT (CCC up to 0.968), whereas other device pairs generally showed CCC values below 0.75. Among texture phantoms, robust agreement was mainly limited to DS-DECT and RS-DECT, with additional consistency between PCD-CT and DL-DECT. In VNC images, robustness between PCD-CT and DECT systems was consistently poor (CCC < 0.75), while agreement among DECT platforms remained acceptable. For VMI, low-energy images (40 keV) significantly improved robustness between PCD-CT and DL-DECT or DS-DECT, whereas robustness at 70 and 100 keV was reduced. Similar robustness patterns were observed across radiation dose levels.
Conclusion
Radiomics feature robustness across modern energy CT platforms is strongly dependent on reconstruction strategy. Low-energy VMI substantially enhances inter-scanner robustness, including for PCD-CT, whereas VNC reconstruction compromises harmonization between PCD-CT and DECT systems. These findings underscore the importance of reconstruction selection and standardization in multicenter thoracic radiomics studies.