Impact of improved spatial resolution on radiomic features using photon-counting-detector CT

Area of Science:

• Medical Imaging
• Radiology
• Quantitative Imaging

Background:

• Radiomics analyzes quantitative features from medical images for disease characterization and outcome prediction.
• Photon-counting-detector CT (PCD-CT) offers superior spatial resolution and dose efficiency compared to energy-integrating-detector CT (EID-CT).
• Improved spatial resolution in CT imaging can influence textural features and radiomic calculations.

Purpose of the Study:

• To quantify the impact of PCD-CT's improved spatial resolution on radiomic features.
• To evaluate the effect of resolution differences on sample classification using radiomics.
• To compare radiomic feature calculations between PCD-CT and EID-CT systems.

Main Methods:

• Organic samples (zucchini, onions, oranges) were scanned using both clinical PCD-CT and EID-CT systems.
• Images were reconstructed with dedicated sharp (PCD-CT) and routine (EID-CT) kernels.
• Noise levels were quantified, and 14 relevant radiomic features were calculated and compared between scanners.

Main Results:

• Thirteen out of 14 radiomic features were significantly altered by the improved resolution of PCD-CT.
• Cluster separation of sample types was improved when using features derived from PCD-CT.
• PCD-CT features, despite higher noise, demonstrated better sample differentiation.

Conclusions:

• High-resolution imaging from PCD-CT substantially impacts radiomic feature calculations.
• PCD-CT derived radiomic features show potential for improved classification and clinical decision-making in radiomics.
• The enhanced detail from PCD-CT necessitates careful consideration in radiomics analysis and validation.