Development, Assessment, and Applications of Novel Micro-CT Imaging Devices

Abstract

Preclinical imaging plays an essential role in biomedical research. Among availablemodalities, micro-computed tomography (micro-CT) offers high spatial resolution with relatively high throughput making it valuable for structural and functional imaging in small animals. However, traditional micro-CT systems are limited by poor soft tissue contrast, lack of spectral specificity, and underutilization of modern computational methods. This dissertation addresses these limitations through the development of novel imaging systems, the application of spectral photon-counting CT (PCCT), and the integration of machine learning for quantitative image analysis. For this work, several hardware platforms were designed and validated, including a dual-detector system combining a photon-counting and energy-integrating detector, a high throughput turn-table perfusion scanner, and a dual detector configuration using both a cadmium telluride and gallium arsenide detector for spectral imaging. Imaging data acquired from these systems were used in combination with deep learning models for anatomical segmentation and with radiomics pipelines for feature extraction. Quantitative imaging metrics were applied to evaluate cardiac function, quantify perfusion, and classify tumors. Additionally, spectral decomposition techniques were used to distinguish and quantify multiple contrast agents within a single scan. The results demonstrate that combining advanced hardware configurations with spectral and computational techniques significantly enhances the information content of preclinical CT imaging. These integrated approaches improve soft tissue contrast, enable functional imaging, and support high-throughput, biologically relevant quantification. The dissertation concludes that the next generation of micro-CT systems—when coupled with spectral detectors and machine learning—has the potential to transform preclinical imaging into a more versatile and quantitative tool for translational research.