Acquisition and registration of 3D optical surface scanning, MRI, and microwave tomographic imaging for improved clinical breast imaging and surgical planning.

摘要

Continual improvements to clinical breast imaging technologies have led to ever earlier detections and increasingly accurate localizations of female breast cancers. While these advancements have certainly improved the prognoses of many patients, challenges remain to accurately acquiring and interpreting diagnostic images of the breast. Specifically, soft tissue deformation within the various imaging environments hinders image reconstruction, multimodal registration, and image-based treatment planning. Consequently, a method of quantifying and modeling breast tissue deformations would significantly improve the clinical utility of breast imaging technologies.;Of particular interest is the use of three-dimensional (3D) optical scanners to track the shape of the breast through various imaging modalities. When combined with a model of tissue deformation, optical scanners become convenient tools for enhancing existing imaging technologies. This thesis discusses the development and assessment of specific optical scanning and deformable registration methods as they relate to advancements in the microwave tomographic imaging, magnetic resonance imaging (MRI), and surgical planning of breast cancer patients.;Image reconstructions of clinical microwave imaging spectroscopy (MIS) data are improved with prior knowledge of the breast surface geometry. This study investigates an optical laser scanning system for recovering the breast geometry during clinical microwave imaging. Sub-millimeter scanner accuracy is demonstrated along with improvements in microwave image quality.;Additionally, preoperative supine breast MRI may be registered to intraoperative optical scanning to improve the localization of cancerous lesions during breast conserving surgeries. Here, an early phase clinical trial explores this localization method along with a model of breast tissue deformation. Analysis of the results leads to an optimized tissue deformation model based on 3D surface registration and finite element biomechanical modeling.;Finally, the application of optical scanning and tissue deformation modeling to MIS/MRI-based multimodal imaging is surveyed. Preliminary results from scans of healthy patients demonstrate the potential clinical benefits of this approach.