Medical image processing and computer-assisted surgery

申请人的研究重点是先进医学影像在计算机辅助手术和癌症诊断中的应用。在本研究中，实时超声成像将发挥重要作用。与其他传统成像技术相比，超声波是一种非电离工具（在对儿童和孕妇进行手术时尤其重要），成本显著降低，并且可以在整个手术过程中实时可视化解剖结构。提出了一种从超声数据中高效、自动分割和配准骨表面的概率框架。这一框架将用于计算机辅助骨科创伤手术新技术的发展。该框架将扩展到生成统计解剖地图集和基于地图集的注册技术，以及用于适应性放射治疗的软组织的可变形注册。申请人研究的另一个方面涉及将计算技术应用于基于超声的前列腺癌诊断。创新的假设是，有一个混沌动态控制的相互作用的超声波和组织。这一假设将通过从前列腺组织收集的原始超声回波信号中提取分形和统计特征来验证。使用先进的分类技术，提取的特征将与癌症的存在和分期相关联。这项跨学科研究将在主要临床医生的密切合作下进行。这些算法的设计将考虑到直接的临床应用，预计将对加拿大的医疗保健部门产生重大影响。该研究项目对研究生具有重要的培训潜力，在那里他们将学习将他们的技术知识转化为具有实际意义的高影响力医学问题。

The applicant's research focuses on the deployment of advanced medical imaging in computer-assisted surgery and cancer diagnosis. In this research, real-time ultrasound imaging will play a significant role. In contrast to other conventional imaging techniques, ultrasound is a non-ionizing tool (which is especially important when performing operations on children and pregnant women), has significantly lower cost, and allows real-time visualization of the anatomy throughout the procedure. A probabilistic framework that could efficiently and automatically segment and register bone surfaces from ultrasound data is proposed. This framework will be utilized in the development of novel techniques in computer-assisted orthopaedic trauma surgery. The framework will be extended to generate statistical anatomical atlases and atlas-based registration techniques, as well as deformable registration of soft tissue for adaptive radiotherapy.Another aspect of the applicant's research involves applying computing techniques to ultrasound-based prostate cancer diagnosis. The innovative hypothesis is that there is a chaotic dynamic governing the interaction of ultrasound and the tissue. This hypothesis will be verified by extracting fractal and statistical features from raw ultrasound echo signals, collected from prostate tissue.  The extracted features will be correlated with the existence and staging of cancer using advanced classification techniques.This interdisciplinary research will be performed in close collaboration with leading clinicians. The algorithms will be designed with direct clinical applications in mind, expected to greatly impact Canada's healthcare sector. The research program has significant training potential for graduate students, where they will learn to translate their technical knowledge to high-impact medical problems with practical implications.