作者: Vijay Rajagopal , Angela Lee , Jae-Hoon Chung , Ruth Warren , Ralph P. Highnam
DOI: 10.1016/J.ACRA.2008.07.017
关键词: Image (mathematics) 、 Image registration 、 Biomechanics 、 Artificial intelligence 、 Biomechanical Phenomena 、 Computer vision 、 Skin surface 、 Surgery 、 Breast cancer 、 Magnetic resonance imaging 、 Computer science 、 Neutral buoyancy 、 Radiology Nuclear Medicine and imaging
摘要: Rationale and Objectives Anatomically realistic biomechanical models of the breast potentially provide a reliable way mapping tissue locations across medical images, such as mammograms, magnetic resonance imaging (MRI), ultrasound. This work presents new modeling framework that enables us to create are customized individual. We demonstrate framework's capabilities by creating left breasts two volunteers tracking their deformations MRIs. Materials Methods generate finite element automatically fitting geometrical segmented data from MRIs, characterizing in vivo mechanical properties (assuming homogeneity) tissues. For each volunteer, we identified unloaded configuration acquiring MRIs under neutral buoyancy (immersed water). Such is clearly not practical clinical setting; however, these previously unavailable with important which validate biomechanics. Internal features were images tracked prone gravity-loaded state using framework. Results The predicted root-mean-square errors 4.2 3.6 mm predicting skin surface for volunteer. mean error 3.7 4.7 Conclusions capture shape internal clinically acceptable accuracy. Further refinement incorporation more anatomic detail will make useful cancer diagnosis.