DEVELOPING OF ULTRASONIC TISSUE CHARACTERIZATION METHODS

本项目旨在提高学生的诊断能力

This Program Project is designed to enhance the ability of diagnostic ultrasound imaging to detect and to characterize tumors in humans. Since ultrasonic imaging is widely used in many medical specialties, and is non- ionizing, any improvement that makes it more useful would have a beneficial impact on medicines. The improvements in oncology would be in detection, diagnosis and treatment monitoring. The program consists of two types of projects: 1. Imaging efforts that take different approaches to finding optimum methods of processing the echo data from current instruments to form images. 2. Physics projects seeking to advance the current state of imaging and to provide better data to the imaging efforts. Combing these two types of project allows direct communication and synergy between the groups. The imaging efforts are based on statistical methods that are largely new in the medical context and do not duplicate the types of signal processing using power spectra that have been widely investigated in this field. Our methods depend on phase information in addition to the amplitude information in power spectra. In future the combination of these new and the existing methods with the power spectrum methods may provide much improve tissue characterization. The physics projects are seeking to develop the wider bandwidth transducers desired by the imagers, as well as to develop improved contrast materials. We provide for laboratory and clinical cores to give each project access to the raw r.f. data, results of biopsy, and support in coordination of their efforts. In addition, the clinical collaborators provide insight based on tissue structure and function, and on the results from conventional ultrasound and other imaging modalities. The cores work together closely in maintaining the ability to collect clinical data under agreed protocols, and in providing advice. A major core effort consists of assessing the possible improvement in imaging by using the receiver operating characteristic (ROC) method for analysis of the results of the different imaging projects. This is a cooperative effort between the cores and involves all of the imaging projects. We are currently investigating the possible improvements in diagnostic accuracy that is possible by using optimum combinations of features from the imaging projects. The study can be expanded to include power spectral methods in future.