EVALUATION OF TUNED APERTURE CT FOR 3-D MAMMOGRAPHY

描述（改编自申请人的摘要）： 应用是模拟和测试一个实际的，高分辨率， 三维（3-D）乳房X射线摄影系统在体外。 过程 利用了一种新的断层合成重建方案，称为调谐孔径 计算机断层扫描（TACT）。 这种方法是基于光学的扩展 孔径理论，将传统的透射射线照相术与 计算机断层扫描。 它从根本上不同于其他层析成像 方法，由于这一事实，所有的投影信息所需的 3-D重建从二维（2-D）投影导出，其 投影几何形状在曝光时是未知的。 这种差异 有利于生产高分辨率的三维成像器， 使用相对简单的技术，从现有的 乳房X光检查单位。 这项研究计划涉及重新配置一个传统的 基于现有数字探测器的固态乳房X线摄影系统 模拟TACT，使用专门开发的 测试对象，以比较3D成像性能与理论 预测，并使用结果信息作为模拟的指导 体外临床性能。 通过这些模拟， 在临床上与传统乳腺X线摄影图像进行比较， 从特别挑选和准备的尸体上获得的性能 标本使用ROC分析病变检测任务和适当的 病变定位任务的参数分析。

DESCRIPTION (Adapted from Applicant's Abstract): The main objective of the application is simulation and testing of a practical, high-resolution, three-dimensional (3-D) mammographic x-ray system in vitro. The process makes use of a new tomosynthetic reconstruction scheme called tuned-aperture computed tomography (TACT). This method is based on an extension of optical aperture theory that links conventional transmission radiography with computed tomography. It differs fundamentally from other tomographic methods by virtue of that fact that all projection information required for 3-D reconstruction is derived from two-dimensional (2-D) projections whose projection geometry is unknown at the time of exposure. This difference facilitates the production of high-resolution 3-D imager over extended periods of time using relatively simple technology derived from existing mammography units. The research plan involves reconfiguring a conventional solid-state mammography system based on an existing digital detector to simulate TACT, undertaking comprehensive analyses using specially-developed test objects to compare imaging performance in 3-D with theoretical predictions, and using resulting information as a guide for simulating clinical performance in vitro. Through these simulations TACT is to be compared clinically with conventional mammographic images for diagnostic performance obtainable from specially selected and prepared cadaveric specimens using ROC analysis for lesion-detection tasks and appropriate parametric analyses for lesion-location tasks.