DEVELOPMENT OF PATIENT-DERIVED ANTHROPOMORPHIC BREAST PHANTOMS FOR AUTOMATED OBJECTIVE ASSESSMENT OF BREAST X-RAY IMAGING SYSTEMS

只要乳腺癌仍然是一个普遍的健康问题，对不断改进的乳腺癌成像技术的需求就会存在。设计变更——纳入新的硬件技术或方法——可能引发临床试验的需要，以确保设备安全性和评估性能。进行临床试验不仅费用昂贵，而且通常需要数年的时间来招募和研究患者。利用物理幻影，而不是病人，会更快，更便宜，减少病人的风险，并可能加快拯救生命的技术进步向公众提供的速度。因为目前可用的压缩和未压缩乳房幻象都不是真实的乳房图像，所以用户没有这个选项。利用我们独特的访问大数据集（超过1100）的三维真实患者乳房图像和一种新的方法，我们建议填补这一需求。我们提出了一个两步的过程，设计，制造和评估收集的现实拟人化的乳房幻影。从我们的大型图像数据集中，我们将生成一系列不同乳房形状、大小、密度和腺体分布的数字和物理模型。我们的产品将提供给制造商、医院和研究人员。

As long as breast cancer remains a pervasive health concern, the need for ever-improving breast cancer imaging technology will exist. Design changes – to incorporate new hardware technologies or methods – may trigger the need for clinical trials to ensure device safety and assess performance. Conducting a clinical trial is not only expensive, but often takes years for patient recruitment and study. Utilizing physical phantoms, rather than patients, would be faster, far less expensive, reduce risk to patients and potentially increase the speed with which life-saving technological advances are made available to the public. Because currently available compressed and uncompressed breast phantoms are not real breast images, the users do not have this option. Utilizing our unique access to a large data set (over 1100) of three-dimensional real patient breast images and a novel approach to phantom development, we propose to fill this need. We propose a two-step process for the design, fabrication and evaluation of a collection of realistic anthropomorphic breast phantoms. From our large image data set we will generate the digital and physical phantoms for a series of phantoms spanning different breast shape, size, density and glandular distribution. Our product will be available to manufacturers, hospitals and researchers.