Breast Cancer Detection Using Electrical Impedance Measurements

描述（由申请人提供）：本项目的目标是开发一种计算机辅助系统，使用电阻抗断层扫描（EIT）测量来检测乳房异常，特别是癌症。本项目旨在开发替代和廉价的筛查工具，以帮助诊断乳房X光检查不确定的情况。我们假设不同的异常具有不同的电磁特性，因此这些特性可以使用电阻抗数据进行识别和分类。虽然各种乳房成像技术是可用的，我们的目标是统计检测偏离正常。因此，我们的项目旨在不仅获得乳房的图像，但提供一个概率评估的事实，即在感兴趣的区域，乳房的电磁特性不同于其他组织或从两侧对称的位置在对侧乳房。我们的项目将利用最近资助的替代乳腺成像模式的计划项目的努力和活动。具体来说，我们将使用EIT患者数据，包括来自该项目的癌症数据。将探索三种方法用于统计乳腺异常检测和区分：（1）直接从Neumann-to-Dirichlet映射导出的乳房特征金字塔，（2）通过将基本的拉普拉斯偏微分方程简化为一组耦合的常微分方程，使用乘积图像的异常定位和检测，（3）图像重建的混合模型方法，其中将从EIT数据估计最佳正则化参数。这些新的方法是基于最近的数学发现有关的魔术Toeplitz矩阵，imageless重建的电磁特性在体内组织使用表面测量，欧姆定律的广义形式，和一个混合模型的方法来解决复杂的逆问题。该项目的目标是开发一个计算机辅助系统，利用电阻抗断层扫描（EIT）测量来检测乳房异常，特别是癌症。拟议的项目旨在开发替代和廉价的筛查工具，以协助诊断乳房X光检查不确定的情况。我们预计，我们的方法将提高癌症检测的灵敏度，特别是对乳腺组织致密的年轻女性。

DESCRIPTION (provided by applicant): The goal of this project is to develop a computer-aided system to detect abnormalities in breasts, cancer particularly, using electrical impedance tomography (EIT) measurements. The present project aims to develop alternative and inexpensive screening tools to assist in diagnosis in situations where mammography is inconclusive. We hypothesize that different abnormalities have different electromagnetic properties and therefore these properties may be identified and classified using electrical impedance data. Although various breast imaging techniques are available, our goal is to statistically detect a deviation from normality. Therefore, our project aims to not only obtain an image of the breast, but to provide a probabilistic assessment of the fact that in the region of interest, the electromagnetic properties of the breast differ from the rest of the tissue or from bilaterally symmetric locations in the contralateral breast. Our project will leverage the effort and activity from a recently funded program project on alternative breast imaging modalities. Specifically we will use, EIT patient data, including cancer data from this project. Three methods will be explored for statistical breast abnormality detection and discrimination: (1) a Breast Signature Pyramid derived directly from the Neumann-to-Dirichlet map, (2) abnormality localization and detection using the Product--Image by reducing the underlying Laplace partial differential equations into a set of coupled ordinary differential equations, (3) the Mixed Model approach to image reconstruction in which optimal regularization parameters will be estimated from the EIT data. These novel approaches are based on recent mathematical discoveries related to the magic Toeplitz matrix, imageless reconstruction of the electromagnetic properties of in vivo tissues using surface measurements, the generalized form of Ohm's law, and a mixed model approach to solving complex inverse problems. PUBLIC HEALTH RELEVANCE The goal of this project is to develop a computer-aided system to detect abnormalities in breasts, cancer particularly, using electrical impedance tomography (EIT) measurements. The proposed project aims to develop alternative and inexpensive screening tools to assist in diagnosis in situations where mammography is inconclusive. We anticipate that our methodology will improve the sensitivity of cancer detection especially for younger women with dense breast tissue.