Image Quality Improvement and Breast Compression Reduction in Breast Tomosynthesi

乳房断层合成中的图像质量改善和乳房压迫减少

• 批准号: 8616734
• 负责人: Ioannis Sechopoulos
• 金额: $ 28.86万
• 依托单位: EMORY UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2012
• 资助国家: 美国
• 起止时间: 2012-03-08 至 2016-02-29
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8616734
• 关键词: Address; Adherence; American Cancer Society; Appearance; Behavior; Breast; Breast Cancer Detection; Breast Cancer Early Detection; Breast-Conserving Surgery; Callback; Canada; Clinical; Computer software; Computer-Assisted Diagnosis; Coupled; Detection; Diagnosis; Digital Breast Tomosynthesis; Dimensions; Disease; Dose; Early Diagnosis; Europe; Future; Guidelines; Image; Image Analysis; Imaging Techniques; Imaging technology; Lesion; Malignant Neoplasms; Mammaplasty; Mammography; Masks; Mastectomy; Medical Imaging; Methods; Nature; Noise; Normal tissue morphology; Pain; Patients; Procedures; Process; Radiation; Reporting; Research Project Grants; Risk Assessment; Screening Result; Sensitivity and Specificity; Specificity; Surveys; Techniques; Technology; Testing; Tissues; Woman; Work; X-Ray Computed Tomography; base; breast cancer diagnosis; cancer imaging; cancer risk; image processing; improved; interest; malignant breast neoplasm; model development; mortality; novel; radiologist; screening; two-dimensional

DESCRIPTION (provided by applicant): Breast cancer is the most prevalent non-skin cancer in women in the US. To increase the early detection of breast cancer, annual screening mammography has been recommended for asymptomatic patients. However, mammography is not without limitations. The superposition of glandular tissue is an inherent consequence of the two dimensional nature of mammography, and may result in the masking of lesions of interest (reducing sensitivity) or the false appearance of overlapping normal tissues as lesions (reducing specificity). Digital tomosynthesis of the breast is a novel x-ray based imaging technique being investigated to replace mammography for the early detection and diagnosis of breast cancer. Tomosynthesis is based on the concept of limited angle computed tomography, in which projections of the imaged object acquired from a limited angular range give enough information to partially reconstruct that object in three dimensions. Reports indicate that tomosynthesis reduces callback rate (increasing specificity) and increases detection rate (increasing sensitivity) compared to mammography. However, in current implementations of breast tomosynthesis, the breast undergoes the same vigorous compression as that used in mammography. It has been shown that the pain resulting from this vigorous breast compression during mammography is an important factor in women not undergoing the recommended annual screening mammography. This reduction in screening adherence results in an increase in breast cancer mortality and in a reduction in treatment options, yielding a higher number of mastectomies as opposed to breast conserving surgeries. The objective of this research project is to develop and test new image acquisition and processing techniques that will allow for the acquisition of breast tomosynthesis images with a substantial reduction in the amount of breast compression used with no loss in image quality or increase in radiation dose. For this, we will (1) develop a method to reduce the impact of x-ray scatter on image quality; (2) develop new image acquisition techniques for reduced compression tomosynthesis that result in the same radiation dose and image noise levels as standard tomosynthesis; (3) compare the image quality between standard tomosynthesis and reduced compression tomosynthesis performed with the techniques developed in this project. Successful completion of this project will result in the ability to perform breast tomosynthesis imaging with substantially reduced breast compression levels, with no increase in radiation dose, and no loss in image quality and tissue coverage. This will result in a reduction in the pain associated with breast cancer screening, resulting in an increase in screening adherence by women, which will decrease breast cancer mortality and increase the number of cases in which breast conserving surgery can be performed.

描述（由申请人提供）：乳腺癌是美国女性中最常见的非皮肤癌。为了提高乳腺癌的早期发现率，建议对无症状患者每年进行一次乳房X线检查。然而，乳房X线照相术并非没有局限性。腺体组织的重叠是乳房X线照相术二维性质的固有结果，并且可能导致感兴趣的病变被掩盖（降低敏感性）或重叠的正常组织被错误地表现为病变（降低特异性）。乳房数字断层合成是一种基于 X 射线的新型成像技术，正在研究取代乳房 X 线摄影，用于乳腺癌的早期检测和诊断。断层合成基于有限角度计算机断层扫描的概念，其中从有限角度范围获取的成像对象的投影提供了足够的信息以在三个维度上部分重建该对象。报告表明，与乳房 X 线摄影相比，断层合成降低了回调率（提高了特异性）并提高了检测率（提高了灵敏度）。然而，在目前的乳房断层合成实施中，乳房受到与乳房X线照相术中使用的相同的剧烈压缩。研究表明，乳房 X 光检查期间剧烈的乳房压迫所产生的疼痛是女性不接受推荐的年度乳房 X 光检查的一个重要因素。筛查依从性的降低导致乳腺癌死亡率增加和治疗选择减少，从而导致乳房切除术数量高于保乳手术。 该研究项目的目标是开发和测试新的图像采集和处理技术，该技术将允许采集乳房断层合成图像，同时大幅减少乳房压缩量，而不会损失图像质量或增加辐射剂量。为此，我们将（1）开发一种方法来减少X射线散射对图像质量的影响； (2) 开发新的图像采集技术，以减少压缩断层合成，从而获得与标准断层合成相同的辐射剂量和图像噪声水平； (3) 比较标准断层合成和使用本项目开发的技术执行的简化压缩断层合成之间的图像质量。 该项目的成功完成将能够在显着降低乳房受压水平的情况下进行乳房断层合成成像，而不会增加辐射剂量，并且不会损失图像质量和组织覆盖范围。这将减少与乳腺癌筛查相关的疼痛，从而提高女性筛查的依从性，从而降低乳腺癌死亡率并增加可以进行保乳手术的病例数量。