Tomosynthesis for Improved Breast Cancer Detection

用于改进乳腺癌检测的断层合成

• 批准号: 7096059
• 负责人: JOSEPH Y LO
• 金额: $ 30.4万
• 依托单位: DUKE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-06-20 至 2010-04-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7096059
• 关键词: breast neoplasms; clinical research; computers; health science profession; mammary gland; mammography; neoplasm /cancer; performance; radiology; tissues

DESCRIPTION (provided by applicant): The purpose of this study is to investigate tomosynthesis mammography to improve the sensitivity and specificity of breast cancer detection. By removing overlapping tissue, this system can allow detection of obscured lesions and permit better 3D characterization of lesions. The specific aims of the proposed study are to: (1) Optimize tomosynthesis technique using specimen and patient images. (2) Implement soft-copy display system for tomosynthesis images and CAD outputs to maximize radiologist productivity and performance. (3) Perform prospective pilot studies to evaluate radiologist performance with breast tomosynthesis. (4) Investigate CAD of breast masses as an integral component of breast tomosynthesis. In preliminary studies, we evaluated the physical characteristics of an investigational digital mammography system modified for tomosynthesis scanning, and optimized the radiographic technique for tomosynthesis. We translated our experience with chest tomosynthesis into mammography, and demonstrated feasibility of our reconstruction algorithm on phantoms, mastectomy specimens, and patients. We developed novel CAD algorithms for the detection of breast masses in mammography, and will parlay that experience into the 3D tomosynthesis data. This proposal presents a complete plan to demonstrate physical and clinical feasibility. Upon the successful completion of each aim, the immediate benefit will be crucial information which will facilitate commercial translation of this breast tomosynthesis system by our industrial partner, as well as the evaluation of similar systems from other academic groups and manufacturers.

描述（由申请人提供）：本研究的目的是探讨乳房x线摄影提高乳腺癌检测的敏感性和特异性。通过去除重叠的组织，该系统可以检测到模糊的病变，并允许更好的病变3D表征。本研究的具体目的是：(1)利用标本和患者图像优化断层合成技术。(2)实施软拷贝显示系统，用于断层合成图像和CAD输出，以最大限度地提高放射科医生的生产力和绩效。(3)进行前瞻性试点研究，以评估放射科医生在乳腺断层合成方面的表现。(4)探讨乳腺肿块的CAD作为乳腺断层合成的一个组成部分。在初步研究中，我们评估了用于断层合成扫描的实验性数字乳房x线摄影系统的物理特性，并优化了断层合成的放射摄影技术。我们将胸部断层合成的经验转化为乳房x线照相术，并证明了我们的重建算法在幻影、乳房切除术标本和患者上的可行性。我们开发了一种新的CAD算法来检测乳房x光检查中的肿块，并将这种经验运用到3D断层合成数据中。这个建议提出了一个完整的计划，以证明物理和临床的可行性。在每个目标成功完成后，直接的好处将是关键的信息，这将促进我们的工业合作伙伴对这种乳房组织合成系统的商业翻译，以及对其他学术团体和制造商的类似系统的评估。