Design and Optimization of Dedicated Computed Tomography of the Breast

描述（由申请人提供）：本生物工程研究资助响应PA-07-279（重新发布PA-06-419）。这是马萨诸塞大学医学院和犹他大学犹他高级成像研究中心的合作项目。在胶片和数字乳房x光检查中，乳房结构的叠加可能由于“掩蔽”效应而导致漏诊，或者可能模拟病变的存在，从而导致额外的成像或活检。乳腺专用计算机断层扫描（CT）成像可以克服这种叠加问题，并提供更好的对比度，从而提高病变的可检测性。此外，乳腺CT可以在低辐射剂量（相当于乳房x线摄影）下进行，没有或只有很小的乳房物理压迫，减轻了患者的不适。它可以提供三维病变形态，可以作为诊断指标，更好地定量评估乳腺腺体含量，这可能是乳腺癌的危险因素。本工作的重点是设计和优化一个专用的乳房CT系统。具体而言，研究计划是解决乳房CT中已知的挑战，例如最大限度地在扫描场中包含乳房组织，考虑数据完整性和重建复杂性确定适当的源-检测器轨迹，以及开发噪声抑制方案以提高微钙化可见性。研究计划包括确定最佳患者体位，以最大限度地扩大乳房组织包膜，实现几种图像采集轨迹和适当的重建算法，全面的数学模拟和关键系统和重建参数的优化。通过独立于观察者的物理指标和基于乳房切除手术标本的观察者研究，将对在台式系统上取得的改善进行完整的表征。本研究的结果将提供符合人体工程学的设计和优化的图像采集和重建技术，可以很容易地转化为临床成像系统。设计良好的乳腺专用CT系统可作为筛查诊断成像、植入成像、三维术前规划、术前治疗监测、微创手术指导的平台技术。我们相信我们的方法将对乳腺癌的检测和治疗产生重大影响。公共卫生相关性：具有专用乳腺CT系统的体积三维成像有潜力成为克服乳房x线照相术中组织重叠问题的有效工具，该问题导致遗漏癌症和不必要地召回患者进行额外的成像，并用于监测治疗的有效性。然而，存在一些已知的挑战，如在扫描过程中最大化乳房组织包裹，由于图像噪声和对比度损失而导致的微钙化可视化，以及锥束伪影。在本研究中，我们建议开发方法来克服这些挑战，并展示使用这些方法获得的改进的图像质量。

DESCRIPTION (provided by applicant): This bioengineering research grant is responsive to PA-07-279 (reissue of PA-06-419). It is a collaboration between the University of Massachusetts Medical School and the Utah Center for Advanced Imaging Research at University of Utah. Superposition of breast structures in screen-film and digital mammography may result in missed cancers due to 'masking' effect, or may mimic the presence of a lesion resulting in additional imaging or biopsy. Dedicated breast Computed Tomographic (CT) imaging of the breast can overcome this superposition problem, and provide much improved contrast, thus improving lesion detectability. Further, breast CT can be performed at low radiation dose (equivalent to mammography) with no or minimal physical compression of the breast, alleviating patient discomfort. It can provide for 3-D lesion morphology, which could serve as a diagnostic indicator, and for better quantitative assessment of breast glandular content, a likely risk factor for breast cancer. This work is focused on design and optimization of a dedicated breast CT system. Specifically, the research plan is to address known challenges in breast CT, such as maximizing the inclusion of breast tissue within scan field, determining appropriate source-detector trajectory with consideration for data completeness and reconstruction complexity, and developing noise suppression schemes to improve microcalcification visibility. The research plan includes determining optimal patient position for maximizing breast tissue inclusion, implementation of several image acquisition trajectories and appropriate reconstruction algorithms, comprehensive mathematical simulation and optimization of critical system and reconstruction parameters. Complete characterization of the improvement achieved on a bench-top system will be performed through observer-independent physical metrics and through surgical mastectomy specimen-based observer studies. The results from this study will provide for an ergonomic design and optimized image acquisition and reconstruction technique that can be readily translated to a clinical imaging system. A well-designed dedicated breast CT system can serve as a platform technology for screening and diagnostic imaging, implant imaging, 3-D presurgical planning, monitoring preoperative treatment, and guidance in minimally invasive surgery. We believe our approach will have a major impact on the detection and management of breast cancer. PUBLIC HEALTH RELEVANCE: Volumetric 3-D imaging with a dedicated breast CT system has the potential to be an effective tool to overcome the tissue superposition problem in mammography which results in missed cancers and unnecessary recall of the patient for additional imaging, and for monitoring the effectiveness of therapeutic treatments. However, there are known challenges such as maximizing breast tissue inclusion during the scan, visualization of microcalcifications due to image noise and loss of contrast, and cone- beam artifacts. In this research, we propose to develop methods to overcome these challenges and demonstrate the improved image quality obtained with such methods.