Optical Imaging Fused with Tomosynthesis for Improved Breast Cancer

描述（由申请人提供）：本申请响应项目公告编号PAR-07-214（体内成像系统和癌症研究方法开发和验证的学术-工业合作伙伴关系），提议在马萨诸塞大学达特茅斯医学院和Hologic， Inc（马萨诸塞州贝德福德）之间建立3-way合作伙伴关系，开发和验证一种用于乳房成像的新型融合技术——近红外（NIR）光谱断层扫描（NIRST）与乳房断层合成（BTS）相结合。目标是创建一个单一的检查平台，将通过NIRST获得的功能参数与BTS提供的高分辨率3D结构信息协同结合，使诊断决策具有优越的ROC特征，并大大提高ppv，而不是当前与乳腺癌监测相关的回呼和诊断实践。该项目充分利用了三个合作伙伴的重要技术、临床和商业优势。具体来说，达特茅斯大学在开发NIRST在乳腺成像中的应用的先进概念方面提供了十多年的经验，包括与其他传统成像模式成功融合的形式。达特茅斯也作为领先机构参与了Hologic系统的BTS早期临床研究。马萨诸塞大学在乳房成像的x射线物理方面取得了成就和专业知识，包括在系统硬件以及图像处理和重建算法方面取得了重大进展，这对拟议的研究至关重要。Hologic公司是开发最先进的BTS技术的商业领导者，在BTS评估性临床试验的设计和实施方面拥有丰富的经验，并准备在不久的将来推出第二代商业化系统，预计将获得FDA的批准。该项目的具体目标是：(1)开发NIRST和BTS的融合，并发展为一个完全集成的平台，在BTS成像过程中，光学成像技术将永久存在；(2)开发x射线图像分割和相关重建算法，以定义乳腺实质模式作为近红外图像形成的结构先验；(3)优化和描述Aim 1和Aim 2中最有前途的硬件和软件结果，并在模型和早期临床试验中对其进行评估，以确定可行性，并以两个相同原型的形式完善其功能；(4)开展一项两中心临床研究，旨在证明所提出的融合平台的有效性，并确立该方法的临床潜力。公共卫生相关性：乳房x线摄影是一种检测乳腺癌的有效方法，也是唯一被证明可以降低该疾病死亡率的技术；然而，它的局限性是众所周知的，包括低灵敏度和阳性预测价值，主要是由于乳房x光片中正常实质组织的重叠，特别是在致密的乳房中。提议的项目代表了三方学术-工业合作伙伴关系，其目标是创建一种创新的近红外（NR）光谱断层扫描（NIRST）与乳腺断层合成（BTS）的融合，该融合已被证明(i)在临床工作流程方面足够稳健，已部署在双中心临床研究中；（ii）在临床潜力方面得到验证，可以比单独的BTS增加诊断价值；（iii）准备进行更大规模的多中心评估性临床试验，足以产生获得FDA批准所需的数据。

DESCRIPTION (provided by applicant): This application responds to Program Announcement Number PAR-07-214 (Academic-Industrial Partnerships for Development and Validation of In Vivo Imaging Systems and Methods for Cancer Investigations) by proposing a 3-way partnership between Dartmouth, University of Massachusetts (UMass) Medical School and Hologic, Inc (Bedford, MA) to develop and validate a new fusion technology for breast imaging - Near Infrared (NIR) spectral tomography (NIRST) integrated with breast tomosynthesis (BTS). The goal is to create a single exam platform that will synergistically combine the functional parameters obtained through NIRST with the high resolution 3D structural information available from BTS to enable diagnostic decisions that exhibit superior ROC characteristics and substantially improved PPVs over current call-back and diagnostic practices associated with breast cancer surveillance. The project capitalizes on the significant technical, clinical and commercial strengths of its three partners. Specifically, Dartmouth provides more than a decade of experience in the development of advanced concepts for application of NIRST in breast imaging including forms that have been successfully fused with other conventional imaging modalities. Dartmouth has also participated as a leading institution in early clinical studies of BTS with the Hologic system. UMass brings accomplishments and expertise in the x-ray physics of breast imaging that includes the development of significant advances in system hardware as well as image processing and reconstruction algorithms which are pivotal to the proposed studies. Hologic, Inc, is the commercial leader in the development of the most advanced BTS technology, has extensive experience in the design and conduct of BTS evaluative clinical trials and is poised to release commercially its second generation system which is expected to gain FDA approval in the near future. The specific aims of the project are: (1) to develop a fusion of NIRST and BTS that progresses to a fully integrated platform where the optical imaging technology is permanently in place during BTS imaging; (2) to develop the x-ray image segmentation and concomitant reconstruction algorithms for defining breast parenchymal patterns as structural priors for NIR image formation; (3) to optimize and characterize the most promising hardware and software outcomes from Aim 1 and Aim 2 and evaluate them in phantoms and early-stage clinical exams to establish feasibility and refine their capabilities in the form of two identical prototypes; (4) to conduct a two- center clinical study designed to demonstrate the validity of the proposed fusion platform and establish the clinical potential of the approach. PUBLIC HEALTH RELEVANCE: Screen-film mammography is an effective method for detecting breast cancer and is the only technique proven to reduce mortality from the disease; however, its limitations are well known and include low sensitivity and positive predictive value, largely caused by the overlap of normal parenchymal tissues in the mammogram, especially in the dense breast. The goal of the proposed project, which represents a three-way academic- industrial partnership, is to create an innovative fusion of Near Infrared (NR) spectral tomography (NIRST) with breast tomosynthesis (BTS) that has been demonstrated to be (i) sufficiently robust with respect to clinical workflow to have been deployed in a two-center clinical study, (ii) validated in terms of its clinical potential to add diagnostic value over BTS alone and (iii) ready for much larger-scale multi-center evaluative clinical trials sufficient for generating the data required for gaining FDA approval.