Spectroscopic Photoacoustic Molecular Imaging for Breast Lesion Characterization

用于乳腺病变表征的光谱光声分子成像

• 批准号: 9314864
• 负责人: Katheryne E Wilson
• 金额: $ 7.6万
• 依托单位: STANFORD UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-06-01 至 2019-05-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9314864
• 关键词: Advisory Committees; Algorithms; Antibodies; Benign; Binding; Binding Proteins; Bioinformatics; Biological Neural Networks; Biopsy; Breast; Breast Cancer Detection; Breast Cancer Model; CD276 gene; Callback; Cancer Biology; Cancer Detection; Cessation of life; Classification; Clinic; Clinical; Contrast Media; Data; Detection; Development; Digital Signal Processing; Dimensions; Early Diagnosis; Education; Ethics; Excision; FDA approved; Faculty; Female; Fluorescence; Fluorescent Dyes; Frozen Sections; Goals; Health; Histologic; Hyperplasia; Image; Individual; Indocyanine Green; Institution; Lesion; Linear Regressions; Machine Learning; Malignant - descriptor; Malignant Neoplasms; Mammary Gland Parenchyma; Mammography; Measures; Medical; Mentors; Meta-Analysis; Methodology; Methods; Modality; Modeling; Molecular; Molecular Target; Mus; Noise; Noninfiltrating Intraductal Carcinoma; Operative Surgical Procedures; Optics; Patients; Postoperative Period; Predictive Value; Preparation; Prospective Studies; ROC Curve; Recurrence; Research; Research Personnel; Scheme; Sensitivity and Specificity; Signal Transduction; Specificity; Statistical Methods; Supervision; Technical Expertise; Techniques; Testing; Therapeutic; Time; Tissue Model; Tissues; Training; Transgenic Mice; Transgenic Organisms; Translating; Ultrasonography; United States; Validation; Woman; absorption; antibody conjugate; base; breast imaging; breast lesion; breast surgery; cancer biomarkers; career; career development; clinically actionable; clinically translatable; contrast enhanced; cost; course development; design; diagnostic accuracy; economic need; fluorescence imaging; formal learning; image processing; imaging agent; imaging approach; imaging modality; improved; in vivo; informal learning; malignant breast neoplasm; molecular imaging; mortality; mouse model; novel; optical imaging; overexpression; prototype; screening; targeted agent; targeted imaging; tool; tumor

PROJECT SUMMARY/ABSTRACT Claiming more than 40,000 lives in the United States in 2015, breast cancer presents an important health focus. Mammography and ultrasound, current screening methods, suffer from low sensitivity and low positive predictive value, respectively, particularly in patients with dense breast tissues. Therefore, a non-invasive method of distinguishing between benign and malignant lesions that could be incorporated with current screening modalities is critically needed. With more advanced screening methods, there is an increase in the detection of early malignant lesions, for which breast-conserving treatment has become more routine. However, intraoperative frozen-section margin assessment is time consuming and suffers from low sensitivity, while post-operative histological analysis leaves potential for positive margins, strongly correlated with reoccurrence. Therefore, a real-time method to detect tumor margins intraoperatively is critically needed. We propose using spectroscopic photoacoustic and fluorescence molecular imaging combined with a clinically- translatable contrast agent targeted to a novel breast cancer marker (B7-H3) to non-invasively distinguish normal from malignant tissues both during screening (aim 1) and intraoperatively during surgical resection (aim 3). The sensitivity of this imaging method will be increased with the use of machine learning post-processing algorithms to autonomously detect molecular B7-H3 signal (aim 2). In summary, this proposal will result in significant change to current clinical breast imaging and surgical resection practice to improve the detection and treatment of focal breast lesions. The training portion of this plan, required to accomplish these research goals, has been designed with trainee mentors with specific technical expertise. Dr. Willmann is an expert in translational molecular imaging and contrast agent use, while Dr. Rubin is an expert in bioinformatics, image processing, and machine learning for medical imagine purposes. Additionally, the project is supported by a diverse advisory committee with experts in clinical breast imaging (Dr. Debra Ikeda), optical imaging and intraoperative guidance (Dr. Christopher Contag), and clinical breast surgery (Dr. Irene Wapnir). To date, the candidate has developed expertise in photoacoustic, ultrasound, and fluorescence molecular imaging and molecular contrast agent development and in vivo use during her graduate and postdoctoral research. Her long term career goals include developing clinically translatable combined spectroscopic photoacoustic and fluorescence molecular imaging methods combined with novel contrast agents for cancer detection and differentiation. Additionally, her research will focus on developing machine learning algorithms for increasing the sensitivity of the molecular imaging approach as well as adapting her method for therapeutic purposes. In preparation for her independent research career, the training plan includes formal education in machine learning, digital signal processing, optical imaging, and cancer biology, as well as in career development classes and ethical conduct of research.

项目总结/摘要 2015年，美国有4万多人死于乳腺癌，这是一个重要的健康问题。 专心点目前的筛查方法乳腺X线摄影和超声检查存在灵敏度低、阳性率低的缺点， 预测值，分别，特别是在致密乳腺组织的患者。因此，非侵入性 一种区分良性和恶性病变的方法，可以结合电流 筛选模式是迫切需要的。随着更先进的筛查方法， 早期恶性病变的检测，因此保乳治疗已成为常规。 然而，术中冰冻切片边缘评估耗时且敏感性低， 虽然术后组织学分析留下了阳性边缘的可能性，但与 重现因此，迫切需要一种术中实时检测肿瘤边缘的方法。我们 建议使用光谱光声和荧光分子成像结合临床- 靶向新型乳腺癌标记物（B7-H3）的可转移造影剂， 在筛选期间（aim 1）和手术切除期间（aim 3）。这种成像方法的灵敏度将随着机器学习后处理的使用而增加 自主检测分子B7-H3信号的算法（目标2）。总而言之，这一提议将导致 对当前临床乳腺成像和手术切除实践进行了重大变更，以提高检测率 和治疗局灶性乳腺病变。 本计划的培训部分，需要完成这些研究目标，已设计与受训者 具有特定技术专长的导师。Willmann博士是翻译分子成像专家， 对比剂的使用，而鲁宾博士是生物信息学，图像处理和机器学习的专家， 医学想象的目的此外，该项目还得到了一个由专家组成的多元化咨询委员会的支持。 在临床乳腺成像（Debra Ikeda博士）、光学成像和术中引导（Christopher博士 Contag）和临床乳房手术（Irene Wapnir博士）。到目前为止，候选人已经发展了以下方面的专门知识： 光声、超声和荧光分子成像和分子造影剂开发， 在她的研究生和博士后研究期间体内使用。她的长期职业目标包括发展 临床上可转换的组合光谱光声和荧光分子成像方法 与用于癌症检测和区分的新型造影剂组合。此外，她的研究将 专注于开发机器学习算法，以提高分子成像的灵敏度 方法以及调整她的方法用于治疗目的。为了准备独立 研究职业，培训计划包括机器学习，数字信号处理， 光学成像和癌症生物学，以及职业发展课程和研究的道德行为。