Minimizing Uncertainty in Breast Ultrasound Imaging with Real-Time Coherence-Based Beamforming

通过基于实时相干的波束形成最大限度地减少乳房超声成像的不确定性

• 批准号: 10679017
• 负责人: Muyinatu A. Lediju Bell
• 金额: $ 34.42万
• 依托单位: JOHNS HOPKINS UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-08-15 至 2026-04-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10679017
• 关键词: Accounting; Acoustics; Anxiety; Benign; Biopsy; Biostatistics Core; Board Certification; Breast; Breast Cancer Detection; Cancer Patient; Clinic; Clinical; Clinical Pathways; Clutterings; Collaborations; Complex; Cyst; Data; Dedications; Developing Countries; Development; Diagnosis; Diagnostic; Disease; Early Diagnosis; Environment; Equipment; Faculty; Goals; Healthcare; Healthcare Systems; Hospitals; Human body; Image; Imaging Techniques; Imaging technology; In Situ; Investigation; Length; Letters; Liquid substance; Liver; Malignant - descriptor; Malignant Neoplasms; Malignant neoplasm of liver; Malignant neoplasm of pancreas; Malignant neoplasm of testis; Mammary Ultrasonography; Mammography; Mass in breast; Methods; Morbidity - disease rate; Organ; Pancreas; Pathology; Pathway interactions; Patients; Performance; Procedures; Publishing; Reader; Real-Time Systems; Recording of previous events; Research; Research Personnel; Resource Allocation; Resources; Series; Solid; Speed; Techniques; Technology; Testing; Time; Ultrasonography; Uncertainty; United States; Universities; Wait Time; Woman; Work; breast malignancies; cancer diagnosis; clinical care; deep learning; deep neural network; experience; follow-up; imaging modality; imaging scientist; improved; innovation; malignant breast neoplasm; men; mortality; neural network architecture; novel; novel strategies; photoacoustic imaging; radiologist; rural area; screening; success; tool; ultrasound

Project Summary Approximately 280,000 women are expected to be diagnosed with breast cancer in the United States in 2021 and more than 40,000 will die from the disease. It is well documented that early detection results in improved morbidity and mortality. Ultrasound imaging is an important screening and diagnostic breast cancer detection tool, particularly for women with dense breasts when mammography tends to be suboptimal. While suspicious findings may be clarified with ultrasound imaging, a subset of ultrasound images yield inconclusive results, ne- cessitating biopsies or follow-up imaging, which increase patient anxiety and places additional burdens on the time available for clinical care and the resource allocations of our healthcare system. One reason for this out- standing challenge is that dense breasts tend to create images with significant acoustic clutter, which confounds the differentiation of an otherwise anechoic mass (which is indicative of a benign cyst) from a truly hypoechoic mass (which could be indicative of malignancy). In addition, it can be difficult to distinguish a complicated cyst (which has internal echoes due to proteinaceous material and is benign) from either a solid mass or a complex cystic and solid mass (which could be malignant) using standard ultrasound imaging methods alone. The objective of this proposal is to develop new, real-time ultrasound imaging technology that will simplify clin- ical workflows by distinguishing fluid-filled masses from solid masses and from complex cystic and solid masses, which all appear hypoechoic in traditional ultrasound B-mode images. Our novel approach, Robust Short-Lag Spatial Coherence (R-SLSC) imaging, has demonstrated feasibility to make this distinction by incorporating coherence-based beamforming to augment existing beamforming methods available in clinical ultrasound scan- ners. Aim 1 will focus on development of a real-time system for implementing matched B-mode and R-SLSC imaging. Aim 2 will evaluate and compare real-time system performance. Aim 3 will assess the ability of our novel methods to distinguish fluid from solid or complex cystic and solid masses utilizing a combination of quanti- tative analyses and task-oriented reader studies. Aim 4 will investigate advanced methods to retrieve coherence information and diagnostic information regarding mass contents from ultrasound channel data, including recently discovered options that rely on coherence lengths and lag-one coherence values without requiring reader input. Successful completion of these aims will lead to a real-time, ultrasound-based tool to confidently distinguish solid from fluid hypoechoic breast masses and provide a more simplified clinical workflow for the most challenging of these cases. In addition, results from the proposed studies will be applicable to clarifying the content of masses that may appear in multiple organs throughout the human body (e.g., testicular, liver, or pancreatic masses).

项目概要 预计 2021 年美国将有约 28 万名女性被诊断出患有乳腺癌 超过 40,000 人将死于这种疾病，有据可查的是，早期发现可以改善病情。 超声成像是乳腺癌筛查和诊断的重要检测方法。 工具，特别是对于乳房致密的女性，当乳房 X 光检查往往不理想时。 超声成像可以澄清研究结果，超声图像的子集产生不确定的结果，ne- 停止活检或后续成像，这会增加患者的焦虑并给医生带来额外的负担 用于临床护理的时间和我们医疗保健系统的资源分配是造成这种情况的一个原因。 长期存在的挑战是，致密的乳房往往会产生带有明显声学杂波的图像，这会混淆 区分无回声肿块（表明良性囊肿）与真正的低回声肿块 此外，很难区分复杂的囊肿。 （由于蛋白质材料而具有内部回声并且是良性的）来自固体块或复合体 仅使用标准超声成像方法即可发现囊性和实性肿块（可能是恶性的）。 该提案的目标是开发新的实时超声成像技术，以简化临床工作 通过区分充满液体的物质和固体物质以及复杂的囊性和固体物质来进行工作流程， 这些在传统超声 B 型图像中均出现低回声，我们的新颖方法是鲁棒短滞后。 空间相干 (R-SLSC) 成像已证明通过合并来实现这种区分的可行性 基于相干的波束形成，以增强临床超声扫描中可用的现有波束形成方法 目标 1 将重点开发用于实现匹配 B 模式和 R-SLSC 的实时系统。 目标 2 将评估和比较实时系统性能，目标 3 将评估我们的能力。 利用定量组合区分流体与固体或复杂囊性和固体物质的新方法 目标 4 将研究检索连贯性的先进方法。 有关来自超声通道数据的大量内容的信息和诊断信息，包括最近 发现了依赖相干长度和滞后一相干值的选项，而不需要读者输入。 成功完成这些目标将产生一种基于超声波的实时工具，可以自信地区分 从液体低回声乳腺肿块中分离出固体，并为最具挑战性的问题提供更简化的临床工作流程 此外，拟议研究的结果将适用于澄清群众的内容。 可能出现在人体多个器官中（例如睾丸、肝脏或胰腺肿块）。

项目成果

Coherence Metrics for Reader-Independent Differentiation of Cystic From Solid Breast Masses in Ultrasound Images.

• DOI: 10.1016/j.ultrasmedbio.2022.08.018
• 发表时间: 2023-01
• 期刊: ULTRASOUND IN MEDICINE AND BIOLOGY
• 影响因子: 2.9
• 作者: Wiacek, Alycen;Oluyemi, Eniola;Myers, Kelly;Ambinder, Emily;Bell, Muyinatu A. Lediju
• 通讯作者: Bell, Muyinatu A. Lediju