Quantitative topographic endoscopy for improved screening of non-polypoid colorectal neoplasms

定量地形内窥镜检查改善非息肉样结直肠肿瘤的筛查

基本信息

批准号：
9769725
负责人：
Nicholas James Durr
金额：
$ 32.75万
依托单位：
JOHNS HOPKINS UNIVERSITY
依托单位国家：
美国
项目类别：
财政年份：
2018
资助国家：
美国
起止时间：
2018-09-01 至 2022-06-30
项目状态：
已结题

来源：
https://reporter.nih.gov/project-details/9769725
关键词：
3-Dimensional Address Adult Algorithms American Area Biopsy Caliber Cancer Etiology Carcinoma Categories Cessation of life Characteristics Classification Clinical Clinical assessments Colon Colonoscopes Colonoscopy Color Colorectal Colorectal Cancer Colorectal Neoplasms Complex Computational Technique Computer Vision Systems Computer software Confocal Microscopy Consumption Custom Data Detection Development Dyes Effectiveness Elements Endoscopes Endoscopy Engineering Foundations Frequencies Future Gastroenterology Goals Height Human Hybrids Image Imaging Techniques Interobserver Variability Knowledge Large Intestine Lesion Light Lighting Malignant - descriptor Maps Measurement Measures Morphology Motion Mucous Membrane Optics Patients Pilot Projects Polypectomy Polypoid Lesion Polyps Premalignant Procedures Research Research Proposals Resected Resolution Sampling Shapes Source Spatial Frequency Domain Imaging Structure Surface Surface Properties System Techniques Technology Testing Texture Three-Dimensional Image Time Tissues Training United States adenoma base chromoscopy colorectal cancer risk computer aided detection contrast imaging human study imaging system improved innovation miniaturize mortality mortality risk novel optical imaging phantom model prevent research clinical testing routine screening screening vector

项目摘要

Project Summary This project will create and assess new optical imaging and computational technologies with the goal of improving the detection rates of precancerous, non-polypoid lesions during colonoscopy screening. Identifying and removing these subtle lesions is critical to improving the protective value of colonoscopy in reducing mortality from colorectal cancer. However, current approaches to non-polypoid lesion detection are largely unsuccessful because they are time-consuming and require specialized training (chromoendoscopy), or they start with poor image contrast (software analysis of conventional video). This project focuses on developing a novel technique, called quantitative topographic endoscopy (QTE), that optically measures colon surface properties via a modified commercial colonoscope. The key innovation in this proposal is to utilize structured illumination and build on concepts from computer vision and optical engineering to acquire high-resolution 3D images of the colon surface through a custom endoscope. The project will be implemented through three specific aims: (1) develop a miniaturized, quantitative, high-resolution topography system, (2) implement QTE in a modified commercial colonoscope ready for clinical testing, and (3) determine the validity of QTE in a phantom model and its clinical feasibility in a pilot human study. QTE systems developed in this project will be tested in tissue-mimicking phantoms with a goal of achieving better than 1-mm height sensitivity, in ex-vivo resected colon samples with a goal of accurately reconstructing surface shapes from a complex tissue, and in a pilot human study with the goal of obtaining surface topography non-polypoid lesions. Beyond increasing non-polypoid lesion detection rates, QTE has the potential to address other limitations of colonoscopy, including preventing missed polypoid lesions, classifying lesions for resect-and-discard strategies, and improving colonoscopy quality metrics. Additionally, the development of a QTE system that is approved for human studies will serve as a platform for future clinical assessment of other optical techniques such as spatial frequency domain imaging and speckle imaging in a variety of gastroenterology applications.

项目概要该项目将创建和评估新的光学成像和计算技术，目标是提高结肠镜筛查期间癌前、非息肉样病变的检出率。识别去除这些细微病变对于提高结肠镜检查在降低死亡率方面的保护价值至关重要来自结直肠癌。然而，当前的非息肉样病变检测方法基本上不成功因为它们非常耗时并且需要专门的培训（色素内窥镜检查），或者一开始就很差图像对比度（传统视频的软件分析）。该项目的重点是开发一项新技术，称为定量地形内窥镜 (QTE)，它通过改进的光学方法测量结肠表面特性商业结肠镜。该提案的关键创新是利用结构化照明并建立在利用计算机视觉和光学工程的概念来获取结肠表面的高分辨率 3D 图像通过定制内窥镜。该项目将通过三个具体目标来实施：（1）制定小型化、定量、高分辨率地形系统，(2) 在修改后的商业中实施 QTE 准备好进行临床测试的结肠镜，以及（3）确定模型模型中 QTE 的有效性及其临床人体试验研究的可行性。该项目开发的 QTE 系统将在组织模拟中进行测试模型的目标是在体外切除的结肠样本中实现超过 1 毫米的高度灵敏度从复杂组织中精确重建表面形状的目标，并在一项试点人体研究中实现了这一目标获得非息肉样病变的表面形貌。除了提高非息肉样病变检出率之外， QTE 有潜力解决结肠镜检查的其他局限性，包括防止遗漏息肉样病变、对病变进行分类以实施切除和丢弃策略，并提高结肠镜检查质量指标。此外，开发批准用于人体研究的 QTE 系统将作为未来临床的平台评估其他光学技术，例如空间频域成像和散斑成像各种胃肠病学应用。

项目成果

期刊论文数量（8）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

Colonoscopy 3D video dataset with paired depth from 2D-3D registration.

结肠镜检查 3D 视频数据集，具有 2D-3D 配准的配对深度。

DOI：
发表时间：
2023-12
期刊：
影响因子：
10.9
作者：
Bobrow, Taylor L;Golhar, Mayank;Vijayan, Rohan;Akshintala, Venkata S;Garcia, Juan R;Durr, Nicholas J
通讯作者：
Durr, Nicholas J

Rapid tissue oxygenation mapping from snapshot structured-light images with adversarial deep learning.

通过对抗性深度学习，从快照结构光图像中快速绘制组织氧合图。