An Intelligent Capsule Endoscopy Video Analysis Software Platform

描述（由申请人提供）：胃肠道恶性肿瘤是世界上最常见的癌症原因。在全球每年1080万癌症患者中，约330万人患有胃肠道恶性肿瘤。直到最近，通过传统的内窥镜检查，唯一可以成像的胃肠道部分是食道、胃和结肠。小肠的初始部分使用推肠镜或放射性成像技术进行成像，但这些方法给患者带来明显的不适，成本高，或缺乏敏感性。胶囊内镜（CE）是一项正在改变胃肠道内镜诊断格局的新技术。CE是一项广受赞誉的GI影像学突破，尤其是小肠影像学。然而，目前的CE视频显示/分析软件（SW）远远落后。它的局限性包括较长的视频观看和解释时间（例如~2小时），并且难以分析（例如可视化）结果。这些SW问题给患者带来了相当大的成本，需要紧急解决。为了解决这个问题，我们与我们的合作者一起，建议开发一种新颖的软件系统，使视频审查过程完全自动化，实现全面准确的疾病分析。提出的软件系统将具有以下临床影响和意义：3/4从CE视频中全自动高效的疾病检测和分类；3/4综合可视化或分析工具，更准确诊断疾病；3/4审查过程中没有遗漏胃肠道疾病领域；3/4低成本SW系统，可更好地为大众接受并有效预防疾病；3/4模块化，可扩展和可扩展技术。所开发的专有软件模块化（功能）技术可以很容易地应用于现有的诊断仪器。千禧研究集团预测，胶囊内窥镜市场仅在美国就将从2004年的4000万美元增长到2009年的1.8亿美元。预计2008 - 2015年的复合年增长率为35.1%。根据给定成像有限公司的数据，全球有超过一百万的患者使用他们的PillCam胶囊进行了胶囊内窥镜检查。胶囊内窥镜是一个快速增长的数十亿美元的市场（预计2015年全球将达到10亿美元大关）。所提出的软件系统和技术的商业化潜力是显著的。

DESCRIPTION (provided by applicant): Gastrointestinal (GI) malignancies are the most common cause of cancer in the world. Of the 10.8 million people in the world each year who develop cancer, approximately 3.3 million have GI malignancies. Until recently, the only sections of the GI tract that could be imaged were the esophagus, stomach and colon through traditional endoscopy. Initial segments of the small bowel were imaged using techniques such as push enteroscopy or radioactive imaging techniques, but these methods present significant discomfort to the patients, costly, or lack sensitivity. Capsule endoscopy (CE) is a new technique that is changing the landscape of GI endoscopic diagnosis. CE is a widely acclaimed breakthrough in GI, especially small bowel imaging. However, current CE video display/analysis software (SW) is falling far behind. It has limitations including lengthy video viewing and interpreting time (e.g. ~2 hours) and is difficult in analyzing (e.g. visualizing) findings. These SW issues have incurred considerable amount of costs to patients, and need to be addressed urgently. To address this problem, together with our collaborators, we propose to develop a novel software system to fully automate the video review process and achieve comprehensive and accurate disease analysis. The proposed software system will have the following clinic impact and significance: 3/4 Fully automated and efficient disease detection and classification from CE videos; 3/4 Comprehensive visualization or analysis tool for more accurate disease diagnosis; 3/4 No more missing GI disorder areas from the review process; 3/4 A low cost SW system for better mass acceptance and effective disease prevention; and 3/4 Modular, expandable, and scalable technology. The proprietary software modular (functionally) technology developed can be easily applied to existing diagnosis instruments. The Millennium Research Group predicted the capsule endoscopy market to generate more than $180 million in 2009 in the United States alone - up from just $40 million in 2004. The predicted CAGR is 35.1% during 2008 - 2015. According to Given Imaging Ltd., more than one million patients worldwide have gone through capsule endoscopy procedure using their PillCam capsules. Capsule endoscopy is a rapidly-growing multi-billion dollar market (predicted to reach billion dollar mark in 2015 worldwide). The commercialization potential of the proposed software system and technology is significant. PUBLIC HEALTH RELEVANCE: The software system and technology developed under this SBIR project will provide great benefit to GI patients by providing cost saving and highly accurate GI tract disease finding and analysis software platform. The proposed software will transform the current manual review method into a fully automatic process. With the extended capability over current endoscopes, the proposed SW system is expected to have a reasonably large market share of this 10 billion dollar industry. Other medical areas benefitting from this research include bronchoscopy, colonoscopy, enteroscopy, cystoscopy, laparoscopy, and capsule endoscope. Non-medical fields to benefit this technology include commercial perimeter surveillance, and water and gas pipe inspection - another multi-billion market.