Low-cost, portable, computer assisted imaging for diabe*

描述(申请人提供)：我们开发了一种便携式、低成本、患者友好的视网膜成像设备，Delft Retinascan，基于双波长扫描激光技术。它不需要药理上的瞳孔放大或闪光灯照明，该设备的原型成本不到1100美元。我们利用它来采集正常受试者的高空间分辨率的视网膜图像，典型的采集时间为0.1s。我们和其他人已经证明，使用数码相机和互联网连接进行大规模早期诊断是安全有效的。我们还表明，计算机辅助诊断算法可以从标准商用眼底相机拍摄的两场非立体彩色眼底照片中评估糖尿病视网膜病变，与视网膜专家高度一致。潜在地，这些技术可以组合成一个低成本、高质量、患者友好的糖尿病视网膜病变早期诊断系统。然而，在这种低成本的相机和数字诊断系统可以在临床上实施之前，还需要进行临床验证。 我们的具体目标是： 1.从技术上验证低成本的Delft Retinascan相机系统对糖尿病视网膜病变的非侵入性成像； 2.通过与威斯康星州阅读中心(WRC)合作，评估和验证代尔夫特视网膜扫描的临床应用。WRC将负责读取和评估使用Retinascan获得的图像，并将它们与ETDRS标准7野立体眼底照片进行比较； 3.评估在Delft Retinascan获得的图像上添加计算机辅助诊断算法对性能的改善。 我们已经组建了一个团队，代表了互联网糖尿病视网膜病变筛查、光学工程、视网膜成像、安全在线医学图像服务器、图像分析和流行病学方面的专业知识，由以研究为导向的眼科医生、物理学家、计算机工程师和生物统计学家组成，他们准备将上述技术转化为临床实用系统。我们的总体目标是验证和优化一种低成本、高质量、患者友好的早期诊断糖尿病视网膜病变的成像系统。

DESCRIPTION (provided by applicant): We have developed a portable, low-cost, patient-friendly, retinal imaging device, the Delft Retinascan, based on dual wavelength scanning laser technology. It does not require pharmacologic pupil dilation or flash illumination, and the prototype costs of the device were less than $1100. We have utilized it to acquire retinal images in normal subjects at high spatial resolution with a typical acquisition time of 0.1s. We and others have shown that large scale early diagnosis using digital cameras and internet connectivity is safe and effective. We have also shown that computer assisted diagnostic algorithms can assess diabetic retinopathy from two-field non-stereo color fundus photographs made with standard commercially available fundus cameras, with a high degree of agreement with retinal specialists. Potentially, these techniques can be combined for a low-cost, high quality, patient friendly, early diagnosis system for diabetic retinopathy. However, clinical validation is required before this low-cost camera and digital diagnostic system can be implemented clinically. Our specific aims are: 1. to technically validate the low-cost Delft Retinascan camera system for non-invasive imaging of diabetic retinopathy; 2. to assess and validate the clinical utility of the Delft Retinascan by collaborating with the Wisconsin Reading Center (WRC). The WRC will be responsible for reading and assessing images acquired with the Retinascan and compare them with ETDRS standard 7-field stereo fundus photographs; 3. and to evaluate the improvement in performance of the addition of computer assisted diagnostic algorithms on the images acquired with the Delft Retinascan. We have assembled a team representing expertise in diabetic retinopathy screening over the internet, optical engineering, retinal imaging, secure online medical image servers, image analysis, and epidemiology, consisting of research oriented ophthalmologists, physicists, computer engineers, and biostatisticians, and who are poised to translate the above techniques into a clinically practical system. Our overall aim is to validate and optimize a low-cost, high quality, patient friendly imaging system for the early diagnosis of diabetic retinopathy.